International Journal of Epidemiology | DeepDyve (2025)

Health in the Anthropocene Epoch—implications for epidemiology

Haines,, Andy

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy257pmid: 30452636

Humanity has entered a new geological epoch during which dramatic environmental trends, on a range of scales from local to global, are transforming natural systems. The predominant influence of Homo sapiens on the global environment has led to the growing use of the term ‘Anthropocene Epoch’1 to distinguish it from the ‘Holocene Epoch’, a climatically generally benign period which lasted around 11 500 years, during which humanity developed from hunter-gatherer and agrarian societies to increasingly urbanized communities. This transformation has been accompanied by pronounced increases in energy and freshwater use to drive economic growth and satisfy the growing demands of humanity for food and natural resources.2,3 It has resulted in unprecedented human progress, including the ‘Escape from Poverty’4 of many millions of people, particularly in Asia and Latin America. Since 1900, the global average life expectancy has more than doubled and is now almost 70 years. All countries in the world have a higher life expectancy than the countries with the highest life expectancy in 1800.5 These striking advances have come at a cost, which has been borne by the Earth’s natural systems.2 Global mean temperatures have increased by about 1°C since pre-industrial times as a result of emissions of greenhouse gases, biodiversity loss is about 100-fold pre-human rates, perturbation of nitrogen cycles is widespread and extensive degradation of marine ecosystems is occurring, together with ocean acidification and plastic pollution. Many complex systems can adapt to considerable change but then reach limits of adaptation which constitute a tipping point, after which, as a result of non-linear dynamics, sudden collapse occurs. The planetary boundaries framework aims to define a ‘safe operating space’ within which humanity can flourish,6 based on the assessment of nine biophysical processes essential for Earth systems’ functioning. A recent update of trends shows that two of them, biogeochemical flows and biosphere integrity, are in the high-risk zone, well beyond the proposed boundaries, and climate change and land system change are in the zone of uncertainty where risks are increasing. This analysis shows that we cannot merely extrapolate the trends of the recent past to understand and address the challenges of the future. Estimates by WHO suggest that about 23% of the current global disease burden is related to environmental factors,7 including air and water pollution. The Lancet Commission on Pollution8 estimated that 9 million premature deaths in 2015 could be attributed to the effects of pollution, with the majority arising from ambient and household air pollution. The combustion of fossil fuels is a major contributor to both air pollution and climate change, and the effects of global environmental change are likely to become increasingly dominant over coming decades and centuries. Environmental change can affect human health through three main types of pathways: (i) direct effects such as those due to increasing exposure to intense heat or changes in the frequency and intensity of extreme climate events; (ii) effects mediated through ecosystems, such as changes in the transmission of vector-borne diseases or zoonotic infections, due for example to land use change (or increased risk of undernutrition from reductions in crop yield due to climate change and other environmental stressors); and; (iii) socially mediated effects, for example as result of increased poverty or population displacement. Our understanding of the complex mechanisms underlying these pathways is still incomplete and is compounded by the methodological challenges of quantifying the magnitude of effects in the presence of multiple, potentially interacting, environmental and socioeconomic changes. An example is the effects of multiple environmental stressors, including climate change, water availability, air pollution and carbon dioxide levels on crop yields and nutritional quality of staple crops, vegetables, legumes,9 fruit, nuts and seeds. These could have profound but as yet incompletely understood effects on human health. The Rockefeller Foundation/Lancet Commission on Planetary Health10 defined planetary health in summary as ‘the health of human civilisation and the state of natural systems on which it depends’. It outlined three challenges that would need to be addressed in order to reduce risks to health in the Anthropocene Epoch. First, there is an imperative to address conceptual or imagination challenges, such as a dominant economic system in which powerful interests do not pay the full economic costs of their activities (such as the costs to society of pollution and ill health). Second is the urgent requirement to address knowledge challenges resulting from both the failure to invest in rigorous research and the methodological challenges of undertaking the transdisciplinary research needed. These knowledge challenges can be broadly subdivided into three categories: (i) improved understanding and quantification of the effects of environmental change on health and of factors influencing vulnerability to such changes; (ii) development and evaluation of effective strategies to promote adaptation and resilience to environmental change; and (iii) development and evaluation of interventions, technologies and policies to mitigate environmental change by profoundly reducing the environmental footprint of societies while protecting and enhancing human health and development, capitalizing on the health co-benefits of low carbon development.11 Advances in knowledge to safeguard humanity through the Anthropocene Epoch, as far as possible, will require transdisciplinary collaboration with epidemiology making a salient but not exclusive contribution. Epidemiology should play a central role, for example by improving exposure assessment and the quantification of disease burdens arising from environmental change and through rigorous evaluation of potential solutions, both adaptation and mitigation. An example of methodological innovation to assess adaptation to environmental change separated the contribution of pure adaptation to increasing temperatures and active changes in susceptibility (non-climate-driven mechanisms) to heat and cold. In order to do so, the investigators compared observed yearly-attributable fractions with predicted trends driven by either: (i) changes in exposure-response function (assuming a constant temperature distribution); or (ii) changes in temperature distribution (assuming constant exposure-response relationships). This comparison provides insights about the potential mechanisms and pace of adaptation.10 Systems approaches are needed in order to better understand complex interactions and feedbacks which may result in irreversible non-linear change in vital natural systems, such as the ‘state shifts’ that occur in some aquatic ecosystems as a result of changes in nutrient inputs, climate and fishing.11 They can also illuminate the effects of policy choices, such as those to mitigate climate change, on national economies.12 A study using optimization modelling, to assess the potential to reduce per person freshwater requirements for irrigation by dietary change in India, showed that diets meeting WHO guidelines could reduce freshwater requirements and greenhouse gas emissions while improving population health.13 In the case of potential solutions, unintended adverse consequences (such as the effects of some types of biofuels on food availability and security) may occur and systems approaches can help to anticipate and prevent these outcomes.14 Future studies will no doubt aim for more comprehensive approaches which integrate health, environmental and socioeconomic outcomes to understand trade-offs and synergies. In order to generate better data to facilitate the research needed and to hold decision makers to account, investment will be needed in linking health and environmental data at different scales using a range of sources, including demographic and health surveillance sites, cohort studies and remote sensing.15 The third category of challenges concerns the implementation of research findings in policy and practice.16 The need to address this challenge has been thrown into sharp relief by growing awareness that scientific evidence is increasingly contested by powerful stakeholders, emphasizing the need for effective governance to protect the health of current and future generations. A recent example is the reversal of Clean Air legislation in the USA.17 Barriers to implementation may range from inadequate efforts by researchers to engage well-intentioned decision makers in the co-creation of relevant knowledge, to organized efforts by vested interests to create disinformation and foment sufficient doubt to undermine political support for enforcement of evidence-based legislation. Designing research tailored to the needs of those who will implement the findings is a prerequisite for effective implementation, and the evaluation of credible strategies to accelerate uptake of research findings is an important component of the evolving research agenda for planetary health. The Anthropocene Epoch poses challenges to our current research paradigms, suggesting the need to collaborate beyond traditional disciplinary silos to engage a range of sectors (including agriculture, energy, urban planning and design, transport etc.) in order to develop and implement solutions which protect the health of today’s population and future generations. Over 20 years ago, the late Tony McMichael presciently warned of the threats to health posed by overloading the Earth’s essential life support systems.18 It is now a matter of urgency that the epidemiological community responds to his clarion call by addressing the burgeoning challenges to health and sustainable development in the Anthropocene Epoch. References 1 Crutzen PJ. Geology of mankind . Nature 2002 ; 415 : 23 . Google Scholar Crossref Search ADS PubMed 2 Steffen W , Broadgate W , Deutsch L , Gaffney O , Ludwig C. The trajectory of the Anthropocene: the great acceleration . Anthrop Rev 2015 ; 2 : 81 – 98 . Google Scholar Crossref Search ADS 3 Waters CN , Zalasiewicz J , Summerhayes C et al. The Anthropocene is functionally and stratigraphically distinct from the Holocene . Science 2016 ; 351 : aad2622. Google Scholar Crossref Search ADS PubMed 4 Deaton A. The Great Escape: Health, Wealth, and the Origins of Inequality . Princeton, NJ : Princeton University Press , 2013 . 5 Roser M. Life Expectancy. Our World in Data. https://ourworldindata.org/life-expectancy (2 July 2018 , date last accessed). 6 Steffen W , Richardson K , Rockström J et al. Planetary boundaries: guiding human development on a changing planet . Science 2015 ; 347 : 1259855 . Google Scholar Crossref Search ADS PubMed 7 Prüss-Ustün A , Wolf J , Corvalán C , Neville T , Bos R , Neira M. Diseases due to unhealthy environments: an updated estimate of the global burden of disease attributable to environmental determinants of health . J Public Health 2017 ; 39 : 464 – 75 . 8 Landrigan PJ , Fuller R , Acosta NJR et al. Lancet commission on pollution and health . Lancet 2017 ; doi:10.1016/S0140-6736(17)32345-0. 9 Scheelbeek P , Bird F , Tuomisto H et al. The effect of environmental change on yields and nutritional quality of vegetables and legumes: A systematic review and meta-analysis . Proc Natl Acad Sci U S A 2018 ; doi:10.1073/pnas.1800442115. 10 Vicedo-Cabrera AM , Sera F , Guo Y et al. A multi-country analysis on potential adaptive mechanisms to cold and heat in a changing climate . Environ Int 2018 ; 111 : 239 – 46 . Google Scholar Crossref Search ADS PubMed 11 Whitmee S , Haines A , Beyrer C et al. Safeguarding human health in the Anthropocene epoch: report of The Rockefeller Foundation-Lancet Commission on planetary health . Lancet 2015 ; 386 : 1973 – 2028 . Google Scholar Crossref Search ADS PubMed 12 Jensen HT , Keogh-Brown M , Smith RD et al. The importance of health co-benefits in macroeconomic assessments of UK Greenhouse Gas emission reduction strategies . Clim Change 2013 ; doi:10.1007/s10584-013-0881-6. 13 Milner J , Joy EJM , Green R et al. Projected health effects of realistic dietary changes to address freshwater constraints in India: a modelling study . Lancet Planet Health 2017 ; 1 : e26 – 32 . Google Scholar Crossref Search ADS PubMed 14 Pongsiri MJ , Gatzweiler F , Bassi AM , Haines A , Demassieux F. The need for a systems approach to planetary health . Lancet Planet Health 2017 ; 1 : e257 – 59 . Google Scholar Crossref Search ADS PubMed 15 Haines A , Hanson C , Ranganathan J. Planetary Health Watch: integrated monitoring in the Anthropocene epoch . Lancet Planet Health 2018 ; 2 : e141 – 43 . Google Scholar Crossref Search ADS PubMed 16 Pattanayak S , Haines A. Implementation of policies to protect planetary health . Lancet Planet Health 2017 ; 1 : e255 – 56 . Google Scholar Crossref Search ADS PubMed 17 Beech E. U.S. EPA reverses policy on ‘major sources’ of pollution Reuters U.S. LEGAL NEWS JANUARY 26, 2018. https://www.reuters.com/article/us-usa-trump-epa/u-s-epa-reverses-policy-on-major-sources-of-pollution-idUSKBN1FF075 (2 July 2018, date last accessed). 18 McMichael AJ. Planetary Overload: Global Environmental Change and the Health of the Human Species . Cambridge, UK : Cambridge University Press , 1993 . © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Open Access Collection

Mobilizing an underused resource: cohort studies for population health intervention research

Edwards, Nancy; Plotnikoff, Ronald C

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy191pmid: 30239808

Introduction Prospective, longitudinal cohort studies involve large, expensive and long-term investments by research funders. Traditionally, these studies have focused on risk factor analysis—contributing substantially to our understanding of disease trends, predisposing and protective influences on illness and injury, and susceptibility during life course transitions. However, we have now entered the era of big data and megacohorts, supported by unprecedented computing and statistical analytical power, and fuelled by biobanks that capture hundreds of thousands of biologically diverse samples to answer questions in the field of epigenetics and genomics.1 To this end, a number of national research funding councils have reviewed existing and planned cohorts, and developed research agendas to maximize these investments.2–4 Whereas we applaud these efforts, we are concerned that this new emphasis on links to biorepositories overlooks the potential for a complementary population health intervention research agenda. We use Hawe and Potvin’s5 definition for this commentary: ‘Population health intervention research uses scientific methods to produce knowledge about policy and program interventions that operate within or outside of the health sector and have the potential to impact health at the population level’ (p. 1-8). The status quo To be fair, the use of cohorts for population health intervention research is not starting from a position of strength. Recent reports2,3,6,7 rarely make reference to population health intervention research emanating from cohort studies. Less than one-third of cohort profiles published in IJE between 2015 and 2017 mention population health intervention research as a primary or secondary purpose of their studies.8 Furthermore, in a review we undertook of all publications reported for two purposefully chosen, large and long-term cohort studies [i.e. the Australian Longitudinal Study of Women’s Health (ALSWH), and the Survey of Health, Ageing and Retirement in Europe (SHARE)], we found that only 1.9% (22/1162) of publications from these cohorts were in the domain of population health intervention research. It is our contention that longitudinal cohorts could be used to buttress several areas of weakness identified in existing population health intervention research.9 First, many of these intervention studies fail to examine longer-term issues of equity, scalability and sustainability.10–13 Second, programme interventions designed by researchers are more frequently examined in population health intervention studies than natural experiments of policy interventions.14,15 Third, population health intervention research studies often have a narrow range of outcomes due to funding constraints, time limitations and efforts to reduce response burden. Thus, examinations of the unintended and longer-term consequences of population health interventions and whether or not they have been successfully scaled up, have often been severely curtailed. Longitudinal cohort studies typically follow populations over several cycles of measurement, providing a means to determine whether the outcomes of intervention scale-up are distributed equitably, whether intervention effects are sustained beyond the research funding period and what longer-term consequences (desired or undesired) have resulted from the introduction of an intervention. Larger-scale cohorts, such as national or multi-national longitudinal studies on ageing, and occupational or birth cohorts, typically sample populations from a range of jurisdictions, providing an innate comparative base for policy change. It is this subset of population health intervention research, examining natural policy experiments, to which we now turn our attention. Priority directions To maximize the use of cohort data for policy-oriented population health intervention research, we suggest three priority directions: enhance jurisdictional diversity, attend to health equity and consider policy exposure. Enhance jurisdictional diversity Whereas population representativeness has been a major consideration in the design of cohort studies, contextual representativeness also warrants attention. Jurisdictional diversity is often essential for questions of policy impact, since policies, their implementation and enforcement may differ substantially between organizational settings (e.g. workplaces, schools) or among and between levels of government (e.g. municipality, state/province or nation). Comparisons across jurisdictions may provide a set of counterfactuals to interrogate the impact of: (i) interacting policies and regulations across governance levels; (ii) different constellations of policies across jurisdictions; and (iii) pervasive structural policies such as universal health or social programmes or discriminatory legislation. Furthermore, cohort studies can be used to examine both health-damaging and health-enhancing policies. When selecting jurisdictions for inclusion, these types of queries need to be taken into account. The adequacy of samples for cross-jurisdictional comparisons and multilevel analysis should be carefully considered during planning stages. For instance, real-life examples of policy-oriented questions can be used to test statistical assumptions for analytical approaches such as hierarchical linear modelling. Attend to health equity The intent of population health interventions is to improve health equity—levelling up health for the entire population while reducing social and class disparities. It is imperative that cohorts include measures of these variables, keeping apace of developments in this field such as more progressive ways of describing gender. Analytical tools such as the health equity concentration index and the Lorenz curve,16 or analytical approaches such as intersectionality, can be used to examine the health equity impact of policies and their enforcement. Data linkage (often to administrative records) is critical, both to expand the range of equity variables that may be used and to potentially augment cohort samples with individuals who may not typically participate in longitudinal cohort studies. The health equity impact of population health interventions can only be assessed by comparing population subgroups, such as those representing the full continuum of the socioeconomic gradient. When prioritizing administrative databases for their linkage to cohorts, these equity considerations should be taken into account. Those who fund and design cohort studies need to make their health equity aims explicit. There are excellent examples of strategic funding opportunities (e.g. Health Life Trajectories Initiative, Canadian Institutes of Health Research) that purposefully ask researchers to consider health equity issues such as the ‘social implications of any new policies, practices, health services or interventions resulting from the study’.17 Consider policy exposure Links to policy frameworks, regulations and enforcement strategies are needed to examine the impact of policy interventions. These are sources of policy-related exposure data—complementary to those already being tapped in data linkage initiatives using registries, administrative databases and electronic medical records. Some of the most successful policy research has been in the field of tobacco control, fuelled by deliberate efforts to use policy as a primary lever for change. There are important lessons to be learned from this field for other population health policy questions. Specifically, approaches are needed to measure policy exposure, since policy itself is a rather blunt intervention that is moderated and/or mediated by self-regulation, formal enforcement, informal efforts to undermine enforcement and social movements. Because of their longitudinal design, cohort studies can help examine multiple and cumulative exposures, providing much needed evidence on how interventions interact either synergistically or antagonistically to produce particular effects. When the evaluation of policy interventions is embedded within cohort studies using randomized controlled trials or other study designs, cohorts may provide researchers with a wider range of potential outcomes, including those reflecting unintended effects in both the shorter and the longer term. This, in turn, may yield important findings concerning whether and how such policy interventions were successfully scaled up and sustained. Realizing the potential of longitudinal cohorts for population health intervention studies Cohort studies represent a major, recurring research investment; maximizing the potential of these platforms for population health intervention studies is an important obligation to the scientific community and the public. Practical yet robust approaches for integrating nested randomized controlled trials and other evaluation strategies within longitudinal cohorts need to be considered. Trade-offs may have to be made including population representativeness, oversampling population subgroups, frequency of follow-up, survey content and response burden. We suggest that exemplary policy questions should be posed before implementation to determine whether and how a cohort study can be better structured and designed for future questions of this type. Important new policy questions will undoubtedly emerge over the life of a cohort, but this initial testing will help ensure that this class of studies are considered during the cohort design phase. Many can contribute to the aims of maximizing the potential population health intervention policy yield of longitudinal cohort studies outlined in this paper. Funding agencies should consider specific funding streams to support this kind of research. Explicitly integrating intervention goals into the overarching aims of cohort studies would help to signal their potential use for this purpose. Universities can offer training on how to access and use longitudinal cohorts for population health intervention policy studies. Journals such as IJE can encourage the publication of population health policy studies linked to cohorts. We contend that longitudinal cohorts can meet the multiple aims of risk factor analysis, genome associations and population health interventions. Cohort design trade-offs for these alternative purposes need to be made explicit as research funding organizations consider how to maximize past and ongoing investments in longitudinal cohorts. Optimizing the use of cohorts for population health intervention research and, in particular, policy research needs to be considered when the utility and efficiency of cohort studies are reviewed by funders. Funding This work was supported by the Canadian Institutes of Health Research [Grant Number: 122510]. Acknowledgements We thank Katie Hoogeveen for her assistance with the review and categorization of publications reported from the Australian Longitudinal Study of Women’s Health (ALSWH) and the Survey of Health, Ageing and Retirement in Europe (SHARE). Conflict of interest: None declared. References 1 Gaziano JM. The evolution of population science: advent of the mega cohort . JAMA 2010 ; 304 : 2288 – 89 . Google Scholar Crossref Search ADS PubMed WorldCat 2 MRC Population Health Sciences Group . Maximising the Value of UK Population Cohorts: MRC Strategic Review of the Largest UK Population Cohort Studies . London/Swindon, UK : Medical Research Council (MRC ), 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 3 National Institutes of Health . Summary of Research Activities by Key Approach and Resource: Epidemiological and Longitudinal Studies .https://report.nih.gov/biennialreport0607/pdf/NIH_BR_Ch3_epidemiological.pdf (4 July 2018, date last accessed). Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 4 Sorlie P , Wei GS. Population-based cohort studies: still relevant? J Am Coll Cardiol 2011 ; 58 : 2010 – 13 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Hawe P , Potvin L. What is population health intervention research? Can J Public Health 2009 ; 100(Suppl): I8 . Google Scholar OpenURL Placeholder Text WorldCat 6 UNICEF Office of Research Innocenti . Symposium on Cohorts and Longitudinal Studies in Low and Middle Income Countries , Florence, Italy : UNICEF Office of Research Innocenti , 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 7 The EU Joint Programme – Neurodegenerative Disease Research (JPND) . Longitudinal Cohort Studies in Neurodegeneration Research . 2013 . http://www.neurodegenerationresearch.eu/uploads/media/JPNDAGLCS_Final_Report_Oct_2013-version_07_01_14.pdf (5 March 2018, date last accessed). 8 Edwards NC , Plotnikoff RC, Hoogeveen K. Enhancing the utility of International Journal of Epidemiology cohort profiles . Int J Epidemiol 2018 ; 47 : 1008 – 09 . Google Scholar Crossref Search ADS WorldCat 9 Relton C , Bissell P, Smith C et al. South Yorkshire cohort: a “cohort trials facility” study of health and weight—protocol for the recruitment phase . BMC Public Health 2011 ; 11 : 640. Google Scholar Crossref Search ADS PubMed WorldCat 10 Östlin P , Schrecker T, Sadana R et al. Priorities for research on equity and health: towards an equity-focused health research agenda . PLoS Med 2011 ; 8 : e1001115. Google Scholar Crossref Search ADS PubMed WorldCat 11 Stirman S , Kimberly J, Cook N, Calloway A, Castro F, Charns M. The sustainability of new programs and innovations: a review of the empirical literature and recommendations for future research . Implement Sci 2012 ; 7 : 17. Google Scholar Crossref Search ADS PubMed WorldCat 12 Milat AJ , Bauman A, Redman S. Narrative review of models and success factors for scaling up public health interventions . Implementation Sci 2015 ; 10 : 113. Google Scholar Crossref Search ADS WorldCat 13 Bonell C , Jamal F, Melendez-Torres GJ, Cummins S. “Dark logic”: theorising the harmful consequences of public health interventions . J Epidemiol Community Health 2015 ; 69 : 95. Google Scholar Crossref Search ADS PubMed WorldCat 14 Craig P , Cooper C, Gunnell D et al. Using natural experiments to evaluate population health interventions: new Medical Research Council guidance . J Epidemiol Community Health 2012 ; 66 : 1182 – 86 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Petticrew M , Rehfuess E, Noyes J et al. Synthesizing evidence on complex interventions: how meta-analytical, qualitative, and mixed-method approaches can contribute . J Clin Epidemiol 2013 ; 66 : 1230 – 43 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Nickel N , Brownell M, Chateau D, Katz A, Burland E, Vehling L. A Tale of Three Interventions: Cautionary Accounts in the Use of Health Equity Measures in Population Health Intervention Research. 2016 . https://umanitoba.ca/faculties/health_sciences/medicine/units/chs/departmental_units/mchp/education/media/Three_Interventions_Sparking_Solutions_Final_April_18_2016.pdf (8 March 2018, date last accessed). 17 ResearchNet. Funding Opportunity Details: Objectives . 2015 . https://www.researchnet-recherchenet.ca/rnr16/vwOpprtntyDtls.do?prog=2367&view=currentOpps&type=EXACT&resultCount=25&sort=program&all=1&masterList=true#objective (4 July 2018, date last accessed). © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [emailprotected] © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association.

Data Resource Profile: The China National Health Survey (CNHS)

He,, Huijing;Pan,, Li;Pa,, Lize;Cui,, Ze;Ren,, Xiaolan;Wang,, Dingming;Liu,, Feng;Wang,, Xianghua;Du,, Jianwei;Wang,, Hailing;Wan,, Shaoping;Zhao,, Jingbo;Peng,, Xia;Wang,, Xiaoyang;Zhang,, Jia;Wang,, Ye;Ren,, Huiru;Yu,, Chengdong;Shan,, Guangliang

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy151pmid: 30124853

Abstract The China National Health Survey (CNHS) is the first nationwide multi-ethnic cross-sectional interview and health examination conducted from 2012 to 2017. The survey is designed to study reference intervals for physiological constants as well as determinants of noncommunicable diseases among different ethnic populations in different areas, so that the data can be used to enhance clinical diagnosis strategies and health promotion. CNHS used a stratified, multistage cluster sampling method to obtain a sample of 53 895 people aged 20-80 years in 10 ethnic groups from 11 provinces or autonomous regions all over China. Blood samples were collected from each participant for the establishment of the China Multi-Ethnic Biobank (CMEB). CNHS collected data on demographic and socioeconomic information, lifestyle factors, anthropometric measures, laboratory tests and clinical profiles. These data provide a comprehensive resource for further study on risk factors of noncommunicable disease among different ethnic groups. Information about the CNHS database, including publication list, introduction of the survey design and methods, and guidelines for submitting electronic forms of data application, is available at [http://www.bmicc.cn/web/share/home]. China, multi-ethnic population, biobank, clinical reference intervals, noncommunicable diseases Data resource basics The China National Health Survey (CNHS) is a cross-sectional study conducted from 2012 to 2017 by the Institute of Basic Medical Sciences (IBMS), Chinese Academy of Medical Sciences (CAMS) and School of Basic Medicine, Peking Union Medical College (PUMC). The survey includes 53 895 people in 10 ethnic groups from 11 provinces of China. For decades, foreign physiological and biochemical standards have been used in clinical diagnosis and health status classification, but evidence from previous studies showed that physiological constants varied among different ethnic groups,1–3 thus leading to overestimation or underestimation in clinical diagnosis or health assessment. CNHS is a programme of multidisciplinary studies designed to study reference intervals for physiological constants in different ethnic populations, so that these data can be used to enhance clinical diagnosis strategies and health promotion.3 This survey is unique in that it combines questionnaire interviews with standardized physical examinations and laboratory tests to obtain information on various health correlates or determinants. Participants gave their blood samples for the establishment of the China Multi-Ethnic Biobank (CMEB). We performed biochemical tests including full blood cell counts, haemoglobin, fasting plasma glucose, serum lipids, HBsAg, liver and kidney function indexes, ABO and Rh blood types, immunological indexes, etc. The CNHS is an invaluable resource for health study and clinical research, since to our best knowledge, it is the first comprehensive multi-ethnic population data collection from different areas in China. In addition, by storing blood samples of more than 53 000 individuals of different ethnicities, it will allow reliable assessment of genetic and other factors as risk factors for noncommunicable diseases (NCDs). Data resource areas and population coverage A cross-sectional study with a multistage, stratified cluster sampling method was used in this programme. The sampling process was stratified according to geographical regions (North, South, East and West China), distribution of minority ethnic populations, degree of urbanization (provincial capitals, midsize cities, county seats and rural townships), and economic development status assessed based on gross domestic product (GDP) for each province. In the first stage of sampling, provinces were selected from geographical regions considering the distribution of ethnic groups. A total of 11 provinces or autonomous regions—Guizhou, Hainan, Xinjiang, Shaanxi, Inner Mongolia, Qinghai, Heilongjiang, Yunnan, Gansu, Sichuan and Hebei—were selected as survey sites. In each province, the predominant ethnic groups were selected as study population; then in the second stage, cities and counties were selected from each province. The first two stages were not random. In the next stage, districts were selected from cities, and rural townships were selected from counties. In the fourth stage, communities were selected from districts in urban areas, whereas villages were selected from townships in rural areas. In the final stage, people resident in the selected areas were all invited to participate in the study (see Figure 1). With this multistage, stratified cluster design, representative samples from 10 ethnic populations in 50 urban communities and 90 villages were selected. Only individuals aged 20–80, who had lived in their current residence for at least 1 year, were eligible to participate. Individual with severe mental or physical illness, pregnant females and military personnel on active service were excluded. Figure 1 View largeDownload slide Flowchart of the multistage stratified sampling procedure in CNHS. Figure 1 View largeDownload slide Flowchart of the multistage stratified sampling procedure in CNHS. Survey frequency The CNHS is conducted each year as from 2012, in the summer season (for convenience in implementing physical examinations). The location of the 11 survey sites is presented in Figure 2. Figure 2 View largeDownload slide Location of 11 survey sites. Figure 2 View largeDownload slide Location of 11 survey sites. We conducted the survey first at Guizhou province (South-west China) in October 2012, where the local minority ethnic group was Bouyei. In 2013, the same survey was conducted in Hainan Province (South China, Li and Han People) and Xinjiang Uyghur Autonomous Region (North-west China, Uyghur and Han People). In sequence, surveys in Shaanxi province (West China, Han People) and Inner Mongolia Autonomous Region (IMAR) (North China, Mongolian and Han People) were conducted in 2014, followed by Heilongjiang (North-east China, Korean and Han People), Qinghai (North-west China, Tibetan and Han People), Yunnan (South-west China, Yi and Han People), and Sichuan (South-west China, Yi People) Provinces in 2015. We conducted the survey in the last two provinces, Gansu (North-west China, Yugur and Han People) and Hebei (Middle China, Manchu and Han People), in 2016 and 2017. A total of 10 ethnic groups were included in the survey: Han, Bouyei, Li, Uyghur, Mongolian, Korean, Tibetan, Yi, Yugur and Manchu. A total of 54 111 questionnaires were collected. After excluding participants aged below 20 or above 80, or with missed data on age, sex or ethnic information, we finally collected 53 895 valid questionnaires for data analysis. The eligibility questionnaire rate is 99.60% (53 895/54 111, see Table 1). Table 1. Survey frequency and eligibility questionnaire rate of CNHS Survey year Province Major ethnic groups Questionnaires submitted Eligibility questionnaires Eligibility questionnaire rate (%)a 2012 Guizhou Bouyei, Han 6073 6038 99.42 2013 Hainan Li, Han 4674 4615 98.74 Xinjiang Uyghur, Han 6684 6658 99.61 2014 Shaanxi Han 5784 5763 99.64 Inner Mongolia Mongolian, Han 3468 3461 99.80 2015 Heilongjiang Korean, Han 3262 3260 99.94 Qinghai Tibetan, Han 5004 4960 99.12 Yunnan Yi, Han 3183 3179 99.87 Sichuan Yi 3340 3331 99.73 2016 Gansu Yugur, Han 6081 6078 99.95 2017 Hebei Manchu, Han 6558 6552 99.91 Total 54 111 53 895 99.60 Survey year Province Major ethnic groups Questionnaires submitted Eligibility questionnaires Eligibility questionnaire rate (%)a 2012 Guizhou Bouyei, Han 6073 6038 99.42 2013 Hainan Li, Han 4674 4615 98.74 Xinjiang Uyghur, Han 6684 6658 99.61 2014 Shaanxi Han 5784 5763 99.64 Inner Mongolia Mongolian, Han 3468 3461 99.80 2015 Heilongjiang Korean, Han 3262 3260 99.94 Qinghai Tibetan, Han 5004 4960 99.12 Yunnan Yi, Han 3183 3179 99.87 Sichuan Yi 3340 3331 99.73 2016 Gansu Yugur, Han 6081 6078 99.95 2017 Hebei Manchu, Han 6558 6552 99.91 Total 54 111 53 895 99.60 a Eligibility questionnaire rate = eligibility questionnaires/questionnaires submitted. Table 1. Survey frequency and eligibility questionnaire rate of CNHS Survey year Province Major ethnic groups Questionnaires submitted Eligibility questionnaires Eligibility questionnaire rate (%)a 2012 Guizhou Bouyei, Han 6073 6038 99.42 2013 Hainan Li, Han 4674 4615 98.74 Xinjiang Uyghur, Han 6684 6658 99.61 2014 Shaanxi Han 5784 5763 99.64 Inner Mongolia Mongolian, Han 3468 3461 99.80 2015 Heilongjiang Korean, Han 3262 3260 99.94 Qinghai Tibetan, Han 5004 4960 99.12 Yunnan Yi, Han 3183 3179 99.87 Sichuan Yi 3340 3331 99.73 2016 Gansu Yugur, Han 6081 6078 99.95 2017 Hebei Manchu, Han 6558 6552 99.91 Total 54 111 53 895 99.60 Survey year Province Major ethnic groups Questionnaires submitted Eligibility questionnaires Eligibility questionnaire rate (%)a 2012 Guizhou Bouyei, Han 6073 6038 99.42 2013 Hainan Li, Han 4674 4615 98.74 Xinjiang Uyghur, Han 6684 6658 99.61 2014 Shaanxi Han 5784 5763 99.64 Inner Mongolia Mongolian, Han 3468 3461 99.80 2015 Heilongjiang Korean, Han 3262 3260 99.94 Qinghai Tibetan, Han 5004 4960 99.12 Yunnan Yi, Han 3183 3179 99.87 Sichuan Yi 3340 3331 99.73 2016 Gansu Yugur, Han 6081 6078 99.95 2017 Hebei Manchu, Han 6558 6552 99.91 Total 54 111 53 895 99.60 a Eligibility questionnaire rate = eligibility questionnaires/questionnaires submitted. Measures In the survey, a standard questionnaire was developed to conduct face-to-face personal interviews. All interviewers and staff completed a training programme that guaranteed their ability to use specific tools and methods. At the training session, detailed instructions for questionnaire administration were given to interviewers. Clinical staff were trained before the formal survey, to familiarize them with the physiological and biochemical tests. The questionnaire consisted of items on: (i) demographic and socioeconomic information; (ii) lifestyle factors including tobacco use, alcohol consumption, physical activity and regular exercise; (iii) medical history of diagnosis and treatment for diabetes, hypertension, cardiovascular events and other diseases; and (iv) family history of noncommunicable diseases and cancer, etc. (see Table 2). Current tobacco use was defined as smoking at least one cigarette per day for at least the past 6 months. Former tobacco use was defined as having quit tobacco smoking for more than the 6 months preceding the survey. Alcohol drinking was defined as the consumption of at least 30 g of alcohol, and had last at least 6 months. Physical activity was classified into three categories as light, moderate or heavy according to intensity. Regular exercise was defined as participation in moderate or vigorous activity for 20 min or more per day at least 3 days a week.4 Table 2. Summary of questionnaire data collected in this programme Part 1: Demographic characteristics Part 2: Personal and family history of disease Sex Personal medical history Ethnicity Date of hypertension diagnosis Parents’ ethnicities Hypertension therapy Birthday Date of diabetes diagnosis Birth place Diabetes therapy Rural/urban residence History of fracture Marital status Other disease history (COPD, heart disease, TB, asthma, cancer, etc.) Education Occupation History of HBV vaccination Income Family history of disease Medical insurance Diagnosis of hypertension, diabetes, cancer and other diseases (parental, grandparental, or sibling level) Age at first menstruationa Menopausal statusa Age at menopausea Part3: Lifestyle factors Tobacco use Alcohol consumption Physical activity Regular exercise. Part 1: Demographic characteristics Part 2: Personal and family history of disease Sex Personal medical history Ethnicity Date of hypertension diagnosis Parents’ ethnicities Hypertension therapy Birthday Date of diabetes diagnosis Birth place Diabetes therapy Rural/urban residence History of fracture Marital status Other disease history (COPD, heart disease, TB, asthma, cancer, etc.) Education Occupation History of HBV vaccination Income Family history of disease Medical insurance Diagnosis of hypertension, diabetes, cancer and other diseases (parental, grandparental, or sibling level) Age at first menstruationa Menopausal statusa Age at menopausea Part3: Lifestyle factors Tobacco use Alcohol consumption Physical activity Regular exercise. COPD, chronic obstructive heart disease; TB, tuberculosis; HBV, hepatitis B virus a Only for female participants. Table 2. Summary of questionnaire data collected in this programme Part 1: Demographic characteristics Part 2: Personal and family history of disease Sex Personal medical history Ethnicity Date of hypertension diagnosis Parents’ ethnicities Hypertension therapy Birthday Date of diabetes diagnosis Birth place Diabetes therapy Rural/urban residence History of fracture Marital status Other disease history (COPD, heart disease, TB, asthma, cancer, etc.) Education Occupation History of HBV vaccination Income Family history of disease Medical insurance Diagnosis of hypertension, diabetes, cancer and other diseases (parental, grandparental, or sibling level) Age at first menstruationa Menopausal statusa Age at menopausea Part3: Lifestyle factors Tobacco use Alcohol consumption Physical activity Regular exercise. Part 1: Demographic characteristics Part 2: Personal and family history of disease Sex Personal medical history Ethnicity Date of hypertension diagnosis Parents’ ethnicities Hypertension therapy Birthday Date of diabetes diagnosis Birth place Diabetes therapy Rural/urban residence History of fracture Marital status Other disease history (COPD, heart disease, TB, asthma, cancer, etc.) Education Occupation History of HBV vaccination Income Family history of disease Medical insurance Diagnosis of hypertension, diabetes, cancer and other diseases (parental, grandparental, or sibling level) Age at first menstruationa Menopausal statusa Age at menopausea Part3: Lifestyle factors Tobacco use Alcohol consumption Physical activity Regular exercise. COPD, chronic obstructive heart disease; TB, tuberculosis; HBV, hepatitis B virus a Only for female participants. Physical examination included: anthropometry of height, weight and body composition; measurement of blood pressure, heart rate, axillary temperature, grip strength, bone mineral density, oxygen saturation of blood and visual acuity. Standing Height and sitting height were measured to the nearest 0.1 cm using a fixed stadiometer. Weight was measured by body composition analyser (TANITA BC-420, Japan), with the accuracy to decimal level.5 Three blood pressure (BP) and heart rate readings were taken by trained personnel using a digital BP measuring device (Omron HEM-907, Japan). Before the BP measurement, all participants were required to have at least 5 min of rest, avoiding exercise, drinking, smoking or tea for at least 30 min before the measurement. Axillary temperature was measured by digital medical thermometer (Omron MC-348-HP, Liaoning, China) while measuring blood pressure. Quantitative ultrasound parameters of the right calcaneus were measured as the proxy of bone mineral density, using a Pegasus device (Medilink, France). Grip strength was measured two times for each participant using a Jamar Hydraulic Hand Evaluation Kit (JAMAR, UK). Oxygen saturation of blood was measured using an oximeter (Masimo Red-57, USA). Monocular visual acuity was estimated by a Logarithm of the Minimum Angle of Resolution (LogMAR) Chart (Wehen Dimensional Scale precision Technology, Zhuhai, Guangdong Province, China) from a distance of 4 m. Participants were wearing their spectacles if they had them. To establish valid spirometry predictive equations for the general population in China, spirometry was performed among non-smokers with no cardiovascular disease onset, who had neither significant self-reported pulmonary disease nor acute respiratory illnesses. In accordance with the European Respiratory Society (ERS) and the American Thoracic Society (ATS) guidelines,6 spirometry was performed by trained investigators using a spirometer (Jaeger, Würzburg, Germany). Strict qualification criteria7 were followed for spirometry measurement, and the largest forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) were recorded.8 For each participant, a 9-ml fasting blood sample (two EDTA vacutainers containing 2 ml blood for each, and one vacutainer without EDTA containing 5 ml blood), after at least 8 h fasting overnight, was collected. For each day of the survey, the samples were kept in a portable, insulated coolbox with ice packs to maintain their temperature at 0–4°C for up to 3 h before being transported to the laboratory of local centre for disease control and prevention (CDC) for immediate processing. After centrifuging and aliquoting, four cryovials (including one DNA-containing buffy coat) from each blood sample were stored in a −20°C refrigerator for a few days, before being transported with dry ice to the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, for storage at −80°C. Full blood cell count, haemoglobin, ABO and Rh blood type, fasting glucose, lipids, serum liver and kidney function indexes, HBsAg, and immunological indexes were tested in the General Hospital of Chinese People’s Liberation Army. Blood test items are listed in Table 3. Table 3. Physiological measurements and laboratory tests Variables No. measurements Equipment used Precision Height (cm) 1 Manufactured instrument 0.1 Sitting height (cm) 1 Manufactured instrument 0.1 Weight (kg) 1 Body Composition Analyzer 0.1 Waist circumference (cm) 3 Standard tape measure 0.1 Hip circumference (cm) 3 Standard tape measure 0.1 Body composition 1 Body Composition Analyzer N/A Sitting blood pressure (mmHg) 3 Digital BP monitor 1 Resting pulse rate (beats/min) 3 Digital BP monitor 1 Axillary temperature (°C) 1 Digital medical thermometer 0.1 Visual acuity 1 LogMAR Chart 0.1 Respiratory functiona 1 Spirometer N/A Grip strength (kg) 2 JAMAR Kit 1 Oxygen saturation (%) 1 Masimo Red-57 1 Bone mineral density 1 Medilink OSTEOSPACE N/A ABO blood type 1 Anti-A and anti-B blood grouping reagents (human monoclonal antibodies) N/A Rh blood type 1 Anti-Rh (D) reagent for blood grouping (human monoclonal antibody) N/A Full blood countb 1 Automated haematology Analyzer (XT-4000i, Japan) N/A Haemoglobin(g/L) 1 Automated haematology Analyzer (XT-4000i, Japan) 1 HBsAg test 1 Diagnostic kit for hepatitis B surface antigen (ELISA) N/A Fasting blood glucose (FBG, mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Lipid testc (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Alanine transaminase (ALT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Glutamyl transferase (GGT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Alkaline phosphatase (ALP, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Aspartate transaminase (AST, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Creatinine (CrE, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Urea (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Uric acid (UA, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Total protein (TP, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Albumin (ALB, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 IgG (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgA (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgM (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Variables No. measurements Equipment used Precision Height (cm) 1 Manufactured instrument 0.1 Sitting height (cm) 1 Manufactured instrument 0.1 Weight (kg) 1 Body Composition Analyzer 0.1 Waist circumference (cm) 3 Standard tape measure 0.1 Hip circumference (cm) 3 Standard tape measure 0.1 Body composition 1 Body Composition Analyzer N/A Sitting blood pressure (mmHg) 3 Digital BP monitor 1 Resting pulse rate (beats/min) 3 Digital BP monitor 1 Axillary temperature (°C) 1 Digital medical thermometer 0.1 Visual acuity 1 LogMAR Chart 0.1 Respiratory functiona 1 Spirometer N/A Grip strength (kg) 2 JAMAR Kit 1 Oxygen saturation (%) 1 Masimo Red-57 1 Bone mineral density 1 Medilink OSTEOSPACE N/A ABO blood type 1 Anti-A and anti-B blood grouping reagents (human monoclonal antibodies) N/A Rh blood type 1 Anti-Rh (D) reagent for blood grouping (human monoclonal antibody) N/A Full blood countb 1 Automated haematology Analyzer (XT-4000i, Japan) N/A Haemoglobin(g/L) 1 Automated haematology Analyzer (XT-4000i, Japan) 1 HBsAg test 1 Diagnostic kit for hepatitis B surface antigen (ELISA) N/A Fasting blood glucose (FBG, mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Lipid testc (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Alanine transaminase (ALT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Glutamyl transferase (GGT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Alkaline phosphatase (ALP, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Aspartate transaminase (AST, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Creatinine (CrE, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Urea (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Uric acid (UA, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Total protein (TP, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Albumin (ALB, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 IgG (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgA (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgM (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 N/A, not available. a Only conducted in non-smokers, with no cardiovascular disease, no significant self-reported pulmonary diseases nor acute respiratory illnesses. b Full blood count includes the numbers of white blood cells (WBC, 109/L), red blood cells (RBC, 1012/L) and platelets (109/L). c Lipid test includes triglyceride (TG, mmol/L), total cholesterol (TC, mmol/L), high-density lipoprotein cholesterol (HDL-C, mmol/L) and low-density lipoprotein cholesterol (LDL-C, mmol/L). Table 3. Physiological measurements and laboratory tests Variables No. measurements Equipment used Precision Height (cm) 1 Manufactured instrument 0.1 Sitting height (cm) 1 Manufactured instrument 0.1 Weight (kg) 1 Body Composition Analyzer 0.1 Waist circumference (cm) 3 Standard tape measure 0.1 Hip circumference (cm) 3 Standard tape measure 0.1 Body composition 1 Body Composition Analyzer N/A Sitting blood pressure (mmHg) 3 Digital BP monitor 1 Resting pulse rate (beats/min) 3 Digital BP monitor 1 Axillary temperature (°C) 1 Digital medical thermometer 0.1 Visual acuity 1 LogMAR Chart 0.1 Respiratory functiona 1 Spirometer N/A Grip strength (kg) 2 JAMAR Kit 1 Oxygen saturation (%) 1 Masimo Red-57 1 Bone mineral density 1 Medilink OSTEOSPACE N/A ABO blood type 1 Anti-A and anti-B blood grouping reagents (human monoclonal antibodies) N/A Rh blood type 1 Anti-Rh (D) reagent for blood grouping (human monoclonal antibody) N/A Full blood countb 1 Automated haematology Analyzer (XT-4000i, Japan) N/A Haemoglobin(g/L) 1 Automated haematology Analyzer (XT-4000i, Japan) 1 HBsAg test 1 Diagnostic kit for hepatitis B surface antigen (ELISA) N/A Fasting blood glucose (FBG, mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Lipid testc (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Alanine transaminase (ALT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Glutamyl transferase (GGT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Alkaline phosphatase (ALP, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Aspartate transaminase (AST, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Creatinine (CrE, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Urea (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Uric acid (UA, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Total protein (TP, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Albumin (ALB, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 IgG (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgA (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgM (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Variables No. measurements Equipment used Precision Height (cm) 1 Manufactured instrument 0.1 Sitting height (cm) 1 Manufactured instrument 0.1 Weight (kg) 1 Body Composition Analyzer 0.1 Waist circumference (cm) 3 Standard tape measure 0.1 Hip circumference (cm) 3 Standard tape measure 0.1 Body composition 1 Body Composition Analyzer N/A Sitting blood pressure (mmHg) 3 Digital BP monitor 1 Resting pulse rate (beats/min) 3 Digital BP monitor 1 Axillary temperature (°C) 1 Digital medical thermometer 0.1 Visual acuity 1 LogMAR Chart 0.1 Respiratory functiona 1 Spirometer N/A Grip strength (kg) 2 JAMAR Kit 1 Oxygen saturation (%) 1 Masimo Red-57 1 Bone mineral density 1 Medilink OSTEOSPACE N/A ABO blood type 1 Anti-A and anti-B blood grouping reagents (human monoclonal antibodies) N/A Rh blood type 1 Anti-Rh (D) reagent for blood grouping (human monoclonal antibody) N/A Full blood countb 1 Automated haematology Analyzer (XT-4000i, Japan) N/A Haemoglobin(g/L) 1 Automated haematology Analyzer (XT-4000i, Japan) 1 HBsAg test 1 Diagnostic kit for hepatitis B surface antigen (ELISA) N/A Fasting blood glucose (FBG, mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Lipid testc (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Alanine transaminase (ALT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Glutamyl transferase (GGT, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Alkaline phosphatase (ALP, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Aspartate transaminase (AST, U/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Creatinine (CrE, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Urea (mmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 Uric acid (UA, μmol/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 1 Total protein (TP, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 Albumin (ALB, g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.1 IgG (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgA (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 IgM (g/L) 1 Chemistry Analyzer (ROCHE Cobas8000C701, USA) 0.01 N/A, not available. a Only conducted in non-smokers, with no cardiovascular disease, no significant self-reported pulmonary diseases nor acute respiratory illnesses. b Full blood count includes the numbers of white blood cells (WBC, 109/L), red blood cells (RBC, 1012/L) and platelets (109/L). c Lipid test includes triglyceride (TG, mmol/L), total cholesterol (TC, mmol/L), high-density lipoprotein cholesterol (HDL-C, mmol/L) and low-density lipoprotein cholesterol (LDL-C, mmol/L). Questionnaires were recovered immediately when participants finished their physical examination. Completeness and correctness of each questionnaire were examined by an epidemiologist by face-to-face re-check with the participant. We scanned the questionnaire using software developed by the programme, which allows real-time transcription from handwritten numbers to electronic format numbers into computer immediately after each day’s survey. Double-checks were implemented by two skilled staff separately, to verify accuracy and completeness of each questionnaire input. Overall, there were 990 staff involved in the survey, among whom 876 were from local health settings. All study participants provided written informed consent, and this study was approved by the Bioethical Committee of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, Beijing, China. Data resource use The database of CNHS as well as CMEB will be used for plenty of related research projects, ranging from epidemiological study to health promotion activities. They have generated more than 20 publications to date ( in both Chinese and English). The survey finished in October 2017, and the pooled data are still under analysis. Publications have been mainly on two topics: first, to investigate disease prevalence proportions and disease correlates among different ethnic groups. For example, several studies have been published focusing on the disparities in disease prevalence, or the varied effect of major risk factors (such as body composition). on noncommunicable disease.5,9–12 The second aim has been to generate clinical standards. For example, Zhang et al.8 used data from 3130 lifetime non-smokers involved in this survey, to develop spirometry reference values for the general population in China; before this, spirometry predictive equations had unfavuorable generalizability to Chinese populations. More importantly, we have adopted a ‘subject-centred’ strategy all through the survey. Considering that a large number of participants resided in rural areas, which made the feedback of hard-copy physical examination reports unpracticable, we developed a smart-phone based application (APP), where participants were able to find their physical and blood tests reports by inputting their unique identification (ID) number. They could also find a health conclusion according to their physical examination results (see Figure 3). Individuals attending the survey who were slightly outside the target age range (20–80) were not turned away, but their information was excluded when we performed data analysis. Figure 3 View largeDownload slide Screen shots of ‘Peking Union Health’ App developed by the CNHS programme. The App can be installed by both the IOS and the Android system. Electronic physical examination report is available for each participant by inputting the unique ID number. Figure 3 View largeDownload slide Screen shots of ‘Peking Union Health’ App developed by the CNHS programme. The App can be installed by both the IOS and the Android system. Electronic physical examination report is available for each participant by inputting the unique ID number. Additionally one of the study sites, Hebei Province, has been added to a large population-based cohort study initiated from 2016. Conducted by the same team from IBMS, CAMS and PUMC, this cohort study covers all the participants in Hebei Province who were enrolled in CNHS. Strengths and weaknesses The CNHS has several unique and special features which make it a valuable resource for research purposes. First, the CNHS is the only health examination and interview survey conducted at national level to obtain lifestyle and health information of different ethnic populations in China. China has a diversity of ethnic groups and cultures, as well as heterogeneity of socioeconomics in different regions. However health data, especially biochemical data, are limited in previous researches. By using multistage sampling methods, we collected representative samples on the national level, so this survey provides a comprehensive database for further study on risk factors of NCDs among different ethnic groups. Second, we collected more than 53 000 blood samples. The storage of both plasma and buffy coat samples enables assessment of the relevance of genetic and other molecular factors as correlates or determinants of various NCDs. Third, from 2012 to 2017 we used the same equipment and stringent and consistent criteria for measurements throughout the survey, which makes the results from different study sites comparable. The CNHS has serval limitations: first, the survey is a cross-sectional study, which may make it difficult to establish the temporal sequence between exposures and outcomes (temporality). Second, we did not collect nutritional intake information, which may play key role in NCDs onset. Minority ethnic groups have very different dietary habits and food diversities from Han people (the majority ethnic group in China), making quantitative measurement of nutrition intake impracticable. Third, owing to funding availability, we could only conduct the survey in 11 provinces, including 10 ethnic groups. However, the nine minority ethnic groups selected in our study contribute over 40% of the total minority population (there are in all 55 minority groups in China, some of which comprise less than 10 000 people) in China, which makes our data and biobank a valuable and representative source for research. Data resource access For the sake of programme management and data sharing, the data of CNHS have been linked to the National Scientific Data Sharing Platform for Population and Health, Biologic Medicine Information Center [http://www.bmicc.cn/web/share/home], on which researchers can find information from the CHNS, including introduction of the database, methods used in the survey, related publications etc. (see Figure 4). Individual-level data of CNHS are electronically available for scientific research if approved by the CNHS programme through official application procedures. An application form and instructions can be found in Chinese only, through [http://cnphd.bmicc.cn/chs/cn/index.php? intro=4]. Further information and enquiries can be submitted to the corresponding author, Guangliang Shan [[emailprotected]]. Figure 4 View largeDownload slide Database of CNHS provided on the China National Scientific Data Sharing Platform for Population and Health Biologic Medicine Information Center [http://www.bmicc.cn/web/share/home]. Figure 4 View largeDownload slide Database of CNHS provided on the China National Scientific Data Sharing Platform for Population and Health Biologic Medicine Information Center [http://www.bmicc.cn/web/share/home]. Funding This work was supported by the National Science and Technology Pillar Program during the Twelfth Five-Year Plan Period (Grant No. 2012BAI37B02), the Key Basic Research Program of the Ministry of Science and Technology of China (Grant No. 2013FY114100) and the National Natural Science Foundation of China (Grant No. 81273158). Acknowledgements We appreciate all the participants of CNHS, and gratefully acknowledge that all staff members from 11 provinces have given considerable time and energy to this survey. Conflict of interest: None declared. References 1 Gu H , Li W , Yang J , Wang Y , Bo J , Liu L. Hypertension prevalence, awareness, treatment and control among Han and four ethnic minorities (Uygur, Hui, Mongolian and Dai) in China . J Hum Hypertens 2015 ; 29 : 555 – 60 . Google Scholar Crossref Search ADS PubMed 2 Wang L , Gao P , Zhang M et al. Prevalence and ethnic pattern of diabetes and prediabetes in China in 2013 . JAMA 2017 ; 317 : 2515 – 23 . Google Scholar Crossref Search ADS PubMed 3 Pan L , Dong F , Wang K et al. Towards better clinical laboratory diagnostic criteria for haemoglobin concentration: results from the Survey on Health Examination and Physiological Constants in China . Lancet 2015 ; 386 : S59. Google Scholar Crossref Search ADS 4 Yang W , Lu J , Weng J et al. Prevalence of diabetes among men and women in China . N Engl J Med 2010 ; 362 : 1090 – 101 . Google Scholar Crossref Search ADS PubMed 5 Li G , Guo G , Wang W et al. Association of prehypertension and cardiovascular risk factor clustering in Inner Mongolia: a cross-sectional study . BMJ Open 2017 ; 7 : e015340 . Google Scholar Crossref Search ADS PubMed 6 Miller MR , Hankinson J , Brusasco V et al.; ATS/ERS Task Force . Standardisation of spirometry . Eur Respir J 2005 ; 26 : 319 – 38 . Google Scholar Crossref Search ADS PubMed 7 National Physique and Health Database. General Design and Methodology of the 2012–2015 Data Survey, 2.0 Lung Function. 2012 . http://cnphd.bmicc.cn/chs/en (January 2018, date last accessed). 8 Zhang J , Hu X , Shan G. Spirometry reference values for population aged 7-80 years in China . Respirology 2017 ; 22 : 1630 –3 6 . Google Scholar Crossref Search ADS PubMed 9 Dong F , Wang D , Pan L et al. Disparities in hypertension prevalence, awareness, treatment and control between Bouyei and Han: results from a bi-ethnic health survey in developing regions from South China . Int J Environ Res Public Health 2016 ; 13 : 233. Google Scholar Crossref Search ADS PubMed 10 Zhu H , Xu Y , Gong F et al. Reference ranges for serum insulin-like growth factor I (IGF-I) in healthy Chinese adults . PLoS One 2017 ; 12 : e0185561. Google Scholar Crossref Search ADS PubMed 11 Gong H , Pa L , Wang K et al. Prevalence of diabetes and associated factors in the Uyghur and Han population in Xinjiang, China . Int J Environ Res Public Health 2015 ; 12 : 12792 – 802 . Google Scholar Crossref Search ADS PubMed 12 Li Y , Wang H , Wang K et al. Optimal body fat percentage cut-off values for identifying cardiovascular risk factors in Mongolian and Han adults: a population-based cross-sectional study in Inner Mongolia, China . BMJ Open 2017 ; 7 : e014675 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Cohort profile: The Childhood Asthma Prevention Study (CAPS)

Garden, Frances, L;Toelle, Brett, G;Mihrshahi,, Seema;Webb, Karen, L;Almqvist,, Catarina;Tovey, Euan, R;Brew, Bronwyn, K;Ayer, Julian, G;Skilton, Michael, R;Jones,, Graham;Ferreira, Manuel A, R;Cowie, Christine, T;Weber-Chrysochoou,, Christina;Britton, Warwick, J;Celermajer, David, S;Leeder, Stephen, R;Peat, Jennifer, K;Marks, Guy, B

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy078pmid: 29800224

Why was the cohort set up? The Childhood Asthma Prevention Study (CAPS) commenced in 1997 in Sydney, Australia, because of concern about the high and increasing prevalence of childhood asthma.1,2 Cross-sectional and ecological studies had shown that exposure to high concentrations of house dust mite (HDM) allergen and being sensitized to HDM were both associated with increased prevalence.3–6 Other studies had indicated that children who regularly consumed oily fish containing high levels of omega-3 fatty acids were less likely to have airway hyper-responsiveness (AHR) and asthma.7 Those who regularly consumed oils and spreads containing polyunsaturated fats with a higher proportion of omega-6 fatty acids had an increased prevalence of asthma-like symptoms.8 The CAPS investigators determined that a randomized controlled trial was required, both to test the causal hypotheses about these environmental and dietary risk factors, and to evaluate the effectiveness of omega-3 supplementation. We decided to test the hypothesis that HDM allergen avoidance and omega-3 supplementation, from birth to 5 years of age in high-risk children, would prevent asthma and other manifestations of allergic illness during the first 5 years of life.9 CAPS began as a randomized controlled trial (RCT), using a factorial design to test the combined and separate effects of HDM avoidance and omega-3 supplement intervention. Details of the study design and interventions were described in 2001.10 A secondary aim was to establish a birth cohort of high-risk children to examine the association, over time, between a range of putative risk factors and the incidence of asthma. Data from the first 5 years of CAPS demonstrated that the study had been successfully established and implemented.9,11–13 Based on these initial results and other studies which emphasized the importance of longer-term follow up of trials in primary prevention of allergic disease,14 follow-up of the cohort was extended beyond 5 years. The participants were re-evaluated at age 8, 3 years after cessation of the intervention, to assess the longer-term effectiveness of the interventions.15 The age of 8 was chosen as this is the age at which important childhood predictors of adult asthma, including atopy, AHR and obstructive spirometric function, can be reliably measured.16,17 Additionally, at age 8, members of the cohort were invited to participate in a subsidiary study examining the childhood determinants of early manifestations of cardiovascular disease. This was initiated because we had successfully acquired information on early life exposures and risk factors relevant to cardiovascular health, including perinatal and postnatal growth, parental smoking, infant and early life nutrition, and socioeconomic data.18,19 The study was further extended through puberty and adolescence (11.5 to 14 years).20 The aims of this period were to examine the relation between puberty and sex-specific changes in respiratory symptoms, lung function, AHR and airway inflammation, and to study the effect of early life and concurrent exposure to environmental risk factors on this relationship. Adolescence is a crucial developmental period during which substantial change is found, with differing prevalences of asthma in males and females.21,22 The CAPS study is based at the Woolcock Institute of Medical Research, University of Sydney, Australia. Who is in the cohort? CAPS was initially designed as a randomized controlled trial using a factorial design to test the combined and separate effects of HDM avoidance and omega-3 supplementation.10 Between September 1997 and November 1999, pregnant women whose unborn children were at high risk of developing asthma, because a parent or a sibling had a current diagnosis of asthma or wheezed frequently, were recruited from the antenatal clinics of six hospitals in Sydney. The selection criteria were: at least one parent or sibling with symptoms of asthma, assessed by screening questionnaire; reasonable fluency in English; a telephone at home; and residence within 30 km of the recruitment centre. Exclusion criteria were: a pet cat at home; the family being on a strict vegetarian diet; a multiple birth; and delivery earlier than 36 weeks’ gestation. Details of the recruitment process were published in 2002.11 Of 7171 pregnant women screened, 2095 (29%) were eligible for inclusion. Of these, 616 (29% of those eligible and 9% of those initially screened) were enrolled (Figure 1). The study was powered to detect a 15% absolute reduction in the prevalence of asthma between active and control groups.10 A survey of 200 eligible non-participants revealed that participating parents had higher levels of tertiary education than non-participants. They did not, however, differ in age, country of birth (Australia versus other), full-time employment or primigravida status.11 Figure 1 View largeDownload slide Flowchart for CAPS to age 14 years. Participants at each period were those who completed at least a questionnaire at the major clinical assessment. Participants were considered withdrawn if they had formally withdrawn from the study at or before the assessment. Figure 1 View largeDownload slide Flowchart for CAPS to age 14 years. Participants at each period were those who completed at least a questionnaire at the major clinical assessment. Participants were considered withdrawn if they had formally withdrawn from the study at or before the assessment. How often have they been followed up? Participants have been assessed on 42 occasions between 36 weeks of gestation and age 14 years. Assessments were performed on the mother at 36 weeks of gestation and on the child at ages 1, 3, 6, 9, and 12 months, every 3 months until aged 5 years, every 6 months until aged 7.5, at ages 8, 9 and 11, and then every 3 months until age 14. A detailed schedule of the data collection times and instruments used is shown in Table 1. During the first 5 years, the study team performed home visits at 36 weeks gestation, then at 1 month after birth, at 3 months, then every 3 months until 12 months, then every 6 months until age 5. A series of interviewer-administered questionnaires were conducted with the participant’s parents or guardians. In addition, anthropometric measurements were performed and a home environmental assessment, including dust collection, made. Telephoned interview questionnaires were administered to parents between the 6-monthly home visits from ages 1 to 5. Clinical examinations were performed by study nurses, blinded to treatment group allocation, at ages 1.5, 3, and 5 years at one of two Sydney hospitals (Westmead and Liverpool). Table 1. CAPS questionnaire and measurement data collection schedule Months (m) Years (y) 1 m before birth 1 m 3, 6, 9, 12 m 1.5 y 2, 2.5 y 3 y 3.5–4.5a y 5 y 5.5–7.5a y 8 y 9 y 11, 11.25 y 11.5 y 11.75–13.75b y 14 y >14.25b y Questionnaire Home environment X X X X X X X X X X X Family history, pregnancy and perinatal X Symptoms and illness X X X X X X X X Diet X X X X X X X Clinical X X X X X X Ethnicity (4.5 years only) X Puberty (annually) X (11 y) X (12, 13 y) X X(14, 15, 16 y) Measurement Dust collection X X X X X X X X X Anthropometric X X X X X X X X X X X X X Dietary intake X X X Physical examination X X X X X X Blood collection X X X X X X Skin prick test X X X X X X Forced oscillation technique X X X X X Spirometry X X X X Methacholine challenge X X X Exhaled nitric oxide X X X Cardiovascular X X Urine X X*(12.5 y) X DNA X X Months (m) Years (y) 1 m before birth 1 m 3, 6, 9, 12 m 1.5 y 2, 2.5 y 3 y 3.5–4.5a y 5 y 5.5–7.5a y 8 y 9 y 11, 11.25 y 11.5 y 11.75–13.75b y 14 y >14.25b y Questionnaire Home environment X X X X X X X X X X X Family history, pregnancy and perinatal X Symptoms and illness X X X X X X X X Diet X X X X X X X Clinical X X X X X X Ethnicity (4.5 years only) X Puberty (annually) X (11 y) X (12, 13 y) X X(14, 15, 16 y) Measurement Dust collection X X X X X X X X X Anthropometric X X X X X X X X X X X X X Dietary intake X X X Physical examination X X X X X X Blood collection X X X X X X Skin prick test X X X X X X Forced oscillation technique X X X X X Spirometry X X X X Methacholine challenge X X X Exhaled nitric oxide X X X Cardiovascular X X Urine X X*(12.5 y) X DNA X X a 6-monthly measurements. b Quarterly measurements. Table 1. CAPS questionnaire and measurement data collection schedule Months (m) Years (y) 1 m before birth 1 m 3, 6, 9, 12 m 1.5 y 2, 2.5 y 3 y 3.5–4.5a y 5 y 5.5–7.5a y 8 y 9 y 11, 11.25 y 11.5 y 11.75–13.75b y 14 y >14.25b y Questionnaire Home environment X X X X X X X X X X X Family history, pregnancy and perinatal X Symptoms and illness X X X X X X X X Diet X X X X X X X Clinical X X X X X X Ethnicity (4.5 years only) X Puberty (annually) X (11 y) X (12, 13 y) X X(14, 15, 16 y) Measurement Dust collection X X X X X X X X X Anthropometric X X X X X X X X X X X X X Dietary intake X X X Physical examination X X X X X X Blood collection X X X X X X Skin prick test X X X X X X Forced oscillation technique X X X X X Spirometry X X X X Methacholine challenge X X X Exhaled nitric oxide X X X Cardiovascular X X Urine X X*(12.5 y) X DNA X X Months (m) Years (y) 1 m before birth 1 m 3, 6, 9, 12 m 1.5 y 2, 2.5 y 3 y 3.5–4.5a y 5 y 5.5–7.5a y 8 y 9 y 11, 11.25 y 11.5 y 11.75–13.75b y 14 y >14.25b y Questionnaire Home environment X X X X X X X X X X X Family history, pregnancy and perinatal X Symptoms and illness X X X X X X X X Diet X X X X X X X Clinical X X X X X X Ethnicity (4.5 years only) X Puberty (annually) X (11 y) X (12, 13 y) X X(14, 15, 16 y) Measurement Dust collection X X X X X X X X X Anthropometric X X X X X X X X X X X X X Dietary intake X X X Physical examination X X X X X X Blood collection X X X X X X Skin prick test X X X X X X Forced oscillation technique X X X X X Spirometry X X X X Methacholine challenge X X X Exhaled nitric oxide X X X Cardiovascular X X Urine X X*(12.5 y) X DNA X X a 6-monthly measurements. b Quarterly measurements. Regular phone calls were made at 6-week intervals to promote adherence to the protocol and to ensure an adequate supply of the goods used in the interventions. Using a 3-day weighed food record and a food frequency questionnaire, respectively, the children’s dietary intake was measured at 18 months and 3 years. Between 18 months and 3 years, whole blood was collected from some parents and most participants for DNA extraction and analysis. After the cessation of the intervention at age 5, telephoned interviewer-administered questionnaires were conducted every 6 months from ages 5 to 8. At age 8, another blinded clinical assessment was performed at hospital and a home visit was conducted. Both assessments were similar to those conducted up to age 5 and, for the first time, included AHR. At around age 9, another dietary intake measurement was taken via a telephoned interviewer-administered 24-h recall questionnaire.23 From age 11 onwards, participants were contacted every 3 months to provide information about puberty and growth. At 3-monthly intervals from around the child’s 11th birthday, the parents were contacted by phone or short-message service (SMS) to measure the child’s height, using a provided wall-mounted stadiometer, with the results sent back via a web-based data collection tool, SMS or phone. Annually, from around age 11, the children were asked to complete a questionnaire to assess pubertal stage. These included Tanner pubertal stages diagrams24–26 and the Pubertal Development Scale.27,28 These were administered as a paper questionnaire, either mailed or via a web-based data collection system. This was the first age at which participants were asked to self-complete a questionnaire. At ages 11.5 and 14, another clinical assessment was performed at the hospitals and various interviewer-administered questionnaires were asked of the parents. Throughout the follow-up period, where a participant was unable to attend a clinical assessment (despite attempts to re-schedule), questionnaires were administered by telephone. If at any time a participant declined to participate but did not formally withdraw, they were re-contacted to participate in the next scheduled assessment. Participants could withdraw at any time and, if so, contact was ceased. During the clinical assessments, not all participants were able to perform all procedures on the day of testing. If so, they were invited to repeat the measurement at a future date. Additionally, not all participants were willing, during clinical assessments, to have blood collected or skin prick tests performed. The number of participants who completed tests is shown in Table 2. Hence, the total number of participants completing questionnaires at the clinical assessment is greater than the number who provided other clinical measurements at that assessment. Table 2. The number of participants in CAPS at the major data collection times who were enrolled in the study, those withdrawn and those who completed questionnaires and the other major measurements Collection time (years) 1.5 3 5 8 11.5 14 n n n n n n Enrolled in study 552 530 518 492 463 436 Withdrawn from studya 64 22 12 26 29 27 Completed:  Questionnaires 550 530 516 450 370 352  Anthropometric measures 536 516 468 449 292 196  Blood tests 374 409 396 316 257 178  Skin prick tests 535 522 488 402 292 195  Dietary intake measures 424 456 222b  Spirometry 381 418 283 190  Methacholine challenge 357 269 179  Exhaled nitric oxide 397 290 191  Cardiovascular assessment 405 193  Urine sample 277 183 Collection time (years) 1.5 3 5 8 11.5 14 n n n n n n Enrolled in study 552 530 518 492 463 436 Withdrawn from studya 64 22 12 26 29 27 Completed:  Questionnaires 550 530 516 450 370 352  Anthropometric measures 536 516 468 449 292 196  Blood tests 374 409 396 316 257 178  Skin prick tests 535 522 488 402 292 195  Dietary intake measures 424 456 222b  Spirometry 381 418 283 190  Methacholine challenge 357 269 179  Exhaled nitric oxide 397 290 191  Cardiovascular assessment 405 193  Urine sample 277 183 a The number withdrawn from the study is the number of participants who withdrew before or at the assessment period. b Dietary intake was measured at around 9 years of age. Table 2. The number of participants in CAPS at the major data collection times who were enrolled in the study, those withdrawn and those who completed questionnaires and the other major measurements Collection time (years) 1.5 3 5 8 11.5 14 n n n n n n Enrolled in study 552 530 518 492 463 436 Withdrawn from studya 64 22 12 26 29 27 Completed:  Questionnaires 550 530 516 450 370 352  Anthropometric measures 536 516 468 449 292 196  Blood tests 374 409 396 316 257 178  Skin prick tests 535 522 488 402 292 195  Dietary intake measures 424 456 222b  Spirometry 381 418 283 190  Methacholine challenge 357 269 179  Exhaled nitric oxide 397 290 191  Cardiovascular assessment 405 193  Urine sample 277 183 Collection time (years) 1.5 3 5 8 11.5 14 n n n n n n Enrolled in study 552 530 518 492 463 436 Withdrawn from studya 64 22 12 26 29 27 Completed:  Questionnaires 550 530 516 450 370 352  Anthropometric measures 536 516 468 449 292 196  Blood tests 374 409 396 316 257 178  Skin prick tests 535 522 488 402 292 195  Dietary intake measures 424 456 222b  Spirometry 381 418 283 190  Methacholine challenge 357 269 179  Exhaled nitric oxide 397 290 191  Cardiovascular assessment 405 193  Urine sample 277 183 a The number withdrawn from the study is the number of participants who withdrew before or at the assessment period. b Dietary intake was measured at around 9 years of age. Loss to follow-up Of the 616 participants recruited at birth, the number participating in the major clinical assessments, as determined by completion of the clinical questionnaire, were: 550/616 (89%) at age 1.5 years, 530/616 (86%) at 3, 516/616 (84%) at 5, 450/616 (73%) at 8, 370/616 (60%) at 11.5 and 352/616 (57%) at 14 (Table 2). The loss to follow-up was minimal in the first 5 years (n  =  100), with the greatest loss occurring in the first 12 to 18 months. Common reasons for early withdrawal were that the participants had moved residence and did not leave any forwarding address or telephone number, had moved out of the study area or were withdrawn for medical reasons.11 The number of withdrawals was similar in each of the randomized groups (see Figure 1). Differences between those who participated at the major clinical assessments at ages 5, 8, 11.5 and 14 years, and those who did not, are described in Table 3. The results show that, compared with non-responders at these assessments, respondent mothers were older, more highly educated, more likely to be in full-time employment, more likely to have breastfed for 6 or more months and less likely to have smoked during pregnancy. Respondent fathers were also older, more highly educated and more likely to be in full-time employment than non-respondent fathers. Table 3. Comparison of participants who participated and those who did not participate in the clinical assessments of CAPS at ages 5, 8, 11.5 and 14 years Participated in the major clinical assessment Participants Original 5 years 8 years 11.5 years 14 years No Yes No Yes No Yes No Yes n  =  616 n  =  100 n  =  516 n  =  166 n  =  450 n  =  246 n  =  370 n  =  264 n  =  352 n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) Child characteristics Gender   Male 312 (51%) 55 (55%) 257 (50%) 84 (51%) 228 (51%) 125 (51%) 187 (51%) 126 (48%) 186 (53%)   Female 304 (49%) 45 (45%) 259 (50%) 82 (49%) 222 (49%) 121 (49%) 183 (49%) 138 (52%) 166 (47%) HDM intervention group   Control 309 (50%) 49 (49%) 260 (50%) 79 (48%) 230 (51%) 128 (52%) 181 (49%) 131 (50%) 178 (51%)   Active 307 (50%) 51 (51%) 256 (50%) 87 (52%) 220 (49%) 118 (48%) 189 (51%) 133 (50%) 174 (49%) Diet intervention group   Control 303 (49%) 54 (54%) 249 (48%) 83 (50%) 220 (49%) 120 (49%) 183 (49%) 134 (51%) 169 (48%)   Active 313 (51%) 46 (46%) 267 (52%) 83 (50%) 230 (51%) 126 (51%) 187 (51%) 130 (49%) 183 (52%) Breastfeeding ≥ 6 months 227 (39%) 15*(23%) 212*(41%) 41*(32%) 186*(41%) 67*(32%) 160*(43%) 70*(31%) 157*(45%) Child has older siblings 422 (69%) 70 (70%) 352 (68%) 116 (70%) 306 (68%) 179 (73%) 243 (66%) 193*(73%) 229*(65%) Parent characteristics at child’s birth Age (years) (mean (± SD))   Mother 28.4 (5.3) 26.2*(5.6) 28.9*(5.2) 27.0*(5.6) 29.0*(5.1) 27.6*(5.6) 29.0*(5.1) 27.9*(5.7) 28.8*(5.0)   Father 30.8 (6.1) 28.8*(6.8) 31.1*(5.9) 29.5*(6.4) 31.2*(5.9) 30.2*(6.3) 31.2*(5.9) 30.4 (6.2) 31.0 (6.0) Australian born   Mother 457 (74%) 82 (82%) 375 (73%) 127 (77%) 330 (73%) 185 (75%) 272 (74%) 192 (73%) 265 (75%)   Father 421 (69%) 69 (70%) 352 (69%) 111 (67%) 310 (69%) 166 (68%) 255 (69%) 172 (65%) 249 (71%) Tertiary educated   Mother 276 (45%) 32*(32%) 244*(47%) 53*(32%) 223*(50%) 85*(35%) 191*(52%) 92*(35%) 184*(52%)   Father 265 (44%) 34 (35%) 231 (45%) 54*(33%) 211*(47%) 90*(37%) 175*(48%) 98*(38%) 167*(48%) Full-time employment   Mother 278 (45%) 43 (43%) 235 (46%) 71 (43%) 207 (46%) 95*(39%) 183*(50%) 104*(39%) 174*(49%)   Father 518 (84%) 77 (78%) 441 (86%) 132 (80%) 386 (86%) 195*(80%) 323*(87%) 207*(79%) 311*(88%) Mother smoked during pregnancy 150 (24%) 31 (31%) 119 (23%) 44 (27%) 106 (24%) 72*(29%) 78*(21%) 76*(29%) 74*(21%) Primigravida 199 (33%) 28 (30%) 171 (33%) 51 (32%) 148 (33%) 68 (28%) 131 (35%) 72*(28%) 127*(36%) Participated in the major clinical assessment Participants Original 5 years 8 years 11.5 years 14 years No Yes No Yes No Yes No Yes n  =  616 n  =  100 n  =  516 n  =  166 n  =  450 n  =  246 n  =  370 n  =  264 n  =  352 n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) Child characteristics Gender   Male 312 (51%) 55 (55%) 257 (50%) 84 (51%) 228 (51%) 125 (51%) 187 (51%) 126 (48%) 186 (53%)   Female 304 (49%) 45 (45%) 259 (50%) 82 (49%) 222 (49%) 121 (49%) 183 (49%) 138 (52%) 166 (47%) HDM intervention group   Control 309 (50%) 49 (49%) 260 (50%) 79 (48%) 230 (51%) 128 (52%) 181 (49%) 131 (50%) 178 (51%)   Active 307 (50%) 51 (51%) 256 (50%) 87 (52%) 220 (49%) 118 (48%) 189 (51%) 133 (50%) 174 (49%) Diet intervention group   Control 303 (49%) 54 (54%) 249 (48%) 83 (50%) 220 (49%) 120 (49%) 183 (49%) 134 (51%) 169 (48%)   Active 313 (51%) 46 (46%) 267 (52%) 83 (50%) 230 (51%) 126 (51%) 187 (51%) 130 (49%) 183 (52%) Breastfeeding ≥ 6 months 227 (39%) 15*(23%) 212*(41%) 41*(32%) 186*(41%) 67*(32%) 160*(43%) 70*(31%) 157*(45%) Child has older siblings 422 (69%) 70 (70%) 352 (68%) 116 (70%) 306 (68%) 179 (73%) 243 (66%) 193*(73%) 229*(65%) Parent characteristics at child’s birth Age (years) (mean (± SD))   Mother 28.4 (5.3) 26.2*(5.6) 28.9*(5.2) 27.0*(5.6) 29.0*(5.1) 27.6*(5.6) 29.0*(5.1) 27.9*(5.7) 28.8*(5.0)   Father 30.8 (6.1) 28.8*(6.8) 31.1*(5.9) 29.5*(6.4) 31.2*(5.9) 30.2*(6.3) 31.2*(5.9) 30.4 (6.2) 31.0 (6.0) Australian born   Mother 457 (74%) 82 (82%) 375 (73%) 127 (77%) 330 (73%) 185 (75%) 272 (74%) 192 (73%) 265 (75%)   Father 421 (69%) 69 (70%) 352 (69%) 111 (67%) 310 (69%) 166 (68%) 255 (69%) 172 (65%) 249 (71%) Tertiary educated   Mother 276 (45%) 32*(32%) 244*(47%) 53*(32%) 223*(50%) 85*(35%) 191*(52%) 92*(35%) 184*(52%)   Father 265 (44%) 34 (35%) 231 (45%) 54*(33%) 211*(47%) 90*(37%) 175*(48%) 98*(38%) 167*(48%) Full-time employment   Mother 278 (45%) 43 (43%) 235 (46%) 71 (43%) 207 (46%) 95*(39%) 183*(50%) 104*(39%) 174*(49%)   Father 518 (84%) 77 (78%) 441 (86%) 132 (80%) 386 (86%) 195*(80%) 323*(87%) 207*(79%) 311*(88%) Mother smoked during pregnancy 150 (24%) 31 (31%) 119 (23%) 44 (27%) 106 (24%) 72*(29%) 78*(21%) 76*(29%) 74*(21%) Primigravida 199 (33%) 28 (30%) 171 (33%) 51 (32%) 148 (33%) 68 (28%) 131 (35%) 72*(28%) 127*(36%) * Indicates a significant difference (emboldened) (P < 0.05) between those who participated and those who did not, based on a chi-square test for categorical variables or a t-test for age. Table 3. Comparison of participants who participated and those who did not participate in the clinical assessments of CAPS at ages 5, 8, 11.5 and 14 years Participated in the major clinical assessment Participants Original 5 years 8 years 11.5 years 14 years No Yes No Yes No Yes No Yes n  =  616 n  =  100 n  =  516 n  =  166 n  =  450 n  =  246 n  =  370 n  =  264 n  =  352 n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) Child characteristics Gender   Male 312 (51%) 55 (55%) 257 (50%) 84 (51%) 228 (51%) 125 (51%) 187 (51%) 126 (48%) 186 (53%)   Female 304 (49%) 45 (45%) 259 (50%) 82 (49%) 222 (49%) 121 (49%) 183 (49%) 138 (52%) 166 (47%) HDM intervention group   Control 309 (50%) 49 (49%) 260 (50%) 79 (48%) 230 (51%) 128 (52%) 181 (49%) 131 (50%) 178 (51%)   Active 307 (50%) 51 (51%) 256 (50%) 87 (52%) 220 (49%) 118 (48%) 189 (51%) 133 (50%) 174 (49%) Diet intervention group   Control 303 (49%) 54 (54%) 249 (48%) 83 (50%) 220 (49%) 120 (49%) 183 (49%) 134 (51%) 169 (48%)   Active 313 (51%) 46 (46%) 267 (52%) 83 (50%) 230 (51%) 126 (51%) 187 (51%) 130 (49%) 183 (52%) Breastfeeding ≥ 6 months 227 (39%) 15*(23%) 212*(41%) 41*(32%) 186*(41%) 67*(32%) 160*(43%) 70*(31%) 157*(45%) Child has older siblings 422 (69%) 70 (70%) 352 (68%) 116 (70%) 306 (68%) 179 (73%) 243 (66%) 193*(73%) 229*(65%) Parent characteristics at child’s birth Age (years) (mean (± SD))   Mother 28.4 (5.3) 26.2*(5.6) 28.9*(5.2) 27.0*(5.6) 29.0*(5.1) 27.6*(5.6) 29.0*(5.1) 27.9*(5.7) 28.8*(5.0)   Father 30.8 (6.1) 28.8*(6.8) 31.1*(5.9) 29.5*(6.4) 31.2*(5.9) 30.2*(6.3) 31.2*(5.9) 30.4 (6.2) 31.0 (6.0) Australian born   Mother 457 (74%) 82 (82%) 375 (73%) 127 (77%) 330 (73%) 185 (75%) 272 (74%) 192 (73%) 265 (75%)   Father 421 (69%) 69 (70%) 352 (69%) 111 (67%) 310 (69%) 166 (68%) 255 (69%) 172 (65%) 249 (71%) Tertiary educated   Mother 276 (45%) 32*(32%) 244*(47%) 53*(32%) 223*(50%) 85*(35%) 191*(52%) 92*(35%) 184*(52%)   Father 265 (44%) 34 (35%) 231 (45%) 54*(33%) 211*(47%) 90*(37%) 175*(48%) 98*(38%) 167*(48%) Full-time employment   Mother 278 (45%) 43 (43%) 235 (46%) 71 (43%) 207 (46%) 95*(39%) 183*(50%) 104*(39%) 174*(49%)   Father 518 (84%) 77 (78%) 441 (86%) 132 (80%) 386 (86%) 195*(80%) 323*(87%) 207*(79%) 311*(88%) Mother smoked during pregnancy 150 (24%) 31 (31%) 119 (23%) 44 (27%) 106 (24%) 72*(29%) 78*(21%) 76*(29%) 74*(21%) Primigravida 199 (33%) 28 (30%) 171 (33%) 51 (32%) 148 (33%) 68 (28%) 131 (35%) 72*(28%) 127*(36%) Participated in the major clinical assessment Participants Original 5 years 8 years 11.5 years 14 years No Yes No Yes No Yes No Yes n  =  616 n  =  100 n  =  516 n  =  166 n  =  450 n  =  246 n  =  370 n  =  264 n  =  352 n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) Child characteristics Gender   Male 312 (51%) 55 (55%) 257 (50%) 84 (51%) 228 (51%) 125 (51%) 187 (51%) 126 (48%) 186 (53%)   Female 304 (49%) 45 (45%) 259 (50%) 82 (49%) 222 (49%) 121 (49%) 183 (49%) 138 (52%) 166 (47%) HDM intervention group   Control 309 (50%) 49 (49%) 260 (50%) 79 (48%) 230 (51%) 128 (52%) 181 (49%) 131 (50%) 178 (51%)   Active 307 (50%) 51 (51%) 256 (50%) 87 (52%) 220 (49%) 118 (48%) 189 (51%) 133 (50%) 174 (49%) Diet intervention group   Control 303 (49%) 54 (54%) 249 (48%) 83 (50%) 220 (49%) 120 (49%) 183 (49%) 134 (51%) 169 (48%)   Active 313 (51%) 46 (46%) 267 (52%) 83 (50%) 230 (51%) 126 (51%) 187 (51%) 130 (49%) 183 (52%) Breastfeeding ≥ 6 months 227 (39%) 15*(23%) 212*(41%) 41*(32%) 186*(41%) 67*(32%) 160*(43%) 70*(31%) 157*(45%) Child has older siblings 422 (69%) 70 (70%) 352 (68%) 116 (70%) 306 (68%) 179 (73%) 243 (66%) 193*(73%) 229*(65%) Parent characteristics at child’s birth Age (years) (mean (± SD))   Mother 28.4 (5.3) 26.2*(5.6) 28.9*(5.2) 27.0*(5.6) 29.0*(5.1) 27.6*(5.6) 29.0*(5.1) 27.9*(5.7) 28.8*(5.0)   Father 30.8 (6.1) 28.8*(6.8) 31.1*(5.9) 29.5*(6.4) 31.2*(5.9) 30.2*(6.3) 31.2*(5.9) 30.4 (6.2) 31.0 (6.0) Australian born   Mother 457 (74%) 82 (82%) 375 (73%) 127 (77%) 330 (73%) 185 (75%) 272 (74%) 192 (73%) 265 (75%)   Father 421 (69%) 69 (70%) 352 (69%) 111 (67%) 310 (69%) 166 (68%) 255 (69%) 172 (65%) 249 (71%) Tertiary educated   Mother 276 (45%) 32*(32%) 244*(47%) 53*(32%) 223*(50%) 85*(35%) 191*(52%) 92*(35%) 184*(52%)   Father 265 (44%) 34 (35%) 231 (45%) 54*(33%) 211*(47%) 90*(37%) 175*(48%) 98*(38%) 167*(48%) Full-time employment   Mother 278 (45%) 43 (43%) 235 (46%) 71 (43%) 207 (46%) 95*(39%) 183*(50%) 104*(39%) 174*(49%)   Father 518 (84%) 77 (78%) 441 (86%) 132 (80%) 386 (86%) 195*(80%) 323*(87%) 207*(79%) 311*(88%) Mother smoked during pregnancy 150 (24%) 31 (31%) 119 (23%) 44 (27%) 106 (24%) 72*(29%) 78*(21%) 76*(29%) 74*(21%) Primigravida 199 (33%) 28 (30%) 171 (33%) 51 (32%) 148 (33%) 68 (28%) 131 (35%) 72*(28%) 127*(36%) * Indicates a significant difference (emboldened) (P < 0.05) between those who participated and those who did not, based on a chi-square test for categorical variables or a t-test for age. What has been measured? The administered questionnaires collected information on family characteristics, pregnancy and perinatal details, the indoor home environment, diet, symptoms, illnesses, health care use, vaccinations, medication use and puberty stages as described in Table 4. Other measurements included house dust mite allergen concentrations in the bed and/or other sites at home, anthropometric measures, dietary intake, physical examination for wheeze and eczema, allergen skin prick tests, spirometric lung function, methacholine challenge tests, forced oscillometry, exhaled nitric oxide (FENO) and blood tests for total and specific IgE, lipids, inflammatory markers, sex hormones and DNA.15,20 Both targeted gene and genome-wide analyses have been conducted on subsets of the cohort.29–33 In addition, when the participants were aged 14, telomere length was estimated on these specimens and further specimens were collected.34 Blood pressure, carotid ultrasound, pulse-wave velocity, and pulse-wave analysis were also conducted,18,19 as described in Table 5. We assessed the cytokine (interleukin-5, IL-13, IL-10 and gamma-interferon) concentration in the supernatant of peripheral blood mononuclear cells (PBMCs) collected at ages 18 months and 3, 5 and 8 years and stimulated in vitro with HDM extract, an indicator of specific Th2-like and Th1-like responsiveness.35 All measurements and assessments were performed by the study team except the cardiovascular measurements taken at ages 8 and 14, which were performed by cardiovascular researchers. When the children were aged 13–15, data linkage between CAPS data and academic performance data from the Australian National Assessment Program Literacy and Numeracy (NAPLAN) test was performed.36 Table 4. Details of the main CAPS questionnaires Questionnaire Information collected Home environment Housing details (house type, age, building material, building foundations), number of home occupants, cooking power source, visible mould, pet ownership, child’s bedroom details (temperature, humidity, number of occupants, heating source, cooling source, flooring type, rugs or mats, visible mould), child’s bed details (type and age of bed, blanket, pillow, mattress, cover), exposure to tobacco smoke Family history, pregnancy and perinatal data questionnaire Mother and father (age, date of birth, country of birth, indigenous status, highest level of education, employment status, history of asthma, eczema, or hayfever) Pregnancy information (asthma diagnosed during pregnancy, medication use, vitamin/supplement use, smoking status, foods avoided during pregnancy, gestational diabetes, pre-eclampsia, hypertension) Perinatal information (gravidity, parity, gestational age, labour complications, cord blood taken, Apgar score, resuscitation required, admission to neonatal intensive care or special care nursery, time of birth, birthweight, birth length, head circumference, meconium aspiration, hyaline membrane disease, other neonatal complications) Symptoms and illness questionnaire Symptoms (sleep disturbed by coughing, wheeze, itchy rash, runny nose, flexural dermatitis), doctor-diagnosed (eczema, allergic rhinitis/hay fever, pneumonia, whooping cough, bronchiolitis, bronchitis, croup, asthma), significant medical or surgical problems, immunizations given, antibiotic use Diet Details of breastfeeding, use of infant formula, use of cow’s milk or other milk substitutes and introduction of solid food; vitamin/dietary supplement use and type; consumption of milk and solid foods (asked of mothers if breastfeeding and of children if started solid foods); use of study capsules, spreads and oils Clinical Details and history of symptoms: cough (ever or past 12/18 months, longest episode, episode lasted a week or more, during sleep, during physical activity, without a cold), wheeze (ever or past 12/18 months, episode for a week or more, longest episode, without a cold, caused difficulty breathing, health care use for wheeze, during sleep, during physical activity), rhinitis (ever or previous 12/18 months, episode for a week or more, longest episode, frequency of episode), eczema (itchy rash ever or previous 12 months); food allergy (asked up to 3 years: status and type of reaction); food avoidance (asked up to 3 years: type and on whose advice); doctor diagnosis and visited a GP, specialist, emergency department or hospital admission for eczema, allergic rhinitis, pneumonia, bronchiolitis, whooping cough, bronchitis, cough, asthma (from 8 years: diabetes or heart problems); medication (use, type, duration and frequency); snoring (at 5 years: ever and frequency; from 8 years: ever, frequency, loudness, stop breathing, struggle breathing during sleep, fall asleep at school, while watching television or during the daytime); television viewing (asked from 8 years: days per week, hours per day on weekday and weekend); parental health (parent or grandparent experienced a heart attack or stroke); childcare attendance and type (asked up to 5 years) Ethnicity Child’s maternal and paternal grandparents’ country of birth Puberty Tanner stages, puberty development scale, date of menarche (girls) Questionnaire Information collected Home environment Housing details (house type, age, building material, building foundations), number of home occupants, cooking power source, visible mould, pet ownership, child’s bedroom details (temperature, humidity, number of occupants, heating source, cooling source, flooring type, rugs or mats, visible mould), child’s bed details (type and age of bed, blanket, pillow, mattress, cover), exposure to tobacco smoke Family history, pregnancy and perinatal data questionnaire Mother and father (age, date of birth, country of birth, indigenous status, highest level of education, employment status, history of asthma, eczema, or hayfever) Pregnancy information (asthma diagnosed during pregnancy, medication use, vitamin/supplement use, smoking status, foods avoided during pregnancy, gestational diabetes, pre-eclampsia, hypertension) Perinatal information (gravidity, parity, gestational age, labour complications, cord blood taken, Apgar score, resuscitation required, admission to neonatal intensive care or special care nursery, time of birth, birthweight, birth length, head circumference, meconium aspiration, hyaline membrane disease, other neonatal complications) Symptoms and illness questionnaire Symptoms (sleep disturbed by coughing, wheeze, itchy rash, runny nose, flexural dermatitis), doctor-diagnosed (eczema, allergic rhinitis/hay fever, pneumonia, whooping cough, bronchiolitis, bronchitis, croup, asthma), significant medical or surgical problems, immunizations given, antibiotic use Diet Details of breastfeeding, use of infant formula, use of cow’s milk or other milk substitutes and introduction of solid food; vitamin/dietary supplement use and type; consumption of milk and solid foods (asked of mothers if breastfeeding and of children if started solid foods); use of study capsules, spreads and oils Clinical Details and history of symptoms: cough (ever or past 12/18 months, longest episode, episode lasted a week or more, during sleep, during physical activity, without a cold), wheeze (ever or past 12/18 months, episode for a week or more, longest episode, without a cold, caused difficulty breathing, health care use for wheeze, during sleep, during physical activity), rhinitis (ever or previous 12/18 months, episode for a week or more, longest episode, frequency of episode), eczema (itchy rash ever or previous 12 months); food allergy (asked up to 3 years: status and type of reaction); food avoidance (asked up to 3 years: type and on whose advice); doctor diagnosis and visited a GP, specialist, emergency department or hospital admission for eczema, allergic rhinitis, pneumonia, bronchiolitis, whooping cough, bronchitis, cough, asthma (from 8 years: diabetes or heart problems); medication (use, type, duration and frequency); snoring (at 5 years: ever and frequency; from 8 years: ever, frequency, loudness, stop breathing, struggle breathing during sleep, fall asleep at school, while watching television or during the daytime); television viewing (asked from 8 years: days per week, hours per day on weekday and weekend); parental health (parent or grandparent experienced a heart attack or stroke); childcare attendance and type (asked up to 5 years) Ethnicity Child’s maternal and paternal grandparents’ country of birth Puberty Tanner stages, puberty development scale, date of menarche (girls) Table 4. Details of the main CAPS questionnaires Questionnaire Information collected Home environment Housing details (house type, age, building material, building foundations), number of home occupants, cooking power source, visible mould, pet ownership, child’s bedroom details (temperature, humidity, number of occupants, heating source, cooling source, flooring type, rugs or mats, visible mould), child’s bed details (type and age of bed, blanket, pillow, mattress, cover), exposure to tobacco smoke Family history, pregnancy and perinatal data questionnaire Mother and father (age, date of birth, country of birth, indigenous status, highest level of education, employment status, history of asthma, eczema, or hayfever) Pregnancy information (asthma diagnosed during pregnancy, medication use, vitamin/supplement use, smoking status, foods avoided during pregnancy, gestational diabetes, pre-eclampsia, hypertension) Perinatal information (gravidity, parity, gestational age, labour complications, cord blood taken, Apgar score, resuscitation required, admission to neonatal intensive care or special care nursery, time of birth, birthweight, birth length, head circumference, meconium aspiration, hyaline membrane disease, other neonatal complications) Symptoms and illness questionnaire Symptoms (sleep disturbed by coughing, wheeze, itchy rash, runny nose, flexural dermatitis), doctor-diagnosed (eczema, allergic rhinitis/hay fever, pneumonia, whooping cough, bronchiolitis, bronchitis, croup, asthma), significant medical or surgical problems, immunizations given, antibiotic use Diet Details of breastfeeding, use of infant formula, use of cow’s milk or other milk substitutes and introduction of solid food; vitamin/dietary supplement use and type; consumption of milk and solid foods (asked of mothers if breastfeeding and of children if started solid foods); use of study capsules, spreads and oils Clinical Details and history of symptoms: cough (ever or past 12/18 months, longest episode, episode lasted a week or more, during sleep, during physical activity, without a cold), wheeze (ever or past 12/18 months, episode for a week or more, longest episode, without a cold, caused difficulty breathing, health care use for wheeze, during sleep, during physical activity), rhinitis (ever or previous 12/18 months, episode for a week or more, longest episode, frequency of episode), eczema (itchy rash ever or previous 12 months); food allergy (asked up to 3 years: status and type of reaction); food avoidance (asked up to 3 years: type and on whose advice); doctor diagnosis and visited a GP, specialist, emergency department or hospital admission for eczema, allergic rhinitis, pneumonia, bronchiolitis, whooping cough, bronchitis, cough, asthma (from 8 years: diabetes or heart problems); medication (use, type, duration and frequency); snoring (at 5 years: ever and frequency; from 8 years: ever, frequency, loudness, stop breathing, struggle breathing during sleep, fall asleep at school, while watching television or during the daytime); television viewing (asked from 8 years: days per week, hours per day on weekday and weekend); parental health (parent or grandparent experienced a heart attack or stroke); childcare attendance and type (asked up to 5 years) Ethnicity Child’s maternal and paternal grandparents’ country of birth Puberty Tanner stages, puberty development scale, date of menarche (girls) Questionnaire Information collected Home environment Housing details (house type, age, building material, building foundations), number of home occupants, cooking power source, visible mould, pet ownership, child’s bedroom details (temperature, humidity, number of occupants, heating source, cooling source, flooring type, rugs or mats, visible mould), child’s bed details (type and age of bed, blanket, pillow, mattress, cover), exposure to tobacco smoke Family history, pregnancy and perinatal data questionnaire Mother and father (age, date of birth, country of birth, indigenous status, highest level of education, employment status, history of asthma, eczema, or hayfever) Pregnancy information (asthma diagnosed during pregnancy, medication use, vitamin/supplement use, smoking status, foods avoided during pregnancy, gestational diabetes, pre-eclampsia, hypertension) Perinatal information (gravidity, parity, gestational age, labour complications, cord blood taken, Apgar score, resuscitation required, admission to neonatal intensive care or special care nursery, time of birth, birthweight, birth length, head circumference, meconium aspiration, hyaline membrane disease, other neonatal complications) Symptoms and illness questionnaire Symptoms (sleep disturbed by coughing, wheeze, itchy rash, runny nose, flexural dermatitis), doctor-diagnosed (eczema, allergic rhinitis/hay fever, pneumonia, whooping cough, bronchiolitis, bronchitis, croup, asthma), significant medical or surgical problems, immunizations given, antibiotic use Diet Details of breastfeeding, use of infant formula, use of cow’s milk or other milk substitutes and introduction of solid food; vitamin/dietary supplement use and type; consumption of milk and solid foods (asked of mothers if breastfeeding and of children if started solid foods); use of study capsules, spreads and oils Clinical Details and history of symptoms: cough (ever or past 12/18 months, longest episode, episode lasted a week or more, during sleep, during physical activity, without a cold), wheeze (ever or past 12/18 months, episode for a week or more, longest episode, without a cold, caused difficulty breathing, health care use for wheeze, during sleep, during physical activity), rhinitis (ever or previous 12/18 months, episode for a week or more, longest episode, frequency of episode), eczema (itchy rash ever or previous 12 months); food allergy (asked up to 3 years: status and type of reaction); food avoidance (asked up to 3 years: type and on whose advice); doctor diagnosis and visited a GP, specialist, emergency department or hospital admission for eczema, allergic rhinitis, pneumonia, bronchiolitis, whooping cough, bronchitis, cough, asthma (from 8 years: diabetes or heart problems); medication (use, type, duration and frequency); snoring (at 5 years: ever and frequency; from 8 years: ever, frequency, loudness, stop breathing, struggle breathing during sleep, fall asleep at school, while watching television or during the daytime); television viewing (asked from 8 years: days per week, hours per day on weekday and weekend); parental health (parent or grandparent experienced a heart attack or stroke); childcare attendance and type (asked up to 5 years) Ethnicity Child’s maternal and paternal grandparents’ country of birth Puberty Tanner stages, puberty development scale, date of menarche (girls) Table 5. Details of CAPS assessment tools and measurements Assessment tool/Measurement Details Dust collection Dust collected from child’s bed (or parents’ bed if child slept there >2 h/day), and child’s play area. House dust mite allergen was extracted from dust samples Anthropometric measurements Birth to 12 months: weight, length, head circumference; 12 months onwards: weight and height, with height measured quarterly from 11 years; 8 years: waist and hip circumference; 11.5 and 14 years: body fat and trunk fat by bioelectrical impedance analysis; mother’s and father’s height and weight (when child aged 8 years only) Dietary intake 18 months: 3-day weighed food record; 3 years: food frequency questionnaire; 9 years: 24-h dietary recall Physical examination Audible wheeze, presence of nasal crusting or discharge, presence of flexural eczema Blood collection Total immunoglobulin E (IgE); specific IgE (at 8 years only: alternaria, cat, rye-grass, house dust mite); fatty acids: plasma omega-3, omega-6 and various fatty acids; lipids: cholesterol (total, high-density lipoprotein, low-density lipoprotein) and triglycerides; cytokines: house dust mite stimulated: IL 4 (18 months only), IL 5, IL 10, IL 13 (3, 5 and 8 years), interferon-gamma; hormones: estradiol (girls at 11.5 years only), testosterone (boys at 11.5 and 14 years), insulin-like growth factor 1 (11.5 and 14 years) Skin prick test Allergens tested include: egg, cow’s milk, salmon, tuna, peanut, house dust mite, cat, dog, cockroach, ryegrass, aspergillus, alternaria, and grass-mix Forced oscillation technique Respiratory reactance (Xrs) and respiratory resistance (Rrs) Spirometry Forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) Methacholine challenge Dose-response ratio, and airway hyper-responsiveness measured by PD20FEV1 Exhaled nitric oxide Measure of airway inflammation Cardiovascular measures Blood pressure, carotid intima media thickness, augmentation index, carotid artery distensibility, carotid pulse pressure, brachial pulse wave velocity, non-fasting blood sample for: total cholesterol, high-density lipoprotein cholesterol, triglycerides, apolipoproteins A1 and B, high-sensitivity C-reactive protein, asymmetric dimethylarginine Overnight urine sample Gonadotropins, follicle-stimulating hormone DNA Single nucleotide polymorphisms (SNPs) at 3 years: IL13, IL14, intergenic, PHF11, CTLA4, filaggrin and CD14; telomere length at 3 and 14 years Assessment tool/Measurement Details Dust collection Dust collected from child’s bed (or parents’ bed if child slept there >2 h/day), and child’s play area. House dust mite allergen was extracted from dust samples Anthropometric measurements Birth to 12 months: weight, length, head circumference; 12 months onwards: weight and height, with height measured quarterly from 11 years; 8 years: waist and hip circumference; 11.5 and 14 years: body fat and trunk fat by bioelectrical impedance analysis; mother’s and father’s height and weight (when child aged 8 years only) Dietary intake 18 months: 3-day weighed food record; 3 years: food frequency questionnaire; 9 years: 24-h dietary recall Physical examination Audible wheeze, presence of nasal crusting or discharge, presence of flexural eczema Blood collection Total immunoglobulin E (IgE); specific IgE (at 8 years only: alternaria, cat, rye-grass, house dust mite); fatty acids: plasma omega-3, omega-6 and various fatty acids; lipids: cholesterol (total, high-density lipoprotein, low-density lipoprotein) and triglycerides; cytokines: house dust mite stimulated: IL 4 (18 months only), IL 5, IL 10, IL 13 (3, 5 and 8 years), interferon-gamma; hormones: estradiol (girls at 11.5 years only), testosterone (boys at 11.5 and 14 years), insulin-like growth factor 1 (11.5 and 14 years) Skin prick test Allergens tested include: egg, cow’s milk, salmon, tuna, peanut, house dust mite, cat, dog, cockroach, ryegrass, aspergillus, alternaria, and grass-mix Forced oscillation technique Respiratory reactance (Xrs) and respiratory resistance (Rrs) Spirometry Forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) Methacholine challenge Dose-response ratio, and airway hyper-responsiveness measured by PD20FEV1 Exhaled nitric oxide Measure of airway inflammation Cardiovascular measures Blood pressure, carotid intima media thickness, augmentation index, carotid artery distensibility, carotid pulse pressure, brachial pulse wave velocity, non-fasting blood sample for: total cholesterol, high-density lipoprotein cholesterol, triglycerides, apolipoproteins A1 and B, high-sensitivity C-reactive protein, asymmetric dimethylarginine Overnight urine sample Gonadotropins, follicle-stimulating hormone DNA Single nucleotide polymorphisms (SNPs) at 3 years: IL13, IL14, intergenic, PHF11, CTLA4, filaggrin and CD14; telomere length at 3 and 14 years Table 5. Details of CAPS assessment tools and measurements Assessment tool/Measurement Details Dust collection Dust collected from child’s bed (or parents’ bed if child slept there >2 h/day), and child’s play area. House dust mite allergen was extracted from dust samples Anthropometric measurements Birth to 12 months: weight, length, head circumference; 12 months onwards: weight and height, with height measured quarterly from 11 years; 8 years: waist and hip circumference; 11.5 and 14 years: body fat and trunk fat by bioelectrical impedance analysis; mother’s and father’s height and weight (when child aged 8 years only) Dietary intake 18 months: 3-day weighed food record; 3 years: food frequency questionnaire; 9 years: 24-h dietary recall Physical examination Audible wheeze, presence of nasal crusting or discharge, presence of flexural eczema Blood collection Total immunoglobulin E (IgE); specific IgE (at 8 years only: alternaria, cat, rye-grass, house dust mite); fatty acids: plasma omega-3, omega-6 and various fatty acids; lipids: cholesterol (total, high-density lipoprotein, low-density lipoprotein) and triglycerides; cytokines: house dust mite stimulated: IL 4 (18 months only), IL 5, IL 10, IL 13 (3, 5 and 8 years), interferon-gamma; hormones: estradiol (girls at 11.5 years only), testosterone (boys at 11.5 and 14 years), insulin-like growth factor 1 (11.5 and 14 years) Skin prick test Allergens tested include: egg, cow’s milk, salmon, tuna, peanut, house dust mite, cat, dog, cockroach, ryegrass, aspergillus, alternaria, and grass-mix Forced oscillation technique Respiratory reactance (Xrs) and respiratory resistance (Rrs) Spirometry Forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) Methacholine challenge Dose-response ratio, and airway hyper-responsiveness measured by PD20FEV1 Exhaled nitric oxide Measure of airway inflammation Cardiovascular measures Blood pressure, carotid intima media thickness, augmentation index, carotid artery distensibility, carotid pulse pressure, brachial pulse wave velocity, non-fasting blood sample for: total cholesterol, high-density lipoprotein cholesterol, triglycerides, apolipoproteins A1 and B, high-sensitivity C-reactive protein, asymmetric dimethylarginine Overnight urine sample Gonadotropins, follicle-stimulating hormone DNA Single nucleotide polymorphisms (SNPs) at 3 years: IL13, IL14, intergenic, PHF11, CTLA4, filaggrin and CD14; telomere length at 3 and 14 years Assessment tool/Measurement Details Dust collection Dust collected from child’s bed (or parents’ bed if child slept there >2 h/day), and child’s play area. House dust mite allergen was extracted from dust samples Anthropometric measurements Birth to 12 months: weight, length, head circumference; 12 months onwards: weight and height, with height measured quarterly from 11 years; 8 years: waist and hip circumference; 11.5 and 14 years: body fat and trunk fat by bioelectrical impedance analysis; mother’s and father’s height and weight (when child aged 8 years only) Dietary intake 18 months: 3-day weighed food record; 3 years: food frequency questionnaire; 9 years: 24-h dietary recall Physical examination Audible wheeze, presence of nasal crusting or discharge, presence of flexural eczema Blood collection Total immunoglobulin E (IgE); specific IgE (at 8 years only: alternaria, cat, rye-grass, house dust mite); fatty acids: plasma omega-3, omega-6 and various fatty acids; lipids: cholesterol (total, high-density lipoprotein, low-density lipoprotein) and triglycerides; cytokines: house dust mite stimulated: IL 4 (18 months only), IL 5, IL 10, IL 13 (3, 5 and 8 years), interferon-gamma; hormones: estradiol (girls at 11.5 years only), testosterone (boys at 11.5 and 14 years), insulin-like growth factor 1 (11.5 and 14 years) Skin prick test Allergens tested include: egg, cow’s milk, salmon, tuna, peanut, house dust mite, cat, dog, cockroach, ryegrass, aspergillus, alternaria, and grass-mix Forced oscillation technique Respiratory reactance (Xrs) and respiratory resistance (Rrs) Spirometry Forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) Methacholine challenge Dose-response ratio, and airway hyper-responsiveness measured by PD20FEV1 Exhaled nitric oxide Measure of airway inflammation Cardiovascular measures Blood pressure, carotid intima media thickness, augmentation index, carotid artery distensibility, carotid pulse pressure, brachial pulse wave velocity, non-fasting blood sample for: total cholesterol, high-density lipoprotein cholesterol, triglycerides, apolipoproteins A1 and B, high-sensitivity C-reactive protein, asymmetric dimethylarginine Overnight urine sample Gonadotropins, follicle-stimulating hormone DNA Single nucleotide polymorphisms (SNPs) at 3 years: IL13, IL14, intergenic, PHF11, CTLA4, filaggrin and CD14; telomere length at 3 and 14 years What has the study found? Key findings and publications To date, there have been 63 peer-reviewed publications reporting CAPS results. With respect to our principal aim, we found that the interventions were successful in reducing HDM allergen concentration in dust collected from beds and in increasing the ratio of omega-3 to omega-6 fatty acids detected in plasma at age 5.9,37 However, neither HDM avoidance nor omega-3 fatty acid supplementation, as implemented from birth to age 5, reduced the prevalence of asthma, atopy or other atopic disorders at age 5, nor at the longer-term follow up ages 8 and 11.5.9,15,20 The CAPS study has reported a number of analyses of the association between risk factors of asthma and allergic disease and the incidence of these diseases. We have found the following. Birthweight below the first tertile was associated with a greater risk of current asthma and poorer lung function at age 8 years.38 Longer duration of breastfeeding (>6 months) was associated with an increased risk of allergic sensitization at ages 5 and 8.39,40 Early childhood eczema, but not early life wheeze or rhinitis, predicted subsequent development of allergen sensitization by age 5.41 Owning a pet cat or dog before age 5 was associated with a reduced risk of being atopic at age 5.42 Exposure to low and high, but not intermediate, levels of HDM allergen was associated with a lower prevalence of HDM atopy and asthma at age 5.43 At age 8, exposure to vehicular traffic, quantified as the weighted road density at the child’s residential address, was positively associated with HDM sensitization and rhinitis.44 The presence of HDM-specific interleukin-5 responses at ages 3, 5, and 8 was associated with the presence of asthma and atopy at age 8.45 We have reported a number of findings about the early life predictors and manifestations of cardiovascular disease. Key findings at age 8 include the following. Compared with boys, girls—independent of height—had lower carotid extra-medial thickness46 and greater arterial wave augmentation.47 Greater carotid intima medial thickness was associated with lower high-density lipoprotein (HDL) cholesterol, higher levels of asymmetric dimethylarginine (ADMA), and higher systolic blood pressure.18 Excessive weight gain in infancy was associated with greater carotid intima medial thickness48 and carotid extra-medial thickness.49 Maternal smoking in pregnancy was associated with significantly lower HDL cholesterol.50 Consuming a large amount of dairy food at age 18 months was associated with lower blood pressure at age 8.51 Omega-3 supplementation during the first 5 years of life did not improve arterial structure and function at age 8;19 it did, however, reverse the inverse association between impaired fetal growth and arterial wall thickness.52 Lower spirometric lung volumes were associated with increased vascular stiffness.53 The presence of asthma and of airway inflammation was not associated with alterations in systemic ADMA or L-arginine levels.54 Shorter telomere length in early childhood was associated with arterial wall thickness.34 Carotid extra-medial thickness, but not carotid intima-media thickness, was associated with local arterial stiffness.55 At age 14, the augmentation index was higher in girls than boys and was closely associated with change in height between ages 8 and 14.56 The dietary intake data have been used to characterize the diet of Australian children and to assess the impact of diet components on weight gain and the development of obesity. We have reported the following. Distribution of types of food, nutrients and portion sizes,57 meat intake58 and the intake of energy-dense, nutrient-poor foods59 among children aged 18 months were described. Higher intakes of protein and meat at age 18 months were positively associated with greater adiposity at age 8,60 and high intakes of meat and carbohydrates were associated with high body mass index from birth to age 11.5 years in boys.61 Adequate dairy consumption at age 9 was associated with diets of higher nutritional quality, but also with higher intakes of energy,23 and energy consumed in liquid form contributed more to the development of obesity.62 Genomic data from participants with asthma in the cohort have contributed to multi-centre (Australian and international) genome-wide association studies (GWAS), as part of the Australian Asthma Genetics Consortium to identify new risk loci for asthma and allergic disease in children.29–33 Advanced statistical techniques have been applied to the longitudinally collected data, with repeated measures to provide new insights into asthma, allergic disease and obesity. Finite mixture models have been used to explore the heterogeneity in asthma, atopy and growth by defining latent subgroups, often called classes or phenotypes. A latent class analysis of allergen skin prick tests performed at ages 1.5 to 8 years revealed four phenotypes: late mixed inhalant sensitization; mixed food and inhalant sensitization; HDM monosensitized; and no atopy.63 All three atopy phenotypes were associated with asthma, eczema and rhinitis, but the strongest association, particularly for asthma, was with the mixed food and inhalant sensitization phenotype, implying that food sensitization in early life might be of greater significance for subsequent risk of asthma than previously thought. Growth mixture models have been applied to body mass index (BMI) data collected from birth to age 11.5 and identified three BMI growth trajectories, differing qualitatively between boys and girls;61 growth mixture models applied to height collected from ages 11–14 showed that girls with asthma at age 8 had a higher probability of belonging to a later growth trajectory.64 A latent transition analysis model was applied to data from age 0–11.5 years to incorporate the longitudinal patterns of several manifestations of asthma into a single model, to simultaneously define phenotypes and examine their transitions over time.65 It provided quantitative support for the view that asthma is a heterogeneous entity, and that some children with wheeze and other respiratory symptoms in early life progress to asthma in mid-childhood, whereas others become asymptomatic. What are the study’s main strengths and weaknesses? The major strengths of this study are: an existing, well-motivated cohort of participants with good retention rates during the first 5-8 years of life; the detailed characterization of early life constitutive and environmental risk factors for asthma and allergic disease, which has extended from the antenatal period to age 14 years, and which is accompanied by equally detailed characterization of allergic and respiratory outcomes, including objective measurements during this period; use of the strongest possible study design, an RCT to test HDM avoidance and dietary fatty acid intervention, both of which were successfully implemented from birth to age 5 years, giving a unique opportunity to assess the long-term outcome of these interventions; the high-risk nature of the cohort, which represents the population most likely to be the target for future interventions to reduce asthma and allergic disease; use of the participants, and of their accompanying early life details, to study other childhood diseases including obesity and cardiovascular disease; and a long-serving, multidisciplinary and committed research team, many of whom have been involved in CAPS since its inception or soon after, allowing a strong engagement with the participants and a greater understanding of the resultant data. The main weakness of CAPS, as for most long-term cohort studies, is attrition as the participants grow older. As the participants have entered and moved through adolescence, they have become less interested in participating in clinical examinations. This has resulted in 32% (196/616) of the original sample completing clinical testing at 14 years. A by-product of this attrition is that the remaining sample in CAPS is from a higher socioeconomic background (Table 3) than those who have withdrawn. Can I get hold of the data? Where can I find out more? The CAPS study team are interested in collaborating with others. Researchers interested in collaborating with CAPS researchers or wishing to access CAPS data are encouraged to contact our Chief Investigator, Dr Brett Toelle [[emailprotected]]. For approval from the Chief Investigator, researchers will be asked to write a short proposal describing the aims of their project and specifying what data would be required. Profile in a nutshell CAPS was established as a randomized controlled trial to test the effectiveness of a house dust mite avoidance intervention and an omega-3 supplement intervention for the primary prevention of asthma; it is now a well-established birth cohort for studying the natural development of asthma. Pregnant women (n = 616), whose unborn children were at risk of developing asthma, were recruited from hospitals in Sydney, Australia, between 1997 and 1999. On 42 separate occasions from 36 weeks of gestation to age 14 years, data have been collected using questionnaires and other clinical measurements and assessments; 352 subjects remain eligible for future follow-up. Information collected and tests performed have included: family history, pregnancy and perinatal details, environmental exposures, dietary intake, child’s symptoms and illnesses, anthropometric measures, DNA, skin prick tests, spirometry, airway hyper-responsiveness, exhaled nitric oxide, lipids, fatty acids, cytokines, pubertal stages, hormones and cardiovascular function. More than 63 peer-reviewed articles have been published. Researchers interested in collaborating can contact Chief Investigator Dr Brett Toelle [[emailprotected]]. Funding The National Health and Medical Research Council of Australia has, through a series of grants, been the major funder. Additional substantive funding has come from the Cooperative Research Centre for Asthma, the New South Wales Department of Health, the Children’s Hospital at Westmead in Sydney and the University of Sydney. Acknowledgements Contributions of goods and services were made by Allergopharma Joachim Ganzer KG Germany, John Sands Australia, Hasbro, Toll refrigerated, AstraZeneca Australia and Nu-Mega Ingredients Pty Ltd. Goods were provided at a reduced cost by Auspharm, Allersearch and Goodman Fielder Foods. Liverpool Hospital and the Children’s Hospital at Westmead provided facilities for the conduct of the study. Conflict of interest: None declared. References 1 Burney PG , Chinn S , Rona RJ. Has the prevalence of asthma increased in children? Evidence from the national study of health and growth 1973-86 . BMJ 1990 ; 300 : 1306 – 10 . Google Scholar Crossref Search ADS PubMed 2 Peat JK , Van Den Berg RH , Green WF , Mellis CM , Leeder SR , Wolcock AJ. Changing prevalence of asthma in Australian children . BMJ 1994 ; 308 : 1591 – 96 . Google Scholar Crossref Search ADS PubMed 3 Platts-Mills TAE , de Weck AL , Aalberse RC et al. Dust mite allergens and asthma—a worldwide problem . J Allergy Clin Immunol 1989 ; 83 : 416 – 27 . Google Scholar Crossref Search ADS PubMed 4 Peat JK , Woolcock AJ. Sensitivity to common allergens: relation to respiratory symptoms and bronchial hyper-responsiveness in children from three different climatic areas of Australia . Clin Exp Allergy 1991 ; 21 : 573 – 81 . Google Scholar Crossref Search ADS PubMed 5 Korsgaard J. Mite asthma and residency. A case-control study on the impact of exposure to house-dust mites in dwellings . Am Rev Respir Dis 1983 ; 128 : 231 – 35 . Google Scholar PubMed 6 Sporik R , Holgate ST , Platts-Mills TA , Cogswell JJ. Exposure to house-dust mite allergen (Der p I) and the development of asthma in childhood. A prospective study . N Engl J Med 1990 ; 323 : 502 – 07 . Google Scholar Crossref Search ADS PubMed 7 Hodge L , Salome CM , Peat JK , Haby MM , Xuan W , Woolcock AJ. Consumption of oily fish and childhood asthma risk . Med J Aust 1996 ; 164 : 137 – 40 . Google Scholar PubMed 8 Haby MM , Peat JK , Marks GB , Woolcock AJ , Leeder SR. Asthma in preschool children: prevalence and risk factors . Thorax 2001 ; 56 : 589 – 95 . Google Scholar Crossref Search ADS PubMed 9 Marks GB , Mihrshahi S , Kemp AS et al. Prevention of asthma during the first 5 years of life: a randomized controlled trial . J Allergy Clin Immunol 2006 ; 118 : 53 – 61 . Google Scholar Crossref Search ADS PubMed 10 Mihrshahi S , Peat JK , Webb K et al. The Childhood Asthma Prevention Study (CAPS): design and research protocol of a randomized trial for the primary prevention of asthma . Control Clin Trials 2001 ; 22 : 333 – 54 . Google Scholar Crossref Search ADS PubMed 11 Mihrshahi S , Vukasin N , Forbes S et al. Are you busy for the next 5 years? Recruitment in the Childhood Asthma Prevention Study (CAPS) . Respirology 2002 ; 7 : 147 – 51 . Google Scholar Crossref Search ADS PubMed 12 Mihrshahi S , Peat JK , Marks GB et al. Eighteen-month outcomes of house dust mite avoidance and dietary fatty acid modification in the Childhood Asthma Prevention Study (CAPS) . J Allergy Clin Immunol 2003 ; 111 : 162 – 68 . Google Scholar Crossref Search ADS PubMed 13 Peat JK , Mihrshahi S , Kemp AS et al. Three-year outcomes of dietary fatty acid modification and house dust mite reduction in the Childhood Asthma Prevention Study . J Allergy Clin Immunol 2004 ; 114 : 807 – 13 . Google Scholar Crossref Search ADS PubMed 14 Sears MR , Greene JM , Willan AR et al. Long-term relation between breastfeeding and development of atopy and asthma in children and young adults: a longitudinal study . Lancet 2002 ; 360 : 901 – 07 . Google Scholar Crossref Search ADS PubMed 15 Toelle BG , Ng KKW , Crisafulli D et al. Eight-year outcomes of the Childhood Asthma Prevention Study . J Allergy Clin Immunol 2010 ; 126 : 388 – 89 . Google Scholar Crossref Search ADS PubMed 16 Toelle BG , Xuan W , Peat JK , Marks GB. Childhood factors that predict asthma in young adulthood . Eur Respir J 2004 ; 23 : 66 – 70 . Google Scholar Crossref Search ADS PubMed 17 Sears MR , Greene JM , Willan AR et al. A longitudinal, population-based, cohort study of childhood asthma followed to adulthood . N Engl J Med 2003 ; 349 : 1414 – 22 . Google Scholar Crossref Search ADS PubMed 18 Ayer JG , Harmer JA , Nakhla S et al. HDL-cholesterol, blood pressure, and asymmetric dimethylarginine are significantly associated with arterial wall thickness in children . Arterioscler Thromb Vasc Biol 2009 ; 29 : 943 – 49 . Google Scholar Crossref Search ADS PubMed 19 Ayer JG , Harmer JA , Xuan W et al. Dietary supplementation with n-3 polyunsaturated fatty acids in early childhood: effects on blood pressure and arterial structure and function at age 8 y . Am J Clin Nutr 2009 ; 90 : 438 – 46 . Google Scholar Crossref Search ADS PubMed 20 Toelle BG , Garden FL , Ng KK et al. Outcomes of the Childhood Asthma Prevention Study at 11.5 years . J Allergy Clin Immunol 2013 ; 132 : 1220 – 22.e3 . Google Scholar Crossref Search ADS PubMed 21 Almqvist C , Worm M , Leynaert B ; for the working group of GALENWPG . Impact of gender on asthma in childhood and adolescence: a GA2LEN review . Allergy 2008 ; 63 : 47 – 57 . Google Scholar PubMed 22 Mandhane PJ , Greene JM , Cowan JO , Taylor DR , Sears MR. Sex differences in factors associated with childhood- and adolescent-onset wheeze . Am J Respir Crit Care Med 2005 ; 172 : 45 – 54 . Google Scholar Crossref Search ADS PubMed 23 Rangan AM , Flood VM , Denyer G , Webb K , Marks GB , Gill TP. Dairy consumption and diet quality in a sample of Australian children . J Am Coll Nutr 2012 ; 31 : 185 – 93 . Google Scholar Crossref Search ADS PubMed 24 Tanner JM. Growth at Adolescence: With a General Consideration of the Effects of Hereditary and Environmental Factors Upon Growth and Maturation From Birth to Maturity . 2 nd edn. Oxford, UK : Blackwell Scientific Publications , 1962 . 25 Marshall WA , Tanner JM. Variations in pattern of pubertal changes in girls . Arch Dis Child 1969 ; 44 : 291 – 303 . Google Scholar Crossref Search ADS PubMed 26 Marshall WA , Tanner JM. Variations in the pattern of pubertal changes in boys . Arch Dis Child 1970 ; 45 : 13 – 23 . Google Scholar Crossref Search ADS PubMed 27 Carskadon MA , Acebo C. A self-administered rating scale for pubertal development . J Adolesc Health 1993 ; 14 : 190 – 95 . Google Scholar Crossref Search ADS PubMed 28 Petersen AC , Crockett L , Richards M , Boxer A. A self-report measure of pubertal status: reliability, validity, and initial norms . J Youth Adolesc 1988 ; 17 : 117 – 33 . Google Scholar Crossref Search ADS PubMed 29 Ferreira MAR , Matheson MC , Duffy DL et al. Identification of IL6R and chromosome 11q13.5 as risk loci for asthma . Lancet 2011 ; 378 : 1006 – 14 . Google Scholar Crossref Search ADS PubMed 30 Ramasamy A , Kuokkanen M , Vedantam S et al. Genome-wide association studies of asthma in population-based cohorts confirm known and suggested loci and identify an additional association near HLA . PLoS One 2012 ; 7 : e44008. Google Scholar Crossref Search ADS PubMed 31 Bonnelykke K , Matheson MC , Pers TH et al. Meta-analysis of genome-wide association studies identifies ten loci influencing allergic sensitization . Nat Genet 2013 ; 45 : 902 – 06 . Google Scholar Crossref Search ADS PubMed 32 Marenholz I , Esparza-Gordillo J , Ruschendorf F et al. Meta-analysis identifies seven susceptibility loci involved in the atopic march . Nat Commun 2015 ; 6 : 8804. Google Scholar Crossref Search ADS PubMed 33 Paternoster L , Standl M , Waage J et al. Multi-ancestry genome-wide association study of 21, 000 cases and 95, 000 controls identifies new risk loci for atopic dermatitis . Nat Genet 2015 ; 47 : 1449 – 56 . Google Scholar Crossref Search ADS PubMed 34 Skilton MR , Nakhla S , Ayer JG et al. Telomere length in early childhood: Early life risk factors and association with carotid intima-media thickness in later childhood . Eur J Prev Cardiolog 2016 ; 23 : 1086 – 92 . Google Scholar Crossref Search ADS 35 Weber-Chrysochoou C , Crisafulli D , Almqvist C et al. IL-5 T-cell responses to house dust mite are associated with the development of allergen-specific IgE responses and asthma in the first 5 years of life . J Allergy Clin Immunol 2007 ; 120 : 286 – 92 . Google Scholar Crossref Search ADS PubMed 36 Brew BK , Toelle BG , Webb KL , Almqvist C , Marks GB. Omega-3 supplementation during the first 5 years of life and later academic performance: a randomised controlled trial . Eur J Clin Nutr 2015 ; 69 : 419 – 24 . Google Scholar Crossref Search ADS PubMed 37 Mihrshahi S , Marks GB , Criss S , Tovey ER , Vanlaar CH , Peat J. Effectiveness of an intervention to reduce house dust mite allergen levels in children’s beds . Allergy 2003 ; 58 : 784 – 89 . Google Scholar Crossref Search ADS PubMed 38 Brew BK , Marks GB. Perinatal factors and respiratory health in children . Clin Exp Allergy 2012 ; 42 : 1621 – 29 . Google Scholar Crossref Search ADS PubMed 39 Mihrshahi S , Ampon R , Webb K et al. The association between infant feeding practices and subsequent atopy among children with a family history of asthma . Clin Exp Allergy 2007 ; 37 : 671 – 79 . Google Scholar Crossref Search ADS PubMed 40 Brew BK , Kull I , Garden F et al. Breastfeeding, asthma, and allergy: a tale of two cities . Pediatr Allergy Immunol 2012 ; 23 : 75 – 82 . Google Scholar Crossref Search ADS PubMed 41 Almqvist C , Li Q , Britton WJ et al. Early predictors for developing allergic disease and asthma: examining separate steps in the ‘allergic march’ . Clin Exp Allergy 2007 ; 37 : 1296 – 302 . Google Scholar Crossref Search ADS PubMed 42 Almqvist C , Garden F , Kemp AS et al. Effects of early cat or dog ownership on sensitisation and asthma in a high-risk cohort without disease-related modification of exposure . Paediatr Perinat Epidemiol 2010 ; 24 : 171 – 78 . Google Scholar Crossref Search ADS PubMed 43 Tovey ER , Almqvist C , Li Q , Crisafulli D , Marks GB. Nonlinear relationship of mite allergen exposure to mite sensitization and asthma in a birth cohort . J Allergy Clin Immunol 2008 ; 122 : 114 – 18. 8 e1–5 . Google Scholar Crossref Search ADS PubMed 44 Hansell AL , Rose N , Cowie CT et al. Weighted road density and allergic disease in children at high risk of developing asthma . PLoS One 2014 ; 9 : e98978. Google Scholar Crossref Search ADS PubMed 45 Weber-Chrysochoou C , Crisafulli D , Kemp AS , Britton WJ , Marks GB ; for the CAPS Investigators . Allergen-specific IL-5 responses in early childhood predict asthma at age eight . PLoS One 2014 ; 9 : e97995. Google Scholar Crossref Search ADS PubMed 46 Skilton MR , Sullivan TR , Ayer JG et al. Carotid extra-medial thickness in childhood: early life effects on the arterial adventitia . Atherosclerosis 2012 ; 222 : 478 – 82 . Google Scholar Crossref Search ADS PubMed 47 Ayer JG , Harmer JA , Marks GB , Avolio A , Celermajer DS. Central arterial pulse wave augmentation is greater in girls than boys, independent of height . J Hypertens 2010 ; 28 : 306 – 13 . Google Scholar Crossref Search ADS PubMed 48 Skilton MR , Marks GB , Ayer JG et al. Weight gain in infancy and vascular risk factors in later childhood . Pediatrics 2013 ; 131 : e1821 – 28 . Google Scholar Crossref Search ADS PubMed 49 Skilton MR , Sullivan TR , Ayer JG et al. Weight gain in infancy is associated with carotid extra-medial thickness in later childhood . Atherosclerosis 2014 ; 233 : 370 – 74 . Google Scholar Crossref Search ADS PubMed 50 Ayer JG , Belousova E , Harmer JA , David C , Marks GB , Celermajer DS. Maternal cigarette smoking is associated with reduced high-density lipoprotein cholesterol in healthy 8-year-old children . Eur Heart J 2011 ; 32 : 2446 – 53 . Google Scholar Crossref Search ADS PubMed 51 Rangan AM , Flood VL , Denyer G et al. The effect of dairy consumption on blood pressure in mid-childhood: CAPS cohort study . Eur J Clin Nutr 2012 ; 66 : 652 – 57 . Google Scholar Crossref Search ADS PubMed 52 Skilton MR , Ayer JG , Harmer JA et al. Impaired fetal growth and arterial wall thickening: a randomized trial of omega-3 supplementation . Pediatrics 2012 ; 129 : e698 – 703 . Google Scholar Crossref Search ADS PubMed 53 Ayer JG , Belousova EG , Harmer JA , Toelle B , Celermajer DS , Marks GB. Lung function is associated with arterial stiffness in children . PLoS One 2011 ; 6 : e26303. Google Scholar Crossref Search ADS PubMed 54 Lau EM , Morgan PE , Belousova EG et al. Asymmetric dimethylarginine and asthma: results from the Childhood Asthma Prevention Study . Eur Respir J 2013 ; 41 : 1234 – 37 . Google Scholar Crossref Search ADS PubMed 55 Cai TY , Sullivan TR , Ayer JG et al. Carotid extramedial thickness is associated with local arterial stiffness in children . J Hypertens 2016 ; 34 : 109 – 15 . Google Scholar Crossref Search ADS PubMed 56 Barraclough JY , Garden FL , Toelle B et al. Sex differences in aortic augmentation index in adolescents . J Hypertens 2017 ; 35 : 2016 – 24 . Google Scholar Crossref Search ADS PubMed 57 Webb K , Rutishauser I , Knezevic N. Foods, nutrients and portions consumed by a sample of Australian children aged 16-24 months . Nutr Diet 2008 ; 65 : 56 – 65 . Google Scholar Crossref Search ADS 58 Webb K , Rutishauser I , Katz T et al. Meat consumption among 18-month-old children participating in the Childhood Asthma Prevention Study . Nutr Diet 2005 ; 62 : 12 – 20 . Google Scholar Crossref Search ADS 59 Webb KL , Lahti-Koski M , Rutishauser I et al. Consumption of “extra” foods (energy-dense, nutrient-poor) among children aged 16-24 months from western Sydney, Australia . Public Health Nutr 2006 ; 9 : 1035 – 44 . Google Scholar Crossref Search ADS PubMed 60 Garden FL , Marks GB , Almqvist C , Simpson JM , Webb KL. Infant and early childhood dietary predictors of overweight at age 8 years in the CAPS population . Eur J Clin Nutr 2011 ; 65 : 454 – 62 . Google Scholar Crossref Search ADS PubMed 61 Garden FL , Marks GB , Simpson JM , Webb KL. Body mass index (BMI) trajectories from birth to 11.5 years: relation to early life food intake . Nutrients 2012 ; 4 : 1382 – 98 . Google Scholar Crossref Search ADS PubMed 62 Zheng M , Allman-Farinelli M , Heitmann BL et al. Liquid versus solid energy intake in relation to body composition among Australian children . J Hum Nutr Diet 2015 ; 28 : 70 – 79 . Google Scholar Crossref Search ADS PubMed 63 Garden FL , Simpson JM , Marks GB. Atopy phenotypes in the Childhood Asthma Prevention Study (CAPS) cohort and the relationship with allergic disease: clinical mechanisms in allergic disease . Clin Exp Allergy 2013 ; 43 : 633 – 41 . Google Scholar PubMed 64 Movin M , Garden FL , Protudjer JL et al. Impact of childhood asthma on growth trajectories in early adolescence: Findings from the Childhood Asthma Prevention Study (CAPS) . Respirology 2017 ; 22 : 460 – 65 . Google Scholar Crossref Search ADS PubMed 65 Garden FL , Simpson JM , Mellis CM , Marks GB ; CAPS Investigators . Change in the manifestations of asthma and asthma-related traits in childhood: a latent transition analysis . Eur Respir J 2016 ; 47 : 499 – 509 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Open Access Collection

Cohort Profile: The Maternal and Infant Nutrition Interventions in Matlab (MINIMat) cohort in Bangladesh

Arifeen, Shams El; Ekström, Eva-Charlotte; Frongillo, Edward A; Hamadani, Jena; Khan, Ashraful I; Naved, Ruchira T; Rahman, Anisur; Raqib, Rubhana; Rasmussen, Kathleen M; Selling, Katarina Ekholm; Wagatsuma, Yukiko; Persson, Lars Åke

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy102pmid: 29868907

Why was the cohort set up? Proportions of malnourished mothers and children remain high, especially in South Asia.1 In countries like India, Pakistan and Bangladesh, underweight and stunted linear growth frequently start in fetal life. This is reflected in a high proportion of children who are small for their gestational age at birth and experience continued stunted growth.2–4 A life cycle perspective is needed to address these intergenerational problems.1,5 Trials that have provided pregnant women in undernourished populations with balanced protein-energy supplementations demonstrated increased birthweight.6 Most food supplementation programmes have reached pregnant women in mid or late pregnancy6 for practical reasons. In the first trimester, organs develop and fetal plasticity is pronounced. A balanced dietary intake including micronutrients is needed beginning in the first trimester, to ensure fetal health, growth and later health.7–10 To prevent anaemia and promote fetal health, the World Health Organization has recommended iron-folate supplementation during pregnancy. Based on the notion that iron and other micronutrient deficiencies often coexist, a group of experts developed the composition of a multiple-micronutrient supplement for trial purposes.11 A series of prenatal multiple micronutrient trials were initiated in different countries, with maternal haemoglobin and birthweight as outcomes.12 Nutritional imbalance or insult in fetal or early life may alter later disease risk. This concept, later labelled the Developmental Origin of Health and Disease (DOHaD), has been underpinned by epidemiological and biomedical studies.13–15 Few prenatal nutritional intervention studies have been done, however, to assess the effect of prevention of nutritional imbalance or insult on developmental trajectories at a stage when developmental plasticity is great.16 Based on these considerations, the designers of the Maternal and Infant Nutrition Interventions in Matlab trial (MINIMat, ClinicalTrials.gov identifier ISRCTN16581394) hypothesized that prenatal multiple-micronutrient supplementation (MMS), as well as an early invitation (around gestational week 9) to receive a daily food supplement, would increase maternal haemoglobin concentration at 30 weeks’ gestation, birthweight, and infant survival. The designers also hypothesized that a combination of these interventions (early invitation with MMS) would further improve these outcomes compared with the usual timing of invitation to food supplementation (around week 20) and supplementation with the standard programme of iron-folic acid supplements. At an early stage in the planning of the MINIMat cohort, the perspective that events in early fetal life might have longer-term consequences provided a rationale for a long-term follow-up. Secondary outcomes listed at the time of the trial registration were: child growth and cognitive development; child micronutrient status; child immune function and morbidity; blood pressure; metabolic markers; and mothers’ anthropometric development up to the next pregnancy, when applicable. The trial was carried out in Matlab, Bangladesh, a rural sub-district 57 km south of the capital Dhaka, a setting where child and maternal undernutrition remain widespread. In this area, the International Centre for Diarrhoeal Disease Research, Bangladesh (iccdr, b) has been running a Health and Demographic Surveillance System (HDSS) since 1966. Through monthly household visits, community health research workers (CHRW) collect data on demographic and selected health information, on a population of about 220 000 in more than 140 villages. The use of a unique identification system allows tracking over time and across studies and databases. The initial funding of the MINIMat trial (up to birth) was provided by the United Nations Children’s Fund (UNICEF), followed by many funders, as described below. Who is in the cohort? Participants were recruited from 11 November 2001 to 30 October 2003 (Figure 1). CHRW visited all households monthly. If a woman of reproductive age reported that her last menstrual period (LMP) was overdue or that she was pregnant, she was offered a pregnancy test and the date of her LMP was recorded. A woman who tested positive was encouraged to visit the icddr, b clinic, where an ultrasound examination was offered. Dating based on ultrasound examination at week 8 was used if LMP date was missing Women were eligible for the study if they had a viable fetus and gestational age less than 14 weeks on ultrasound, no severe illness and provided written consent for participation. In total, 5880 women were assessed for eligibility, 1444 were excluded, and 4436 women were randomized to the different nutrition interventions. There were 3625 live births. The MINIMat cohort study and the Health and Demographic Surveillance System in the area have closely monitored these children. Figure 1. Open in new tabDownload slide Study flow. Figure 1. Open in new tabDownload slide Study flow. The MINIMat trial participants were randomized into six groups in a 3 x 2 design. A double-masked supplementation with capsules of 30 mg iron and 400 µg of folic acid, 60 mg of iron and 400 µg of folic acid or multiple micronutrients (MMS) containing a daily allowance of 15 micronutrients, including 30 mg of iron and 400 µg of folic acid, was combined with food supplementation (608 kcal/day on 6 days per week) randomized to either early invitation (9 weeks’ gestation) or usual invitation (20 weeks’ gestation). Primary outcomes were maternal haemoglobin concentration at 30 weeks’ gestation, birthweight, gestational age at birth and infant mortality. The pregnant women who participated in the MINIMat trial were also randomly allocated in the third trimester to receive either the usual health messages from the antenatal services or to receive exclusive breastfeeding counselling. At baseline, the pregnant women in this cohort had an average weight of 45 kg and mean height of 150 cm,17 see Table 1. One-third of these women were in their first pregnancy and four out of 10 were in their third or later pregnancies. One-third had no formal education, and slightly more than half had attended school for 5 years or more. One-quarter had some surplus in the perceived status of income compared with expenditure, and one-fifth had some deficit. The attrition during the follow-up period did not change the general characteristics and proportions allocated to the different interventions of the population being studied, see Table 1. Table 1. Background characteristics of the pregnant women at recruitment, for those with a live birth, and for those where the child also participated at follow-up at 12–14 years. MINIMat cohort, Bangladesh, 2001–17 Characteristic . Levels . Recruitment . Live birth . 12-14 years . (n 3625) . (n 2307) . Total 4415 3587 2306 Age at recruitment <20 711 (16) 563 (16) 331 (14) 20–29 2521 (57) 2067 (57) 1312 (57) >30 1183 (27) 957 (27) 663 (29) 4416 3619 2307 Parity First 1479 (33) 1184 (33) 662 (29) Second or more 2937 (67) 2435 (67) 1645 (71) Total 4462 3580 2300 BMI kg/m2 <18.5 1228 (28) 1000 (28) 669 (29) >18.5 3234 (72) 2580 (72) 1631 (71) Total 4431 3591 2307 Educational level No schooling 1427 (32) 1135 (32) 795 (34) 1–4 years 479 (11) 400 (11) 298 (13) 5 or more years 2525 (57) 2506 (57) 1214 (53) Total 4431 3591 2307 Household asset score at recruitment Lowest third 1476 (33) 1195 (33) 821 (36) Middle 1483 (34) 1209 (34) 812 (35) Highest third 1472 (33) 1187 (33) 674 (29) Total 4429 3589 2305 Household surplus at recruitment Surplus 1199 (27) 976 (27) 585 (25) Equal 2367 (53) 1917 (53) 1254 (54) Occasional deficit 727 (16) 586 (16) 393 (17) Constant deficit 125 (3) 102 (3) 66 (3) Don’t know 11 (0.2) 8 (0.2) 7 (0.3) Total 4436 3625 2307 Trial randomizationa E+Fe30 739 (16.7) 608 (16.8) 389 (16.9) E+Fe60 738 (16.6) 610 (16.8) 387 (16.8) E+MMS 740 (16.7) 595 (16.4) 404 (17.5) U+Fe30 741 (16.7) 605 (16.7) 374 (16.2) U+Fe60 738 (16.6) 612 (16.9) 387 (16.8) U+MMS 740 (16.7) 595 (16.4) 366 (15.9) Characteristic . Levels . Recruitment . Live birth . 12-14 years . (n 3625) . (n 2307) . Total 4415 3587 2306 Age at recruitment <20 711 (16) 563 (16) 331 (14) 20–29 2521 (57) 2067 (57) 1312 (57) >30 1183 (27) 957 (27) 663 (29) 4416 3619 2307 Parity First 1479 (33) 1184 (33) 662 (29) Second or more 2937 (67) 2435 (67) 1645 (71) Total 4462 3580 2300 BMI kg/m2 <18.5 1228 (28) 1000 (28) 669 (29) >18.5 3234 (72) 2580 (72) 1631 (71) Total 4431 3591 2307 Educational level No schooling 1427 (32) 1135 (32) 795 (34) 1–4 years 479 (11) 400 (11) 298 (13) 5 or more years 2525 (57) 2506 (57) 1214 (53) Total 4431 3591 2307 Household asset score at recruitment Lowest third 1476 (33) 1195 (33) 821 (36) Middle 1483 (34) 1209 (34) 812 (35) Highest third 1472 (33) 1187 (33) 674 (29) Total 4429 3589 2305 Household surplus at recruitment Surplus 1199 (27) 976 (27) 585 (25) Equal 2367 (53) 1917 (53) 1254 (54) Occasional deficit 727 (16) 586 (16) 393 (17) Constant deficit 125 (3) 102 (3) 66 (3) Don’t know 11 (0.2) 8 (0.2) 7 (0.3) Total 4436 3625 2307 Trial randomizationa E+Fe30 739 (16.7) 608 (16.8) 389 (16.9) E+Fe60 738 (16.6) 610 (16.8) 387 (16.8) E+MMS 740 (16.7) 595 (16.4) 404 (17.5) U+Fe30 741 (16.7) 605 (16.7) 374 (16.2) U+Fe60 738 (16.6) 612 (16.9) 387 (16.8) U+MMS 740 (16.7) 595 (16.4) 366 (15.9) Data are n/n (%). BMI, body mass index; U = start of food supplementation at the standard programme start (week 20); Fe30= 30 mg Fe with 400 µg folic acid; Fe60 = 60 mg Fe with 400 µg folic acid; MMS= 15 micronutrients including 30 mg Fe and 400 µg folic acid. a E = Early invitation to daily prenatal food supplementation (gestational week 9). Open in new tab Table 1. Background characteristics of the pregnant women at recruitment, for those with a live birth, and for those where the child also participated at follow-up at 12–14 years. MINIMat cohort, Bangladesh, 2001–17 Characteristic . Levels . Recruitment . Live birth . 12-14 years . (n 3625) . (n 2307) . Total 4415 3587 2306 Age at recruitment <20 711 (16) 563 (16) 331 (14) 20–29 2521 (57) 2067 (57) 1312 (57) >30 1183 (27) 957 (27) 663 (29) 4416 3619 2307 Parity First 1479 (33) 1184 (33) 662 (29) Second or more 2937 (67) 2435 (67) 1645 (71) Total 4462 3580 2300 BMI kg/m2 <18.5 1228 (28) 1000 (28) 669 (29) >18.5 3234 (72) 2580 (72) 1631 (71) Total 4431 3591 2307 Educational level No schooling 1427 (32) 1135 (32) 795 (34) 1–4 years 479 (11) 400 (11) 298 (13) 5 or more years 2525 (57) 2506 (57) 1214 (53) Total 4431 3591 2307 Household asset score at recruitment Lowest third 1476 (33) 1195 (33) 821 (36) Middle 1483 (34) 1209 (34) 812 (35) Highest third 1472 (33) 1187 (33) 674 (29) Total 4429 3589 2305 Household surplus at recruitment Surplus 1199 (27) 976 (27) 585 (25) Equal 2367 (53) 1917 (53) 1254 (54) Occasional deficit 727 (16) 586 (16) 393 (17) Constant deficit 125 (3) 102 (3) 66 (3) Don’t know 11 (0.2) 8 (0.2) 7 (0.3) Total 4436 3625 2307 Trial randomizationa E+Fe30 739 (16.7) 608 (16.8) 389 (16.9) E+Fe60 738 (16.6) 610 (16.8) 387 (16.8) E+MMS 740 (16.7) 595 (16.4) 404 (17.5) U+Fe30 741 (16.7) 605 (16.7) 374 (16.2) U+Fe60 738 (16.6) 612 (16.9) 387 (16.8) U+MMS 740 (16.7) 595 (16.4) 366 (15.9) Characteristic . Levels . Recruitment . Live birth . 12-14 years . (n 3625) . (n 2307) . Total 4415 3587 2306 Age at recruitment <20 711 (16) 563 (16) 331 (14) 20–29 2521 (57) 2067 (57) 1312 (57) >30 1183 (27) 957 (27) 663 (29) 4416 3619 2307 Parity First 1479 (33) 1184 (33) 662 (29) Second or more 2937 (67) 2435 (67) 1645 (71) Total 4462 3580 2300 BMI kg/m2 <18.5 1228 (28) 1000 (28) 669 (29) >18.5 3234 (72) 2580 (72) 1631 (71) Total 4431 3591 2307 Educational level No schooling 1427 (32) 1135 (32) 795 (34) 1–4 years 479 (11) 400 (11) 298 (13) 5 or more years 2525 (57) 2506 (57) 1214 (53) Total 4431 3591 2307 Household asset score at recruitment Lowest third 1476 (33) 1195 (33) 821 (36) Middle 1483 (34) 1209 (34) 812 (35) Highest third 1472 (33) 1187 (33) 674 (29) Total 4429 3589 2305 Household surplus at recruitment Surplus 1199 (27) 976 (27) 585 (25) Equal 2367 (53) 1917 (53) 1254 (54) Occasional deficit 727 (16) 586 (16) 393 (17) Constant deficit 125 (3) 102 (3) 66 (3) Don’t know 11 (0.2) 8 (0.2) 7 (0.3) Total 4436 3625 2307 Trial randomizationa E+Fe30 739 (16.7) 608 (16.8) 389 (16.9) E+Fe60 738 (16.6) 610 (16.8) 387 (16.8) E+MMS 740 (16.7) 595 (16.4) 404 (17.5) U+Fe30 741 (16.7) 605 (16.7) 374 (16.2) U+Fe60 738 (16.6) 612 (16.9) 387 (16.8) U+MMS 740 (16.7) 595 (16.4) 366 (15.9) Data are n/n (%). BMI, body mass index; U = start of food supplementation at the standard programme start (week 20); Fe30= 30 mg Fe with 400 µg folic acid; Fe60 = 60 mg Fe with 400 µg folic acid; MMS= 15 micronutrients including 30 mg Fe and 400 µg folic acid. a E = Early invitation to daily prenatal food supplementation (gestational week 9). Open in new tab How often have they been followed up? After recruitment at around gestational week 8, women were assessed at gestational weeks 9, 14, 19 and 30 (measurements listed in Table 2). Mothers and newborn children were assessed at birth, followed by monthly examinations of the dyads up to 12 months. During the second year of life, the children were assessed quarterly. The next follow-ups were performed at 4.5 and 10 years of age. At 12.3 to 14.5 years of age, assessments of pubertal development, anthropometry and body composition were performed. Table 2. Measurements in the MINIMat trial and follow-up . Pregnancy . Birth . 0–24 months . 4.5 years . 10 years . 12–14 years . Interventions  Compliance food supplementation +  Compliance micronutrients (eDEM) + Anthropometry etc.  Gestational age (LMP, ultrasound) +  Maternal anthropometry + + + + +  Fetal growth (ultrasound) +  Child anthropometry + + + + +  Skinfolds, body composition + + + Child development  Motor and cognitive development + + +  Language development + +  Motor milestones +  Mother-child interaction +  IQ + +  Home environment + + + Infections, immune function  Morbidity + + + + +  Thymus size (ultrasound) + Food, diet, feeding  Food security + + + + +  Diet + + +  Diet diversity + +  Breastfeeding + + Reproductive history  Previous pregnancies, outcomes +  Follow-up to next pregnancy + + Social conditions  Household asset score + + + +  Parents’ education +  Parents’ occupation + + +  Marital status  Partner violence + +  Depressive symptoms/distress + + + Biomarkers  Haematology + + + +  Micronutrients + + + +  Oxidative stress +  Toxic exposure (urine) + + + +  Metabolic markers + +  Blood pressure + + +  Salivary cortisol + + . Pregnancy . Birth . 0–24 months . 4.5 years . 10 years . 12–14 years . Interventions  Compliance food supplementation +  Compliance micronutrients (eDEM) + Anthropometry etc.  Gestational age (LMP, ultrasound) +  Maternal anthropometry + + + + +  Fetal growth (ultrasound) +  Child anthropometry + + + + +  Skinfolds, body composition + + + Child development  Motor and cognitive development + + +  Language development + +  Motor milestones +  Mother-child interaction +  IQ + +  Home environment + + + Infections, immune function  Morbidity + + + + +  Thymus size (ultrasound) + Food, diet, feeding  Food security + + + + +  Diet + + +  Diet diversity + +  Breastfeeding + + Reproductive history  Previous pregnancies, outcomes +  Follow-up to next pregnancy + + Social conditions  Household asset score + + + +  Parents’ education +  Parents’ occupation + + +  Marital status  Partner violence + +  Depressive symptoms/distress + + + Biomarkers  Haematology + + + +  Micronutrients + + + +  Oxidative stress +  Toxic exposure (urine) + + + +  Metabolic markers + +  Blood pressure + + +  Salivary cortisol + + Open in new tab Table 2. Measurements in the MINIMat trial and follow-up . Pregnancy . Birth . 0–24 months . 4.5 years . 10 years . 12–14 years . Interventions  Compliance food supplementation +  Compliance micronutrients (eDEM) + Anthropometry etc.  Gestational age (LMP, ultrasound) +  Maternal anthropometry + + + + +  Fetal growth (ultrasound) +  Child anthropometry + + + + +  Skinfolds, body composition + + + Child development  Motor and cognitive development + + +  Language development + +  Motor milestones +  Mother-child interaction +  IQ + +  Home environment + + + Infections, immune function  Morbidity + + + + +  Thymus size (ultrasound) + Food, diet, feeding  Food security + + + + +  Diet + + +  Diet diversity + +  Breastfeeding + + Reproductive history  Previous pregnancies, outcomes +  Follow-up to next pregnancy + + Social conditions  Household asset score + + + +  Parents’ education +  Parents’ occupation + + +  Marital status  Partner violence + +  Depressive symptoms/distress + + + Biomarkers  Haematology + + + +  Micronutrients + + + +  Oxidative stress +  Toxic exposure (urine) + + + +  Metabolic markers + +  Blood pressure + + +  Salivary cortisol + + . Pregnancy . Birth . 0–24 months . 4.5 years . 10 years . 12–14 years . Interventions  Compliance food supplementation +  Compliance micronutrients (eDEM) + Anthropometry etc.  Gestational age (LMP, ultrasound) +  Maternal anthropometry + + + + +  Fetal growth (ultrasound) +  Child anthropometry + + + + +  Skinfolds, body composition + + + Child development  Motor and cognitive development + + +  Language development + +  Motor milestones +  Mother-child interaction +  IQ + +  Home environment + + + Infections, immune function  Morbidity + + + + +  Thymus size (ultrasound) + Food, diet, feeding  Food security + + + + +  Diet + + +  Diet diversity + +  Breastfeeding + + Reproductive history  Previous pregnancies, outcomes +  Follow-up to next pregnancy + + Social conditions  Household asset score + + + +  Parents’ education +  Parents’ occupation + + +  Marital status  Partner violence + +  Depressive symptoms/distress + + + Biomarkers  Haematology + + + +  Micronutrients + + + +  Oxidative stress +  Toxic exposure (urine) + + + +  Metabolic markers + +  Blood pressure + + +  Salivary cortisol + + Open in new tab Of the 5880 pregnant women who were initially assessed, 4436 were eligible and consented to participate in the trial (Figure 1). These women had 3625 live births, of which 3560 were singleton live births. For the follow-up at 4.5 years of age, 2851 children participated (80% of the single live births). For the follow-up at 12.3–14.5 years of age, 2307 children participated (64% of the singleton live births). Out-migration and unavailability because of other activities were the main explanations for non-participation by these schoolchildren. Children who did not attend the latest follow-up more frequently came from wealthier households (asset score in the highest tertile, 42% lost to follow up, as compared with the lower two tertiles with 31% attrition). They also more frequently had mothers with some formal education (37% attrition) in comparison with children of mothers with no formal schooling (30% attrition). What has been measured? This cohort was established within the Matlab HDSS, which implies that the cohort data may be linked to family, sociodemographic and reproductive information on mothers and grandmothers if these relatives stayed within the Matlab HDSS the years before. Valid data on important demographic and health-related events are available, such as last menstrual period, date of birth, out-migration, in-migration, death and causes of death. The pregnancy cohort was followed from about gestational week 9 (Table 1). Fetal growth was monitored by repeated ultrasound assessments. Infant and young child anthropometric measurements were carefully monitored up to 2 years of age and after that at ages 4.5 and 10.0 years. Body composition was assessed by bioelectrical impedance measurements at 4.5, 10.0 and 12.3–14.5 years of age. The quality of maternal-infant interaction was measured at 4 months of age. Child development was assessed when the participants were 7 and 18 months as well as 5 and 10 years old. These assessments included Bayley Scales of Infant Development, behaviour ratings, language development, IQ and motor milestones. Morbidity was assessed by 7-day recalls throughout infancy up to 2 years of age, and after that at 4.5 and 10 years of age. At 8, 24 and 52 weeks of age, thymic volume was assessed by ultrasound. The social conditions of the members of the MINIMat cohort have been carefully and repeatedly assessed. These indicators include household characteristics (including asset scores), food security, educational level of parents and a wide range of characteristics of the mother and the child. Participating women also responded to questions regarding exposure to physical, sexual and emotional violence and levels of controlling behaviour. Women’s levels of stress were measured. Haemoglobin has been measured repeatedly in mothers and their children. In sub-samples, assessments have been made of iron and other micronutrient status, markers of oxidative stress, toxic exposure including arsenic, metabolic markers and salivary cortisol (Table 2). What has been found? Key findings and publications To date, more than 100 scientific publications and 20 PhD dissertations have been completed based on data from the MINIMat cohort. See [http://www.kbh.uu.se/Research/International+Child+Health+and+Nutrition/main-fields-of-research-and-projects/global-nutrition] for a complete publication list. There was a pronounced effect on infant survival in line with the hypothesis. With early invitation to food supplementation with MMS, this group had an infant mortality rate of 16.8 per 1000 live births, versus 44.1 per 1000 live births among the group with the usual later invitation to food supplementation with 60 mg of iron and 400 µg of folic acid [hazard ratio (HR) 0.38, 95% confidence interval (CI) 0.18–0.78], Figure 2.17 Adjusted maternal haemoglobin concentrations at 30 weeks’ gestation were 115.0 g/L with no differences among treatment groups. Mean birthweight was 2694 g with no difference among groups. An equity analysis revealed that the intervention reduced the gap in child survival between social groups.18 The interventions were also judged to be cost-effective in reducing mortality.19 The early invitation to food supplementation in pregnancy reduced the occurrence of stunting at 0–54 months of age.20 In contrast, prenatal MMS increased the proportion of stunting,20 a finding that also was supported by corresponding IgF1 levels.21 Figure 2. Open in new tabDownload slide Infant deaths per 1000 live births in the MINIMat trial with 95% confidence intervals. E = early (around week 9) invitation to daily balanced food supplementation, 600 kcal, 6 days per week; U = the usual timing (around week 20) of start of this food supplementation; Fe30Fol = 30 mg iron and 400 µg of folic acid; Fe60Fol = 60 mg of iron and 400 µg of folic acid; MMS = multiple micronutrients containing a daily allowance of 15 micronutrients, including 30 mg of iron and 400 µg of folic acid. Figure 2. Open in new tabDownload slide Infant deaths per 1000 live births in the MINIMat trial with 95% confidence intervals. E = early (around week 9) invitation to daily balanced food supplementation, 600 kcal, 6 days per week; U = the usual timing (around week 20) of start of this food supplementation; Fe30Fol = 30 mg iron and 400 µg of folic acid; Fe60Fol = 60 mg of iron and 400 µg of folic acid; MMS = multiple micronutrients containing a daily allowance of 15 micronutrients, including 30 mg of iron and 400 µg of folic acid. The prenatal supplementations had also effects on metabolic markers at 4.5 years of age.21 These effects were judged to be of public health importance and suggest programming effects in early fetal life. What are the main strengths and weaknesses? The MINIMat cohort is a well-characterized cohort with information from early pregnancy to puberty. Plans are under way for a new update at 15 years of age. The randomized prenatal food and micronutrient interventions have enabled analyses of short- and medium-term effects and will provide opportunities to address the question whether long-term (adulthood) negative consequences of early nutritional insults may be prevented by this type of prenatal intervention. The context of this cohort is characterized by widespread undernutrition of mothers and children, and almost two-thirds of these children were born small for their gestational-age. This is important to consider when making inferences to other low-income settings. There was no follow-up between 5 and 10 years. For some rare outcomes, the sample size may be too small. Can I get hold of the data? Where can I find out more? Several researchers and research groups have so far analysed MINIMat data to address new research questions that could be answered by the available databases. If you are interested, please contact Qazi Sadeq-ur Rahman at [[emailprotected]]. Profile in a nutshell The rationale for the MINIMat cohort in Bangladesh was the high prevalence of maternal malnutrition and intrauterine growth restriction. It was hypothesized that prenatal multiple micronutrient supplementation, as well as an early invitation to a daily food supplementation, would increase maternal haemoglobin concentration, birthweight and infant survival, and that a combination of these interventions would further improve these outcomes. Secondary outcomes were linked to the Developmental Origin of Health and Disease perspective. A total of 4436 women were recruited from 2001 to 2003. There were 3625 live births, and 2307 children participated in the latest follow-up at 12.3–14.5 years of age. Fetal growth was monitored. Further assessments include child growth, body composition, child development, morbidity and mortality. Sociodemographic information included household characteristics, food security and educational levels. Haemoglobin was measured repeatedly. Sub-sample information is available on iron and other micronutrients, markers of oxidative stress, toxic exposure, metabolic markers and salivary cortisol. The early invitation to prenatal food supplementation combined with multiple micronutrients lowered infant mortality substantially. Among the 100 publications to date, several reports address feeding, growth, development, micronutrients, immunological development, social conditions and the Developmental Origins of Adult Disease perspective. If interested in the MINIMat cohort data, please contact [[emailprotected]]. Funding The MINIMat cohort was funded by icddr, b, the United Nations Children's Fund, the Swedish International Development Cooperation Agency, the UK Medical Research Council, the Swedish Research Council, the United Kingdom Department for International Development, the Japan Society for the Promotion of Science, the Child Health and Nutrition Research Initiative, Uppsala University and the United States Agency for International Development. The funders had no role in any aspect of the study design or resulting papers. Acknowledgements We are grateful to families, mothers and children for their long-term commitment to the MINIMat study. Conflict of interest: None declared. References 1 Underwood BA. Health and nutrition in women, infants, and children: overview of the global situation and the Asian enigma . Nutr Rev 2002 ; 60(5 Pt 2) : S7 – 13 . Google Scholar Crossref Search ADS WorldCat 2 Shrimpton R , Victora CG, de Onis M, Lima RC, Blössner M, Clugston G. Worldwide timing of growth faltering: implications for nutritional interventions . Pediatrics 2001 ; 107 : E75. Google Scholar Crossref Search ADS PubMed WorldCat 3 Christian P , Lee SE, Donahue Angel M et al. Risk of childhood undernutrition related to small-for-gestational age and preterm birth in low- and middle-income countries . Int J Epidemiol 2013 ; 42: 1340 – 55 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Victora CG , de Onis M, Hallal PC, Blössner M, Shrimpton R. Worldwide timing of growth faltering: revisiting implications for interventions . Pediatrics 2010 ; 125 : e473 – 80 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Leroy JL , Ruel M, Habicht J-P, Frongillo EA. Linear growth deficit continues to accumulate beyond the first 1000 days in low- and middle-income countries: global evidence from 51 national surveys . J Nutr 2014 ; 144 : 1460 – 66 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Imdad A , Yakoob MY, Bhutta ZA. The effect of folic acid, protein energy and multiple micronutrient supplements in pregnancy on stillbirths . BMC Public Health 2011 ; 11 : S4. Google Scholar Crossref Search ADS PubMed WorldCat 7 Bukowski R , Smith GCS, Malone FD et al. Fetal growth in early pregnancy and risk of delivering low birth weight infant: prospective cohort study . BMJ 2007 ; 334 : 836. Google Scholar Crossref Search ADS PubMed WorldCat 8 Merialdi M , Carroli G, Villar J et al. Nutritional interventions during pregnancy for the prevention or treatment of impaired fetal growth: an overview of randomized controlled trials . J Nutr 2003 ; 133 : 1626 – 31S . Google Scholar Crossref Search ADS WorldCat 9 Ramakrishnan U. Nutrition and low birthweight: from research to practice . Am J Clin Nutr 2004 ; 79 : 17 – 21 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Cetin I , Alvino G. Intrauterine growth restriction: implications for placental metabolism and transport. A review . Placenta 2009 ; 30 : 77 – 82 . Google Scholar Crossref Search ADS WorldCat 11 UNICEF, WHO, UNU . Composition of a Multiple-micronutrient Supplement to Be Used in Pilot Programmes Among Pregnant Women in Developing Countries. New York, NY: UNICEF, 1999 . 12 Margetts BM , Fall CH, Ronsmans C, Allen LH, Fisher DJ. Multiple micronutrient supplementation during pregnancy in low-income countries: review of methods and characteristics of studies included in the meta-analyses . Food Nutr Bull 2009 ; 30 : S517 – 26 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Barker DJ. Early growth and cardiovascular disease . Arch Dis Child 1999 ; 80 : 305 – 07 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Barouki R , Gluckman PD, Grandjean P, Hanson M, Heindel JJ. Developmental origins of non-communicable disease: implications for research and public health . Environ Health 2012 ; 11 : 42. Google Scholar Crossref Search ADS PubMed WorldCat 15 Gluckman PD , Hanson MA, Buklijas T. A conceptual framework for the developmental origins of health and disease . J Dev Orig Health Dis 2010 ; 1 : 6 – 18 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Hanson M , Godfrey KM, Lillycrop KA, Burdge GC, Gluckman PD. Developmental plasticity and developmental origins of non-communicable disease: theoretical considerations and epigenetic mechanisms . Prog Biophys Mol Biol 2011 ; 106 : 272 – 80 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Persson LÅ , Arifeen S, Ekström E-C et al. Effects of prenatal micronutrient and early food supplementation on maternal hemoglobin, birth weight, and infant mortality among children in Bangladesh . JAMA 2012 ; 307 : 2050 – 59 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Shaheen R , Streatfield PK, Naved RT, Lindholm L, Persson LÅ. Equity in adherence to and effect of prenatal food and micronutrient supplementation on child mortality: results from the MINIMat randomized trial, Bangladesh . BMC Public Health 2014 ; 14 : 5. Google Scholar Crossref Search ADS PubMed WorldCat 19 Shaheen R , Persson LÅ, Ahmed S, Streatfield PK, Lindholm L. Cost-effectiveness of invitation to food supplementation early in pregnancy combined with multiple micronutrients on infant survival: analysis of data from MINIMat randomized trial, Bangladesh . BMC Pregnancy Childbirth 2015 ; 15 : 125. Google Scholar Crossref Search ADS PubMed WorldCat 20 Khan AI , Kabir I, Ekstrom E-C et al. Effects of prenatal food and micronutrient supplementation on child growth from birth to 54 months of age: a randomized trial in Bangladesh . Nutr J 2011 ; 10 : 134. Google Scholar Crossref Search ADS PubMed WorldCat 21 Ekstrom E-C , Lindström E, Raqib R et al. Effects of prenatal micronutrient and early food supplementation on metabolic status of the offspring at 4.5 years of age. The MINIMat randomized trial in rural Bangladesh . Int J Epidemiol 2016 ; 45 : 1656 – 67 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [emailprotected] © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association.

Cohort profile: The Boston Hospital Workers Health Study (BHWHS)

Sabbath, Erika, L;Hashimoto,, Dean;Boden, Leslie, I;Dennerlein, Jack, T;Williams, Jessica A, R;Hopcia,, Karen;Orechia,, Theresa;Tripodis,, Yorghos;Stoddard,, Anne;Sorensen,, Glorian

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy164pmid: 30107500

Profile in a nutshell The Boston Hospital Workers Health Study (BHWHS) is a longitudinal, integrated open cohort consisting of employer data and survey data on approximately 16 000 hospital-based patient care workers aged 18-72, based at two hospitals in Massachusetts, USA. It represents an employer-researcher intellectual partnership between Partners HealthCare, and the Harvard T.H. Chan Centre for Work, Health, and Wellbeing. Data include individual-level, employer-provided data on health care spending, workplace injury, staffing, human resources and workload, and survey data on a subset of workers (on working conditions, self-reported health conditions and health behaviours). Workers are grouped into work-groups or units, allowing multilevel in addition to longitudinal analyses. Although the data are restricted in how they can be distributed (although the team is open to collaboration from outside researchers), the BHWHS is a model of the public health benefits of data collaborations between researchers and employers. This cohort profile serves as a guide to other researcher-employer teams looking to establish a similar database and collaboration. Why was the cohort set up ? Protecting and promoting the health of hospital workers is both an occupational health priority and a public health imperative. Health care workers are the fastest-growing segment of the US labour force.1,2 Their working conditions present many health risks, which can affect them, their families, their patients and their employers.3–6 Though hospitals routinely collect employee payroll, injury, health and survey data, these databases are seldom integrated with each other and are rarely available to researchers outside the organization. This disconnect impedes efficient and organizationally relevant occupational health research and evidence-based practice regarding this high-priority workforce. To address that gap, the Boston Hospital Workers Health Study (BHWHS), established in 2006, integrates several employee databases with worker surveys in two large hospitals that are part of the same health system. BHWHS resulted from a partnership between the Harvard T.H. Chan School of Public Health [National Institute for Occupational Safety and Health (NIOSH)-funded Centre for Work, Health & Well-being (hereafter referred to as ‘the Centre’)] and two academic and teaching hospitals that are part of Partners HealthCare (hereafter referred to as ‘Partners’). The BHWHS is funded by NIOSH and is based in Boston, Massachusetts, USA. BHWHS was created as a way for both the Centre and Partners to mutually advance their research and practice goals. At the study’s inception in 2006 (systematic enrolment of new employees into the database was not fully realized until 2009), Partners was adopting more data-driven practices to inform decision making on organizational change, including the occupational health department. As a result of the changing focus, Partners created sophisticated employee databases. Concurrently, the newly formed Centre aimed to increase the evidence for a more holistic approach to occupational safety and health, by integrating traditional worker health protection functions with other worker health promotion and disease prevention activities.7 Partners provided a rich source of data that could address the Centre’s core research questions around work organization, worker health and safety, and integration of health protection and promotion activities locally. The Centre provided Partners with the opportunity to have employee data analysed by occupational and social epidemiologists to inform policy and practice at the hospital. With this alignment of incentives, the chief of occupational safety and health at Partners (D.H.) and the principal investigator of the Centre (G.S.) formed a partnership to create the study. A key feature of the BHWHS continues to be close collaboration between Partners and Centre; Partners investigators (D.H. and K.H.) are full intellectual partners and co-investigators on the study. BHWHS consists of a longitudinal, multidisciplinary database that integrates existing employer data with surveys of a subset of workers. Data are linked temporally at the individual and work-group levels. Using precise dates within each dataset, the team can conduct analyses of time-varying exposures and outcomes. The research questions for the BHWHS have evolved over the three funding periods of the Centre. For the first two periods, from 2007 to 2011 and from 2011 to 2016, the BHWHS aimed to determine the factors in the work organization and environment that increased risk of musculoskeletal disorder (MSD) symptoms and risk-related health behaviours and, through pilot interventions, to estimate the feasibility and efficacy of integrated policies, programmes, and practices that reduce these risks. In the current phase (2016–21), the research questions relate to the role of the work organization in shaping outcomes for workers, patients and the employer. Who is in the cohort? Overall study BHWHS is a prospective cohort study. All employees at the two hospitals who are classified as ‘patient care services’ workers are automatically enrolled in the cohort and are followed throughout their tenure. If employment is terminated or employees leave, their data remain in BHWHS; if they are re-hired, their original record is re-opened. Thus, the number of participants in BHWHS grows annually even as the number of active employees remains relatively stable (Figure 1). As of 2017, the cohort has 8200 active employees and 15 965 total employees. The study has been approved by the human subjects committee at Partners. Because all patient care services workers are automatically enrolled in BHWHS, participation is 100%, and attrition or loss to follow-up takes place only through employment separation. Figure 1 View largeDownload slide Number of participants in BHWHS, from 2009-present and projected through 2021. ‘Continuing active’ refers to workers who remain employed since the previous year, ‘new’ refers to workers added to BHWHS that year, ‘lost that year’ refers to workers who separated from the employer during that year, and ‘previously lost’ refers to workers who in past years have separated from the employer but nevertheless remain in the database. Workers accruing in the latter category over time drives the year-over-year increase in cohort size. (see separate file for figure). Figure 1 View largeDownload slide Number of participants in BHWHS, from 2009-present and projected through 2021. ‘Continuing active’ refers to workers who remain employed since the previous year, ‘new’ refers to workers added to BHWHS that year, ‘lost that year’ refers to workers who separated from the employer during that year, and ‘previously lost’ refers to workers who in past years have separated from the employer but nevertheless remain in the database. Workers accruing in the latter category over time drives the year-over-year increase in cohort size. (see separate file for figure). BHWHS has a natural multilevel structure with the patient care unit as a critical organizational feature. Each unit has its own subculture with its own leaders and a consistent set of workers. Across the two hospitals, workers are nested within 106 units. Our data allow us to group workers to their assigned unit, which allows us to evaluate the impact of both individual- and unit-level factors on outcomes. Social and demographic characteristics of BHWHS participants have remained stable over time (Table 1). Table 1. Social and demographic characteristics of active members of the BHWHS database at baseline and in 2016, using data from the human resources portion of the database 2010 (n = 8076) 2016 (n = 9172) Mean age 41 41 Sex  Male 840 (10%) 1014 (11%)  Female 7236 (90%) 8158 (89%) Job titlea  Nurse 5114 (63%) 5836 (61%)  PCA 1178 (15%) 1278 (13%)  Other 1800 (22%) 2465 (26%) 2010 (n = 8076) 2016 (n = 9172) Mean age 41 41 Sex  Male 840 (10%) 1014 (11%)  Female 7236 (90%) 8158 (89%) Job titlea  Nurse 5114 (63%) 5836 (61%)  PCA 1178 (15%) 1278 (13%)  Other 1800 (22%) 2465 (26%) a One employee can have multiple job categories for the year if they change jobs during the year or work two different jobs, so the sum here is slightly different than the overall total. PCA, patient care associate. Table 1. Social and demographic characteristics of active members of the BHWHS database at baseline and in 2016, using data from the human resources portion of the database 2010 (n = 8076) 2016 (n = 9172) Mean age 41 41 Sex  Male 840 (10%) 1014 (11%)  Female 7236 (90%) 8158 (89%) Job titlea  Nurse 5114 (63%) 5836 (61%)  PCA 1178 (15%) 1278 (13%)  Other 1800 (22%) 2465 (26%) 2010 (n = 8076) 2016 (n = 9172) Mean age 41 41 Sex  Male 840 (10%) 1014 (11%)  Female 7236 (90%) 8158 (89%) Job titlea  Nurse 5114 (63%) 5836 (61%)  PCA 1178 (15%) 1278 (13%)  Other 1800 (22%) 2465 (26%) a One employee can have multiple job categories for the year if they change jobs during the year or work two different jobs, so the sum here is slightly different than the overall total. PCA, patient care associate. Survey subsample A subset of BHWHS workers are periodically sampled and asked to complete surveys to measure constructs not available in the employee databases. These survey responses are linked to BHWHS administrative data at the individual worker level. Thus far, three surveys have been completed (in 2009, 2012 and 2014) and a fourth is planned for 2018 (Table 2). Table 2. Response characteristics and rates for periodic survey subsamples 2009 2012 2014 2018 (planned) Sampled 2000 2133 1968a; 1840 longitudinal, 128 new 2800 (est.) Completed 1572 1595 Overall 1409; TBD 1301 longitudinal, 108 new Response rate 79% 75% Overall 72%; TBD 71% longitudinal, 84% new Type of sampling Cross-sectional Cross-sectional Longitudinal from 2012 + 128 new participants Cross-sectional + follow-up of longitudinal participantsb 2009 2012 2014 2018 (planned) Sampled 2000 2133 1968a; 1840 longitudinal, 128 new 2800 (est.) Completed 1572 1595 Overall 1409; TBD 1301 longitudinal, 108 new Response rate 79% 75% Overall 72%; TBD 71% longitudinal, 84% new Type of sampling Cross-sectional Cross-sectional Longitudinal from 2012 + 128 new participants Cross-sectional + follow-up of longitudinal participantsb a At the time of the 2014 follow-up survey, 287 members of the original survey cohort no longer met the eligibility criteria and were not included. TBD, to be decided. b Random sample of 2000, plus follow-up of still-eligible longitudinal participants not captured by random sample (estimated n=800). Table 2. Response characteristics and rates for periodic survey subsamples 2009 2012 2014 2018 (planned) Sampled 2000 2133 1968a; 1840 longitudinal, 128 new 2800 (est.) Completed 1572 1595 Overall 1409; TBD 1301 longitudinal, 108 new Response rate 79% 75% Overall 72%; TBD 71% longitudinal, 84% new Type of sampling Cross-sectional Cross-sectional Longitudinal from 2012 + 128 new participants Cross-sectional + follow-up of longitudinal participantsb 2009 2012 2014 2018 (planned) Sampled 2000 2133 1968a; 1840 longitudinal, 128 new 2800 (est.) Completed 1572 1595 Overall 1409; TBD 1301 longitudinal, 108 new Response rate 79% 75% Overall 72%; TBD 71% longitudinal, 84% new Type of sampling Cross-sectional Cross-sectional Longitudinal from 2012 + 128 new participants Cross-sectional + follow-up of longitudinal participantsb a At the time of the 2014 follow-up survey, 287 members of the original survey cohort no longer met the eligibility criteria and were not included. TBD, to be decided. b Random sample of 2000, plus follow-up of still-eligible longitudinal participants not captured by random sample (estimated n=800). The initial survey (2009) was a random sample of employees. Beginning with the 2012 survey, the study follows a second random sample of workers longitudinally while also refreshing the sample with workers randomly selected at each wave. Refreshing the sample mitigates healthy worker survivor effects8 and makes for a sample that is more reflective of the distribution of organizational tenure within the hospitals. For the survey, sampled workers are invited to participate via e-mail and are offered a gift card as an incentive to complete the survey electronically. E-mail reminders are sent to non-respondents at prescribed intervals, and paper versions are thereafter sent to non-respondents. All workers provide informed consent before survey participation: see Table 2 for survey response rates and Table 3 for worker characteristics at each survey point. Table 3. Social, demographic and occupational characteristics (n/percent) of the subset of workers who were surveyed at each wave. Note that 2009 was a cross-sectional survey. Starting in 2012, workers were followed longitudinally and the sample was refreshed with newly hired workers in 2014. Due to missing data on some variables, not all categories sum to the overall n Surveyed in 2009 (n = 1572) Surveyed in 2012 (n = 1595) Surveyed in 2014 (n = 1409) Age (mean/SD) 41.3 (11.7) 40.9 (11.9) 41.5 (11.9) Sex  Male 143 (10%) 112 (7%) 89 (6%)  Female 1369 (90%) 1462 (93%) 1298 (94%) Race/ethnicity  Non-Hispanic White 1185 (79%) 1273 (81%) 1122 (80%)  Non-Hispanic Black 159 (11%) 136 (9%) 109 (8%)  Hispanic 65 (4%) 64 (4%) 55 (4%)  Mixed race/other 89 (6%) 107 (7%) 112 (8%) Job title  Staff nurse 1103 (70%) 1321 (83%) 1207 (86%)  Patient care associate 127 (8%) 141 (9%) 142 (10%)  Other 335 (21%) 129 (8%) 56 (4%) Surveyed in 2009 (n = 1572) Surveyed in 2012 (n = 1595) Surveyed in 2014 (n = 1409) Age (mean/SD) 41.3 (11.7) 40.9 (11.9) 41.5 (11.9) Sex  Male 143 (10%) 112 (7%) 89 (6%)  Female 1369 (90%) 1462 (93%) 1298 (94%) Race/ethnicity  Non-Hispanic White 1185 (79%) 1273 (81%) 1122 (80%)  Non-Hispanic Black 159 (11%) 136 (9%) 109 (8%)  Hispanic 65 (4%) 64 (4%) 55 (4%)  Mixed race/other 89 (6%) 107 (7%) 112 (8%) Job title  Staff nurse 1103 (70%) 1321 (83%) 1207 (86%)  Patient care associate 127 (8%) 141 (9%) 142 (10%)  Other 335 (21%) 129 (8%) 56 (4%) SD, standard deviation. Table 3. Social, demographic and occupational characteristics (n/percent) of the subset of workers who were surveyed at each wave. Note that 2009 was a cross-sectional survey. Starting in 2012, workers were followed longitudinally and the sample was refreshed with newly hired workers in 2014. Due to missing data on some variables, not all categories sum to the overall n Surveyed in 2009 (n = 1572) Surveyed in 2012 (n = 1595) Surveyed in 2014 (n = 1409) Age (mean/SD) 41.3 (11.7) 40.9 (11.9) 41.5 (11.9) Sex  Male 143 (10%) 112 (7%) 89 (6%)  Female 1369 (90%) 1462 (93%) 1298 (94%) Race/ethnicity  Non-Hispanic White 1185 (79%) 1273 (81%) 1122 (80%)  Non-Hispanic Black 159 (11%) 136 (9%) 109 (8%)  Hispanic 65 (4%) 64 (4%) 55 (4%)  Mixed race/other 89 (6%) 107 (7%) 112 (8%) Job title  Staff nurse 1103 (70%) 1321 (83%) 1207 (86%)  Patient care associate 127 (8%) 141 (9%) 142 (10%)  Other 335 (21%) 129 (8%) 56 (4%) Surveyed in 2009 (n = 1572) Surveyed in 2012 (n = 1595) Surveyed in 2014 (n = 1409) Age (mean/SD) 41.3 (11.7) 40.9 (11.9) 41.5 (11.9) Sex  Male 143 (10%) 112 (7%) 89 (6%)  Female 1369 (90%) 1462 (93%) 1298 (94%) Race/ethnicity  Non-Hispanic White 1185 (79%) 1273 (81%) 1122 (80%)  Non-Hispanic Black 159 (11%) 136 (9%) 109 (8%)  Hispanic 65 (4%) 64 (4%) 55 (4%)  Mixed race/other 89 (6%) 107 (7%) 112 (8%) Job title  Staff nurse 1103 (70%) 1321 (83%) 1207 (86%)  Patient care associate 127 (8%) 141 (9%) 142 (10%)  Other 335 (21%) 129 (8%) 56 (4%) SD, standard deviation. How often have they been followed up? Employee data on current study participants are updated quarterly or annually by the data manager (T.O.) from each of the Partners data sources; the data sources themselves accrue data continuously. As all records within each dataset are dated, once the sources are integrated within the BHWHS database, each employee has complete longitudinal data for the period in which they are or were employed. Table 4 presents characteristics of those who have remained, left and joined the hospital workforces over the years. Among those who were employed in 2009 (study inception), those who left the workforce between 2010 and 2016 are roughly the same age as those who remained (respective birth years 1969 versus 1968). However, those who were hired between 2010 and 2016 and remain employed are younger on average (mean birth year 1984) than those who were hired during this period but left (mean birth year 1976). Among those in the baseline cohort who left, 13% retired, 43% separated either voluntarily or involuntarily, less than 1% died and the rest are unknown. The proportion of ‘unknown’ terminations increased from those employed at baseline to those leaving during the next 6 years and is a limitation of the administrative data. Table 4. Characteristics of workers remaining in the cohort and those lost to follow-up, using data from the human resources portion of the database Employed continuously 2010-16 (n = 4496, 56%) Original cohort members who left 2010–16 (n = 3599, 44%) Hired between 2010 and 2016, remain employed (n = 3626, 56%) Hired between 2010 and 2016 and left (n = 2886, 44%) Birth year 1968 1969 1984 1982 Percent female 92% 87% 87% 83% Reason left − −  Retireda 452 (13%)a 23 (1%)a  Terminated 1539 (43%) 1262 (44%)  Died 30 (<1%) 2 (<1%)  Unknown 1578 (44%) 1589 (55%) Employed continuously 2010-16 (n = 4496, 56%) Original cohort members who left 2010–16 (n = 3599, 44%) Hired between 2010 and 2016, remain employed (n = 3626, 56%) Hired between 2010 and 2016 and left (n = 2886, 44%) Birth year 1968 1969 1984 1982 Percent female 92% 87% 87% 83% Reason left − −  Retireda 452 (13%)a 23 (1%)a  Terminated 1539 (43%) 1262 (44%)  Died 30 (<1%) 2 (<1%)  Unknown 1578 (44%) 1589 (55%) a Including those who terminated for unknown reason but were born before 1950, so we assume they have retired. Table 4. Characteristics of workers remaining in the cohort and those lost to follow-up, using data from the human resources portion of the database Employed continuously 2010-16 (n = 4496, 56%) Original cohort members who left 2010–16 (n = 3599, 44%) Hired between 2010 and 2016, remain employed (n = 3626, 56%) Hired between 2010 and 2016 and left (n = 2886, 44%) Birth year 1968 1969 1984 1982 Percent female 92% 87% 87% 83% Reason left − −  Retireda 452 (13%)a 23 (1%)a  Terminated 1539 (43%) 1262 (44%)  Died 30 (<1%) 2 (<1%)  Unknown 1578 (44%) 1589 (55%) Employed continuously 2010-16 (n = 4496, 56%) Original cohort members who left 2010–16 (n = 3599, 44%) Hired between 2010 and 2016, remain employed (n = 3626, 56%) Hired between 2010 and 2016 and left (n = 2886, 44%) Birth year 1968 1969 1984 1982 Percent female 92% 87% 87% 83% Reason left − −  Retireda 452 (13%)a 23 (1%)a  Terminated 1539 (43%) 1262 (44%)  Died 30 (<1%) 2 (<1%)  Unknown 1578 (44%) 1589 (55%) a Including those who terminated for unknown reason but were born before 1950, so we assume they have retired. What has been measured? Participant data for the BHWHS integrates existing hospital databases (Table 5), including employee occupational health services records, human resources, payrolls and employee health care use. As BHWHS has grown, the study team has developed partnerships across the hospitals to add additional types and sources of data. For example, the health system is self-insured, with a third-party provider acting as the group health plan. Through this arrangement, the hospitals have access to de-identified data on plan members’ use and expenditures which has become part of BHWHS (2010 and onwards). Currently, the Partners team is working with administrators at both hospitals to add data on patient outcomes such as adverse clinical events, length of stay and patient incidents, with the hope of temporally linking such data to the conditions of work on the unit where patients were treated. Table 5. Sources of administrative data in BHWHS and key variables that have been created or constructed from those data Source Type of data Key variables created or constructed from data Occupational health services Occupational injury Injury timing, location, type, body part, nature, cause (including patient violence incident reports) Days away from work; worker’s compensation costs (medical care utilization, indemnity claim litigation) Employee health insurancea,b Health care use and expenditures Prescriptions and associated costs Physician visits and inpatient care by diagnosis and associated costs Mental health, allied health, physical and occupational therapy Human resources Demographics.work hours Paid hours of work, overtime, benefits Sociodemographic worker characteristics Date of hire, date left employment Patient acuity Workload Number of patients per unit (hourly data available) Acuity of patients Number and mix of nurses and PCAs working Employee scheduling and payrolla Days and times worked Exact time and place scheduled and shifts worked Absenteeism and whether scheduled How shift was paid and by whom (paid, unpaid, sick, vacation, call) Source Type of data Key variables created or constructed from data Occupational health services Occupational injury Injury timing, location, type, body part, nature, cause (including patient violence incident reports) Days away from work; worker’s compensation costs (medical care utilization, indemnity claim litigation) Employee health insurancea,b Health care use and expenditures Prescriptions and associated costs Physician visits and inpatient care by diagnosis and associated costs Mental health, allied health, physical and occupational therapy Human resources Demographics.work hours Paid hours of work, overtime, benefits Sociodemographic worker characteristics Date of hire, date left employment Patient acuity Workload Number of patients per unit (hourly data available) Acuity of patients Number and mix of nurses and PCAs working Employee scheduling and payrolla Days and times worked Exact time and place scheduled and shifts worked Absenteeism and whether scheduled How shift was paid and by whom (paid, unpaid, sick, vacation, call) a Use of these data is restricted to computers within the Partners network. b Available only for employees who are members of the hospital’s group health insurance plan. Table 5. Sources of administrative data in BHWHS and key variables that have been created or constructed from those data Source Type of data Key variables created or constructed from data Occupational health services Occupational injury Injury timing, location, type, body part, nature, cause (including patient violence incident reports) Days away from work; worker’s compensation costs (medical care utilization, indemnity claim litigation) Employee health insurancea,b Health care use and expenditures Prescriptions and associated costs Physician visits and inpatient care by diagnosis and associated costs Mental health, allied health, physical and occupational therapy Human resources Demographics.work hours Paid hours of work, overtime, benefits Sociodemographic worker characteristics Date of hire, date left employment Patient acuity Workload Number of patients per unit (hourly data available) Acuity of patients Number and mix of nurses and PCAs working Employee scheduling and payrolla Days and times worked Exact time and place scheduled and shifts worked Absenteeism and whether scheduled How shift was paid and by whom (paid, unpaid, sick, vacation, call) Source Type of data Key variables created or constructed from data Occupational health services Occupational injury Injury timing, location, type, body part, nature, cause (including patient violence incident reports) Days away from work; worker’s compensation costs (medical care utilization, indemnity claim litigation) Employee health insurancea,b Health care use and expenditures Prescriptions and associated costs Physician visits and inpatient care by diagnosis and associated costs Mental health, allied health, physical and occupational therapy Human resources Demographics.work hours Paid hours of work, overtime, benefits Sociodemographic worker characteristics Date of hire, date left employment Patient acuity Workload Number of patients per unit (hourly data available) Acuity of patients Number and mix of nurses and PCAs working Employee scheduling and payrolla Days and times worked Exact time and place scheduled and shifts worked Absenteeism and whether scheduled How shift was paid and by whom (paid, unpaid, sick, vacation, call) a Use of these data is restricted to computers within the Partners network. b Available only for employees who are members of the hospital’s group health insurance plan. The periodic worker survey measures worker perceptions of organizational policies and practices, features of the psychosocial work environment, worker health behaviours and worker characteristics (Table 6). Most measures are based on validated instruments, although some were developed by the research team. Table 6. Categories of variables and specific constructs measured by the periodic surveys of subsamples of the larger BHWHS cohort Category Constructs measured Worker health and well-being Musculoskeletal symptoms and functional limitations,34 pain severity,35 self-reported injuries, work limitations,36 psychological distress,37 job satisfaction,38 self-reported chronic health conditions, self-reported height and weight Worker health behaviours Physical activity,39 dietary patterns,40 sleep,41 self-efficacy to maintain healthy behaviours Organizational policies and practices Safety practices,42 ergonomic practices,42 people-oriented culture,42 flexibility,43 safe patient-handling norms, break practices, meal breaks,43 shift scheduling and control21 Psychosocial occupational exposures Supervisor support,44 coworker support,18 harassment,45 bullying46 job demands and control44 Physical occupational exposures Self-reported injuries,18 physical activity at work,29 safe patient-handling practices20,31 Social, occupational and demographic worker characteristics Age, gender, race, job title, immigrant status, family characteristics, financial distress,47 hours worked, shifts worked Category Constructs measured Worker health and well-being Musculoskeletal symptoms and functional limitations,34 pain severity,35 self-reported injuries, work limitations,36 psychological distress,37 job satisfaction,38 self-reported chronic health conditions, self-reported height and weight Worker health behaviours Physical activity,39 dietary patterns,40 sleep,41 self-efficacy to maintain healthy behaviours Organizational policies and practices Safety practices,42 ergonomic practices,42 people-oriented culture,42 flexibility,43 safe patient-handling norms, break practices, meal breaks,43 shift scheduling and control21 Psychosocial occupational exposures Supervisor support,44 coworker support,18 harassment,45 bullying46 job demands and control44 Physical occupational exposures Self-reported injuries,18 physical activity at work,29 safe patient-handling practices20,31 Social, occupational and demographic worker characteristics Age, gender, race, job title, immigrant status, family characteristics, financial distress,47 hours worked, shifts worked Table 6. Categories of variables and specific constructs measured by the periodic surveys of subsamples of the larger BHWHS cohort Category Constructs measured Worker health and well-being Musculoskeletal symptoms and functional limitations,34 pain severity,35 self-reported injuries, work limitations,36 psychological distress,37 job satisfaction,38 self-reported chronic health conditions, self-reported height and weight Worker health behaviours Physical activity,39 dietary patterns,40 sleep,41 self-efficacy to maintain healthy behaviours Organizational policies and practices Safety practices,42 ergonomic practices,42 people-oriented culture,42 flexibility,43 safe patient-handling norms, break practices, meal breaks,43 shift scheduling and control21 Psychosocial occupational exposures Supervisor support,44 coworker support,18 harassment,45 bullying46 job demands and control44 Physical occupational exposures Self-reported injuries,18 physical activity at work,29 safe patient-handling practices20,31 Social, occupational and demographic worker characteristics Age, gender, race, job title, immigrant status, family characteristics, financial distress,47 hours worked, shifts worked Category Constructs measured Worker health and well-being Musculoskeletal symptoms and functional limitations,34 pain severity,35 self-reported injuries, work limitations,36 psychological distress,37 job satisfaction,38 self-reported chronic health conditions, self-reported height and weight Worker health behaviours Physical activity,39 dietary patterns,40 sleep,41 self-efficacy to maintain healthy behaviours Organizational policies and practices Safety practices,42 ergonomic practices,42 people-oriented culture,42 flexibility,43 safe patient-handling norms, break practices, meal breaks,43 shift scheduling and control21 Psychosocial occupational exposures Supervisor support,44 coworker support,18 harassment,45 bullying46 job demands and control44 Physical occupational exposures Self-reported injuries,18 physical activity at work,29 safe patient-handling practices20,31 Social, occupational and demographic worker characteristics Age, gender, race, job title, immigrant status, family characteristics, financial distress,47 hours worked, shifts worked What has it found? Key findings and publications BHWHS studies are rooted in a common conceptual model in which the conditions of work are central determinants of workers' proximal and distal health outcomes as well as enterprise outcomes9 (Figure 2). Figure 2 View largeDownload slide Conceptual model for BHWHS9,48 with citations for BHWHS studies that have tested the given pathways (see separate file for figure). Figure 2 View largeDownload slide Conceptual model for BHWHS9,48 with citations for BHWHS studies that have tested the given pathways (see separate file for figure). Several themes have emerged from the findings. The first theme is that injury and musculoskeletal pain are rooted in the conditions of work, in terms of both work organization and workplace psychosocial environment.10–19 The second theme is that the psychosocial work environment—including supervisor support, work-family conflict, job flexibility, scheduling, harassment and work organizational factors—shape safety and health behaviours and downstream health outcomes.12–14,16,20–27 The third theme is that health behaviours have origins in the conditions of work, suggesting that working conditions need to be addressed in order for worker health behaviours to change.23–26,28,29 The final theme of our findings is that, in a hospital setting, the workgroup or unit is a key unit of analysis and intervention,13,14,21 but that interventions are most effective in changing behaviour when messages come from the highest levels of the organization.30,31 One of the many studies to emerge from BHWHS which highlight its capacity to illuminate the relationship between workplace factors and worker health, is a programme evaluation of hospital-wide safe patient-handling initiatives undertaken by one of the hospitals and analysed by the BHWHS team.31 In an analysis using the non-intervention hospital as a comparison group, workers’ self-reported ergonomic practices and safe patient-handling practices at the intervention hospital improved between baseline and follow-up (survey data), as did laundry services’ weekly reports of laundered slings used for patient lifting (integrated database). Additionally, comparing the intervention with the comparison hospital and using the injury database, we observed lower odds of lifting/exertion injuries [post-intervention odds ratio (OR) 0.73, 95% confidence interval (CI) 0.60, 0.89] and neck/shoulder injuries (OR 0.68, 95% CI 0.46, 1.00). BHWHS has also been used to highlight health disparities. Another study18 compared workers’ self-reported injuries in the survey with injuries found in the administrative database, to test whether injury under-reporting occurred differentially by racial and ethnic group. The survey asked workers whether they had been seriously injured during the past year; researchers then searched the database for injuries for each worker in the year preceding the date they completed the survey. The study found that Black workers had greater odds (OR 1.91, 95% CI 1.04, 3.49) of self-reported injury than White workers, but no greater odds (OR 1.22, 95% CI 0.52, 2.77) of administratively-reported injury than White workers. This finding suggests that Black workers may systematically under-report their injuries, and thus administrative injury databases—upon which policies are often made—may not capture disparities. A third study tested whether some of the medical expenditures related to workplace injuries were borne by group health insurance rather than by workers’ compensation.19 At the individual worker level, the study merged: data on occupational health services and data on injuries; health care use data on total outpatient health expenditures before and after injury; payroll and staffing data to develop denominators; and human resources data for demographic information. The study compared injured workers with workers who were not injured, controlling for pre-injury expenditures, age and job category. Injured workers had USD $587 (95% CI 167, 1140) more in non-workers’ compensation health expenditures in the 6 months following injury than their non-injured peers. What are the main strengths and weaknesses? The main strength of BHWHS is its innovative creation of an integrated health and safety database of employees and their work environment. This results from the shared mission of the Centre and Partners, which is specific to this cohort (and thus may be difficult to replicate). Several practices have facilitated that relationship. The database is physically housed at Partners and is managed by a Partners employee (T.O.) who is partially funded by the study grant and is jointly supervised by investigators at Partners and the Centre. All data are de-identified and the team has honoured Partners’ request that the most sensitive data remain on hospital networked computers for analysis. There is close collaboration between the human subjects committees at Partners and the Harvard T.H. Chan School of Public Health, with each institution ceding review to the other at points; currently, the primary protocol is housed at Partners. All these factors build trust between various departments at Partners and the BHWHS team, which in turn has led to the addition of more sensitive information, such as the health expenditures data, to the database. Communications with the managers of all database components at Partners is handled exclusively by the Partners investigators on the study team (D.H. and K.H.), to further build those relationships and control study-related demands on those managers’ time and effort. Over time, the relationship between the Centre and Partners has led to a true intellectual partnership between the two groups. Partners investigators (mainly D.H. and K.H.) attend and participate in all BHWHS team meetings. They contribute substantively to all manuscripts that come out of BHWHS, and thus are co-authors. The Centre’s close communication with the Partners team permits analysis of programmes, policies and practices already taking place at the hospitals,31 a mutually beneficial arrangement. For example, the team’s surveys in 2012 and 2014 were timed to occur before and after a safe patient-handling initiative being planned by one of the hospitals, independent of BHWHS.31 The robustness of that study design would not have been possible without established channels of communication between the study team and the occupational health department. The Partners team brings disciplinary perspectives that complement those of the Centre team. D.H. is an occupational medicine physician and a law professor, and thus contributes expertise in both policy and occupational health practice. K.H. holds a doctorate in environmental health and is an occupational health nurse; her research focuses on nursing organization and practice. Both K.H. and D.H. contribute substantive knowledge around patient care, health insurance, occupational health practice and the structure and culture of the hospitals. The two groups engage in joint priority-setting, in which the Partners collaborators bring to the BHWHS team issues that are of particular interest to the two hospitals (e.g. bullying, mental health of employees) because those issues are likely of interest to other hospitals as well, increasing utility of findings. Researchers also share with Partners the research questions they are planning to ask in a given funding cycle, and Partners investigators provide feedback on how to make those questions most salient to current hospital concerns. Methodologically, the study has several strengths that differentiate it from other occupational health studies. The size and breadth of this database are rare.32,33 The automatic enrolment of participants—with careful de-identification to protect human subjects—captures all patient care workers with less selection bias and loss to follow-up than traditional cohort designs. The study enables multilevel analysis by linking individual and work-group data. With its longitudinal design—as of 2018, 10 years of data and counting—we can test complex hypotheses and account for temporal ordering of exposures and outcomes. The breadth of the database is supplemented with surveys. Furthermore, given that workers are clustered within units and many factors measured with the survey inherently occur at the group level, we can measure both group- and individual-level determinants of health. The study does have limitations. The biggest is external validity; the very fact that the health system collects such detailed data on everyday operations, and that they wish to have such data analysed and published, is an indicator of an unusual employer. Other weaknesses are that survey data are available only on a subset of workers overall, in a smaller subset of workers longitudinally and at inconsistent time intervals. Can I get hold of the data? Where can I find out more? Because many of the datasets contain sensitive employee information, the study team maintains tight control of the data and its distribution. As with many occupational cohorts, the data cannot be made open access. However, the team is open to partnering with external collaborators who can travel to Boston and use the data on site. Those interested can contact the study principal investigator, Dr Erika Sabbath, at [[emailprotected]] for further discussion and to obtain an application. Certain data, namely the health claims and payroll data, are restricted to computers in the Partners network and cannot be distributed to external collaborators. Additionally, all manuscript proposals must be submitted to the BHWHS publications committee and approved by both Centre and Partners investigators before data analyses can commence. Although the data have restrictions in how they can be distributed, our database serves as a model of the public health benefits of data collaborations between researchers and employers. This cohort profile serves as a guide to other researcher-employer teams looking to establish a similar database and collaboration, and the BHWHS team is available for further consultation to such teams. Funding This work was supported by the U.S. Centers for Disease Control and Prevention National Institute for Occupational Safety and Health [grant numbers 5U19 OH008861, to G.S., and 7K01 OH010673, to E.L.S.]. Acknowledgements BHWHS would not be possible without the participation of Partners Health care System and leadership from Joseph Cabral and Kurt Westerman. The authors thank Partners Health care Occupational Health Services; Lisa DiMarino and Rachel Kolbin-Gupp at Partners Health care Human Resources; individuals at each of the hospitals, including Jeanette Ives Erickson, Jackie Somerville, Dawn Tenney and Deborah Mulloy in Patient Care Services leadership; and Jeff Davis and Julie Celano in Human Resources. The authors thank Truven Health Analytics for providing access to the health insurance data. They also thank Eddie Tan, Mario Dashi and Shari Weingarten for assistance with supporting databases, and Christopher Kenwood (formerly at New England Research Institute) and Na Wang and Lily Chen (at Boston University) for data preparation. Conflict of interest: None declared. References 1 Bureau of Labor Statistics . Workplace Injuries and Illnesses—2016. https://www.bls.gov/iif/osch0060.pdf (23 May 2018 , date last accessed). 2 Bureau of Labor Statistics . Spotlight on Statistics: Health Care. http://www.bls.gov/spotlight/2009/health_care/ (16 August 2013 , date last accessed). 3 Rodríguez-Acosta RL , Richardson DB , Lipscomb HJ , Chen JC , Dement JM , Myers DJ. Occupational injuries among aides and nurses in acute care . Am J Ind Med 2009 ; 52 : 953 – 64 . Google Scholar Crossref Search ADS PubMed 4 Bridges CB , Katz JM , Seto WH et al. Risk of influenza A (H5N1) infection among health care workers exposed to patients with influenza A (H5N1), Hong Kong . J Infect Dis 2000 ; 181 : 344 – 48 . Google Scholar Crossref Search ADS PubMed 5 Bronner G , Peretz C , Ehrenfeld M. Sexual harassment of nurses and nursing students . J Adv Nurs 2003 ; 42 : 637 – 44 . Google Scholar Crossref Search ADS PubMed 6 Sauter SL , Brightwell W , Colligan M et al. The Changing Organization of Work and the Safety and Health of Working People . Washington, DC : National Institute for Occupational Safety and Health , 2002 . 7 Schill AL , Chosewood LC. The NIOSH Total Worker Health program: an overview . J Occup Environ Med 2013 ; 55 : S8 – 11 . Google Scholar Crossref Search ADS PubMed 8 Li C-Y , Sung F-C. A review of the healthy worker effect in occupational epidemiology . Occup Med (Lond) 1999 ; 49 : 225 – 29 . Google Scholar Crossref Search ADS PubMed 9 Sorensen G , McLellan D , Dennerlein J et al. A conceptual model for guiding integrated interventions and research: pathways through the conditions of work. In: Hudson H , Nigam J , Sauter S , Chosewood L , Schill A , Howard J (eds). Total Worker Health: Integrative Approaches to Worker Safety, Health, and Well-Being . Washington, DC : APA Press , 2017 . 10 Boden LI , Petrofsky YV , Hopcia K , Wagner GR , Hashimoto D. Understanding the hospital sharps injury reporting pathway . Am J Ind Med 2015 ; 58 : 282 – 89 . Google Scholar Crossref Search ADS PubMed 11 Boden LI , Sembajwe G , Tveito TH et al. Occupational injuries among nurses and aides in a hospital setting . Am J Ind Med 2012 ; 55 : 117 – 26 . Google Scholar Crossref Search ADS PubMed 12 Reme SE , Dennerlein JT , Hashimoto D , Sorensen G. Musculoskeletal pain and psychological distress in hospital patient care workers . J Occup Rehabil 2012 ; 22 : 503 – 10 . Google Scholar Crossref Search ADS PubMed 13 Sabbath EL , Hurtado DA , Okechukwu CA et al. Occupational injury among hospital patient‐care workers: what is the association with workplace verbal abuse? Am J Ind Med 2014 ; 57 : 222 – 32 . Google Scholar Crossref Search ADS PubMed 14 Tveito T , Sembajwe G , Boden L et al. Impact of organizational policies and practices on workplace injuries in a hospital setting . J Occup Environ Med 2014 ; 56 : 802 – 08 . Google Scholar Crossref Search ADS PubMed 15 Dennerlein JT , Hopcia K , Sembajwe G et al. Ergonomic practices within patient care units are associated with musculoskeletal pain and limitations . Am J Ind Med 2012 ; 55 : 107 – 16 . Google Scholar Crossref Search ADS PubMed 16 Kim SS , Okechukwu CA , Buxton OM et al. Association between work–family conflict and musculoskeletal pain among hospital patient care workers . Am J Ind Med 2013 ; 56 : 488 – 95 . Google Scholar Crossref Search ADS PubMed 17 Kim SS , Okechukwu CA , Dennerlein JT et al. Association between perceived inadequate staffing and musculoskeletal pain among hospital patient care workers . Int Arch Occup Environ Health 2014 ; 87 : 323 – 30 . Google Scholar Crossref Search ADS PubMed 18 Sabbath EL , Boden LI , Williams JA , Hashimoto D , Hopcia K , Sorensen G. Obscured by administrative data? Racial disparities in occupational injury . Scand J Work Environ Health 2017 ; 43 : 155 – 62 . Google Scholar Crossref Search ADS PubMed 19 Williams JAR , Sorensen G , Hashimoto D , Hopcia K , Wagner GR , Boden LI. Impact of occupational injuries on nonworkers’ compensation medical costs of patient-care workers . J Occup Environ Med 2017 ; 59 : e119 – 24 . Google Scholar Crossref Search ADS PubMed 20 Caspi CE , Dennerlein JT , Kenwood C et al. Results of a pilot intervention to improve health and safety for health care workers . J Occup Environ Med 2013 ; 55 : 1449 . Google Scholar Crossref Search ADS PubMed 21 Hurtado DA , Glymour M , Reme SE , Berkman L , Hashimoto D , Sorensen G. Schedule control and mental health: The relevance of coworkers' reports . Community Work Fam 2015 ; 18 : 416 – 34 . Google Scholar Crossref Search ADS 22 Hurtado DA , Nelson CC , Hashimoto D , Sorensen G. Supervisors’ support for nurses’ meal breaks and mental health . Workplace Health Saf 2015 ; 63 : 107 – 15 . Google Scholar Crossref Search ADS PubMed 23 Jacobsen HB , Reme SE , Sembajwe G et al. Work stress, sleep deficiency, and predicted 10‐year cardiometabolic risk in a female patient care worker population . Am J Ind Med 2014 ; 57 : 940 . Google Scholar Crossref Search ADS PubMed 24 Jacobsen HB , Reme SE , Sembajwe G et al. Work-family conflict, psychological distress, and sleep deficiency among patient care workers . Workplace Health Saf 2014 ; 62 : 282 – 91 . Google Scholar PubMed 25 Nelson CC , Wagner GR , Caban-Martinez AJ et al. Physical activity and body mass index: the contribution of age and workplace characteristics . Am J Prev Med 2014 ; 46 : S42 – 51 . Google Scholar Crossref Search ADS PubMed 26 Sorensen G , Stoddard AM , Stoffel S et al. The role of the work context in multiple wellness outcomes for hospital patient care workers . J Occup Environ Med 2011 ; 53 : 899. Google Scholar Crossref Search ADS PubMed 27 Sembajwe G , Tveito TH , Hopcia K et al. Psychosocial stress and multi-site musculoskeletal pain: a cross-sectional survey of patient care workers . Workplace Health Saf 2013 ; 61 : 117 – 25 . Google Scholar PubMed 28 Buxton OM , Hopcia NP , Sembajwe G et al. Relationship of sleep deficiency to perceived pain and functional limitations in hospital patient care workers . J Occup Environ Med 2012 ; 54 : 851. Google Scholar Crossref Search ADS PubMed 29 Umukoro PE , Arias OE , Stoffel SD , Hopcia K , Sorensen G , Dennerlein JT. Physical activity at work contributes little to patient care workers' weekly totals . J Occup Environ Med 2013 ; 55 : S63 – 68 . Google Scholar Crossref Search ADS PubMed 30 Sorensen G , Nagler E , Hashimoto D et al. Implementing an integrated health protection/health promotion intervention in the hospital setting: lessons learned from the Be Well, Work Well Study . J Occup Environ Med 2016 ; 58 : 185 – 94 . Google Scholar Crossref Search ADS PubMed 31 Dennerlein J , Kenwood C , Stoddard A et al. Lifting and exertion injuries decrease after implementation of an integrated hospital-wide safe patient handling and mobilisation programme . Occup Environ Med 2017 ; 74 : 336 – 43 . Google Scholar Crossref Search ADS PubMed 32 Dement JM , Pompeii LA , Ostbye T et al. An integrated comprehensive occupational surveillance system for health care workers . Am J Ind Med 2004 ; 45 : 528 – 38 . Google Scholar Crossref Search ADS PubMed 33 Pollack KM , Agnew J , Slade MD et al. Use of employer administrative databases to identify systematic causes of injury in aluminum manufacturing . Am J Ind Med 2007 ; 50 : 676. Google Scholar Crossref Search ADS PubMed 34 Kuorinka I , Jonsson B , Kilbom A et al. Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms . Appl Ergon 1987 ; 18 : 233 – 37 . Google Scholar Crossref Search ADS PubMed 35 Hudak PL , Amadio PC , Bombardier C. Development of an upper extremity outcome measure: the DASH . Am J Ind Med 1996 ; 29 : 602 – 08 . Google Scholar Crossref Search ADS PubMed 36 Lerner D , Amick BC III , Rogers WH , Malspeis S , Bungay K , Cynn D. The work limitations questionnaire . Med Care 2001 ; 39 : 72 – 85 . Google Scholar Crossref Search ADS PubMed 37 Kessler RC , Andrews G , Colpe LJ et al. Short screening scales to monitor population prevalences and trends in non-specific psychological distress . Psychol Med 2002 ; 32 : 959 – 76 . Google Scholar Crossref Search ADS PubMed 38 Quinn RP , Mangione TW , Seashore SE. 1972-73 Quality of Employment Survey: ISR Social Science Archive . Frankfurt am Main, Germany : Institute for Social Research , 1975 . 39 Lee PH , Macfarlane DJ , Lam T , Stewart SM. Validity of the international physical activity questionnaire short form (IPAQ-SF): a systematic review . Int J Behav Nutr Phys Act 2011 ; 8 : 115. Google Scholar Crossref Search ADS PubMed 40 Harley AE , Yang M , Stoddard AM et al. Patterns and predictors of health behaviors among racially/ethnically diverse residents of low-income housing developments . Am J Health Promot 2014 ; 29 : 59 – 67 . Google Scholar Crossref Search ADS PubMed 41 Buysse DJ , Reynolds CF III , Monk TH , Berman SR , Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research . Psychiatry Res 1989 ; 28 : 193 – 213 . Google Scholar Crossref Search ADS PubMed 42 Amick BC , Habeck RV , Hunt A et al. Measuring the impact of organizational behaviors on work disability prevention and management . J Occup Rehabil 2000 ; 10 : 21 – 38 . Google Scholar Crossref Search ADS 43 Thomas LT , Ganster DC. Impact of family-supportive work variables on work-family conflict and strain: a control perspective . J Appl Psychol 1995 ; 80 : 6. Google Scholar Crossref Search ADS 44 Quick TL , Theorell T . Healthy work: stress, productivity, and the reconstruction of working life: Basic Books . Natl Prod Rev 1990 ; 9 : 475. Google Scholar Crossref Search ADS 45 Richman JA , Rospenda KM , Nawyn SJ et al. Sexual harassment and generalized workplace abuse among university employees: Prevalence and mental health correlates . Am J Public Health 1999 ; 89 : 358 – 63 . Google Scholar Crossref Search ADS PubMed 46 Simons SR , Stark RB , DeMarco RF. A new, four-item instrument to measure workplace bullying . Res Nurs Health 2011 ; 34 : 132 – 40 . Google Scholar PubMed 47 Szanton SL , Allen JK , Thorpe RJ , Seeman T , Bandeen-Roche K , Fried LP. Effect of financial strain on mortality in community-dwelling older women . J Gerontol B Psychol Sci Soc Sci 2008 ; 63 : S369 – 74 . Google Scholar Crossref Search ADS PubMed 48 Sorensen G , McLellan D , Sabbath E et al. Integrating worksite health protection and health promotion: a conceptual model for intervention and research . Prev Med 2016 ; 91 : 188 – 96 . Google Scholar Crossref Search ADS PubMed 49 Reme SE , Shaw WS , Boden LI et al. Worker assessments of organizational practices and psychosocial work environment are associated with musculoskeletal injuries in hospital patient care workers . Am J Ind Med 2014 ; 57 : 810 – 18 . Google Scholar Crossref Search ADS PubMed 50 Hurtado DA , Kim S-S , Subramanian SV et al. Nurses' but not supervisors' safety practices are linked with job satisfaction . J Nurs Manag 2017 ; 25 : 491 – 97 . Google Scholar Crossref Search ADS PubMed 51 Hopcia K , Dennerlein JT , Hashimoto D , Orechia MT , Sorensen G. A case-control study of occupational injuries for consecutive and cumulative shifts among hospital registered nurses and patient care associates . Workplace Health Saf 2012 ; 60 : 437. Google Scholar PubMed 52 Sorensen G , McLellan D , Dennerlein JT et al. Integration of health protection and health promotion: rationale, indicators, and metrics . J Occup Environ Med 2013 ; 55 : S12 – 18 . Google Scholar Crossref Search ADS PubMed 53 Sabbath EL , Williams JAR , Boden LI et al. Mental health expenditures: Association with workplace incivility and bullying among hospital patient care workers . J Occup Environ Med 2018 , Mar 13. doi: 10.1097/JOM.0000000000001322. 54 Sabbath EL , Sparer EH , Boden LI et al. Preventive care utilization: association with individual- and workgroup-level policy and practice perceptions . Prev Med 2018 ; 111 : 235 – 40 . Google Scholar Crossref Search ADS PubMed 55 Sparer EH , Boden LI , Sorensen G et al. The relationship between organizational policies and practices and work limitations among hospital patient care workers. Am J Ind Med 2018 ; 61 : 691 – 8 . © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Open Access Collection

Cohort Profile: The Epidemiology of Chronic Diseases Cohort (EpiDoC)

Dias, Sara Simões; Rodrigues, Ana Maria; Gregório, Maria João; de Sousa, Rute Dinis; Branco, Jaime Cunha; Canhão, Helena

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy185pmid: 30212889

Why was the cohort set up? Non-communicable chronic diseases are the leading cause of death and the main contributor to disease burden worldwide, accounting for 86% of all deaths in Portugal.1 Several modifiable behavioural risk factors, such as unhealthy dietary habits, physical inactivity, tobacco use and harmful use of alcohol, are the main risk factors for these diseases. Thus, the existence of epidemiological data on chronic diseases and their determinants (i.e. socioeconomic and demographic factors), associated factors and consequences are important public health tools for designing and developing strategies to tackle the burden of non-communicable diseases. In 2011, a prospective cohort study called Epidemiology of Chronic Diseases (EpiDoC) aimed to create a large population database for medical and health-related research in Portugal. To our knowledge, the EpiDoC study constitutes one of the first Portuguese prospective large cohort studies, including a representative sample of the Portuguese population, with the primary aim of examining the health determinants and outcomes of chronic non-communicable diseases and their impact on health care resource consumption. The EpiDoC study was designed by researchers from NOVA Medical School in Lisbon with close collaboration between social and biomedical scientists, ensuring a thorough multidisciplinary approach. The first wave of this cohort study, named EpiDoC 1 (EpiReumaPt), occurred between September 2011 and December 2013. Its primary aim was to assess rheumatic and musculoskeletal disease (RMD) prevalence and its burden in Portugal. This wave had two phases: the first consisted of a face-to-face interview, and the second included a detailed clinical evaluation of RMD performed by rheumatologists. This baseline assessment also enabled the creation of a population-based biobank (i.e. DNA, serum and total blood samples) for identifying genetic predictors and serum risk factors for chronic diseases. Musculoskeletal imaging data were also collected, in particular peripheral dual energy X-ray (DXA) in all second phase participants and X-ray of the affected joint(s). Similar to other cohort studies,1,2 the scope of the EpiDoC study has expanded over time. So far, two subsequent waves have been completed: EpiDoC 2 (March 2013–July 2015) and EpiDoC 3 (September 2015–July 2016). In both waves, data were collected through a phone interview. EpiDoC 2 (CoReumaPt) focused on lifestyle behaviours and their determinants, with a secondary goal of identifying innovative patient solutions for coping with disability. EpiDoC 3 (Saúde.Come) assessed inequalities in access to healthy food and health services, with a focus on food insecurity and its determinants and health consequences. The EpiDoC study was performed according to the principles established by the Declaration of Helsinki and revised in 2013 in Fortaleza. Ethical approval was obtained from the National Committee for Data Protection (Comissão Nacional de Proteção de Dados) and NOVA Medical School Ethics Committee. Ethical committees of regional health authorities also approved the study. Who is in the cohort? Setting EpiDoC is a prospective closed cohort study including a nationally representative sample of adults (≥18 years old) who were non-institutionalized and living in private households in Portugal Mainland and Islands (Azores and Madeira).3 Portugal is a south-western European country with a resident population of 10 562 178, of whom 8 million are adults (4 072 122 men and 4 585 118 women).4 During the past two decades, life expectancy in Portugal has been increasing. Data from the World Health Organization indicate that life expectancy in Portugal was 83.9 years for women and 78.2 years for men in 2015. In addition, as in other European countries, the Portuguese population has been undergoing demographic changes. The Portuguese population pyramid shows an increasing number of individuals at the top and a decreasing number at the bottom, indicating a new structure of the Portuguese population with fewer young people and more elderly. In 2015, the old-age dependency ratio was 31.1 per 100 persons of working age, which is the ratio between the number of persons aged ≥65 years (i.e. when individuals are generally economically inactive) and those aged 15–64 years.5 Portugal is divided into seven regions according to the Nomenclature of Territorial Units for Statistics II (NUTS II): Norte, Centro, Lisboa e Vale do Tejo, Alentejo, Algarve, Região Autónoma dos Açores (the Azores) and Região Autónoma da Madeira (Madeira). At the NUTS II level, the Norte region has the largest population density (34.7%), followed by Lisboa e Vale do Tejo (26.6%) and Centro (22.4%) (Figure 1). The other NUTS II regions (Alentejo, Algarve, the Azores, and Madeira) encompass small towns and villages with lower population densities and higher desertification rates. Figure 1. Open in new tabDownload slide Portuguese population density distribution according to the 7 NUTS II. Participant recruitment Considering the primary aim of EpiDoC 1, the sample size was calculated based on the estimated prevalence of rheumatic diseases with a 95% confidence interval (CI), and standardized for age and sex according to the total adult population of the studied areas. Assuming that the expected prevalence of rheumatic diseases was between 0.5% and 1%, and expecting a drop-out rate of 50%, it was estimated that a total of 9000 individuals should be recruited. To obtain regional representativeness, the sample size was stratified according to dimensions and characteristics of the seven Portuguese regions. Population recruitment was conducted by Centro de Estudos e Sondagens de Opinião da Universidade Católica Portuguesa (CESOP-UCP), and multistage random sampling was used for participant selection. In EpiDoC 1, candidates for participation were visited at their homes by a team of trained interviewers. Locations were selected as the primary unit of sampling according to the Census 2001. Selected households and their addresses were identified using a random selection of points in the map of each location, where the interviewer began a systematic step count (defined for each locality based on its size). Each selected household was visited, with no previous contact, up to three times (including evenings and weekends) if no candidate participant was present during the first visit. In each household, an individual ≥18 years old with permanent residence and the most recently completed birthday was selected to be a participant in the EpiDoC study. Before participant interviews, the EpiDoC team gave information about study details and aims at local churches, primary care centres and municipalities. Local priests, health providers and municipality employers helped us to spread the information and motivate participation. EpiDoC 1 (2011–13) EpiDoC 1 enrolled 10 661 participants and was primarily designed to estimate the prevalence of RMDs. To provide a comprehensive understanding of the burden of RMDs, this wave had the secondary aim of evaluating quality of life, physical function, mental health, work status and health care resource consumption, with the purpose of identifying differences in these and other outcomes between individuals with and without RMDs.3 EpiDoC 1 data collection consisted of two phases. Phase 1 involved face-to-face interviews conducted by a team of trained interviewers (non-physicians) through door-to-door visits. Phase 2 involved clinical observations with physical examination performed by rheumatologists, for participants identified as potentially having an RMD (using a screening questionnaire applied at Phase 1) and 20% of asymptomatic individuals. All procedures occurred between September 2011 and December 2013. Of the 10 661 participants selected in Phase 1, 7451 had a positive RMD screening and 3210 had a negative RMD screening. A total of 8152 participants were contacted in Phase 2: 7451 with a positive RMD and 701 (∼20%) without an RMD as previously defined in the study protocol. Of these, 4275 did not attend a clinical observation by a rheumatologist. Therefore, at the end of Phase 2, there were 3877 clinical observations with physical examination performed by rheumatologists; 3198 participants received validation of an RMD diagnosis and 679 did not have an RMD diagnosis. In Phase 1, a structured questionnaire using a computer-assisted personal interview (CAPI) system was used to collect data. Questions on rheumatic symptoms were asked, and an algorithm for screening each RMD was applied. An individual was considered to have a positive screening: if he/she mentioned a previously known RMD; if any of the specific disease algorithms in the screening questionnaires were positive; or if the participant reported muscle, vertebral or peripheral joint pain in the previous 4 weeks.3 Phase 2 was performed by rheumatologists at the local primary care centre for all participants who were identified as having a positive RMD screening. All clinical laboratory and imaging data were verified by a team of three experienced rheumatologists, and diagnoses were confirmed according to validated criteria.3 All participants enrolled in EpiDoC 1 (10 661 participants) were invited to participate in a follow-up study, of whom 10 153 (95.2%) signed consent forms and agreed to participate. For follow-up waves (EpiDoC 2 and 3), data were collected using a structured questionnaire administered by phone call interviews using a CAPI system. A core questionnaire was used in each EpiDoC wave, with additional questions added according to the focus of each wave. In EpiDoC 2 and 3, when a participant was not available, additional attempts were made at different times up to a maximum of six attempts. The last contact attempt had to follow the previous contact by least 1 month; only then was the contact attempt abandoned. EpiDoC 2 (2013–15) EpiDoC 2 was the first follow-up wave, with data collected between March 2013 and July 2015. EpiDoC 2 included 7591 participants (out of 10 153 eligible participants) representative of the adult Portuguese population, resulting in a response rate of 71.2% from EpiDoC 1. Considering that the main risk factors for non-communicable diseases are unhealthy lifestyle behaviours, EpiDoC 2 employed the core structured questionnaire but included more detailed questions on lifestyle behaviours, such as physical activity, dietary habits, tobacco and alcohol use and sleeping habits. Questions regarding innovative patient solutions for coping with disability were also included. EpiDoC 3 (2015–16) EpiDoC 3 occurred between September 2015 and July 2016 and included 5653 participants, resulting in a response rate of 55.7% from EpiDoC 1. This wave continued to employ the core structured questionnaire but included questions on food insecurity, its determinants and its health consequences. This particular interest in food insecurity was based on a growing awareness of social inequalities in health and modifiable risk factors for chronic diseases, such as dietary patterns, as well as the economic crisis faced by Portugal in previous years. Cohort characteristics The participation rate declined from EpiDoC 1 to EpiDoC 3, similar to most other population-based studies.2,6Table 1 presents the characteristics of participants in the cohort. There were no significant differences in any categories of variables between the three waves. Table 1. Characteristics of the participants in the cohort . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . Census 2011 . Sex n = 10 661 n = 7591 n = 5653 n = 8 657 240  Female 6551 (52.6%) 4784 (52.2%) 3607 (52.5%) 4 585 118 (53.0%) Age group n = 10 661 n = 7591 n = 5648  18–29 1182 (22.1%) 621 (18.4%) 355 (15.4%) 1 470 782 (17.0%)  30–39 1511 (18.8%) 975 (18.7%) 605 (19.1%) 1 598 250 (18.5%)  40–49 1906 (17.3%) 1437 (18.2%) 1049 (18.3%) 1 543 392 (17.8%)  50–59 1801 (14.8%) 1437 (16.2%) 1143 (15.9%) 1 400 011 (16.2%)  60–69 1915 (12.9%) 1440 (13.2%) 1112 (13.7%) 1 186 442 (13.7%)  70–74 849 (5.8%) 645 (6.2%) 491 (6.7%) 496 438 (5.7%)  ≥75 1497 (8.4%) 1036 (9.1%) 893 (11.0%) 961 925 (11.1%) Ethnicity/race n = 10 629 n = 7574 n = 5638  Caucasian 10 342 (96.0%) 7423 (97.1%) 5536 (97.2%) No comparable data  Black 221 (3.4%) 119 (2.5%) 81 (2.3%)  Asian 8 (0.1%) 3 (0.0%) 2 (0.1%)  Romany 20 (0.3%) 7 (0.1%) 5 (0.1%)  Other 38 (0.3%) 22 (0.3%) 14 (0.3%) Years of education  (mean ± SD) 7.41  ±  4.1 8.66  ±  3.90 8.80  ±  3.94 Education level n = 10 585 n = 7546 n = 5615  0–4 years 4726 (33.2%) 3272 (31.7%) 2392 (30.9%) 3 239 724 (37.4%)  5–9 years 2175 (22.6%) 1547 (21.3%) 1122 (19.6%) 2 134 401 (24.6%)  10–12 years 1920 (23.8%) 1391 (24.8%) 1049 (25.6%) 1 560 958 (18.0%)  >12 years 1764 (20.4%) 1336 (22.2%) 1052 (24.0%) 1 741 567 (20.1%) NUTS II n = 10 661 n = 7591 n = 5648  Norte 3122 (34.9%) 2240 (35.8%) 1659 (36.5%) 3 007 823 (34.7%)  Centro 1997 (22.8%) 1504 (23.3%) 1087 (23.2%) 1 938 815 (22.4%)  Lisboa 2484 (26.7%) 1588 (25.4%) 1131 (24.8%) 2 300 053 (26.6%)  Alentejo 669 (7.3%) 422 (7.2%) 320 (7.2%) 633 691 (7.3%)  Algarve 352 (3.8%) 245 (3.8%) 183 (3.7%) 370 704 (4.3%)  Azores 1029 (2.2%) 793 (2.1%) 657 (2.5%) 192 357 (2.2%)  Madeira 1008 (2.3%) 799 (2.4%) 611 (2.4%) 213 797 (2.5%) Marital status n = 10 652 n = 7586 n = 5644  Single 1935 (29.4%) 1285 (28.4%) 922 (28.5%) No comparable data  Married 6111 (50.2%) 4591 (53.2%) 3457 (53.4%)  Divorced 810 (7.4%) 556 (6.8%) 391 (6.1%)  Widow(er) 1414 (8.2%) 910 (7.3%) 697 (7.6%)  Consensual union 382 (4.8%) 244 (4.2%) 177 (4.4%) BMI n = 10 109 n = 6922 n = 5174  Underweight 167 (2.2%) 111 (2.0%) 88 (2.1%) No comparable data  Normal 4063 (45.5%) 2670 (45.5%) 2009 (44.5%)  Overweight 3799 (35.1%) 2788 (37.1%) 2098 (37.7%)  Obese 2080 (17.1%) 1353 (15.4%) 979 (15.7%) Monthly household income n = 7613 n = 5558 n = 4167  <500€ 1994 (19.9%) 1331 (18.0%) 945 (16.66%) No comparable data  501€ to 750€ 1707 (21.7%) 1257 (20.8%) 949 (20.91%)  751€ to 1000€ 1268 (18.8%) 943 (19.0%) 717 (19.89%)  1001€ to 1500€ 1141 (17.2%) 852 (17.5%) 638 (16.97%)  1501€ to 2000€ 657 (9.9%) 511 (10.9%) 386 (11.08%)  2001€ to 2500€ 379 (5.9%) 295 (5.7%) 246 (6.37%)  2501€ to 3000€ 222 (3.0%) 188 (3.8%) 148 (3.98%)  3001€ to 4000€ 146 (1.8%) 108 (2.1%) 83 (1.94%)  >4000€ 99 (1.9%) 73 (2.2%) 55 (2.20%) . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . Census 2011 . Sex n = 10 661 n = 7591 n = 5653 n = 8 657 240  Female 6551 (52.6%) 4784 (52.2%) 3607 (52.5%) 4 585 118 (53.0%) Age group n = 10 661 n = 7591 n = 5648  18–29 1182 (22.1%) 621 (18.4%) 355 (15.4%) 1 470 782 (17.0%)  30–39 1511 (18.8%) 975 (18.7%) 605 (19.1%) 1 598 250 (18.5%)  40–49 1906 (17.3%) 1437 (18.2%) 1049 (18.3%) 1 543 392 (17.8%)  50–59 1801 (14.8%) 1437 (16.2%) 1143 (15.9%) 1 400 011 (16.2%)  60–69 1915 (12.9%) 1440 (13.2%) 1112 (13.7%) 1 186 442 (13.7%)  70–74 849 (5.8%) 645 (6.2%) 491 (6.7%) 496 438 (5.7%)  ≥75 1497 (8.4%) 1036 (9.1%) 893 (11.0%) 961 925 (11.1%) Ethnicity/race n = 10 629 n = 7574 n = 5638  Caucasian 10 342 (96.0%) 7423 (97.1%) 5536 (97.2%) No comparable data  Black 221 (3.4%) 119 (2.5%) 81 (2.3%)  Asian 8 (0.1%) 3 (0.0%) 2 (0.1%)  Romany 20 (0.3%) 7 (0.1%) 5 (0.1%)  Other 38 (0.3%) 22 (0.3%) 14 (0.3%) Years of education  (mean ± SD) 7.41  ±  4.1 8.66  ±  3.90 8.80  ±  3.94 Education level n = 10 585 n = 7546 n = 5615  0–4 years 4726 (33.2%) 3272 (31.7%) 2392 (30.9%) 3 239 724 (37.4%)  5–9 years 2175 (22.6%) 1547 (21.3%) 1122 (19.6%) 2 134 401 (24.6%)  10–12 years 1920 (23.8%) 1391 (24.8%) 1049 (25.6%) 1 560 958 (18.0%)  >12 years 1764 (20.4%) 1336 (22.2%) 1052 (24.0%) 1 741 567 (20.1%) NUTS II n = 10 661 n = 7591 n = 5648  Norte 3122 (34.9%) 2240 (35.8%) 1659 (36.5%) 3 007 823 (34.7%)  Centro 1997 (22.8%) 1504 (23.3%) 1087 (23.2%) 1 938 815 (22.4%)  Lisboa 2484 (26.7%) 1588 (25.4%) 1131 (24.8%) 2 300 053 (26.6%)  Alentejo 669 (7.3%) 422 (7.2%) 320 (7.2%) 633 691 (7.3%)  Algarve 352 (3.8%) 245 (3.8%) 183 (3.7%) 370 704 (4.3%)  Azores 1029 (2.2%) 793 (2.1%) 657 (2.5%) 192 357 (2.2%)  Madeira 1008 (2.3%) 799 (2.4%) 611 (2.4%) 213 797 (2.5%) Marital status n = 10 652 n = 7586 n = 5644  Single 1935 (29.4%) 1285 (28.4%) 922 (28.5%) No comparable data  Married 6111 (50.2%) 4591 (53.2%) 3457 (53.4%)  Divorced 810 (7.4%) 556 (6.8%) 391 (6.1%)  Widow(er) 1414 (8.2%) 910 (7.3%) 697 (7.6%)  Consensual union 382 (4.8%) 244 (4.2%) 177 (4.4%) BMI n = 10 109 n = 6922 n = 5174  Underweight 167 (2.2%) 111 (2.0%) 88 (2.1%) No comparable data  Normal 4063 (45.5%) 2670 (45.5%) 2009 (44.5%)  Overweight 3799 (35.1%) 2788 (37.1%) 2098 (37.7%)  Obese 2080 (17.1%) 1353 (15.4%) 979 (15.7%) Monthly household income n = 7613 n = 5558 n = 4167  <500€ 1994 (19.9%) 1331 (18.0%) 945 (16.66%) No comparable data  501€ to 750€ 1707 (21.7%) 1257 (20.8%) 949 (20.91%)  751€ to 1000€ 1268 (18.8%) 943 (19.0%) 717 (19.89%)  1001€ to 1500€ 1141 (17.2%) 852 (17.5%) 638 (16.97%)  1501€ to 2000€ 657 (9.9%) 511 (10.9%) 386 (11.08%)  2001€ to 2500€ 379 (5.9%) 295 (5.7%) 246 (6.37%)  2501€ to 3000€ 222 (3.0%) 188 (3.8%) 148 (3.98%)  3001€ to 4000€ 146 (1.8%) 108 (2.1%) 83 (1.94%)  >4000€ 99 (1.9%) 73 (2.2%) 55 (2.20%) SD, standard deviation. Open in new tab Table 1. Characteristics of the participants in the cohort . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . Census 2011 . Sex n = 10 661 n = 7591 n = 5653 n = 8 657 240  Female 6551 (52.6%) 4784 (52.2%) 3607 (52.5%) 4 585 118 (53.0%) Age group n = 10 661 n = 7591 n = 5648  18–29 1182 (22.1%) 621 (18.4%) 355 (15.4%) 1 470 782 (17.0%)  30–39 1511 (18.8%) 975 (18.7%) 605 (19.1%) 1 598 250 (18.5%)  40–49 1906 (17.3%) 1437 (18.2%) 1049 (18.3%) 1 543 392 (17.8%)  50–59 1801 (14.8%) 1437 (16.2%) 1143 (15.9%) 1 400 011 (16.2%)  60–69 1915 (12.9%) 1440 (13.2%) 1112 (13.7%) 1 186 442 (13.7%)  70–74 849 (5.8%) 645 (6.2%) 491 (6.7%) 496 438 (5.7%)  ≥75 1497 (8.4%) 1036 (9.1%) 893 (11.0%) 961 925 (11.1%) Ethnicity/race n = 10 629 n = 7574 n = 5638  Caucasian 10 342 (96.0%) 7423 (97.1%) 5536 (97.2%) No comparable data  Black 221 (3.4%) 119 (2.5%) 81 (2.3%)  Asian 8 (0.1%) 3 (0.0%) 2 (0.1%)  Romany 20 (0.3%) 7 (0.1%) 5 (0.1%)  Other 38 (0.3%) 22 (0.3%) 14 (0.3%) Years of education  (mean ± SD) 7.41  ±  4.1 8.66  ±  3.90 8.80  ±  3.94 Education level n = 10 585 n = 7546 n = 5615  0–4 years 4726 (33.2%) 3272 (31.7%) 2392 (30.9%) 3 239 724 (37.4%)  5–9 years 2175 (22.6%) 1547 (21.3%) 1122 (19.6%) 2 134 401 (24.6%)  10–12 years 1920 (23.8%) 1391 (24.8%) 1049 (25.6%) 1 560 958 (18.0%)  >12 years 1764 (20.4%) 1336 (22.2%) 1052 (24.0%) 1 741 567 (20.1%) NUTS II n = 10 661 n = 7591 n = 5648  Norte 3122 (34.9%) 2240 (35.8%) 1659 (36.5%) 3 007 823 (34.7%)  Centro 1997 (22.8%) 1504 (23.3%) 1087 (23.2%) 1 938 815 (22.4%)  Lisboa 2484 (26.7%) 1588 (25.4%) 1131 (24.8%) 2 300 053 (26.6%)  Alentejo 669 (7.3%) 422 (7.2%) 320 (7.2%) 633 691 (7.3%)  Algarve 352 (3.8%) 245 (3.8%) 183 (3.7%) 370 704 (4.3%)  Azores 1029 (2.2%) 793 (2.1%) 657 (2.5%) 192 357 (2.2%)  Madeira 1008 (2.3%) 799 (2.4%) 611 (2.4%) 213 797 (2.5%) Marital status n = 10 652 n = 7586 n = 5644  Single 1935 (29.4%) 1285 (28.4%) 922 (28.5%) No comparable data  Married 6111 (50.2%) 4591 (53.2%) 3457 (53.4%)  Divorced 810 (7.4%) 556 (6.8%) 391 (6.1%)  Widow(er) 1414 (8.2%) 910 (7.3%) 697 (7.6%)  Consensual union 382 (4.8%) 244 (4.2%) 177 (4.4%) BMI n = 10 109 n = 6922 n = 5174  Underweight 167 (2.2%) 111 (2.0%) 88 (2.1%) No comparable data  Normal 4063 (45.5%) 2670 (45.5%) 2009 (44.5%)  Overweight 3799 (35.1%) 2788 (37.1%) 2098 (37.7%)  Obese 2080 (17.1%) 1353 (15.4%) 979 (15.7%) Monthly household income n = 7613 n = 5558 n = 4167  <500€ 1994 (19.9%) 1331 (18.0%) 945 (16.66%) No comparable data  501€ to 750€ 1707 (21.7%) 1257 (20.8%) 949 (20.91%)  751€ to 1000€ 1268 (18.8%) 943 (19.0%) 717 (19.89%)  1001€ to 1500€ 1141 (17.2%) 852 (17.5%) 638 (16.97%)  1501€ to 2000€ 657 (9.9%) 511 (10.9%) 386 (11.08%)  2001€ to 2500€ 379 (5.9%) 295 (5.7%) 246 (6.37%)  2501€ to 3000€ 222 (3.0%) 188 (3.8%) 148 (3.98%)  3001€ to 4000€ 146 (1.8%) 108 (2.1%) 83 (1.94%)  >4000€ 99 (1.9%) 73 (2.2%) 55 (2.20%) . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . Census 2011 . Sex n = 10 661 n = 7591 n = 5653 n = 8 657 240  Female 6551 (52.6%) 4784 (52.2%) 3607 (52.5%) 4 585 118 (53.0%) Age group n = 10 661 n = 7591 n = 5648  18–29 1182 (22.1%) 621 (18.4%) 355 (15.4%) 1 470 782 (17.0%)  30–39 1511 (18.8%) 975 (18.7%) 605 (19.1%) 1 598 250 (18.5%)  40–49 1906 (17.3%) 1437 (18.2%) 1049 (18.3%) 1 543 392 (17.8%)  50–59 1801 (14.8%) 1437 (16.2%) 1143 (15.9%) 1 400 011 (16.2%)  60–69 1915 (12.9%) 1440 (13.2%) 1112 (13.7%) 1 186 442 (13.7%)  70–74 849 (5.8%) 645 (6.2%) 491 (6.7%) 496 438 (5.7%)  ≥75 1497 (8.4%) 1036 (9.1%) 893 (11.0%) 961 925 (11.1%) Ethnicity/race n = 10 629 n = 7574 n = 5638  Caucasian 10 342 (96.0%) 7423 (97.1%) 5536 (97.2%) No comparable data  Black 221 (3.4%) 119 (2.5%) 81 (2.3%)  Asian 8 (0.1%) 3 (0.0%) 2 (0.1%)  Romany 20 (0.3%) 7 (0.1%) 5 (0.1%)  Other 38 (0.3%) 22 (0.3%) 14 (0.3%) Years of education  (mean ± SD) 7.41  ±  4.1 8.66  ±  3.90 8.80  ±  3.94 Education level n = 10 585 n = 7546 n = 5615  0–4 years 4726 (33.2%) 3272 (31.7%) 2392 (30.9%) 3 239 724 (37.4%)  5–9 years 2175 (22.6%) 1547 (21.3%) 1122 (19.6%) 2 134 401 (24.6%)  10–12 years 1920 (23.8%) 1391 (24.8%) 1049 (25.6%) 1 560 958 (18.0%)  >12 years 1764 (20.4%) 1336 (22.2%) 1052 (24.0%) 1 741 567 (20.1%) NUTS II n = 10 661 n = 7591 n = 5648  Norte 3122 (34.9%) 2240 (35.8%) 1659 (36.5%) 3 007 823 (34.7%)  Centro 1997 (22.8%) 1504 (23.3%) 1087 (23.2%) 1 938 815 (22.4%)  Lisboa 2484 (26.7%) 1588 (25.4%) 1131 (24.8%) 2 300 053 (26.6%)  Alentejo 669 (7.3%) 422 (7.2%) 320 (7.2%) 633 691 (7.3%)  Algarve 352 (3.8%) 245 (3.8%) 183 (3.7%) 370 704 (4.3%)  Azores 1029 (2.2%) 793 (2.1%) 657 (2.5%) 192 357 (2.2%)  Madeira 1008 (2.3%) 799 (2.4%) 611 (2.4%) 213 797 (2.5%) Marital status n = 10 652 n = 7586 n = 5644  Single 1935 (29.4%) 1285 (28.4%) 922 (28.5%) No comparable data  Married 6111 (50.2%) 4591 (53.2%) 3457 (53.4%)  Divorced 810 (7.4%) 556 (6.8%) 391 (6.1%)  Widow(er) 1414 (8.2%) 910 (7.3%) 697 (7.6%)  Consensual union 382 (4.8%) 244 (4.2%) 177 (4.4%) BMI n = 10 109 n = 6922 n = 5174  Underweight 167 (2.2%) 111 (2.0%) 88 (2.1%) No comparable data  Normal 4063 (45.5%) 2670 (45.5%) 2009 (44.5%)  Overweight 3799 (35.1%) 2788 (37.1%) 2098 (37.7%)  Obese 2080 (17.1%) 1353 (15.4%) 979 (15.7%) Monthly household income n = 7613 n = 5558 n = 4167  <500€ 1994 (19.9%) 1331 (18.0%) 945 (16.66%) No comparable data  501€ to 750€ 1707 (21.7%) 1257 (20.8%) 949 (20.91%)  751€ to 1000€ 1268 (18.8%) 943 (19.0%) 717 (19.89%)  1001€ to 1500€ 1141 (17.2%) 852 (17.5%) 638 (16.97%)  1501€ to 2000€ 657 (9.9%) 511 (10.9%) 386 (11.08%)  2001€ to 2500€ 379 (5.9%) 295 (5.7%) 246 (6.37%)  2501€ to 3000€ 222 (3.0%) 188 (3.8%) 148 (3.98%)  3001€ to 4000€ 146 (1.8%) 108 (2.1%) 83 (1.94%)  >4000€ 99 (1.9%) 73 (2.2%) 55 (2.20%) SD, standard deviation. Open in new tab How often have they been followed up? The EpiDoC study employed cross-sectional and longitudinal study designs (Figure 2). As it used a closed cohort, no new participants were added in any wave. Table 2 presents the attrition rates between EpiDoC 1 and 2, EpiDoC 1 and 3, and EpiDoC 2 and 3. Table 2. Characteristics of the participants in the cohort (attrition rate) . EpiDoC 1 vs EpiDoC2 (attrition rate) . EpiDoC 2 vs EpiDoC3 (attrition rate) . EpiDoC 1 vs EpiDoC3 (attrition rate) . Total 28.80% 25.53% 46.97% Sex  Female 26.97% 24.60% 44.94% Age group  18–29 47.46% 42.83% 69.97%  30–39 35.47% 37.95% 59.96%  40–49 24.61% 27.00% 44.96%  50–59 20.21% 20.46% 36.54%  60–69 24.80% 22.78% 41.93%  70–74 24.03% 23.88% 42.17%  ≥75 30.79% 13.80% 40.35% Ethnicity/race  Caucasian 28.22% 25.42% 46.47%  Black 46.15% 31.93% 63.35%  Asian 62.50% 33.33% 75.00%  Romany 65.00% 28.57% 75.00%  Other 42.11% 36.36% 63.16% Education level  0–4 years 30.77% 26.89% 49.39%  5–9 years 28.87% 27.47% 48.41%  10–12 years 27.55% 24.59% 45.36%  >12 years 24.26% 21.26% 40.36% NUTS II  Norte 28.25% 25.94% 46.86%  Centro 24.69% 27.73% 45.57%  Lisboa 36.07% 28.78% 54.47%  Alentejo 36.92% 24.17% 52.17%  Algarve 30.40% 25.31% 48.01%  Azores 22.93% 17.15% 36.15%  Madeira 20.73% 23.53% 39.38% Marital status  Single 33.59% 28.25% 52.35%  Married 24.87% 24.70% 43.43%  Divorced 31.36% 29.68% 51.73%  Widow(er) 35.64% 23.41% 50.71%  Consensual union 36.13% 27.46% 53.66% BMI  Underweight 33.53% 20.72% 47.31%  Normal 34.29% 24.76% 50.55%  Overweight 26.61% 24.75% 44.77%  Obese 34.95% 27.64% 52.93% Monthly household income  <500€ 33.25% 29.00% 52.61%  501€ to 750€ 26.36% 24.50% 44.41%  751€ to 1000€ 25.63% 23.97% 43.45%  1001€ to 1500€ 25.33% 25.12% 44.08%  1501€ to 2000€ 22.22% 24.46% 41.25%  2001€ to 2500€ 22.16% 16.61% 35.09%  2501€ to 3000€ 15.32% 21.28% 33.33%  3001€ to 4000€ 26.03% 23.15% 43.15%  >4000€ 26.26% 24.66% 44.44% . EpiDoC 1 vs EpiDoC2 (attrition rate) . EpiDoC 2 vs EpiDoC3 (attrition rate) . EpiDoC 1 vs EpiDoC3 (attrition rate) . Total 28.80% 25.53% 46.97% Sex  Female 26.97% 24.60% 44.94% Age group  18–29 47.46% 42.83% 69.97%  30–39 35.47% 37.95% 59.96%  40–49 24.61% 27.00% 44.96%  50–59 20.21% 20.46% 36.54%  60–69 24.80% 22.78% 41.93%  70–74 24.03% 23.88% 42.17%  ≥75 30.79% 13.80% 40.35% Ethnicity/race  Caucasian 28.22% 25.42% 46.47%  Black 46.15% 31.93% 63.35%  Asian 62.50% 33.33% 75.00%  Romany 65.00% 28.57% 75.00%  Other 42.11% 36.36% 63.16% Education level  0–4 years 30.77% 26.89% 49.39%  5–9 years 28.87% 27.47% 48.41%  10–12 years 27.55% 24.59% 45.36%  >12 years 24.26% 21.26% 40.36% NUTS II  Norte 28.25% 25.94% 46.86%  Centro 24.69% 27.73% 45.57%  Lisboa 36.07% 28.78% 54.47%  Alentejo 36.92% 24.17% 52.17%  Algarve 30.40% 25.31% 48.01%  Azores 22.93% 17.15% 36.15%  Madeira 20.73% 23.53% 39.38% Marital status  Single 33.59% 28.25% 52.35%  Married 24.87% 24.70% 43.43%  Divorced 31.36% 29.68% 51.73%  Widow(er) 35.64% 23.41% 50.71%  Consensual union 36.13% 27.46% 53.66% BMI  Underweight 33.53% 20.72% 47.31%  Normal 34.29% 24.76% 50.55%  Overweight 26.61% 24.75% 44.77%  Obese 34.95% 27.64% 52.93% Monthly household income  <500€ 33.25% 29.00% 52.61%  501€ to 750€ 26.36% 24.50% 44.41%  751€ to 1000€ 25.63% 23.97% 43.45%  1001€ to 1500€ 25.33% 25.12% 44.08%  1501€ to 2000€ 22.22% 24.46% 41.25%  2001€ to 2500€ 22.16% 16.61% 35.09%  2501€ to 3000€ 15.32% 21.28% 33.33%  3001€ to 4000€ 26.03% 23.15% 43.15%  >4000€ 26.26% 24.66% 44.44% Open in new tab Table 2. Characteristics of the participants in the cohort (attrition rate) . EpiDoC 1 vs EpiDoC2 (attrition rate) . EpiDoC 2 vs EpiDoC3 (attrition rate) . EpiDoC 1 vs EpiDoC3 (attrition rate) . Total 28.80% 25.53% 46.97% Sex  Female 26.97% 24.60% 44.94% Age group  18–29 47.46% 42.83% 69.97%  30–39 35.47% 37.95% 59.96%  40–49 24.61% 27.00% 44.96%  50–59 20.21% 20.46% 36.54%  60–69 24.80% 22.78% 41.93%  70–74 24.03% 23.88% 42.17%  ≥75 30.79% 13.80% 40.35% Ethnicity/race  Caucasian 28.22% 25.42% 46.47%  Black 46.15% 31.93% 63.35%  Asian 62.50% 33.33% 75.00%  Romany 65.00% 28.57% 75.00%  Other 42.11% 36.36% 63.16% Education level  0–4 years 30.77% 26.89% 49.39%  5–9 years 28.87% 27.47% 48.41%  10–12 years 27.55% 24.59% 45.36%  >12 years 24.26% 21.26% 40.36% NUTS II  Norte 28.25% 25.94% 46.86%  Centro 24.69% 27.73% 45.57%  Lisboa 36.07% 28.78% 54.47%  Alentejo 36.92% 24.17% 52.17%  Algarve 30.40% 25.31% 48.01%  Azores 22.93% 17.15% 36.15%  Madeira 20.73% 23.53% 39.38% Marital status  Single 33.59% 28.25% 52.35%  Married 24.87% 24.70% 43.43%  Divorced 31.36% 29.68% 51.73%  Widow(er) 35.64% 23.41% 50.71%  Consensual union 36.13% 27.46% 53.66% BMI  Underweight 33.53% 20.72% 47.31%  Normal 34.29% 24.76% 50.55%  Overweight 26.61% 24.75% 44.77%  Obese 34.95% 27.64% 52.93% Monthly household income  <500€ 33.25% 29.00% 52.61%  501€ to 750€ 26.36% 24.50% 44.41%  751€ to 1000€ 25.63% 23.97% 43.45%  1001€ to 1500€ 25.33% 25.12% 44.08%  1501€ to 2000€ 22.22% 24.46% 41.25%  2001€ to 2500€ 22.16% 16.61% 35.09%  2501€ to 3000€ 15.32% 21.28% 33.33%  3001€ to 4000€ 26.03% 23.15% 43.15%  >4000€ 26.26% 24.66% 44.44% . EpiDoC 1 vs EpiDoC2 (attrition rate) . EpiDoC 2 vs EpiDoC3 (attrition rate) . EpiDoC 1 vs EpiDoC3 (attrition rate) . Total 28.80% 25.53% 46.97% Sex  Female 26.97% 24.60% 44.94% Age group  18–29 47.46% 42.83% 69.97%  30–39 35.47% 37.95% 59.96%  40–49 24.61% 27.00% 44.96%  50–59 20.21% 20.46% 36.54%  60–69 24.80% 22.78% 41.93%  70–74 24.03% 23.88% 42.17%  ≥75 30.79% 13.80% 40.35% Ethnicity/race  Caucasian 28.22% 25.42% 46.47%  Black 46.15% 31.93% 63.35%  Asian 62.50% 33.33% 75.00%  Romany 65.00% 28.57% 75.00%  Other 42.11% 36.36% 63.16% Education level  0–4 years 30.77% 26.89% 49.39%  5–9 years 28.87% 27.47% 48.41%  10–12 years 27.55% 24.59% 45.36%  >12 years 24.26% 21.26% 40.36% NUTS II  Norte 28.25% 25.94% 46.86%  Centro 24.69% 27.73% 45.57%  Lisboa 36.07% 28.78% 54.47%  Alentejo 36.92% 24.17% 52.17%  Algarve 30.40% 25.31% 48.01%  Azores 22.93% 17.15% 36.15%  Madeira 20.73% 23.53% 39.38% Marital status  Single 33.59% 28.25% 52.35%  Married 24.87% 24.70% 43.43%  Divorced 31.36% 29.68% 51.73%  Widow(er) 35.64% 23.41% 50.71%  Consensual union 36.13% 27.46% 53.66% BMI  Underweight 33.53% 20.72% 47.31%  Normal 34.29% 24.76% 50.55%  Overweight 26.61% 24.75% 44.77%  Obese 34.95% 27.64% 52.93% Monthly household income  <500€ 33.25% 29.00% 52.61%  501€ to 750€ 26.36% 24.50% 44.41%  751€ to 1000€ 25.63% 23.97% 43.45%  1001€ to 1500€ 25.33% 25.12% 44.08%  1501€ to 2000€ 22.22% 24.46% 41.25%  2001€ to 2500€ 22.16% 16.61% 35.09%  2501€ to 3000€ 15.32% 21.28% 33.33%  3001€ to 4000€ 26.03% 23.15% 43.15%  >4000€ 26.26% 24.66% 44.44% Open in new tab Figure 2. Open in new tabDownload slide Flowchart of EpiDoC study. Loss to follow-up The participation rate in EpiDoC 3 was of 53.03%. The attrition was most pronounced in younger adults (18–29 years old). Of the 10 661 participants in EpiDoC 1,509 (4.8%) refused to sign the consent form for follow-up. Of the resulting 10 153 eligible participants for EpiDoC 2, 79 (0.8%) had died, 179 (1.8%) wished to leave the study and 917 (9.0%) had an invalid contact. Thus, a total of 1639 participants were lost to follow-up; these subjects had a mean age of 55 years, and 962 (58.7%) were women. Between EpiDoC 2 and 3, 51 (0.6%) participants had died, 232 (2.6%) wished to leave the study and 721 (8.0%) had an invalid contact. Thus, a total of 1004 participants were lost to follow-up; these individuals had a mean age of 56 years, and 620 (61.8%) were women. Figure 1 shows the flowchart of the EpiDoC study. What has been measured? Data collection included measures for five domains that were central to the longitudinal study: sociodemographic characteristics, lifestyle characteristics, health and clinical characteristics, health care resource consumption, a population-based biobank (total blood, serum and DNA) and imaging data (peripheral DXA and X-ray of the affected joint) (Table 3). For reasons of longitudinal comparison, most measurement tools were used consistently across waves. However, some measurement tools were updated or improved, new measurement tools were added and old measurement tools were removed as needed. Table 3. Data collected over EpiDoC study . EpiDoC 1 EpiReumaPt (CESOP) 10 661 . EpiDoC 1 EpiReumaPt (medical appointments) 3877 . EpiDoC 2 CoReumaPt 7591 . EpiDoC 3 Saúde.Come 5653 . Sociodemographic and economic data  Sex X  Age X  Ethnicity X  Nationality X  Years of education and educational level X  Marital status X  Employment status X X X  Household income X X  Household composition X X  Number of people <18 y in household X X  Number of people >65 y in household X  Region (NUT II) X  Location and district X  Home & neighbourhood characteristics X  Single-parent families X  Income perception X Anthropometric data  Self-reported height (in cm) X X X X  Self-reported weight (in kg) X X X X  Body mass index (kg/m2) X X X X Self-reported chronic diseases  High blood pressure, diabetes, high cholesterol level, pulmonary disease, cardiac disease, gastrointestinal disease, neurological disease, allergies, mental disease, neoplastic disease, thyroid and parathyroid disease, hyperuricaemia and urinary disease X X X X Rheumatic diseases  Rheumatoid arthritis, spondyloarthritis, psoriatic arthritis, osteoarthritis, osteoporosis, gout, polymyalgia rheumatica, systemic lupus erythematosus, fibromyalgia, periarticular diseases, low back pain, inflammatory low back pain, chondrocalcinosis and other RMD X X X X  Who diagnosed RMD X X X  Rheumatic complaints X X X X  Medical history X  Physical examination X Anxiety, depression, physical function and quality of life  Hospital Anxiety and Depression Scale (HADS) X X  Health Assessment Questionnaire (HAQ) X X X  Short Form Health Survey (SF-36) X  European Quality of Life questionnaire (EQ-5D-3L) X X X Falls and bone fractures  Suffered any fall, where the fall happened (home, street, work), number of falls (home, street, work), suffered any bone fracture, number of bone fractures and location of bone fracture X X X Health and employment  Retired due to disease, retired due to RMD, work absenteeism due to disease, work disabled due to RMD, unemployed due to disease, unemployed due to RMD, number working h/week and changed employment status (past year) due to RMD X X X Health and economic  Chronic disease management difficulties, medication non-adherence due to economic constraints, and reduction in visits to medical appointments due to economic constraints X Hospitalizations, home care assistance and medical appointments  Was hospitalized (past 12 months/since last contact), reason and duration of hospitalization, home care assistance (past 12 months/since last contact, currently), who provides and who pays for home care assistance, medical appointments (past 12 months/since last contact), number private/public medical appointments, private medical appointments with/without insurance, public medical appointments in hospital/health care centre, number private/public medical appointments by specialty, health care system (ADSE, subsystems, private insurance), medications and other treatments, medicine(s) currently taking, other treatments (physical and rehabilitation medicine, behavioural therapy etc.) and alternative treatments (acupuncture, homeopathy etc.) X X X Lifestyle data  Smoking habits (current/past smoker, number of cigarettes, smoking duration) X X X  Alcohol intake (frequency, number of units) X X X  Coffee intake X  Physical exercise (frequency, type, age when started) X X X  Sleep habits (h/day) X X  Frequency of watching TV X X  Frequency of using computer/videogames/tablets X X  Frequency of using internet X X Dietary intake and behaviours  Frequency of soup, vegetables, fruit, meat, fish, milk/dairy, water consumption X X  Adherence to Mediterranean diet X Food insecurity X Patient innovation to cope with disability X Biobank and imaging data  Serum, whole blood, DNA, peripheral BMD (wrist), X-ray of the affected joint (hand, hip, knee), calcaneus and wrist BMA X . EpiDoC 1 EpiReumaPt (CESOP) 10 661 . EpiDoC 1 EpiReumaPt (medical appointments) 3877 . EpiDoC 2 CoReumaPt 7591 . EpiDoC 3 Saúde.Come 5653 . Sociodemographic and economic data  Sex X  Age X  Ethnicity X  Nationality X  Years of education and educational level X  Marital status X  Employment status X X X  Household income X X  Household composition X X  Number of people <18 y in household X X  Number of people >65 y in household X  Region (NUT II) X  Location and district X  Home & neighbourhood characteristics X  Single-parent families X  Income perception X Anthropometric data  Self-reported height (in cm) X X X X  Self-reported weight (in kg) X X X X  Body mass index (kg/m2) X X X X Self-reported chronic diseases  High blood pressure, diabetes, high cholesterol level, pulmonary disease, cardiac disease, gastrointestinal disease, neurological disease, allergies, mental disease, neoplastic disease, thyroid and parathyroid disease, hyperuricaemia and urinary disease X X X X Rheumatic diseases  Rheumatoid arthritis, spondyloarthritis, psoriatic arthritis, osteoarthritis, osteoporosis, gout, polymyalgia rheumatica, systemic lupus erythematosus, fibromyalgia, periarticular diseases, low back pain, inflammatory low back pain, chondrocalcinosis and other RMD X X X X  Who diagnosed RMD X X X  Rheumatic complaints X X X X  Medical history X  Physical examination X Anxiety, depression, physical function and quality of life  Hospital Anxiety and Depression Scale (HADS) X X  Health Assessment Questionnaire (HAQ) X X X  Short Form Health Survey (SF-36) X  European Quality of Life questionnaire (EQ-5D-3L) X X X Falls and bone fractures  Suffered any fall, where the fall happened (home, street, work), number of falls (home, street, work), suffered any bone fracture, number of bone fractures and location of bone fracture X X X Health and employment  Retired due to disease, retired due to RMD, work absenteeism due to disease, work disabled due to RMD, unemployed due to disease, unemployed due to RMD, number working h/week and changed employment status (past year) due to RMD X X X Health and economic  Chronic disease management difficulties, medication non-adherence due to economic constraints, and reduction in visits to medical appointments due to economic constraints X Hospitalizations, home care assistance and medical appointments  Was hospitalized (past 12 months/since last contact), reason and duration of hospitalization, home care assistance (past 12 months/since last contact, currently), who provides and who pays for home care assistance, medical appointments (past 12 months/since last contact), number private/public medical appointments, private medical appointments with/without insurance, public medical appointments in hospital/health care centre, number private/public medical appointments by specialty, health care system (ADSE, subsystems, private insurance), medications and other treatments, medicine(s) currently taking, other treatments (physical and rehabilitation medicine, behavioural therapy etc.) and alternative treatments (acupuncture, homeopathy etc.) X X X Lifestyle data  Smoking habits (current/past smoker, number of cigarettes, smoking duration) X X X  Alcohol intake (frequency, number of units) X X X  Coffee intake X  Physical exercise (frequency, type, age when started) X X X  Sleep habits (h/day) X X  Frequency of watching TV X X  Frequency of using computer/videogames/tablets X X  Frequency of using internet X X Dietary intake and behaviours  Frequency of soup, vegetables, fruit, meat, fish, milk/dairy, water consumption X X  Adherence to Mediterranean diet X Food insecurity X Patient innovation to cope with disability X Biobank and imaging data  Serum, whole blood, DNA, peripheral BMD (wrist), X-ray of the affected joint (hand, hip, knee), calcaneus and wrist BMA X Open in new tab Table 3. Data collected over EpiDoC study . EpiDoC 1 EpiReumaPt (CESOP) 10 661 . EpiDoC 1 EpiReumaPt (medical appointments) 3877 . EpiDoC 2 CoReumaPt 7591 . EpiDoC 3 Saúde.Come 5653 . Sociodemographic and economic data  Sex X  Age X  Ethnicity X  Nationality X  Years of education and educational level X  Marital status X  Employment status X X X  Household income X X  Household composition X X  Number of people <18 y in household X X  Number of people >65 y in household X  Region (NUT II) X  Location and district X  Home & neighbourhood characteristics X  Single-parent families X  Income perception X Anthropometric data  Self-reported height (in cm) X X X X  Self-reported weight (in kg) X X X X  Body mass index (kg/m2) X X X X Self-reported chronic diseases  High blood pressure, diabetes, high cholesterol level, pulmonary disease, cardiac disease, gastrointestinal disease, neurological disease, allergies, mental disease, neoplastic disease, thyroid and parathyroid disease, hyperuricaemia and urinary disease X X X X Rheumatic diseases  Rheumatoid arthritis, spondyloarthritis, psoriatic arthritis, osteoarthritis, osteoporosis, gout, polymyalgia rheumatica, systemic lupus erythematosus, fibromyalgia, periarticular diseases, low back pain, inflammatory low back pain, chondrocalcinosis and other RMD X X X X  Who diagnosed RMD X X X  Rheumatic complaints X X X X  Medical history X  Physical examination X Anxiety, depression, physical function and quality of life  Hospital Anxiety and Depression Scale (HADS) X X  Health Assessment Questionnaire (HAQ) X X X  Short Form Health Survey (SF-36) X  European Quality of Life questionnaire (EQ-5D-3L) X X X Falls and bone fractures  Suffered any fall, where the fall happened (home, street, work), number of falls (home, street, work), suffered any bone fracture, number of bone fractures and location of bone fracture X X X Health and employment  Retired due to disease, retired due to RMD, work absenteeism due to disease, work disabled due to RMD, unemployed due to disease, unemployed due to RMD, number working h/week and changed employment status (past year) due to RMD X X X Health and economic  Chronic disease management difficulties, medication non-adherence due to economic constraints, and reduction in visits to medical appointments due to economic constraints X Hospitalizations, home care assistance and medical appointments  Was hospitalized (past 12 months/since last contact), reason and duration of hospitalization, home care assistance (past 12 months/since last contact, currently), who provides and who pays for home care assistance, medical appointments (past 12 months/since last contact), number private/public medical appointments, private medical appointments with/without insurance, public medical appointments in hospital/health care centre, number private/public medical appointments by specialty, health care system (ADSE, subsystems, private insurance), medications and other treatments, medicine(s) currently taking, other treatments (physical and rehabilitation medicine, behavioural therapy etc.) and alternative treatments (acupuncture, homeopathy etc.) X X X Lifestyle data  Smoking habits (current/past smoker, number of cigarettes, smoking duration) X X X  Alcohol intake (frequency, number of units) X X X  Coffee intake X  Physical exercise (frequency, type, age when started) X X X  Sleep habits (h/day) X X  Frequency of watching TV X X  Frequency of using computer/videogames/tablets X X  Frequency of using internet X X Dietary intake and behaviours  Frequency of soup, vegetables, fruit, meat, fish, milk/dairy, water consumption X X  Adherence to Mediterranean diet X Food insecurity X Patient innovation to cope with disability X Biobank and imaging data  Serum, whole blood, DNA, peripheral BMD (wrist), X-ray of the affected joint (hand, hip, knee), calcaneus and wrist BMA X . EpiDoC 1 EpiReumaPt (CESOP) 10 661 . EpiDoC 1 EpiReumaPt (medical appointments) 3877 . EpiDoC 2 CoReumaPt 7591 . EpiDoC 3 Saúde.Come 5653 . Sociodemographic and economic data  Sex X  Age X  Ethnicity X  Nationality X  Years of education and educational level X  Marital status X  Employment status X X X  Household income X X  Household composition X X  Number of people <18 y in household X X  Number of people >65 y in household X  Region (NUT II) X  Location and district X  Home & neighbourhood characteristics X  Single-parent families X  Income perception X Anthropometric data  Self-reported height (in cm) X X X X  Self-reported weight (in kg) X X X X  Body mass index (kg/m2) X X X X Self-reported chronic diseases  High blood pressure, diabetes, high cholesterol level, pulmonary disease, cardiac disease, gastrointestinal disease, neurological disease, allergies, mental disease, neoplastic disease, thyroid and parathyroid disease, hyperuricaemia and urinary disease X X X X Rheumatic diseases  Rheumatoid arthritis, spondyloarthritis, psoriatic arthritis, osteoarthritis, osteoporosis, gout, polymyalgia rheumatica, systemic lupus erythematosus, fibromyalgia, periarticular diseases, low back pain, inflammatory low back pain, chondrocalcinosis and other RMD X X X X  Who diagnosed RMD X X X  Rheumatic complaints X X X X  Medical history X  Physical examination X Anxiety, depression, physical function and quality of life  Hospital Anxiety and Depression Scale (HADS) X X  Health Assessment Questionnaire (HAQ) X X X  Short Form Health Survey (SF-36) X  European Quality of Life questionnaire (EQ-5D-3L) X X X Falls and bone fractures  Suffered any fall, where the fall happened (home, street, work), number of falls (home, street, work), suffered any bone fracture, number of bone fractures and location of bone fracture X X X Health and employment  Retired due to disease, retired due to RMD, work absenteeism due to disease, work disabled due to RMD, unemployed due to disease, unemployed due to RMD, number working h/week and changed employment status (past year) due to RMD X X X Health and economic  Chronic disease management difficulties, medication non-adherence due to economic constraints, and reduction in visits to medical appointments due to economic constraints X Hospitalizations, home care assistance and medical appointments  Was hospitalized (past 12 months/since last contact), reason and duration of hospitalization, home care assistance (past 12 months/since last contact, currently), who provides and who pays for home care assistance, medical appointments (past 12 months/since last contact), number private/public medical appointments, private medical appointments with/without insurance, public medical appointments in hospital/health care centre, number private/public medical appointments by specialty, health care system (ADSE, subsystems, private insurance), medications and other treatments, medicine(s) currently taking, other treatments (physical and rehabilitation medicine, behavioural therapy etc.) and alternative treatments (acupuncture, homeopathy etc.) X X X Lifestyle data  Smoking habits (current/past smoker, number of cigarettes, smoking duration) X X X  Alcohol intake (frequency, number of units) X X X  Coffee intake X  Physical exercise (frequency, type, age when started) X X X  Sleep habits (h/day) X X  Frequency of watching TV X X  Frequency of using computer/videogames/tablets X X  Frequency of using internet X X Dietary intake and behaviours  Frequency of soup, vegetables, fruit, meat, fish, milk/dairy, water consumption X X  Adherence to Mediterranean diet X Food insecurity X Patient innovation to cope with disability X Biobank and imaging data  Serum, whole blood, DNA, peripheral BMD (wrist), X-ray of the affected joint (hand, hip, knee), calcaneus and wrist BMA X Open in new tab Of the measurements obtained across all three waves, lifestyle variables included smoking habits, alcohol intake and physical exercise. Health variables included anthropometric measures, self-reported chronic diseases, rheumatic diseases, a health assessment questionnaire (HAQ)7 and the European Quality of Life Survey with five dimensions and three levels (EQ-5D-3L).8,9 Employment variables included employment status, retirement due to disease, retirement due to RMD, work absenteeism due to disease, work disability due to RMD, unemployment due to disease, unemployment due to RMD, number of working hours/week and changed employment status due to RMD. Health care resource variables included hospitalization events (in previous 12 months since last contact), their reason and their duration. Concerning falls and bone fractures, variables included any falls or bone fractures and the number and location of bone fractures. Sociodemographic data, including sex, age, ethnicity, years of education and education level, and marital status, were collected only in EpiDoC 1, based on the assumption that these characteristics would not change over time. Other information obtained only in EpiDoC 1 were household income, household composition, coffee intake and health information from the 36-item Short Form Survey (SF-36).10 Information obtained only in EpiDoC 1 and 2 were the Hospital Anxiety and Depression Scale (HADS)11 and home care assistance (in previous 12 months or since last contact), its provider and its payer. Information obtained only in EpiDoC 2 and 3 included: sleep habits; frequency of watching TV, using computer/videogames/tablets and using the internet; number of meals per day; frequency of soup, vegetable, fruit, meat, fish, milk/dairy and water consumption; numbers of medical appointments (in previous 12 months or since last contact), private versus public medical appointments, private medical appointments with or without insurance, public medical appointments in a hospital/health care centre and private or public medical appointments by specialty. Information obtained only in EpiDoC 3 were frequency of olive oil, wine, beans, fat and sugar consumption; attitudes toward food; a food insecurity scale; and characteristics of food acquisition and preparation. Population-based biobank and imaging data were collected in EpiDoC 1 during medical appointments at the local primary care centre. Blood samples were collected from 3608 participants (DNA, serum and whole blood). Taking into consideration the imaging reservoir, there were a total of 3342 participants who had a forearm bone mineral density evaluation through peripheral DXA. Also, bone mineral assessment (BMA) using a high-resolution digital X-ray machine (D3A, France) was collected from 2422 wrists and 2228 calcaneus bones. Simple X-rays were performed to examine 438 hands, 122 hips, 479 knees, 1265 lumbar spines, 691 thoracic spines and 206 cervical spines, according to participants’ musculoskeletal complaints. All data collected, including biobank and imaging data, are detailed in Table 3. What has it found? Key findings and publications Over 24 peer-reviewed journal publications based on EpiDoC data have been published to date, covering a wide range of scientific domains. A full list of publications can be found on our website[http://cedoc.unl.pt/epidoc-unit/]. Sample overviews of study data are shown in Tables 1, 2 and 4. Here, we summarize key findings. Table 4. Prevalence and 95% of confidence interval of reported chronic diseases and lifestyle habits . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . n = 10 661 . n = 7591 . n = 5653 . Reported diseases Chronic diseases n = 10 661 95% CI n = 7591 95% CI n = 5653 95% CI  High blood pressure 3369 (23.1%) 21.9–24.9 2538 (24.1%) 22.7–25.5 1872 (24.8%) 23.1–26.7  Diabetes 1217 (8.3%) 7.6–9.1 877 (8.6%) 7.8–9.5 690 (9.2%) 8.1–10.4  High cholesterol level 3360 (24.4%) 23.2–.25.7 2595 (25.9%) 24.5–27.4 1831 (25.3%) 23.6–27.2  Lung disease 637 (5.4%) 4.6–6.3 496 (5.7%) 4.8–6.7 213 (2.8%) 2.4–3.3  Cardiac disease 1366 (10.5%) 9.4–11.6 1034 (11.9%) 10.5–13.4 704 (9.8%) 8.7–11.1  Gastrointestinal disease 1837 (14.9%) 13.8–16.1 1411 (16.1%) 14.7–17.6 544 (8.8%) 7.6–10.3  Neurological disease 418 (3.3%) 2.8–3.9 311 (3.4%) 2.8–4.1 212 (2.9%) 2.4–3.4  Allergies 2287 (21.2%) 19.9–22.7 1720 (22.8%) 21.2–24.5 548 (10.3%) 8.6–12.3  Mental disease 1619 (12.9%) 11.7–14.1 1274 (14.1%) 12.4–16.0 1008 (13.4%) 12.3–14.5  Cancer 439 (3.4%) 2.8–4.2 364 (4.0%) 3.3–4.9 318 (4.6%) 3.8–5.5  Hyperuricaemia 690 (5.2%) 4.7–5.8 514 (5.4%) 4.8–5.9 130 (1.9%) 1.5–2.4  Renal colic 885 (7.0%) 6.4–7.8 716 (8.4%) 7.3–9.6 250 (4.3%) 3.4–5.4  Rheumatic disease 2994 (21.2%) 20.0–22.5 2552 (25.5%) 24.0–27.1 2096 (29.5%) 27.5–31.5 Lifestyle habits Alcohol  Never 4625 (37.2%) 35.6–38.8 3150 (37.1%) 35.2–39.2 1945 (30.6%) 28.2–33.2  Occasionally 3967 (42.6%) 40.9–44.3 2437 (39.6%) 37.7–41.7 2020 (39.6%) 37.4–42.0  Daily 2050 (20.2%) 18.9–21.6 1693 (23.2%) 21.7–24.8 1565 (29.8%) 27.7–31.9 Smoking habits  Never/occasionally 8800 (76.8%) 75.1–78.4 4447 (54.4%) 52.3–56.4 3584 (58.8%) 56.3–61.3  Past smokinga Not applicable 1522 (21.1%) 19.6–22.6 1149 (21.1%) 19.4–23.0  Present smoker 1854 (23.2%) 21.6–24.9 1289 (24.5%) 22.4–26.7 802 (20.0%) 17.6–22.7 Physical activity  Regular 3499 (37.0%) 35.3–38.6 3442 (50.1%) 48.1–52.1 2147 (40.8%) 38.5–43.2  Not regular 7155 (63.0%) 61.3–64.6 3976 (49.8%) 47.9–51.9 3498 (59.2%) 56.8–61.5 . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . n = 10 661 . n = 7591 . n = 5653 . Reported diseases Chronic diseases n = 10 661 95% CI n = 7591 95% CI n = 5653 95% CI  High blood pressure 3369 (23.1%) 21.9–24.9 2538 (24.1%) 22.7–25.5 1872 (24.8%) 23.1–26.7  Diabetes 1217 (8.3%) 7.6–9.1 877 (8.6%) 7.8–9.5 690 (9.2%) 8.1–10.4  High cholesterol level 3360 (24.4%) 23.2–.25.7 2595 (25.9%) 24.5–27.4 1831 (25.3%) 23.6–27.2  Lung disease 637 (5.4%) 4.6–6.3 496 (5.7%) 4.8–6.7 213 (2.8%) 2.4–3.3  Cardiac disease 1366 (10.5%) 9.4–11.6 1034 (11.9%) 10.5–13.4 704 (9.8%) 8.7–11.1  Gastrointestinal disease 1837 (14.9%) 13.8–16.1 1411 (16.1%) 14.7–17.6 544 (8.8%) 7.6–10.3  Neurological disease 418 (3.3%) 2.8–3.9 311 (3.4%) 2.8–4.1 212 (2.9%) 2.4–3.4  Allergies 2287 (21.2%) 19.9–22.7 1720 (22.8%) 21.2–24.5 548 (10.3%) 8.6–12.3  Mental disease 1619 (12.9%) 11.7–14.1 1274 (14.1%) 12.4–16.0 1008 (13.4%) 12.3–14.5  Cancer 439 (3.4%) 2.8–4.2 364 (4.0%) 3.3–4.9 318 (4.6%) 3.8–5.5  Hyperuricaemia 690 (5.2%) 4.7–5.8 514 (5.4%) 4.8–5.9 130 (1.9%) 1.5–2.4  Renal colic 885 (7.0%) 6.4–7.8 716 (8.4%) 7.3–9.6 250 (4.3%) 3.4–5.4  Rheumatic disease 2994 (21.2%) 20.0–22.5 2552 (25.5%) 24.0–27.1 2096 (29.5%) 27.5–31.5 Lifestyle habits Alcohol  Never 4625 (37.2%) 35.6–38.8 3150 (37.1%) 35.2–39.2 1945 (30.6%) 28.2–33.2  Occasionally 3967 (42.6%) 40.9–44.3 2437 (39.6%) 37.7–41.7 2020 (39.6%) 37.4–42.0  Daily 2050 (20.2%) 18.9–21.6 1693 (23.2%) 21.7–24.8 1565 (29.8%) 27.7–31.9 Smoking habits  Never/occasionally 8800 (76.8%) 75.1–78.4 4447 (54.4%) 52.3–56.4 3584 (58.8%) 56.3–61.3  Past smokinga Not applicable 1522 (21.1%) 19.6–22.6 1149 (21.1%) 19.4–23.0  Present smoker 1854 (23.2%) 21.6–24.9 1289 (24.5%) 22.4–26.7 802 (20.0%) 17.6–22.7 Physical activity  Regular 3499 (37.0%) 35.3–38.6 3442 (50.1%) 48.1–52.1 2147 (40.8%) 38.5–43.2  Not regular 7155 (63.0%) 61.3–64.6 3976 (49.8%) 47.9–51.9 3498 (59.2%) 56.8–61.5 a Past smoker was not included at baseline. Open in new tab Table 4. Prevalence and 95% of confidence interval of reported chronic diseases and lifestyle habits . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . n = 10 661 . n = 7591 . n = 5653 . Reported diseases Chronic diseases n = 10 661 95% CI n = 7591 95% CI n = 5653 95% CI  High blood pressure 3369 (23.1%) 21.9–24.9 2538 (24.1%) 22.7–25.5 1872 (24.8%) 23.1–26.7  Diabetes 1217 (8.3%) 7.6–9.1 877 (8.6%) 7.8–9.5 690 (9.2%) 8.1–10.4  High cholesterol level 3360 (24.4%) 23.2–.25.7 2595 (25.9%) 24.5–27.4 1831 (25.3%) 23.6–27.2  Lung disease 637 (5.4%) 4.6–6.3 496 (5.7%) 4.8–6.7 213 (2.8%) 2.4–3.3  Cardiac disease 1366 (10.5%) 9.4–11.6 1034 (11.9%) 10.5–13.4 704 (9.8%) 8.7–11.1  Gastrointestinal disease 1837 (14.9%) 13.8–16.1 1411 (16.1%) 14.7–17.6 544 (8.8%) 7.6–10.3  Neurological disease 418 (3.3%) 2.8–3.9 311 (3.4%) 2.8–4.1 212 (2.9%) 2.4–3.4  Allergies 2287 (21.2%) 19.9–22.7 1720 (22.8%) 21.2–24.5 548 (10.3%) 8.6–12.3  Mental disease 1619 (12.9%) 11.7–14.1 1274 (14.1%) 12.4–16.0 1008 (13.4%) 12.3–14.5  Cancer 439 (3.4%) 2.8–4.2 364 (4.0%) 3.3–4.9 318 (4.6%) 3.8–5.5  Hyperuricaemia 690 (5.2%) 4.7–5.8 514 (5.4%) 4.8–5.9 130 (1.9%) 1.5–2.4  Renal colic 885 (7.0%) 6.4–7.8 716 (8.4%) 7.3–9.6 250 (4.3%) 3.4–5.4  Rheumatic disease 2994 (21.2%) 20.0–22.5 2552 (25.5%) 24.0–27.1 2096 (29.5%) 27.5–31.5 Lifestyle habits Alcohol  Never 4625 (37.2%) 35.6–38.8 3150 (37.1%) 35.2–39.2 1945 (30.6%) 28.2–33.2  Occasionally 3967 (42.6%) 40.9–44.3 2437 (39.6%) 37.7–41.7 2020 (39.6%) 37.4–42.0  Daily 2050 (20.2%) 18.9–21.6 1693 (23.2%) 21.7–24.8 1565 (29.8%) 27.7–31.9 Smoking habits  Never/occasionally 8800 (76.8%) 75.1–78.4 4447 (54.4%) 52.3–56.4 3584 (58.8%) 56.3–61.3  Past smokinga Not applicable 1522 (21.1%) 19.6–22.6 1149 (21.1%) 19.4–23.0  Present smoker 1854 (23.2%) 21.6–24.9 1289 (24.5%) 22.4–26.7 802 (20.0%) 17.6–22.7 Physical activity  Regular 3499 (37.0%) 35.3–38.6 3442 (50.1%) 48.1–52.1 2147 (40.8%) 38.5–43.2  Not regular 7155 (63.0%) 61.3–64.6 3976 (49.8%) 47.9–51.9 3498 (59.2%) 56.8–61.5 . EpiDoC 1 . EpiDoC 2 . EpiDoC 3 . n = 10 661 . n = 7591 . n = 5653 . Reported diseases Chronic diseases n = 10 661 95% CI n = 7591 95% CI n = 5653 95% CI  High blood pressure 3369 (23.1%) 21.9–24.9 2538 (24.1%) 22.7–25.5 1872 (24.8%) 23.1–26.7  Diabetes 1217 (8.3%) 7.6–9.1 877 (8.6%) 7.8–9.5 690 (9.2%) 8.1–10.4  High cholesterol level 3360 (24.4%) 23.2–.25.7 2595 (25.9%) 24.5–27.4 1831 (25.3%) 23.6–27.2  Lung disease 637 (5.4%) 4.6–6.3 496 (5.7%) 4.8–6.7 213 (2.8%) 2.4–3.3  Cardiac disease 1366 (10.5%) 9.4–11.6 1034 (11.9%) 10.5–13.4 704 (9.8%) 8.7–11.1  Gastrointestinal disease 1837 (14.9%) 13.8–16.1 1411 (16.1%) 14.7–17.6 544 (8.8%) 7.6–10.3  Neurological disease 418 (3.3%) 2.8–3.9 311 (3.4%) 2.8–4.1 212 (2.9%) 2.4–3.4  Allergies 2287 (21.2%) 19.9–22.7 1720 (22.8%) 21.2–24.5 548 (10.3%) 8.6–12.3  Mental disease 1619 (12.9%) 11.7–14.1 1274 (14.1%) 12.4–16.0 1008 (13.4%) 12.3–14.5  Cancer 439 (3.4%) 2.8–4.2 364 (4.0%) 3.3–4.9 318 (4.6%) 3.8–5.5  Hyperuricaemia 690 (5.2%) 4.7–5.8 514 (5.4%) 4.8–5.9 130 (1.9%) 1.5–2.4  Renal colic 885 (7.0%) 6.4–7.8 716 (8.4%) 7.3–9.6 250 (4.3%) 3.4–5.4  Rheumatic disease 2994 (21.2%) 20.0–22.5 2552 (25.5%) 24.0–27.1 2096 (29.5%) 27.5–31.5 Lifestyle habits Alcohol  Never 4625 (37.2%) 35.6–38.8 3150 (37.1%) 35.2–39.2 1945 (30.6%) 28.2–33.2  Occasionally 3967 (42.6%) 40.9–44.3 2437 (39.6%) 37.7–41.7 2020 (39.6%) 37.4–42.0  Daily 2050 (20.2%) 18.9–21.6 1693 (23.2%) 21.7–24.8 1565 (29.8%) 27.7–31.9 Smoking habits  Never/occasionally 8800 (76.8%) 75.1–78.4 4447 (54.4%) 52.3–56.4 3584 (58.8%) 56.3–61.3  Past smokinga Not applicable 1522 (21.1%) 19.6–22.6 1149 (21.1%) 19.4–23.0  Present smoker 1854 (23.2%) 21.6–24.9 1289 (24.5%) 22.4–26.7 802 (20.0%) 17.6–22.7 Physical activity  Regular 3499 (37.0%) 35.3–38.6 3442 (50.1%) 48.1–52.1 2147 (40.8%) 38.5–43.2  Not regular 7155 (63.0%) 61.3–64.6 3976 (49.8%) 47.9–51.9 3498 (59.2%) 56.8–61.5 a Past smoker was not included at baseline. Open in new tab In EpiDoC 1, we characterized socioeconomic features of the Portuguese adult population. From a social and health point of view, an alarming finding was that one-fifth of the adult Portuguese population had a monthly family income of <500€.3 Indeed, data from EpiDoC 2 showed that poverty and a low education level are associated with an unhealthy lifestyle and higher prevalence of chronic diseases.12 Social inequality in health is a major concern within public health, with food insecurity being one of its main drivers. Food insecurity is defined as a difficulty in achieving a healthy diet due to economic constraints, and is a well-known determinant of health. EpiDoC 3 showed a high prevalence of food insecurity and its associations and unhealthy dietary behaviours. Food insecurity was associated with several non-communicable diseases, lower quality of life and higher health care resource consumption.13 Publications using EpiDoC data have raised questions and informed policy makers about the need to reduce food insecurity, not only to improve individual health status but also to reduce public health costs. Considering health and health-related characteristics, high blood pressure, high cholesterol level, allergies and RMDs were frequently self-reported among the Portuguese adult population. The prevalence of RMDs in Portugal is similar to that reported in other countries,14–19 namely Portugal’s close neighbour Spain.20 Another interesting finding was the high proportion of individuals presenting typical features of one or more RMDs, who did not have a previous diagnosis (1532 out of 3877 participants).21 This could be explained by the scarce number of rheumatologists in Portugal (1: 100 000 inhabitants)22 and the lack of awareness among the population about these diseases, as RMD symptoms are frequently accepted as part of the normal ageing process. These results helped support a new national network for hospital reference of rheumatology, developed by the National Directorate General of Health in collaboration with the EpiDoC research team. The RMD with the highest prevalence in Portugal was low back pain (26.4%; 95% CI, 23.3-29.5%), which was significantly more frequent in women than in men (29.6% vs 22.8%; P = 0.040). Low back pain increased with age, and its prevalence was highest in the 46–55-year age group (27.7%; 95% CI 23.1-32.4%).21 Regarding the impact of RMDs on health-related quality of life, physical function and mental health among the Portuguese population, EpiDoC data showed that patients with RMDs have more health care resource consumption, were more often hospitalized and had more homecare support needs in the previous 12 months, compared with participants with no RMDs.12,21,23 In EpiDoC 1, a meaningful number (n = 488, 30.9%) of people claimed to have retired prematurely due to RMDs.24 This translates to many years of working life already lost and many others still potentially lost. Indirect costs due to self-reported RMDs are also substantial, equivalent to at least 0.5% of the gross domestic product.21 These results emphasize the burden of RMDs and the need to develop RMD awareness, which is a strong argument encouraging policy makers to increase the amount of resources allocated to the treatment of rheumatic patients. EpiDoC 1 also showed a high prevalence of other chronic diseases among Portuguese adults such as dyslipidaemia (24.4–25.9%), hypertension (23.1–24.8%) and diabetes (7.6–9.1%). The elderly are a particularly vulnerable population for chronic diseases, among whom the coexistence of two or more chronic diseases is particularly high (78.3%), leading to low quality of life and disability.25 The most common chronic diseases in the elderly were hypertension (57.3%), rheumatic disease (51.9%), hypercholesterolaemia (49.4%) and diabetes (22.7%). Among older adults, 66.6% were physically inactive and 22.3% were obese, particularly among Azoreans (33.0%). Similar results were found for Portuguese adults, of whom more than half did not exercise (63.0%) and more than 15% were obese.25 EpiDoC 2 estimated a prevalence of anxiety and depression among Portuguese elderly of 9.6% and 11.8%, respectively. Seniors with anxiety or depression were more likely to self-report higher levels of physical disability and lower quality of life.26 Biological and clinical data have been used in national studies of older adult lifestyles,23,27 the impact of falls and fractures, vitamin D level, sun exposure, dairy consumption and oral health, as well as international collaborative projects on mitochondrial DNA and BMA and bone texture in osteoarthritis.28 In conclusion, EpiDoC publications have improved our understanding of socioeconomic and health inequalities among Portuguese adults, particularly the elderly. These studies demonstrate that unhealthy lifestyles are more prevalent among the most socioeconomically vulnerable groups and are associated with a higher prevalence of chronic non-communicable diseases and higher health care resource consumption. The EpiDoC study has also shed light on the burden of rheumatic diseases in Portugal. It shows a need to rethink the rheumatology support network and to provide better care to rheumatic patients. EpiDoC ongoing work is aimed at revealing the determinants and burden of multimorbidity and other chronic non-communicable diseases, namely mental and cardiovascular diseases. Particular attention will be directed at better understanding unmet elderly health needs. What are the main strengths and weaknesses? The main strengths of the EpiDoC study are its general population base and sample size, availability of repeated measures and extensive biobank blood collection. Another strength is its interdisciplinary research cooperation, with a team comprising physicians, psychologists, epidemiologists, nutritionists, statisticians, laboratory technicians and others. The different purposes of the three waves are also a strength, as they have expanded the scope of the EpiDoC study to become a more complete cohort study. The EpiDoC study also has some weaknesses, such as its attrition rate, which is similar to that of other studies29,30 and was not significantly different between the three waves. In EpiDoC 2 and EpiDoC 3, data were collected by phone interviews; however, we attempted to reduce attrition bias by using reminders for scheduled visits and sending periodic newsletters and reminders to all participants. Another limitation is that diseases were self-reported, although a detailed and comprehensive questionnaire included a screening for RMD symptoms. All measurement tools (HADS, EQ-5D-3L, SF-36 and HAQ) were validated and the screening of RMDs was validated by an algorithm supplemented by expert rheumatologist opinion. Each wave survey was composed of a structured comprehensive questionnaire which was tested for feasibility, participant comprehension and language.3,12 Can I get hold of the data? Where can I find more information? The EpiDoC Unit promotes research networking—both national and international—and develops collaborative projects. Data from our cohort studies and projects are freely available for researchers who submit a research proposal to the scientific committee. More details about questionnaire content and clinical measurements can be found on our website [http://cedoc.unl.pt/epidoc-unit/]. A research proposal editable form can be downloaded and sent to [[emailprotected]]. An EpiDoC steering committee will evaluate all proposals for future studies and collaborations, to access data and use of biological samples. Profile in a nutshell EpiDoC is a prospective population-based closed cohort study that collects health information. The study primarily aimed to address rheumatic diseases, but its scope has broadened to other chronic diseases, namely cardiovascular, gastroenterological, pulmonary, anxiety and depression and neurological diseases. Three health surveys of the general adult population (aged ≥18 years) in Portugal were completed: EpiDoC 1 (September 2011–December 2013), EpiDoC 2 (March 2013–July 2015) and EpiDoC 3 (September 2015–July 2016). EpiDoC surveys have spanned a total of 5 years, with an attrition rate of approximately 25%. EpiDoC 1, 2 and 3 had 10 661, 7591 and 5663 participants, respectively. The EpiDoC sample is representative of the Portuguese population. In EpiDoC 1, 6551 (52.6%) participants were women, and most were Caucasian (n = 10 342, 96.0%) and married (n = 6111, 50.2%). In EpiDoC 2, 4784 (52.2%) participants were women, and the mean age of all participants was 48.0 ± 18.0 years. In EpiDoC 3, 3607 (52.5%) participants were women, and the mean age of all participants was 49.64 ± 18.11 years. EpiDoC data are available to researchers who submit research proposals to the scientific committee. More details can be found on our website [http://cedoc.unl.pt/epidoc-unit/]. Funding The EpiDoC study has had several sources of financial support. EpiDoC 1 and EpiDoC 2 were supported by grants from the Portuguese Directorate-General of Health and sponsorships from Fundação Calouste Gulbenkian, Fundação Champalimaud, Fundação AstraZeneca, Abbott, Merck Sharp & Dohme, Pfizer, Roche, Servier, Bial, D3A Medical Systems, Happybrands, Centro de Medicina Laboratorial Germano de Sousa, CAL-Clínica, Galp Energia, Açoreana Seguros and individual rheumatologists. EpiDoC 3 was supported by the Public Health Initiatives Programme (PT06) financed by EEA Grants Financial Mechanism 2009–14. Conflict of interest: None declared. References 1 Green MA , Li J, Relton C et al. Cohort Profile: The Yorkshire Health Study . Int J Epidemiol 2016 ; 45 : 707 – 12 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Krokstad S , Langhammer A, Hveem K et al. Cohort Profile: The HUNT study, Norway . Int J Epidemiol 2013 ; 42 : 968 – 77 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Rodrigues AM , Gouveia N, da Costa LP et al. EpiReumaPt—the study of rheumatic and musculoskeletal diseases in Portugal: a detailed view of the methodology . Acta Reumatol Port 2015 ; 40 : 110 – 24 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 4 Statistics Portugal INE . Censos—Resultados Definitivos . Lisbon : INE , 2012 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 5 Eurostat . Projected Old-Age Dependency Ratio. 2016 . http://ec.europa.eu/eurostat/tgm/table.do?tab=table&init=1&language=en&pcode=tps00200&plugin=1 (26 April 2018, date last accessed). 6 Galea S , Tracy M. Participation rates in epidemiologic studies . Ann Epidemiol 2007 ; 17 : 643 – 53 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Fries JF , Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis . Arthritis Rheum 1980 ; 23 : 137 – 45 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Ferreira LN , Ferreira PL, Pereira LN, Oppe M. EQ-5D Portuguese population norms . Qual Life Res 2014 ; 23 : 425 – 30 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Ferreira LN , Ferreira PL, Pereira LN, Oppe M. The valuation of the EQ-5D in Portugal . Qual Life Res 2014 ; 23 : 413 – 23 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item selection . Med Care 1992 ; 2 : 98 – 83 . Google Scholar OpenURL Placeholder Text WorldCat 11 Pais-Ribeiro J , Silva I, Ferreira T, Martins A, Meneses R, Baltar M. Validation study of a Portuguese version of the hospital anxiety and depression scale . Psychol Health Med 2007 ; 12 : 225 – 37 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Gregório MJ , Rodrigues AM, Eusebio M et al. Dietary patterns characterized by high meat consumption are associated with other unhealthy life styles and depression symptoms . Front Nutr 2017 ; 4 . Google Scholar OpenURL Placeholder Text WorldCat 13 Gregório MJ , Rodrigues AM, Graça P et al. Food insecurity is associated with low adherence to the mediterranean diet and adverse health conditions in Portuguese adults . Front Public Health 2018 ; 6 . Google Scholar OpenURL Placeholder Text WorldCat 14 Hoy D , March L, Brooks P et al. The global burden of low back pain: estimates from the Global Burden of Disease 2010 study . Ann Rheum Dis 2014 ; 73 : 968 – 74 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Gabriel SE , Michaud K. Epidemiological studies in incidence, prevalence, mortality, and comorbidity of the rheumatic diseases . Arthritis Res Ther 2009 ; 11 : 229. Google Scholar Crossref Search ADS PubMed WorldCat 16 Helmick CG , Felson DT, Lawrence RC et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States: Part I . Arthritis Rheum 2008 ; 58 : 15 – 25 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Anagnostopoulos I , Zinzaras E, Alexiou I et al. The prevalence of rheumatic diseases in central Greece: a population survey . BMC Musculoskelet Disord 2010 ; 11 : 98 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Peláez-Ballestas I , Sanin LH, Moreno-Montoya J et al. Epidemiology of the rheumatic diseases in Mexico. A study of 5 regions based on the COPCORD methodology . J Rheumatol 2011 ; 86 : 3 – 8 . Google Scholar OpenURL Placeholder Text WorldCat 19 Lawrence RC , Felson DT, Helmick CG et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States: Part II . Arthritis Rheum 2008 ; 58 : 26 – 35 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Carmona L , Ballina J, Gabriel R, Laffon A. The burden of musculoskeletal diseases in the general population of Spain: results from a national survey . Ann Rheum Dis 2001 ; 60 : 1040 – 45 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Branco JC , Rodrigues AM, Gouveia N et al. Prevalence of rheumatic and musculoskeletal diseases and their impact on health-related quality of life, physical function and mental health in Portugal: results from EpiReumaPt – a national health survey . RMD Open 2016 ; 2 : e000166 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Central Administration of the Health System (ACSS) . Actuais e Futuras Necessidades Previsionais de Médicos (SNS) . Lisboa : ACSS , 2011 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 23 Fernandes S , Rodrigues AM, Nunes C et al. Food insecurity in older adults: results from the EpiDoC3 study . Front Med 2018 ; 5 : 203 . Google Scholar Crossref Search ADS WorldCat 24 Laires P. The Impact of Rheumatic Diseases on Early Retirement . Lisboa : Universidade de Lisboa , 2017 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 25 Rodrigues AM , Gregório MJ, Sousa RD et al. Challenges of ageing in Portugal—data from the EpiDoC cohort . Acta Med Port 2018 ; 31 : 80 – 93 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Sousa RD , Rodrigues AM, Gregório MJ et al. Anxiety and depression in the Portuguese older adults: prevalence and associated factors . Front Med 2017 ; 4 . Google Scholar OpenURL Placeholder Text WorldCat 27 Duarte N , Rodrigues AM, Branco JC, Canhão H, Hughes SL, Paúl C. Health and lifestyles factors associated with osteoarthritis among older adults in Portugal . Front Med 2017 ; 4 . Google Scholar OpenURL Placeholder Text WorldCat 28 Hladůvka J , Phuong BTM, Ljuhar R et al. Femoral ROIs and entropy for texture-based detection of osteoarthritis from high-resolution knee radiographs . Comput Vis Pattern Recognit 2017 ; arXiv:1703.09296. Google Scholar OpenURL Placeholder Text WorldCat 29 Huisman M , Poppelaars J, van der Horst M et al. Cohort Profile: The Longitudinal Aging Study Amsterdam . Int J Epidemiol 2011 ; 40 : 868 – 76 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Hasselhorn HM , Peter R, Rauch A et al. Cohort Profile: The LIDA Cohort Study—a German cohort study on work, age, health and work participation . Int J Epidemiol 2014 ; 43 : 1736 – 49 . Google Scholar Crossref Search ADS PubMed WorldCat Author notes Sara Simões Dias and Ana Maria Rodrigues authors contributed equally. © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [emailprotected] © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association.

Cohort Profile: The Cohort of Universities of Minas Gerais (CUME)

Gomes Domingos, Ana, Luiza;Miranda, Aline Elizabeth da, Silva;Pimenta, Adriano, Marçal;Hermsdorff, Helen Hermana, Miranda;Oliveira, Fernando Luiz Pereira, de;dos Santos, Luana, Caroline;Lopes, Aline Cristine, Souza;Martínez González, Miguel, Ángel;Bressan,, Josefina

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy152pmid: 30060144

Why was the cohort set up? In Brazil, noncommunicable disease (NCD) accounts for 72% of deaths, representing a serious public health concern.1,2 In this context, the Ministry of Health proposed a strategic action plan to confront these diseases between 2011 and 2022, with emphasis on cardiovascular diseases, cancer, chronic respiratory diseases and diabetes, as well as their main risk factors—smoking, harmful alcohol consumption, physical inactivity, inadequate diet and obesity.3 In this context, studies have evaluated the relationship between dietary/behavioural factors and the prevalence of NCD. Thus, high consumption of fruits and vegetables has been associated with low concentrations of homocysteine, a compound involved in the development of cardiovascular and cerebrovascular diseases as well as with reduction of oxidative stress and DNA damage markers.4,5 In turn, high consumption of red meat, a common food habit among the Brazilian population, was positively associated with central obesity, hypertriglyceridaemia, metabolic syndrome and insulin resistance.6 However, there is still a lack of longitudinal studies from Brazil examining the association between lifestyle and NCD which are capable of making causal inference. Thus, the Cohort of Universities of Minas Gerais (CUME) aimed to assess the impact of the Brazilian dietary pattern and nutrition transition on the occurrence of NCD among graduates and postgraduates of federal higher education institutions located in the state of Minas Gerais, Brazil. The CUME project is an open concurrent cohort restricted to a population group of alumni of universities located in the state of Minas Gerais, Brazil. Who is in the cohort? The cohort’s baseline (Q_0) was established in 2016, having as participants Universidade Federal de Viçosa (UFV) and Universidade Federal de Minas Gerais (UFMG) alumni who graduated between 1994 and 2014. This research was conducted in accordance with the ethical principles stated in the Declaration of Helsinki, and was approved by the Human Research Ethics Committees of the UFV and the UFMG (Protocol No. 596,741-0/2013). All participants read the informed consent form and indicated online agreement before responding to the questionnaire.7 In order to contact as many graduates and postgraduates as possible, the alumni association and postgraduate programmes of UFV provided the registration data of potential participants. To update information, an internet search on professional networking sites was performed. At the UFMG, the Information and Technology Directorate offered to send the research questionnaires to potential participants registered in its database. As a disclosure strategy, a logo was created for the cohort, as well as a website and a social media page to disseminate the project. The logo comprises symbols representing the mountains of the state of Minas Gerais, and it was used on all dissemination materials, social media, reports, questionnaires and the material used by the team. Before sending the invitation, the project was publicized by means of interviews with researchers on social media and on the website of the project. The invitation to participate in the study was e-mailed to each potential participant. In addition, a registration form was made available on the cohort’s website at for those graduates and postgraduates who initially did not receive the e-mail. The filling out of the baseline questionnaire was divided into two parts. The second part was sent 1 week after the first part was completed. Five e-mail invitations without any response was considered refusal to participate in the study. Likewise, incomplete questionnaires were defined as those that the participant did not fully complete after five e-mail reminders. For the development of the questionnaire in the virtual environment, an information technology specialist was hired to configure the software and assist the research team in solving technical problems through e-mail, instant messaging and telephone. Additionally, this technician was responsible for submitting questionnaires via e-mail and exporting the data generated in a format compatible with Microsoft Excel®, containing variable labels and a frequency script. Then, questionnaires were sent and completed between March and August 2016. In total, 4987 alumni responded to the online questionnaire. However, surveys that did not present some demographic data (n = 516), participants of other nationalities (n = 19) and residents abroad (n = 161) were excluded (Figure 1). After these exclusions, 4291 alumni remained in the baseline cohort, of whom 2915 (67.9%) were women and 1376 (32.1%) were men. Figure 1 View largeDownload slide Flowchart of baseline data collection of the Cohort of Universities of Minas Gerais (CUME), 2016. UFMG, Universidade Federal de Minas Gerais; UFV, Universidade Federal de Viçosa. Figure 1 View largeDownload slide Flowchart of baseline data collection of the Cohort of Universities of Minas Gerais (CUME), 2016. UFMG, Universidade Federal de Minas Gerais; UFV, Universidade Federal de Viçosa. The cohort included graduates and postgraduates in all Brazilian states and the Federal District. The largest concentration of participants were in the South-east region (n = 3804; 88.7%), followed by the Mid-west region (n = 189, 4.4%), North-east (n = 166, 3.9%), North (n = 68, 1.6%) and South (n = 64, 1.5%) (Figure 2). Interestingly, 78.1% of the alumni reported living in the state of Minas Gerais, thus making it possible to collect face-to-face data for validation studies at the research centres located in the cities of Belo Horizonte and Viçosa. Figure 2 View largeDownload slide Distribution of the place of residence of the baseline participants of the Cohort of Universities of Minas Gerais (CUME) throughout Brazil, 2016 (n = 4291). North region, a: Acre, b: Rondonia, c: Amazonas, d: Roraima, e: Amapa, f: Para, g: Tocantins. Mid-west region, h: Mato Grosso, i: Mato Grosso do Sul, j: Goias, k: Federal District. North-east region, l: Maranhao, m: Piaui, n: Ceara, o: Rio Grande do Norte, p: Paraiba, q: Pernambuco, r: Alagoas, s: Sergipe, t: Bahia. South-east region, u: Minas Gerais, v: Espirito Santo, w: Rio de Janeiro, x: Sao Paulo. South region, y: Parana, z: Santa Catarina, α: Rio Grande do Sul. Figure 2 View largeDownload slide Distribution of the place of residence of the baseline participants of the Cohort of Universities of Minas Gerais (CUME) throughout Brazil, 2016 (n = 4291). North region, a: Acre, b: Rondonia, c: Amazonas, d: Roraima, e: Amapa, f: Para, g: Tocantins. Mid-west region, h: Mato Grosso, i: Mato Grosso do Sul, j: Goias, k: Federal District. North-east region, l: Maranhao, m: Piaui, n: Ceara, o: Rio Grande do Norte, p: Paraiba, q: Pernambuco, r: Alagoas, s: Sergipe, t: Bahia. South-east region, u: Minas Gerais, v: Espirito Santo, w: Rio de Janeiro, x: Sao Paulo. South region, y: Parana, z: Santa Catarina, α: Rio Grande do Sul. How often have they been followed up? Having established the cohort’s baseline in 2016 with UFV and UFMG alumni who graduated between 1994 and 2014 as participants, we will perform evaluation waves every 2 years in a virtual environment. We intend to apply the first follow-up questionnaire (Q_2) in 2018, and to extend the study to other universities from Minas Gerais State (Brazil); then the CUME project can be considered a concurrent open cohort. What has been measured? The first questionnaire sent to the participants comprised 83 questions related to lifestyle, sociodemographics, anthropometrics, biochemical and clinical data, physical activity practice, individual and family morbidity and personal history of preventive checkups. The participants were invited to report their current weight and height for further calculation of body mass index—(BMI: kg/m2), in addition to results of the past 2 years of the following biochemical tests: total cholesterol, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol, triglycerides, glucose and systolic and diastolic blood pressure. Data on the current use of medicines, passive and active smoking habits and alcohol consumption were collected. Binge drinking was defined as five or more alcoholic drinks for males or four or more alcoholic drinks for females on the same occasion on at least 1 day in the past month.8 Physical activity was investigated using a list of 23 leisure activities and the time/frequency spent on them, as well as a second part that included questions about the time spent in sedentary activities.9 Individuals with ≥150 min/week of moderate-intensity activity or ≥75 min/week of vigorous-intensity activity were considered active. Physical inactivity was defined as the absence of leisure time physical activity.10 Questions regarding the means of commuting to work were also included. Results of medical examinations or preventive checkups (without illness been previously diagnosed), i.e. ultrasound, endoscopy, medical review, intraocular pressure, digital rectal examination and mammography, were registered. At this stage of the questionnaire, the participant was allowed to mark more than one type of examination and provide an age range (<25, 25–39, 40–59, ≥60 years). We also investigated groups of diseases and the age group when diagnosed. The diseases investigated were cardiovascular, gastrointestinal, respiratory, renal, cancer or tumours, infections, other diseases or injuries and traffic accidents. Subsequently, family history of illness was reported. For women, diagnosis of benign disease or malignant tumour in the breast and reproductive history (gestations and the type of feeding offered to their first child during the child’s first year) were also investigated. At the end of this first stage, participants received information related to their reported blood pressure and BMI as well as any cardiovascular risk identified and calculated from the Framingham Heart Study.11 To calculate the prevalence of diseases in the base population of the CUME project, the following criteria were considered: report of previous medical diagnosis; use of medication; and/or results of clinical, biochemical and anthropometric tests. The outcomes investigated were: obesity, overweight, hypertension, type 2 diabetes, elevated total cholesterol and triacylglycerol, and depression. Individuals who reported the diagnosis of obesity or who had a BMI ≥30 kg/m² were classified as obese, and those with a BMI value ≥25 kg/m² were considered overweight.12 The classification of hypertension was based on medical diagnosis, use of antihypertensive medications and systolic blood pressure values ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg.13 Type 2 diabetes was estimated based on reports of the use of oral hypoglycaemic medications, insulin, medical diagnosis and/or fasting blood glucose ≥126 mg/dL.14 The diagnosis of high total cholesterol (TC) was defined by the use of TC-lowering medications, medical diagnosis and/or TC ≥200 mg/dL.15 A diagnosis of high triacylglycerol (TG) was based on reports of the use of TG-reducing medications, medical diagnosis and/or TG ≥150 mg/dL.15 Finally, depression was defined as that diagnosed by a physician. In the second stage, participants were asked to fill out the quantitative Food Frequency Questionnaire (FFQ) and to report their eating habits and practices. The survey was previously validated for the Brazilian population and contains 144 food items grouped into dairy, meat and fish, cereals and legumes, fats and oils, fruits, vegetables and greens, beverages and other foods.16 Thus, each participant reported the frequency (daily, weekly, monthly or annual) of the consumption of a given food in the previous year and the portion consumed. To adapt the instrument to the virtual environment, a photo album of the food portions was provided aiming to assist in the visualization of the portions of the foods and to improve the reliability of the answers. To elucidate the habits and food practices of the alumni, questions were asked about the consumption of visible meat fat and chicken skin, organic foods, probiotics, prebiotics, addition of salt to salad and sugar in beverages. To help subjects understand the questions related to the organic foods, probiotics and prebiotics, explanatory notes with simple language containing concepts and examples of everyday use were used. Issues relating to food environment were also addressed, such as the type of food establishment where the participant has the habit of having lunch (a la carte, self-service, university restaurants, bakeries and cafeterias), average amount spent on meals and how many blocks walked to the food establishment. Also, whether the establishment offered various options of salads and vegetables, fresh fruits and fruit salads, fresh natural juices or frozen fruit pulp-based juice, nutritional information on the preparations served and whether 300 mL of natural juice or prepared fruit juice from frozen pulp cost more than canned soda (350 mL) or a cup (300 mL) of soda. In return for their participation, participants received a list of information on the number of daily meals eaten; consumption of fruits, greens, legumes and sodas; removal or consumption of visible meat fat or chicken skin; increased sugar in drinks and salt in salad. In addition, a specially designed newsletter was sent which included important aspects of diet including the Food Guide for the Brazilian Population, as material to be consulted online.17 In order to evaluate the food consumption of the participants, the consumption frequencies of each food were transformed into daily consumption, followed by grams or millilitres, and finally nutrients or compounds. For the calculation of daily intake of calories and nutrients, Brazilian tables of nutritional composition of foods were used and, in the absence of such information, the United States Department of Agriculture table was used.18,19 Since the objective of the CUME project is to evaluate the impact of Brazilian dietary pattern and nutrition transition on NCD, participants of other nationalities residing in Brazil, residents abroad, those who did not complete the FFQ and those with estimated daily energy intake [<500 kcal/day (n = 1) or >6000 kcal/day (n = 92)] were excluded from the analyses.20,21 In order to characterize the baseline profile of the participants of the CUME project, absolute and relative frequencies of the variables of interest were presented according to gender. Statistical differences were evaluated with the chi-square Pearson test (χ2). The values of macronutrients consumed were previously adjusted by daily energy intake using the residual nutrient method.22 Energy and macronutrients intake were compared between men and women through a Mann-Whitney test, according to the normality of the variables, which in turn was verified by the Shapiro-Wilk test. All analyses were conducted with Stata® Software (version 13.0), with a significance level of 5%. What has CUME found? Key findings and publications As described in Table 1, participants of CUME project hold degrees from different fields of study, predominantly Applied Social Sciences (28.7%) and Health Sciences (26.7%). Of the total number of participants, 72.9 % had concluded some postgraduate study. At baseline, most individuals were aged between 30 and 39 years (46.5%), White (64.4%), were legally married or in a stable union (52%), received less than five times the minimum wage (49.5%) and were employed (80.5%). Regarding lifestyle, 56.5% reported binge consumption of alcohol, 8.9% smoked tobacco and 46.3% were insufficiently active or were inactive. Men and women differed in relation to all characteristics except skin colour, with emphasis on binge consumption of alcohol and smoking among men. Table 1. Sociodemographic characteristics of baseline participants of the Cohort of Universities of Minas Gerais (CUME), according to sex, 2016 Female Male Total P-value* n % n % n % Area of study (n = 4291) <0.001*  Exact and Earth Sciences 258 8.9 201 14.6 459 10.7  Biological Sciences 242 8.3 82 6.0 324 7.6  Engineering 195 6.7 253 18.4 448 10.4  Health 932 32.0 212 15.4 1144 26.7  Applied Social Sciences 900 30.9 332 24.1 1232 28.7  Agricultural Sciences 195 6.7 248 18.0 443 10.3  Linguistics, Language Studies and Arts 193 6.6 48 3.5 241 5.6 Age group (n = 4289) <0.001*  20–29 years 821 28.2 227 20.1 1098 25.6  30–39 years 1327 45.5 668 48.6 1995 46.5  40–49 years 471 16.2 279 20.3 750 17.5  50–59 years 233 8.0 103 7.5 336 7.8  ≥60 years 62 2.1 48 3.5 110 2.6 Individual income (n = 3480)a <0.001*  <5 times the minimum wage 1353 57.0 370 33.4 1723 49.5  ≥5 to <10 times the minimum wage 741 31.2 409 37.0 1150 33.0  ≥10 times the minimum wage 279 11.8 328 29.6 607 17.4 Skin colour (n = 4291) 0.734  White 1889 64.8 876 63.7 2765 64.4  Black/Brown 996 34.2 484 35.2 1480 34.5  Yellow/Indigenous 30 1.0 16 1.2 46.0 1.1 Level of education (n = 4291) <0.001*  Bachelor’s degree 781 26.8 381 27.7 1162 27.1  Specialization degree 782 26.8 319 23.2 1101 25.7  Master’s degree 846 29.0 353 25.7 1199 27.9  Doctorate/post-doctorate 506 17.4 323 23.5 829 19.3 Marital status (n = 4291) <0.001*  Legally married/stable union/other 1429 49.0 802 58.3 2231 52.0  Single 1299 44.6 511 37.1 1810 42.2  Separated/divorced 171 5.9 60 4.4 231 5.4  Widowed 16 0.5 3 0.2 19 0.4 Professional situation (n = 4291) <0.001*  Full time/part time/informal 2288 78.5 1165 84.7 3453 80.5  Student 389 13.3 135 9, 8 524 12.2  Retired/home duties 81 2.8 28 2.0 109 2.5  Unemployed 157 5.4 48 3.5 205 4.8 Physical activity (n = 4289) <0.001*  Inactive 764 26.2 324 23.6 1088 25.4  Insufficiently active 662 22.7 235 17.1 897 20.9  Active 1488 51.1 816 59.3 2304 53.7 Smoking habit (n = 4287) <0.001*  No 2372 81.4 996 72.5 3368 78.6  Former smoker 337 11.6 199 14.5 536 12.5  Yes 204 7.0 179 13.0 383 8.9 Binge drinking (n = 3130) <0.001*  No 947 47.0 415 37.3 1362 43.5  Yes 1070 53.0 698 62.7 1768 56.5 Female Male Total P-value* n % n % n % Area of study (n = 4291) <0.001*  Exact and Earth Sciences 258 8.9 201 14.6 459 10.7  Biological Sciences 242 8.3 82 6.0 324 7.6  Engineering 195 6.7 253 18.4 448 10.4  Health 932 32.0 212 15.4 1144 26.7  Applied Social Sciences 900 30.9 332 24.1 1232 28.7  Agricultural Sciences 195 6.7 248 18.0 443 10.3  Linguistics, Language Studies and Arts 193 6.6 48 3.5 241 5.6 Age group (n = 4289) <0.001*  20–29 years 821 28.2 227 20.1 1098 25.6  30–39 years 1327 45.5 668 48.6 1995 46.5  40–49 years 471 16.2 279 20.3 750 17.5  50–59 years 233 8.0 103 7.5 336 7.8  ≥60 years 62 2.1 48 3.5 110 2.6 Individual income (n = 3480)a <0.001*  <5 times the minimum wage 1353 57.0 370 33.4 1723 49.5  ≥5 to <10 times the minimum wage 741 31.2 409 37.0 1150 33.0  ≥10 times the minimum wage 279 11.8 328 29.6 607 17.4 Skin colour (n = 4291) 0.734  White 1889 64.8 876 63.7 2765 64.4  Black/Brown 996 34.2 484 35.2 1480 34.5  Yellow/Indigenous 30 1.0 16 1.2 46.0 1.1 Level of education (n = 4291) <0.001*  Bachelor’s degree 781 26.8 381 27.7 1162 27.1  Specialization degree 782 26.8 319 23.2 1101 25.7  Master’s degree 846 29.0 353 25.7 1199 27.9  Doctorate/post-doctorate 506 17.4 323 23.5 829 19.3 Marital status (n = 4291) <0.001*  Legally married/stable union/other 1429 49.0 802 58.3 2231 52.0  Single 1299 44.6 511 37.1 1810 42.2  Separated/divorced 171 5.9 60 4.4 231 5.4  Widowed 16 0.5 3 0.2 19 0.4 Professional situation (n = 4291) <0.001*  Full time/part time/informal 2288 78.5 1165 84.7 3453 80.5  Student 389 13.3 135 9, 8 524 12.2  Retired/home duties 81 2.8 28 2.0 109 2.5  Unemployed 157 5.4 48 3.5 205 4.8 Physical activity (n = 4289) <0.001*  Inactive 764 26.2 324 23.6 1088 25.4  Insufficiently active 662 22.7 235 17.1 897 20.9  Active 1488 51.1 816 59.3 2304 53.7 Smoking habit (n = 4287) <0.001*  No 2372 81.4 996 72.5 3368 78.6  Former smoker 337 11.6 199 14.5 536 12.5  Yes 204 7.0 179 13.0 383 8.9 Binge drinking (n = 3130) <0.001*  No 947 47.0 415 37.3 1362 43.5  Yes 1070 53.0 698 62.7 1768 56.5 a Minimum wage (R$880.00 in 2016). * P-values from Pearson chi-square test. Table 1. Sociodemographic characteristics of baseline participants of the Cohort of Universities of Minas Gerais (CUME), according to sex, 2016 Female Male Total P-value* n % n % n % Area of study (n = 4291) <0.001*  Exact and Earth Sciences 258 8.9 201 14.6 459 10.7  Biological Sciences 242 8.3 82 6.0 324 7.6  Engineering 195 6.7 253 18.4 448 10.4  Health 932 32.0 212 15.4 1144 26.7  Applied Social Sciences 900 30.9 332 24.1 1232 28.7  Agricultural Sciences 195 6.7 248 18.0 443 10.3  Linguistics, Language Studies and Arts 193 6.6 48 3.5 241 5.6 Age group (n = 4289) <0.001*  20–29 years 821 28.2 227 20.1 1098 25.6  30–39 years 1327 45.5 668 48.6 1995 46.5  40–49 years 471 16.2 279 20.3 750 17.5  50–59 years 233 8.0 103 7.5 336 7.8  ≥60 years 62 2.1 48 3.5 110 2.6 Individual income (n = 3480)a <0.001*  <5 times the minimum wage 1353 57.0 370 33.4 1723 49.5  ≥5 to <10 times the minimum wage 741 31.2 409 37.0 1150 33.0  ≥10 times the minimum wage 279 11.8 328 29.6 607 17.4 Skin colour (n = 4291) 0.734  White 1889 64.8 876 63.7 2765 64.4  Black/Brown 996 34.2 484 35.2 1480 34.5  Yellow/Indigenous 30 1.0 16 1.2 46.0 1.1 Level of education (n = 4291) <0.001*  Bachelor’s degree 781 26.8 381 27.7 1162 27.1  Specialization degree 782 26.8 319 23.2 1101 25.7  Master’s degree 846 29.0 353 25.7 1199 27.9  Doctorate/post-doctorate 506 17.4 323 23.5 829 19.3 Marital status (n = 4291) <0.001*  Legally married/stable union/other 1429 49.0 802 58.3 2231 52.0  Single 1299 44.6 511 37.1 1810 42.2  Separated/divorced 171 5.9 60 4.4 231 5.4  Widowed 16 0.5 3 0.2 19 0.4 Professional situation (n = 4291) <0.001*  Full time/part time/informal 2288 78.5 1165 84.7 3453 80.5  Student 389 13.3 135 9, 8 524 12.2  Retired/home duties 81 2.8 28 2.0 109 2.5  Unemployed 157 5.4 48 3.5 205 4.8 Physical activity (n = 4289) <0.001*  Inactive 764 26.2 324 23.6 1088 25.4  Insufficiently active 662 22.7 235 17.1 897 20.9  Active 1488 51.1 816 59.3 2304 53.7 Smoking habit (n = 4287) <0.001*  No 2372 81.4 996 72.5 3368 78.6  Former smoker 337 11.6 199 14.5 536 12.5  Yes 204 7.0 179 13.0 383 8.9 Binge drinking (n = 3130) <0.001*  No 947 47.0 415 37.3 1362 43.5  Yes 1070 53.0 698 62.7 1768 56.5 Female Male Total P-value* n % n % n % Area of study (n = 4291) <0.001*  Exact and Earth Sciences 258 8.9 201 14.6 459 10.7  Biological Sciences 242 8.3 82 6.0 324 7.6  Engineering 195 6.7 253 18.4 448 10.4  Health 932 32.0 212 15.4 1144 26.7  Applied Social Sciences 900 30.9 332 24.1 1232 28.7  Agricultural Sciences 195 6.7 248 18.0 443 10.3  Linguistics, Language Studies and Arts 193 6.6 48 3.5 241 5.6 Age group (n = 4289) <0.001*  20–29 years 821 28.2 227 20.1 1098 25.6  30–39 years 1327 45.5 668 48.6 1995 46.5  40–49 years 471 16.2 279 20.3 750 17.5  50–59 years 233 8.0 103 7.5 336 7.8  ≥60 years 62 2.1 48 3.5 110 2.6 Individual income (n = 3480)a <0.001*  <5 times the minimum wage 1353 57.0 370 33.4 1723 49.5  ≥5 to <10 times the minimum wage 741 31.2 409 37.0 1150 33.0  ≥10 times the minimum wage 279 11.8 328 29.6 607 17.4 Skin colour (n = 4291) 0.734  White 1889 64.8 876 63.7 2765 64.4  Black/Brown 996 34.2 484 35.2 1480 34.5  Yellow/Indigenous 30 1.0 16 1.2 46.0 1.1 Level of education (n = 4291) <0.001*  Bachelor’s degree 781 26.8 381 27.7 1162 27.1  Specialization degree 782 26.8 319 23.2 1101 25.7  Master’s degree 846 29.0 353 25.7 1199 27.9  Doctorate/post-doctorate 506 17.4 323 23.5 829 19.3 Marital status (n = 4291) <0.001*  Legally married/stable union/other 1429 49.0 802 58.3 2231 52.0  Single 1299 44.6 511 37.1 1810 42.2  Separated/divorced 171 5.9 60 4.4 231 5.4  Widowed 16 0.5 3 0.2 19 0.4 Professional situation (n = 4291) <0.001*  Full time/part time/informal 2288 78.5 1165 84.7 3453 80.5  Student 389 13.3 135 9, 8 524 12.2  Retired/home duties 81 2.8 28 2.0 109 2.5  Unemployed 157 5.4 48 3.5 205 4.8 Physical activity (n = 4289) <0.001*  Inactive 764 26.2 324 23.6 1088 25.4  Insufficiently active 662 22.7 235 17.1 897 20.9  Active 1488 51.1 816 59.3 2304 53.7 Smoking habit (n = 4287) <0.001*  No 2372 81.4 996 72.5 3368 78.6  Former smoker 337 11.6 199 14.5 536 12.5  Yes 204 7.0 179 13.0 383 8.9 Binge drinking (n = 3130) <0.001*  No 947 47.0 415 37.3 1362 43.5  Yes 1070 53.0 698 62.7 1768 56.5 a Minimum wage (R$880.00 in 2016). * P-values from Pearson chi-square test. The current population of Brazil is 207.7 million inhabitants, according to the report of the Population Count conducted by Brazilian Institute of Geography and Statistics (IBGE).23 Since Census 2010, Brazil has been an adult country, in the transition phase to becoming an old country in the year 2050.24 Participants of the CUME study are younger, mainly women (68%), smoke less tobacco (8.9% vs 10.2%) and have a higher prevalence of heavy episodic drinking (56.5% vs 20.4%) compared with the general Brazilian population.25 Furthermore, according to World Health Organization statistics, 49.2% of the Brazilian population does not reach the recommendations on physical activity,26 and CUME participants have a similar activity profile. Table 2 describes the most prevalent diseases in the population, with obesity identified in 14.9% and overweight in 40.8% of the respondents. In addition, the prevalence of hypertension was 11.6%, with high total cholesterol in 22.6% and high triacylglycerol in 12.7%; all disease was more frequent in males. Among the latest results published by the Surveillance System for Risk and Protective Factors for Chronic Diseases by Telephone Inquiry (VIGITEL) for Brazilians,25 the increase in the prevalence of overweight in adults over the past 10 years stands out, corresponding to from 42.6% in 2006 to 53.8% in 2016. Obesity currently affects 18.9% of the population,25 which is higher than that found at baseline in the CUME project. The prevalence of depression was 12.8% in the CUME project, being two times higher than that in the Brazilian population (5.8%), which just represents the highest prevalence in Latin America and the second highest in the Americas.27 Table 2. Diseases prevalence of baseline participants of the Cohort of Universities of Minas Gerais (CUME), according to sex, 2016 CUME Brazil Female Male Total P-value* Total n % n % n % % Obesity (n = 4288) <0.001*  No 2534 87.0 1114 81.0 3648 85.1 –  Yes 379 13.0 261 19.0 640 14.9 18.9a Overweight (n = 4286) <0.001*  No 1955 67.2 583 42.4 2538 59.2 –  Yes 956 32.8 792 57.6 1748 40.8 53.8a Type 2 diabetes (n = 4102) 0.011*  No 2716 97.2 1250 95.6 3966 96.7 –  Yes 79 2.8 57 4.4 136 3.3 8.9a High cholesterol (n = 4103) 0.007*  No 2197 78.6 978 74.8 3175 77.4 –  Yes 599 21.4 329 25.2 928 22.6 22.6a,** High triglycerides (n = 4107) <0.001*  No 2509 89.7 1075 82.1 3584 87.3 –  Yes 289 10.3 234 17.9 523 12.7 22.6a,** Hypertension (n = 4102) <0.001*  No 2523 90.3 1102 84.3 3625 88.4 –  Yes 272 9.7 205 15.7 477 11.6 25.7a Depression (n = 3987) <0.001*  No 2314 85.1 1164 91.7 3478 87.2 –  Yes 404 14.9 105 8.3 509 12.8 5.8b CUME Brazil Female Male Total P-value* Total n % n % n % % Obesity (n = 4288) <0.001*  No 2534 87.0 1114 81.0 3648 85.1 –  Yes 379 13.0 261 19.0 640 14.9 18.9a Overweight (n = 4286) <0.001*  No 1955 67.2 583 42.4 2538 59.2 –  Yes 956 32.8 792 57.6 1748 40.8 53.8a Type 2 diabetes (n = 4102) 0.011*  No 2716 97.2 1250 95.6 3966 96.7 –  Yes 79 2.8 57 4.4 136 3.3 8.9a High cholesterol (n = 4103) 0.007*  No 2197 78.6 978 74.8 3175 77.4 –  Yes 599 21.4 329 25.2 928 22.6 22.6a,** High triglycerides (n = 4107) <0.001*  No 2509 89.7 1075 82.1 3584 87.3 –  Yes 289 10.3 234 17.9 523 12.7 22.6a,** Hypertension (n = 4102) <0.001*  No 2523 90.3 1102 84.3 3625 88.4 –  Yes 272 9.7 205 15.7 477 11.6 25.7a Depression (n = 3987) <0.001*  No 2314 85.1 1164 91.7 3478 87.2 –  Yes 404 14.9 105 8.3 509 12.8 5.8b a VIGITEL survey, 2016. b WHO, 2017. * P-values from Pearson’s chi-square test; **prevalence of dyslipidaemia. Table 2. Diseases prevalence of baseline participants of the Cohort of Universities of Minas Gerais (CUME), according to sex, 2016 CUME Brazil Female Male Total P-value* Total n % n % n % % Obesity (n = 4288) <0.001*  No 2534 87.0 1114 81.0 3648 85.1 –  Yes 379 13.0 261 19.0 640 14.9 18.9a Overweight (n = 4286) <0.001*  No 1955 67.2 583 42.4 2538 59.2 –  Yes 956 32.8 792 57.6 1748 40.8 53.8a Type 2 diabetes (n = 4102) 0.011*  No 2716 97.2 1250 95.6 3966 96.7 –  Yes 79 2.8 57 4.4 136 3.3 8.9a High cholesterol (n = 4103) 0.007*  No 2197 78.6 978 74.8 3175 77.4 –  Yes 599 21.4 329 25.2 928 22.6 22.6a,** High triglycerides (n = 4107) <0.001*  No 2509 89.7 1075 82.1 3584 87.3 –  Yes 289 10.3 234 17.9 523 12.7 22.6a,** Hypertension (n = 4102) <0.001*  No 2523 90.3 1102 84.3 3625 88.4 –  Yes 272 9.7 205 15.7 477 11.6 25.7a Depression (n = 3987) <0.001*  No 2314 85.1 1164 91.7 3478 87.2 –  Yes 404 14.9 105 8.3 509 12.8 5.8b CUME Brazil Female Male Total P-value* Total n % n % n % % Obesity (n = 4288) <0.001*  No 2534 87.0 1114 81.0 3648 85.1 –  Yes 379 13.0 261 19.0 640 14.9 18.9a Overweight (n = 4286) <0.001*  No 1955 67.2 583 42.4 2538 59.2 –  Yes 956 32.8 792 57.6 1748 40.8 53.8a Type 2 diabetes (n = 4102) 0.011*  No 2716 97.2 1250 95.6 3966 96.7 –  Yes 79 2.8 57 4.4 136 3.3 8.9a High cholesterol (n = 4103) 0.007*  No 2197 78.6 978 74.8 3175 77.4 –  Yes 599 21.4 329 25.2 928 22.6 22.6a,** High triglycerides (n = 4107) <0.001*  No 2509 89.7 1075 82.1 3584 87.3 –  Yes 289 10.3 234 17.9 523 12.7 22.6a,** Hypertension (n = 4102) <0.001*  No 2523 90.3 1102 84.3 3625 88.4 –  Yes 272 9.7 205 15.7 477 11.6 25.7a Depression (n = 3987) <0.001*  No 2314 85.1 1164 91.7 3478 87.2 –  Yes 404 14.9 105 8.3 509 12.8 5.8b a VIGITEL survey, 2016. b WHO, 2017. * P-values from Pearson’s chi-square test; **prevalence of dyslipidaemia. CUME baseline data also show the diagnosis of chronic respiratory diseases (35.3%), cardiovascular diseases (16.3%), type 2 diabetes (3.3%) and cancer (2.2%) (data not shown). In Brazil, NCD are responsible for 72% of the causes of death, with the main causes being cardiovascular diseases (31.3%), cancer (16.3%), type 2 diabetes (5.2%) and chronic respiratory disease (5.8%).28 Considering this national scenario, the devolvement of a longitudinal study on NCD and eating and lifestyle habits, such as the CUME project, is a crucial epidemiological strategy. FFQs to assess food consumption and nutrient intake were returned by 3045 alumni, 2136 women (70.1%) and 909 men (29.9%). Table 3 presents the description of daily energy intake and macronutrients in the CUME and Brazilian populations.29 We found that only the total energy intake and lipids differed between the sexes, being greater among men. Table 3. Daily intake of calories and macronutrients of baseline participants of the Cohort of Universities of Minas Gerais (CUME), according to sex, 2016 CUME Brazila Female Male Total P-value* Female Male n = 2136 n = 909 n = 3045 Energy intake (kcal/day) 2130 2524 2242 <0.001* 1710 2163 (1683–2678) (1993–3225) (1755–2857) Carbohydrate (EI %) 47.7 46.4 47.3 0.214 56.2 54.8 (42.7–53.2) (40.9–51.9) (42.1–52.7) Protein (EI %) 17.3 17.6 17.4 0.227 16.4 16.9 (15.0–19.8) (15.3–19.8) (15.1–19.8) Lipids (EI %) 33.1 33.3 33.1 0.002* 27.5 27.2 (28.9–37.1) (29.1–37.2) (28.9–37.2) Saturated fat (EI %) 11.4 11.7 11.5 0.514 9.7 9.2 (9.6–13.3) (10.0–13.5) (9.8–13.4) CUME Brazila Female Male Total P-value* Female Male n = 2136 n = 909 n = 3045 Energy intake (kcal/day) 2130 2524 2242 <0.001* 1710 2163 (1683–2678) (1993–3225) (1755–2857) Carbohydrate (EI %) 47.7 46.4 47.3 0.214 56.2 54.8 (42.7–53.2) (40.9–51.9) (42.1–52.7) Protein (EI %) 17.3 17.6 17.4 0.227 16.4 16.9 (15.0–19.8) (15.3–19.8) (15.1–19.8) Lipids (EI %) 33.1 33.3 33.1 0.002* 27.5 27.2 (28.9–37.1) (29.1–37.2) (28.9–37.2) Saturated fat (EI %) 11.4 11.7 11.5 0.514 9.7 9.2 (9.6–13.3) (10.0–13.5) (9.8–13.4) Data are median (25th–75th percentile). All values presented are energy adjusted. EI, energy intake. a Average daily energy and macronutrients consumption according to sex and age (19-59 years) [Brazilian Family Budgets Survey (2008–09)]. * P-values from Mann-Whitney test. Table 3. Daily intake of calories and macronutrients of baseline participants of the Cohort of Universities of Minas Gerais (CUME), according to sex, 2016 CUME Brazila Female Male Total P-value* Female Male n = 2136 n = 909 n = 3045 Energy intake (kcal/day) 2130 2524 2242 <0.001* 1710 2163 (1683–2678) (1993–3225) (1755–2857) Carbohydrate (EI %) 47.7 46.4 47.3 0.214 56.2 54.8 (42.7–53.2) (40.9–51.9) (42.1–52.7) Protein (EI %) 17.3 17.6 17.4 0.227 16.4 16.9 (15.0–19.8) (15.3–19.8) (15.1–19.8) Lipids (EI %) 33.1 33.3 33.1 0.002* 27.5 27.2 (28.9–37.1) (29.1–37.2) (28.9–37.2) Saturated fat (EI %) 11.4 11.7 11.5 0.514 9.7 9.2 (9.6–13.3) (10.0–13.5) (9.8–13.4) CUME Brazila Female Male Total P-value* Female Male n = 2136 n = 909 n = 3045 Energy intake (kcal/day) 2130 2524 2242 <0.001* 1710 2163 (1683–2678) (1993–3225) (1755–2857) Carbohydrate (EI %) 47.7 46.4 47.3 0.214 56.2 54.8 (42.7–53.2) (40.9–51.9) (42.1–52.7) Protein (EI %) 17.3 17.6 17.4 0.227 16.4 16.9 (15.0–19.8) (15.3–19.8) (15.1–19.8) Lipids (EI %) 33.1 33.3 33.1 0.002* 27.5 27.2 (28.9–37.1) (29.1–37.2) (28.9–37.2) Saturated fat (EI %) 11.4 11.7 11.5 0.514 9.7 9.2 (9.6–13.3) (10.0–13.5) (9.8–13.4) Data are median (25th–75th percentile). All values presented are energy adjusted. EI, energy intake. a Average daily energy and macronutrients consumption according to sex and age (19-59 years) [Brazilian Family Budgets Survey (2008–09)]. * P-values from Mann-Whitney test. Due to the nutrition transition experienced by Brazilians and its association with the incidence and prevalence of NCD, the food consumption profile of the baseline participants of the CUME project was analysed. The daily energy intake (median) of the study population was 2242 kcal/day, being 47.3%, 17.4% and 33.1% of energy intake from carbohydrate, protein and lipids, respectively. Comparing with data from the Brazilian population, the participants of both sexes in our cohort study had a higher consumption of energy and lipids (Table 3). Further, the median intake of saturated fat among the alumni was 11.5%, which is a value higher than the recommended value for adults without NCD (<10% of daily energy intake), and for those who already have some NCD (<7% of daily calorie intake).30 In addition, the younger and more educated population of the CUME project, as previously presented, seems to present distinct dynamics in the occurrence of NCD. Thus, the longitudinal results of the CUME project will be of great importance in elucidating the temporal relationship between dietary patterns and NCD in this growing population group in Brazil. What are the main strengths and weaknesses? The main strengths of this study involve the longitudinal design that enables evaluation of associations between diseases and exposures. Also, this study uses an online platform, which has been a growing line of research in the field of nutritional epidemiology.31,32 Apart from the lower cost of the online platform compared with the printed version, it allows flexibility as to day and time to fill the questionnaire, which may favour the recruitment of participants from different locations, as verified by this project. The main disadvantage is that CUME is an open concurrent cohort restricted to a high educational-level population group. Therefore, this particular sample cannot represent the Brazilian population. However, the inclusion of individuals with high educational levels in this study is fundamental to providing reliable exposure data and outcomes, as well as to verifying how these highly educated individuals behave over time. Can I get hold of the data? Where can I find out more? The CUME project is conducted by the Universidade Federal de Viçosa and Universidade Federal de Minas Gerais, Brazil. Further information can be obtained at [http://www.projetocume.com.br] and [[emailprotected]]. Profile in a nutshell CUME is an open concurrent cohort restricted to a high educational-level population group. Initiated in 2016, CUME reached graduates in all Brazilian states and the Federal District. A total of 4291 alumni were eligible for the baseline data collection, mostly women (68%), young adults (72%, 20-39 years old) and postgraduate degree holders (80%). The cohort consists of waves of evaluation that will occur every 2 years in a virtual environment. Among the participants, 40.8% reported being overweight, 22.6% had high total cholesterol, 11.6% had hypertension and 3.3% had type 2 diabetes; all of these frequencies were lower compared with the general Brazilian adult population. However, CUME baseline participants already have chronic diseases, although they are younger and have a higher educational level than the general population, demonstrating the importance of the epidemiological scenario of these groups of diseases in Brazil. The prevalence of depression (12.8%) in our study was two times higher than the national prevalence (5.8%), which could be related to the contemporary lifestyle of our participants marked by an exhausting work routine, physical inactivity, binge drinking and unhealthy eating habits (e.g. saturated fat intake), despite their greater access to information and health services. Funding This study is being supported by Minas Gerais Research Foundation—FAPEMIG (Grants numbers: CDS-APQ-00571/13, CDS-APQ-02407/16, CDS-APQ-00424–17). Acknowledgements The authors are especially grateful to all the participants of the study, without whom this research would not be possible. We also thank: the funding agency, Minas Gerais Research Foundation (FAPEMIG), for the grants; the Department of Nutrition and Health of UFV; the Department of Maternal-Child Nursing and Public Health; and the Department of Nutrition of the Nursing School of UFMG, for the human and physical resources required to carry out the project. J.B., H.H. and F.O. are research productivity fellows of CNPq (Ministry of Science and Technology, Brazil). This work was conducted during a visiting scholar period at University of Navarra, sponsored by the Capes Foundation within the Ministry of Education, Brazil (Grant no. 88881.135299/2016–01). Conflict of interest: None declared. References 1 Malta DC , de Moura L , do Prado RR , Escalante JC , Schmidt MI , Duncan BB. Chronic non-communicable disease mortality in Brazil and its regions, 2000-2011 . Rev Epidemiol e Serviços Saúde 2014 ; 23 : 599 – 608 . Google Scholar Crossref Search ADS 2 Schmidt MI , Duncan BB , Azevedo G , Menezes AM , Monteiro CA , Barreto SM. Health in Brazil: four chronic non-communicable diseases in Brazil: burden and current challenges . Lancet 2011 ; 377 : 1949 – 61 . Google Scholar Crossref Search ADS PubMed 3 Malta DC , Morais Neto OL , Silva Junior JB. Presentation of the strategic action plan for coping with chronic diseases in Brazil from 2011 to 2022 . Epidemiol Serv Saúde 2011 ; 20 : 425 – 38 . Google Scholar Crossref Search ADS 4 Verly E Jr , Steluti J , Fisberg RM , Marchioni DML. A quantile regression approach can reveal the effect of fruit and vegetable consumption on plasma homocysteine levels . PLoS One 2014 ; 9 : e111619.9 . 5 Cocate PG , Natali AJ , de Oliveira A et al. Fruit and vegetable intake and related nutrients are associated with oxidative stress markers in middle-aged men . Nutrition 2014 ; 30 : 660 – 65 . Google Scholar Crossref Search ADS PubMed 6 Cocate PG , Natali AJ , Oliveira AD et al. Red but not white meat consumption is associated with metabolic syndrome, insulin resistance and lipid peroxidation in Brazilian middle-aged men . Eur J Prev Cardiol 2015 ; 22 : 223 – 30 . Google Scholar Crossref Search ADS PubMed 7 Ministério da Saúde . Conselho Nacional de Saúde. Resolução n 466 de 12 de dezembro de 2012 Aprova as diretrizes e normas regulamentadoras de pesquisas envolvendo seres humanos . Brasilia : Ministério da Saúde , 2012 . Brasil. National Council of Health. Resolution No. 466, of 12 December 2012. Available from: <http://conselho.saude.gov.br/resolucoes/2012/466_english.pdf> (15 July 2017, date last accessed). 8 National Institute on Alcohol Abuse and Alcoholism (NIAAA) . Drinking Levels Defined. 2015 . https://www.niaaa.nih.gov/alcohol-health/overview-alcohol-consumption/moderate-binge-drinking (2 April 2018 , date last accessed). 9 Martinez-Gonzalez MA , Lopez-Fontana C , Varo JJ , Sanchez-Villegas A , Martinez JA. Validation of the Spanish version of the physical activity questionnaire used in the Nurses’ Health Study and the Health Professionals’ Follow-up Study . Public Health Nutr 2005 ; 8 : 920 – 27 . Google Scholar Crossref Search ADS PubMed 10 World Health Organization . Global Recommendations on Physical Activity for Health . Geneva : WHO , 2010 . 11 D'Agostino RB , Vasan RS , Pencina MJ et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study . Circulation 2008 ; 117 : 743 – 53 . Google Scholar Crossref Search ADS PubMed 12 World Health Organization . Obesity: Preventing and Managing the Global Epidemic . Geneva : WHO , 2000 . 13 Malachias M , Plavnik FL , Machado CA et al. 7th Brazilian Guideline of Arterial Hypertension . Arq Bras Cardiol 2016 ; 107 : 1 – 83 . Google Scholar PubMed 14 de Oliveira JEP , Vencio S (eds); Sociedade Brasileira de Diabetes . Diretrizes Sociedade Brasileira de Diabetes (2015–2016) (Guidelines of Brazilian Diabetes Society). São Paulo, Brazil : A.C. Farmacêutica , 2016 . 15 Xavier HT , Izar MC , Faria Neto JR , Assad MH , Rocha VZ , Sposito AC. V Diretriz Brasileira de Dislipidemias e Prevenção da Aterosclerose (V Brazilian Guidelines on Dyslipidemias and Prevention of Atherosclerosis). Arq Bras Cardiol 2013 ; 101 : 1 – 20 . Google Scholar Crossref Search ADS 16 Henn RL , Fuchs SC , Moreira LB , Fuchs FD. Development and validation of a food frequency questionnaire (FFQ-Porto Alegre) for adolescent, adult and elderly populations from Southern Brazil . Cad Saúde Pública 2010 ; 26 : 2068 – 79 . Google Scholar Crossref Search ADS PubMed 17 Ministry of Health of Brazil , Primary Health Care Department. Dietary Guidelines for the Brazilian Population (Dietary Guidelines for the Brazilian Population) . Brasilia : Ministério da Saúde , 2014 . 18 Núcleo de Estudos e Pesquisas em Alimentação. Tabela Brasileira de Composição de Alimentos. Campinas, Brazil: Universidade Estadual de Campinas, 2011 . 19 Rodriguez-Amaya DB , Kimura M , Amaya-Farfán J. Fontes Brasileiras de Carotenóides. Tabela Brasileira de Composição de Carotenóides em Alimentos . Brasilia: Ministério de Meio Ambiente , 2008 . 20 Schmidt MI , Duncan BB , Mill JG et al. Cohort Profile: Longitudinal Study of Adult Health (ELSA-Brasil) . Int J Epidemiol 2015 ; 44 : 68 – 75 . Google Scholar Crossref Search ADS PubMed 21 Teixeira MG , Mill JG , Pereira AC , Molina M , del C. Dietary intake of antioxidant in ELSA-Brasil population: baseline results . Rev Bras Epidemiol 2016 ; 19 : 149 – 59 . Google Scholar Crossref Search ADS PubMed 22 Willett W , Stampfer MJ. Total energy intake: implications for epidemiologic analyses . Am J Epidemiol 1986 ; 124 : 17 – 27 . Google Scholar Crossref Search ADS PubMed 23 Instituto Brasileiro de Geografia e Estatística . Estimativas de População | Estatísticas | IBGE:: Instituto Brasileiro de Geografia e Estatística [Internet]. 2017. https://www.ibge.gov.br/estatisticas-novoportal/sociais/populacao/9103-estimativas-de-populacao.html?=&t=o-que-e (5 April 2018 , date last accessed). 24 Instituto Brasileiro de Geografia e Estatística (IBGE) . IBGE Censo 2010. https://censo2010.ibge.gov.br/ (5 April 2018 , date last accessed). 25 Ministry of Health . Vigitel Brazil 2016. Private Health Insurance and Plans Beneficiaries: Protective and Risk Factors for Chronic Diseases by Telephone Survey . Brasilia : Ministério da Saúde , 2017 . 26 Hallal PC , Andersen LB , Bull FC , Guthold R , Haskell W , Ekelund U. Global physical activity levels: surveillance progress, pitfalls, and prospects . Lancet 2012 ; 380 : 247 – 57 . Google Scholar Crossref Search ADS PubMed 27 WHO . Depression and Other Common Mental Disorders: Global Health Estimates . Geneva : World Health Organization , 2017 . 28 Ministry of Health . Strategic Action Plan to Tackle Noncommunicable Diseases (NCD) in Brazil 2011-2022 . Brasilia : Ministério da Saúde , 2011 *. 29 Brazilian Institute of Geography and Statistics . Consumer Expenditure Survey (POF) 2008-2009 , Rio de Janeiro, Brazil : Brazilian Institute of Geography and Statistics , 2011 . 30 Santos R , Gagliardi A , Xavier H et al. I Diretriz sobre o consumo de Gorduras e Saúde Cardiovascular (First guidelines on fat consumption and cardiovascular health). Arq Bras Cardiol 2013 ; 100 : 1 – 40 . 31 Apovian CM , Murphy MC , Cullum-Dugan D et al. Validation of a web-based dietary questionnaire designed for the DASH (dietary approaches to stop hypertension) diet: the DASH online questionnaire . Public Health Nutr 2010 ; 13 : 615 – 22 . Google Scholar Crossref Search ADS PubMed 32 Kristal AR , Kolar AS , Fisher JL et al. Evaluation of web-based, self-administered, graphical food frequency questionnaire . J Acad Nutr Diet 2014 ; 114 : 613 – 21 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Open Access Collection

Past and current asbestos exposure and future mesothelioma risks in Britain: The Inhaled Particles Study (TIPS)

Gilham, Clare; Rake, Christine; Hodgson, John; Darnton, Andrew; Burdett, Garry; Peto Wild, James; Newton, Michelle; Nicholson, Andrew G; Davidson, Leslie; Shires, Mike; Treasure, Tom; Peto, Julian; TIPS Collaboration ; Duncan, Andrew; Dusmet, Michael; Edwards, John G; Lim, Eric; Milton, Richard; Morgan, Ian; O’Keefe, P; Power, Danielle; Rajesh, P B; Rathinam, Sridhar; Rassl, Doris M; Routledge, T; Shackcloth, Michael; De Soyza, Anthony

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyx276pmid: 29534192

Abstract Background Occupational and environmental airborne asbestos concentrations are too low and variable for lifetime exposures to be estimated reliably, and building workers and occupants may suffer higher exposure when asbestos in older buildings is disturbed or removed. Mesothelioma risks from current asbestos exposures are therefore not known. Methods We interviewed and measured asbestos levels in lung samples from 257 patients treated for pneumothorax and 262 with resected lung cancer, recruited in England and Wales. Average lung burdens in British birth cohorts from 1940 to 1992 were estimated for asbestos-exposed workers and the general population. Results Regression analysis of British mesothelioma death rates and average lung burdens in birth cohorts born before 1965 suggests a lifetime mesothelioma risk of approximately 0.01% per fibre/mg of amphiboles in the lung. In those born since 1965, the average lung burden is ∼1 fibre/mg among those with no occupational exposure. Conclusions The average lifetime mesothelioma risk caused by recent environmental asbestos exposure in Britain will be about 1 in 10 000. The risk is an order of magnitude higher in a subgroup of exposed workers and probably in occupants in the most contaminated buildings. Further data are needed to discover whether asbestos still present in buildings, particularly schools, is a persistent or decreasing hazard to workers who disturb it and to the general population, and whether environmental exposure occurs predominantly in childhood or after beginning work. Similar studies are needed in other countries to estimate continuing environmental and occupational mesothelioma hazards worldwide, including the contribution from chrysotile. Key Messages Occupational and environmental mesothelioma risks from asbestos in older buildings are not known. Airborne concentrations are too low and variable for lifetime exposures to be estimated reliably, and mesothelioma rarely develops within 35 years of beginning asbestos exposure. British mesothelioma death rates are proportional to the population’s average amphibole asbestos lung burden (lifetime risk 0.01% per fibre/mg). Occupational and environmental risks can therefore be predicted from the distribution of asbestos lung burdens in people who began work since the 1980s, when asbestos was no longer used. The lifetime mesothelioma risk from environmental exposure among people born since 1965 will be ∼1 in 10 000, 10-fold less than in older people and almost 1000-fold less than in carpenters born in the 1940s. The risk is an order of magnitude higher in a subgroup of exposed workers. Further data are needed to discover whether asbestos in buildings, particularly schools, is a persisting or decreasing hazard. Introduction Britain’s mesothelioma rate is the highest worldwide and is still rising above age 70.1 Former construction workers, particularly carpenters, plumbers and electricians, are the main high-risk group.2 Most mesotheliomas develop more than 35 years after first asbestos exposure, so almost all recent cases are due to exposure before 1980 when asbestos was widely used, and only three of the 2542 mesothelioma deaths in Britain in 2015 were born after 1975.1 Building workers may still suffer substantial exposure when asbestos in older buildings is disturbed or removed, and the general population are potentially exposed in such buildings. However, the resulting mesothelioma risks are not known, as current occupational and environmental airborne concentrations are too low and variable for lifetime exposures to be estimated reliably. The aims of The Inhaled Particle Study (TIPS) were to determine whether the linear relationship between mesothelioma risk and asbestos lung burden in individuals3 is also seen in national mesothelioma death rates and population average burdens, and hence to predict future occupational and environmental mesothelioma rates from the lung burdens of exposed workers and of the general population born since 1965 who started work after 1980, when use of asbestos had virtually ceased in the UK. Chrysotile (white asbestos) fibres are ignored in our analyses which are based on amphibole fibres, mainly amosite (brown asbestos) and crocidolite (blue asbestos). Chrysotile causes a much lower mesothelioma risk than the amphiboles,4–6 but its effect cannot be estimated from our data because its half-life in the lung is too short3 for lung burden to reflect lifetime exposure. Chrysotile constituted 88% of UK asbestos imports between 1955 and 1990 but only 2% of asbestos fibres in the lungs of men with mesothelioma or lung cancer, born 1940–64.3 Whatever its effect, therefore, the dose-response estimate based on all asbestos fibres in the lung would be virtually the same as our estimate for amphiboles. Materials and Methods The study was approved by South Thames Multicentre Research Ethics Committee. Lifetime occupational histories were obtained by telephone interview from resected lung cancer and mesothelioma patients in a national case-control study as previously described,2,7 and also from 1005 unselected pneumothorax patients (648 men, 357 women) born between 1918 and 1996, recruited from 13 hospital centres in England and Wales. All eligible pneumothorax patients (aged 18 or over, with retained lung samples obtained at operation within the past 10 years) identified in these centres were invited by the local clinician to take part in a telephone interview. Overall 42% replied agreeing to be interviewed, of whom 91% gave consent for their lung material to be analysed. The lung burden study was restricted to participants born in 1940 or later. Normal lung tissue for transmission electron microscopy (TEM) analysis was excised from residual stored material from 262 lung cancers resected in 1999–2010 and at subsequent postmortem from 133 pleural mesothelioma patients in a previous study,3 and from 271 pneumothorax patients surgically treated in 2002–10 (a random sample of 251 stratified by year of birth, sex and centre and 20 additional men born since 1965 who had worked in construction). Asbestos fibres longer than 5 µm were counted by transmission electron microscopy (TEM). The analytical detection limit (lung burden per counted fibre) was reduced from 10 to 3.3 f/mg (fibres per milligram of dry lung) for the 165 (90%) pneumothorax patients born since 1965 with sufficient material available. Job titles were assigned to Standard Occupational Classification 1990 (SOC 90) and grouped into categories of similar mesothelioma risk, as in our case-control study.2 Subjects were assigned to the highest risk job category they had worked in irrespective of duration. We classified those who had ever worked in any of the five categories with elevated mesothelioma odds ratios in our case-control study,2 as having occupational exposure (carpenters; plumbers, electricians and painters; other construction workers; other high-risk work; and medium risk). Those who worked in none of these jobs are referred to as environmentally exposed, which includes any exposures from buildings they worked in. The Health and Safety Executive provided cumulative mesothelioma mortality rates to age 50 years in England, Scotland and Wales for each birth cohort from 1940–44 to 1960–64. Statistical methods The distribution of lung burden is approximately lognormal (Figure 1) and fibre counts are modelled as Poisson. Mean population lung burdens in different subgroups in Tables 1 and 2 were therefore estimated by maximizing the Poisson-lognormal likelihood. Mean asbestos lung burdens in the general population born before 1965 were estimated using samples from lung cancer and pneumothorax cases. Asbestos increases lung cancer risk, so our analysis adjusts for this, using the previously estimated3 increase in lung cancer risk ratio (RR) with lung burden (0.00255 per f/mg) to estimate mean lung burden in the population from the observed levels in lung cancer patients. The linear relationship between cumulative mesothelioma mortality to age 50 and population mean lung burden was also estimated by maximum likelihood. To estimate the increase per f/mg in lifetime risk (defined as the actuarial probability of dying of mesothelioma by age 90), the slope was multiplied by 51.8, the ratio of projected lifetime risk to observed risk by age 50 in men. The statistical appendix gives further details. All tables, figures and analyses are restricted to amphibole fibres, except Table 3 and Figure 3 which also show chrysotile lung burdens. Table 1. British mesothelioma mortality up to age 50 and population average amphibole lung burdens (f/mg) in the unselected sample by sex and year of birth . Males . Females . . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . 1940–44 184 302 62.2b (42.9, 91.8) 551/74 153/9 33 54 18.3 (11.2, 30.4) 87/26 0/1 1945–49 148 294 41.7 (30.5, 58.0) 394/66 54/13 29 58 13.3 (8.6, 21.2) 53/32 19/7 1950–54 99 180 30.8 (19.6, 49.0) 98/31 45/10 23 42 13.5 (7.1, 25.7) 19/15 11/6 1955–59 58 111 13.5 (5.8, 31.4) 25/7 6/7 22 44 10.8 (4.7, 25.2) 3/4 15/8 1960–64 35 63 10.9 (3.6, 32.0) 6/3 13/7 16 27 8.6 (3.6, 21.0) 8/3 8/7 1965–69 7.2 (2.3, 21.6) 1/1 9/8 1.2 (0.2, 4.4) 3/11 1970–74 3.3 (1.5, 7.0) 22/24 4.3 (1.7, 10.6) 14/11 1975–79 1.0 (0.3, 2.7) 6/21 1.2 (0.3, 3.3) 5/15 1980–84 3.2 (1.1, 9.1) 11/12 0.8 (0.2, 2.9) 3/12 1985–89 0.5 (0.1, 1.6) 3/21 1.0 (0.3, 2.7) 5/17 1990–92 0.0 (0.0, 2.4) 0/5 0.7 (0.03, 4.7) 1/5 Total 1075/182 322/137 170/80 84/100 . Males . Females . . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . 1940–44 184 302 62.2b (42.9, 91.8) 551/74 153/9 33 54 18.3 (11.2, 30.4) 87/26 0/1 1945–49 148 294 41.7 (30.5, 58.0) 394/66 54/13 29 58 13.3 (8.6, 21.2) 53/32 19/7 1950–54 99 180 30.8 (19.6, 49.0) 98/31 45/10 23 42 13.5 (7.1, 25.7) 19/15 11/6 1955–59 58 111 13.5 (5.8, 31.4) 25/7 6/7 22 44 10.8 (4.7, 25.2) 3/4 15/8 1960–64 35 63 10.9 (3.6, 32.0) 6/3 13/7 16 27 8.6 (3.6, 21.0) 8/3 8/7 1965–69 7.2 (2.3, 21.6) 1/1 9/8 1.2 (0.2, 4.4) 3/11 1970–74 3.3 (1.5, 7.0) 22/24 4.3 (1.7, 10.6) 14/11 1975–79 1.0 (0.3, 2.7) 6/21 1.2 (0.3, 3.3) 5/15 1980–84 3.2 (1.1, 9.1) 11/12 0.8 (0.2, 2.9) 3/12 1985–89 0.5 (0.1, 1.6) 3/21 1.0 (0.3, 2.7) 5/17 1990–92 0.0 (0.0, 2.4) 0/5 0.7 (0.03, 4.7) 1/5 Total 1075/182 322/137 170/80 84/100 a Lung burdens are adjusted for the effect of asbestos on lung cancer risk (see Statistical Methods). Respective unadjusted mean burdens in those born in 1940–44, 1945–49, 1950–54, 1955–59 and 1960–64 were 154.4, 52.0, 36.6, 14.8 and 11.7 f/mg in men and 20.2, 14.4, 14.6, 11.5 and 9.1 f/mg in women; respective unadjusted means based only on pneumothorax patients were 121.8, 17.8, 80.8, 1.6 and 15.0 f/mg in men and 0.0, 10.0, 16.2, 10.6 and 3.4 f/mg in women. b Including a lung cancer with 22 000 fibre/mg. Open in new tab Table 1. British mesothelioma mortality up to age 50 and population average amphibole lung burdens (f/mg) in the unselected sample by sex and year of birth . Males . Females . . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . 1940–44 184 302 62.2b (42.9, 91.8) 551/74 153/9 33 54 18.3 (11.2, 30.4) 87/26 0/1 1945–49 148 294 41.7 (30.5, 58.0) 394/66 54/13 29 58 13.3 (8.6, 21.2) 53/32 19/7 1950–54 99 180 30.8 (19.6, 49.0) 98/31 45/10 23 42 13.5 (7.1, 25.7) 19/15 11/6 1955–59 58 111 13.5 (5.8, 31.4) 25/7 6/7 22 44 10.8 (4.7, 25.2) 3/4 15/8 1960–64 35 63 10.9 (3.6, 32.0) 6/3 13/7 16 27 8.6 (3.6, 21.0) 8/3 8/7 1965–69 7.2 (2.3, 21.6) 1/1 9/8 1.2 (0.2, 4.4) 3/11 1970–74 3.3 (1.5, 7.0) 22/24 4.3 (1.7, 10.6) 14/11 1975–79 1.0 (0.3, 2.7) 6/21 1.2 (0.3, 3.3) 5/15 1980–84 3.2 (1.1, 9.1) 11/12 0.8 (0.2, 2.9) 3/12 1985–89 0.5 (0.1, 1.6) 3/21 1.0 (0.3, 2.7) 5/17 1990–92 0.0 (0.0, 2.4) 0/5 0.7 (0.03, 4.7) 1/5 Total 1075/182 322/137 170/80 84/100 . Males . Females . . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . Mortality to age 50 . Mean lung burden (fibres/mg) . Fibres counted/ subjects . . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . Rate per million . No. of deaths . Meana . 95% CI . Lung cancer . Pneumothorax . 1940–44 184 302 62.2b (42.9, 91.8) 551/74 153/9 33 54 18.3 (11.2, 30.4) 87/26 0/1 1945–49 148 294 41.7 (30.5, 58.0) 394/66 54/13 29 58 13.3 (8.6, 21.2) 53/32 19/7 1950–54 99 180 30.8 (19.6, 49.0) 98/31 45/10 23 42 13.5 (7.1, 25.7) 19/15 11/6 1955–59 58 111 13.5 (5.8, 31.4) 25/7 6/7 22 44 10.8 (4.7, 25.2) 3/4 15/8 1960–64 35 63 10.9 (3.6, 32.0) 6/3 13/7 16 27 8.6 (3.6, 21.0) 8/3 8/7 1965–69 7.2 (2.3, 21.6) 1/1 9/8 1.2 (0.2, 4.4) 3/11 1970–74 3.3 (1.5, 7.0) 22/24 4.3 (1.7, 10.6) 14/11 1975–79 1.0 (0.3, 2.7) 6/21 1.2 (0.3, 3.3) 5/15 1980–84 3.2 (1.1, 9.1) 11/12 0.8 (0.2, 2.9) 3/12 1985–89 0.5 (0.1, 1.6) 3/21 1.0 (0.3, 2.7) 5/17 1990–92 0.0 (0.0, 2.4) 0/5 0.7 (0.03, 4.7) 1/5 Total 1075/182 322/137 170/80 84/100 a Lung burdens are adjusted for the effect of asbestos on lung cancer risk (see Statistical Methods). Respective unadjusted mean burdens in those born in 1940–44, 1945–49, 1950–54, 1955–59 and 1960–64 were 154.4, 52.0, 36.6, 14.8 and 11.7 f/mg in men and 20.2, 14.4, 14.6, 11.5 and 9.1 f/mg in women; respective unadjusted means based only on pneumothorax patients were 121.8, 17.8, 80.8, 1.6 and 15.0 f/mg in men and 0.0, 10.0, 16.2, 10.6 and 3.4 f/mg in women. b Including a lung cancer with 22 000 fibre/mg. Open in new tab Table 2. Average amphibole lung burdena (fibres/mg) and 95% CI by occupation and year of birth in unselected lung cancer and pneumothorax patients and additional 20 construction workers with pneumothorax. (Number of fibres counted/number of subjects shown in parentheses.) The lower part shows the distribution of lung burdens by occupation and year of birth . Occupational exposure . Environmental exposure only . . Men . Women . Men . Women . Both sexes . . Carpenter . Plumber, electrician or painter . Other construction worker . High risk . Medium risk . Any occupational exposure . Any occupational exposure . . . Observed . Predicted scenario Ab . Predicted scenario Bc . Mesothelioma OR v. population controlsd 34.2 15.9 5.1 17.5 4.1 2.4 1.0 (ref) 1.0 (ref) Year of birth  1940–54 154.3 87.6 29.7 59.8 49.2 56.4 13.5 19.6 15.2 16.9 18.5 19.6 68.3–346.8 48.9–156.6 20.4–46.4 34.4–103.7 29.9–81.7 43.9–73.4 8.4–21.4 13.6–28.7 10.7–21.6 13.2–22.1 (217/12) (264/25) (204/48) (297/31) (207/41) (1189/157) (66/31) (106/46) (123/56) (229/102)  1955–64 78.0 15.6 2.1 0.0 11.7 22.7 8.9 5.9 9.4 7.9 6.3 7.4 18.8–323.9 4.1–57.6 0.2–17.7 3.0–41.2 8.4–60.2 1.6–37.8 2.2–14.5 5.0–17.5 4.8–13.3 (20/2) (11/4) (2/2) (0/1) (7/4) (40/13) (3/4) (10/11) (31/18) (41/29)  1965–74 1.8 9.1 4.1 3.0 6.2 4.0 1.0 2.4 1.7 1.1 1.9 0.2–9.1 3.7–21.7 1.5–10.7 1.0–7.9 3.0–12.8 1.0–13.7 0.3–3.1 1.0–5.4 0.9–3.4 (2/4) (19/9) (10/9) (9/10) (40/32) (6/5) (4/12) (11/17) (15/29)  1975–84 1.7 9.1 1.4 0.5 2.9 2.5 1.2 0.9 1.0 1.1 1.0 0.1–16.1 2.6–31.2 0.4–4.6 0.0–3.3 1.1–7.5 0.1–24.3 0.4–2.8 0.3–2.1 0.5–1.9 (1/2) (9/4) (4/9) (1/6) (15/21) (1/2) (7/19) (7/25) (14/44)  1985–92 0.0 1.8 0.0 0.5 0.5 0.9 0.7 1.1 0.3 0.1–16.9 0.0–4.7 0.1–1.4 0.3–2.2 0.3–1.4 (0/3) (1/2) (0/2) (1/7) (3/21) (6/22) (9/43) Lung fibre concentration f/mg Born 1940–64  < 5 1 5 15 12 12 45 15 22 30 52  5–24 2 9 17 7 17 52 14 21 29 50  25–199 7 10 17 8 14 56 6 14 14 28  ≥ 200 4 5 1 5 2 17 0 0 1 1 Born 1965–92  < 5 6 8 16 16 46 6 51 60 111  5–24 0 6 3 2 11 1 1 4 5  25–60 0 2 1 0 3 0 0 0 0 . Occupational exposure . Environmental exposure only . . Men . Women . Men . Women . Both sexes . . Carpenter . Plumber, electrician or painter . Other construction worker . High risk . Medium risk . Any occupational exposure . Any occupational exposure . . . Observed . Predicted scenario Ab . Predicted scenario Bc . Mesothelioma OR v. population controlsd 34.2 15.9 5.1 17.5 4.1 2.4 1.0 (ref) 1.0 (ref) Year of birth  1940–54 154.3 87.6 29.7 59.8 49.2 56.4 13.5 19.6 15.2 16.9 18.5 19.6 68.3–346.8 48.9–156.6 20.4–46.4 34.4–103.7 29.9–81.7 43.9–73.4 8.4–21.4 13.6–28.7 10.7–21.6 13.2–22.1 (217/12) (264/25) (204/48) (297/31) (207/41) (1189/157) (66/31) (106/46) (123/56) (229/102)  1955–64 78.0 15.6 2.1 0.0 11.7 22.7 8.9 5.9 9.4 7.9 6.3 7.4 18.8–323.9 4.1–57.6 0.2–17.7 3.0–41.2 8.4–60.2 1.6–37.8 2.2–14.5 5.0–17.5 4.8–13.3 (20/2) (11/4) (2/2) (0/1) (7/4) (40/13) (3/4) (10/11) (31/18) (41/29)  1965–74 1.8 9.1 4.1 3.0 6.2 4.0 1.0 2.4 1.7 1.1 1.9 0.2–9.1 3.7–21.7 1.5–10.7 1.0–7.9 3.0–12.8 1.0–13.7 0.3–3.1 1.0–5.4 0.9–3.4 (2/4) (19/9) (10/9) (9/10) (40/32) (6/5) (4/12) (11/17) (15/29)  1975–84 1.7 9.1 1.4 0.5 2.9 2.5 1.2 0.9 1.0 1.1 1.0 0.1–16.1 2.6–31.2 0.4–4.6 0.0–3.3 1.1–7.5 0.1–24.3 0.4–2.8 0.3–2.1 0.5–1.9 (1/2) (9/4) (4/9) (1/6) (15/21) (1/2) (7/19) (7/25) (14/44)  1985–92 0.0 1.8 0.0 0.5 0.5 0.9 0.7 1.1 0.3 0.1–16.9 0.0–4.7 0.1–1.4 0.3–2.2 0.3–1.4 (0/3) (1/2) (0/2) (1/7) (3/21) (6/22) (9/43) Lung fibre concentration f/mg Born 1940–64  < 5 1 5 15 12 12 45 15 22 30 52  5–24 2 9 17 7 17 52 14 21 29 50  25–199 7 10 17 8 14 56 6 14 14 28  ≥ 200 4 5 1 5 2 17 0 0 1 1 Born 1965–92  < 5 6 8 16 16 46 6 51 60 111  5–24 0 6 3 2 11 1 1 4 5  25–60 0 2 1 0 3 0 0 0 0 a Lung burden estimates are adjusted for the effect of asbestos on lung cancer risk, see Table 1 footnote a. b Scenario A: annual accumulation of 0.1 f/mg per year from ages 5 to 16 from 1945 to the present, followed after age 16 by 1 f/mg per year until 1980 and zero since 1980. c Scenario B: negligible exposure until age 16, followed after age 16 by 1 f/mg per year until 1980 and 0.1 f/mg per from 1980 until lung samples were obtained. For both scenarios, the calculation was based on individual years of birth and years of operation among those reporting no occupational exposure. d ORs (odds ratios) from the case-control study.2 Open in new tab Table 2. Average amphibole lung burdena (fibres/mg) and 95% CI by occupation and year of birth in unselected lung cancer and pneumothorax patients and additional 20 construction workers with pneumothorax. (Number of fibres counted/number of subjects shown in parentheses.) The lower part shows the distribution of lung burdens by occupation and year of birth . Occupational exposure . Environmental exposure only . . Men . Women . Men . Women . Both sexes . . Carpenter . Plumber, electrician or painter . Other construction worker . High risk . Medium risk . Any occupational exposure . Any occupational exposure . . . Observed . Predicted scenario Ab . Predicted scenario Bc . Mesothelioma OR v. population controlsd 34.2 15.9 5.1 17.5 4.1 2.4 1.0 (ref) 1.0 (ref) Year of birth  1940–54 154.3 87.6 29.7 59.8 49.2 56.4 13.5 19.6 15.2 16.9 18.5 19.6 68.3–346.8 48.9–156.6 20.4–46.4 34.4–103.7 29.9–81.7 43.9–73.4 8.4–21.4 13.6–28.7 10.7–21.6 13.2–22.1 (217/12) (264/25) (204/48) (297/31) (207/41) (1189/157) (66/31) (106/46) (123/56) (229/102)  1955–64 78.0 15.6 2.1 0.0 11.7 22.7 8.9 5.9 9.4 7.9 6.3 7.4 18.8–323.9 4.1–57.6 0.2–17.7 3.0–41.2 8.4–60.2 1.6–37.8 2.2–14.5 5.0–17.5 4.8–13.3 (20/2) (11/4) (2/2) (0/1) (7/4) (40/13) (3/4) (10/11) (31/18) (41/29)  1965–74 1.8 9.1 4.1 3.0 6.2 4.0 1.0 2.4 1.7 1.1 1.9 0.2–9.1 3.7–21.7 1.5–10.7 1.0–7.9 3.0–12.8 1.0–13.7 0.3–3.1 1.0–5.4 0.9–3.4 (2/4) (19/9) (10/9) (9/10) (40/32) (6/5) (4/12) (11/17) (15/29)  1975–84 1.7 9.1 1.4 0.5 2.9 2.5 1.2 0.9 1.0 1.1 1.0 0.1–16.1 2.6–31.2 0.4–4.6 0.0–3.3 1.1–7.5 0.1–24.3 0.4–2.8 0.3–2.1 0.5–1.9 (1/2) (9/4) (4/9) (1/6) (15/21) (1/2) (7/19) (7/25) (14/44)  1985–92 0.0 1.8 0.0 0.5 0.5 0.9 0.7 1.1 0.3 0.1–16.9 0.0–4.7 0.1–1.4 0.3–2.2 0.3–1.4 (0/3) (1/2) (0/2) (1/7) (3/21) (6/22) (9/43) Lung fibre concentration f/mg Born 1940–64  < 5 1 5 15 12 12 45 15 22 30 52  5–24 2 9 17 7 17 52 14 21 29 50  25–199 7 10 17 8 14 56 6 14 14 28  ≥ 200 4 5 1 5 2 17 0 0 1 1 Born 1965–92  < 5 6 8 16 16 46 6 51 60 111  5–24 0 6 3 2 11 1 1 4 5  25–60 0 2 1 0 3 0 0 0 0 . Occupational exposure . Environmental exposure only . . Men . Women . Men . Women . Both sexes . . Carpenter . Plumber, electrician or painter . Other construction worker . High risk . Medium risk . Any occupational exposure . Any occupational exposure . . . Observed . Predicted scenario Ab . Predicted scenario Bc . Mesothelioma OR v. population controlsd 34.2 15.9 5.1 17.5 4.1 2.4 1.0 (ref) 1.0 (ref) Year of birth  1940–54 154.3 87.6 29.7 59.8 49.2 56.4 13.5 19.6 15.2 16.9 18.5 19.6 68.3–346.8 48.9–156.6 20.4–46.4 34.4–103.7 29.9–81.7 43.9–73.4 8.4–21.4 13.6–28.7 10.7–21.6 13.2–22.1 (217/12) (264/25) (204/48) (297/31) (207/41) (1189/157) (66/31) (106/46) (123/56) (229/102)  1955–64 78.0 15.6 2.1 0.0 11.7 22.7 8.9 5.9 9.4 7.9 6.3 7.4 18.8–323.9 4.1–57.6 0.2–17.7 3.0–41.2 8.4–60.2 1.6–37.8 2.2–14.5 5.0–17.5 4.8–13.3 (20/2) (11/4) (2/2) (0/1) (7/4) (40/13) (3/4) (10/11) (31/18) (41/29)  1965–74 1.8 9.1 4.1 3.0 6.2 4.0 1.0 2.4 1.7 1.1 1.9 0.2–9.1 3.7–21.7 1.5–10.7 1.0–7.9 3.0–12.8 1.0–13.7 0.3–3.1 1.0–5.4 0.9–3.4 (2/4) (19/9) (10/9) (9/10) (40/32) (6/5) (4/12) (11/17) (15/29)  1975–84 1.7 9.1 1.4 0.5 2.9 2.5 1.2 0.9 1.0 1.1 1.0 0.1–16.1 2.6–31.2 0.4–4.6 0.0–3.3 1.1–7.5 0.1–24.3 0.4–2.8 0.3–2.1 0.5–1.9 (1/2) (9/4) (4/9) (1/6) (15/21) (1/2) (7/19) (7/25) (14/44)  1985–92 0.0 1.8 0.0 0.5 0.5 0.9 0.7 1.1 0.3 0.1–16.9 0.0–4.7 0.1–1.4 0.3–2.2 0.3–1.4 (0/3) (1/2) (0/2) (1/7) (3/21) (6/22) (9/43) Lung fibre concentration f/mg Born 1940–64  < 5 1 5 15 12 12 45 15 22 30 52  5–24 2 9 17 7 17 52 14 21 29 50  25–199 7 10 17 8 14 56 6 14 14 28  ≥ 200 4 5 1 5 2 17 0 0 1 1 Born 1965–92  < 5 6 8 16 16 46 6 51 60 111  5–24 0 6 3 2 11 1 1 4 5  25–60 0 2 1 0 3 0 0 0 0 a Lung burden estimates are adjusted for the effect of asbestos on lung cancer risk, see Table 1 footnote a. b Scenario A: annual accumulation of 0.1 f/mg per year from ages 5 to 16 from 1945 to the present, followed after age 16 by 1 f/mg per year until 1980 and zero since 1980. c Scenario B: negligible exposure until age 16, followed after age 16 by 1 f/mg per year until 1980 and 0.1 f/mg per from 1980 until lung samples were obtained. For both scenarios, the calculation was based on individual years of birth and years of operation among those reporting no occupational exposure. d ORs (odds ratios) from the case-control study.2 Open in new tab Table 3. Number and percentage of fibres counted by asbestos fibre type, year of birth, sex and occupation . . Number of fibres counted . % of fibres counted . Average lung burden f/mgb . . . Amphiboles . Chrysotile . Amphiboles . Chrysotile . Amphiboles . Chrysotile . . Fibre typea . am . cr . tr . an . ac . ua . . am . cr . tr + an + ac + ua . . am + cr . tr+ an + ac . . . n persons . . Men born since 1965  Environmental only 52 6 2 4 1 1 0 5 31.6 10.5 31.6 26.3 0.5 0.3 0.3  Carpenter 6 3 0 0 0 0 0 0 100.0 0.0 0.0 0.0 1.6 0.0 0.0  Plumber, electrician, painter 16 26 1 0 0 1 0 0 92.9 3.6 3.6 0.0 6.8 0.9 0.0  Other construction workers 20 10 0 1 1 3 0 1 62.5 0.0 31.3 6.3 3.1 0.8 0.3  Medium-risk 18 7 0 1 1 1 0 2 58.3 0.0 25.0 16.7 1.3 0.6 0.4  Total 112 52 3 6 3 6 0 8 66.7 3.8 19.2 10.3 1.9 0.4 0.2 Women born since 1965  Environmental only 64 14 2 3 4 2 0 5 46.7 6.7 30.0 16.7 0.8 0.5 0.3  Medium-risk 7 6 0 1 0 0 0 0 85.7 0.0 14.3 0.0 2.9 0.5 0.0  Total 71 20 2 4 4 2 0 5 54.1 5.4 27.0 13.5 1.0 0.5 0.2 Men born 1940–64  Environmental only 57 62 14 12 21 5 3 11 48.4 10.9 32.0 8.6 11.4 4.8 1.7  High-risk 32 243 43 1 4 4 2 8 79.7 14.1 3.6 2.6 343.4 2.9 2.0  Carpenter 14 216 11 1 4 1 4 1 90.8 4.6 4.2 0.4 173.0 3.2 0.5  Plumber, electrician, painter 29 203 55 4 5 4 4 4 72.8 19.7 6.1 1.4 129.1 3.1 1.0  Other construction workers 50 170 17 1 8 5 5 13 77.6 7.8 8.7 5.9 30.0 3.1 2.5  Medium-risk 45 149 43 5 11 2 4 5 68.0 19.6 10.0 2.3 80.1 2.6 0.6  Total 227 1043 183 24 53 21 22 42 75.1 13.2 8.6 3.0 70.7 3.1 1.5 Women born 1940–64  Environmental only 74 88 23 13 24 3 3 29 48.1 12.6 23.5 15.8 12.1 3.0 1.8  Medium-risk 35 46 8 2 12 0 1 7 60.5 10.5 19.7 9.2 12.2 3.2 1.3  Total 109 134 31 15 36 3 4 36 51.7 12.0 22.4 13.9 12.1 3.0 1.4 . . Number of fibres counted . % of fibres counted . Average lung burden f/mgb . . . Amphiboles . Chrysotile . Amphiboles . Chrysotile . Amphiboles . Chrysotile . . Fibre typea . am . cr . tr . an . ac . ua . . am . cr . tr + an + ac + ua . . am + cr . tr+ an + ac . . . n persons . . Men born since 1965  Environmental only 52 6 2 4 1 1 0 5 31.6 10.5 31.6 26.3 0.5 0.3 0.3  Carpenter 6 3 0 0 0 0 0 0 100.0 0.0 0.0 0.0 1.6 0.0 0.0  Plumber, electrician, painter 16 26 1 0 0 1 0 0 92.9 3.6 3.6 0.0 6.8 0.9 0.0  Other construction workers 20 10 0 1 1 3 0 1 62.5 0.0 31.3 6.3 3.1 0.8 0.3  Medium-risk 18 7 0 1 1 1 0 2 58.3 0.0 25.0 16.7 1.3 0.6 0.4  Total 112 52 3 6 3 6 0 8 66.7 3.8 19.2 10.3 1.9 0.4 0.2 Women born since 1965  Environmental only 64 14 2 3 4 2 0 5 46.7 6.7 30.0 16.7 0.8 0.5 0.3  Medium-risk 7 6 0 1 0 0 0 0 85.7 0.0 14.3 0.0 2.9 0.5 0.0  Total 71 20 2 4 4 2 0 5 54.1 5.4 27.0 13.5 1.0 0.5 0.2 Men born 1940–64  Environmental only 57 62 14 12 21 5 3 11 48.4 10.9 32.0 8.6 11.4 4.8 1.7  High-risk 32 243 43 1 4 4 2 8 79.7 14.1 3.6 2.6 343.4 2.9 2.0  Carpenter 14 216 11 1 4 1 4 1 90.8 4.6 4.2 0.4 173.0 3.2 0.5  Plumber, electrician, painter 29 203 55 4 5 4 4 4 72.8 19.7 6.1 1.4 129.1 3.1 1.0  Other construction workers 50 170 17 1 8 5 5 13 77.6 7.8 8.7 5.9 30.0 3.1 2.5  Medium-risk 45 149 43 5 11 2 4 5 68.0 19.6 10.0 2.3 80.1 2.6 0.6  Total 227 1043 183 24 53 21 22 42 75.1 13.2 8.6 3.0 70.7 3.1 1.5 Women born 1940–64  Environmental only 74 88 23 13 24 3 3 29 48.1 12.6 23.5 15.8 12.1 3.0 1.8  Medium-risk 35 46 8 2 12 0 1 7 60.5 10.5 19.7 9.2 12.2 3.2 1.3  Total 109 134 31 15 36 3 4 36 51.7 12.0 22.4 13.9 12.1 3.0 1.4 aAm, amosite; cr, crocidolite; tr, tremolite; an, anthophyllite; ac, actinolite; ua, untyped amphibole. bAverage lung burdens unadjusted for the effect of asbestos on lung cancer (see Statistical Methods). Open in new tab Table 3. Number and percentage of fibres counted by asbestos fibre type, year of birth, sex and occupation . . Number of fibres counted . % of fibres counted . Average lung burden f/mgb . . . Amphiboles . Chrysotile . Amphiboles . Chrysotile . Amphiboles . Chrysotile . . Fibre typea . am . cr . tr . an . ac . ua . . am . cr . tr + an + ac + ua . . am + cr . tr+ an + ac . . . n persons . . Men born since 1965  Environmental only 52 6 2 4 1 1 0 5 31.6 10.5 31.6 26.3 0.5 0.3 0.3  Carpenter 6 3 0 0 0 0 0 0 100.0 0.0 0.0 0.0 1.6 0.0 0.0  Plumber, electrician, painter 16 26 1 0 0 1 0 0 92.9 3.6 3.6 0.0 6.8 0.9 0.0  Other construction workers 20 10 0 1 1 3 0 1 62.5 0.0 31.3 6.3 3.1 0.8 0.3  Medium-risk 18 7 0 1 1 1 0 2 58.3 0.0 25.0 16.7 1.3 0.6 0.4  Total 112 52 3 6 3 6 0 8 66.7 3.8 19.2 10.3 1.9 0.4 0.2 Women born since 1965  Environmental only 64 14 2 3 4 2 0 5 46.7 6.7 30.0 16.7 0.8 0.5 0.3  Medium-risk 7 6 0 1 0 0 0 0 85.7 0.0 14.3 0.0 2.9 0.5 0.0  Total 71 20 2 4 4 2 0 5 54.1 5.4 27.0 13.5 1.0 0.5 0.2 Men born 1940–64  Environmental only 57 62 14 12 21 5 3 11 48.4 10.9 32.0 8.6 11.4 4.8 1.7  High-risk 32 243 43 1 4 4 2 8 79.7 14.1 3.6 2.6 343.4 2.9 2.0  Carpenter 14 216 11 1 4 1 4 1 90.8 4.6 4.2 0.4 173.0 3.2 0.5  Plumber, electrician, painter 29 203 55 4 5 4 4 4 72.8 19.7 6.1 1.4 129.1 3.1 1.0  Other construction workers 50 170 17 1 8 5 5 13 77.6 7.8 8.7 5.9 30.0 3.1 2.5  Medium-risk 45 149 43 5 11 2 4 5 68.0 19.6 10.0 2.3 80.1 2.6 0.6  Total 227 1043 183 24 53 21 22 42 75.1 13.2 8.6 3.0 70.7 3.1 1.5 Women born 1940–64  Environmental only 74 88 23 13 24 3 3 29 48.1 12.6 23.5 15.8 12.1 3.0 1.8  Medium-risk 35 46 8 2 12 0 1 7 60.5 10.5 19.7 9.2 12.2 3.2 1.3  Total 109 134 31 15 36 3 4 36 51.7 12.0 22.4 13.9 12.1 3.0 1.4 . . Number of fibres counted . % of fibres counted . Average lung burden f/mgb . . . Amphiboles . Chrysotile . Amphiboles . Chrysotile . Amphiboles . Chrysotile . . Fibre typea . am . cr . tr . an . ac . ua . . am . cr . tr + an + ac + ua . . am + cr . tr+ an + ac . . . n persons . . Men born since 1965  Environmental only 52 6 2 4 1 1 0 5 31.6 10.5 31.6 26.3 0.5 0.3 0.3  Carpenter 6 3 0 0 0 0 0 0 100.0 0.0 0.0 0.0 1.6 0.0 0.0  Plumber, electrician, painter 16 26 1 0 0 1 0 0 92.9 3.6 3.6 0.0 6.8 0.9 0.0  Other construction workers 20 10 0 1 1 3 0 1 62.5 0.0 31.3 6.3 3.1 0.8 0.3  Medium-risk 18 7 0 1 1 1 0 2 58.3 0.0 25.0 16.7 1.3 0.6 0.4  Total 112 52 3 6 3 6 0 8 66.7 3.8 19.2 10.3 1.9 0.4 0.2 Women born since 1965  Environmental only 64 14 2 3 4 2 0 5 46.7 6.7 30.0 16.7 0.8 0.5 0.3  Medium-risk 7 6 0 1 0 0 0 0 85.7 0.0 14.3 0.0 2.9 0.5 0.0  Total 71 20 2 4 4 2 0 5 54.1 5.4 27.0 13.5 1.0 0.5 0.2 Men born 1940–64  Environmental only 57 62 14 12 21 5 3 11 48.4 10.9 32.0 8.6 11.4 4.8 1.7  High-risk 32 243 43 1 4 4 2 8 79.7 14.1 3.6 2.6 343.4 2.9 2.0  Carpenter 14 216 11 1 4 1 4 1 90.8 4.6 4.2 0.4 173.0 3.2 0.5  Plumber, electrician, painter 29 203 55 4 5 4 4 4 72.8 19.7 6.1 1.4 129.1 3.1 1.0  Other construction workers 50 170 17 1 8 5 5 13 77.6 7.8 8.7 5.9 30.0 3.1 2.5  Medium-risk 45 149 43 5 11 2 4 5 68.0 19.6 10.0 2.3 80.1 2.6 0.6  Total 227 1043 183 24 53 21 22 42 75.1 13.2 8.6 3.0 70.7 3.1 1.5 Women born 1940–64  Environmental only 74 88 23 13 24 3 3 29 48.1 12.6 23.5 15.8 12.1 3.0 1.8  Medium-risk 35 46 8 2 12 0 1 7 60.5 10.5 19.7 9.2 12.2 3.2 1.3  Total 109 134 31 15 36 3 4 36 51.7 12.0 22.4 13.9 12.1 3.0 1.4 aAm, amosite; cr, crocidolite; tr, tremolite; an, anthophyllite; ac, actinolite; ua, untyped amphibole. bAverage lung burdens unadjusted for the effect of asbestos on lung cancer (see Statistical Methods). Open in new tab Figure 1 Open in new tabDownload slide Approximately lognormal distribution of amphibole lung burdens in male mesothelioma, lung cancer and pneumothorax patients born 1940–64. Values < 5 f/mg are recoded as 0.01 f/mg, including 5/106 mesothelioma, 35/181 lung cancer and 14/46 pneumothorax samples in which no fibres were counted. Figure 1 Open in new tabDownload slide Approximately lognormal distribution of amphibole lung burdens in male mesothelioma, lung cancer and pneumothorax patients born 1940–64. Values < 5 f/mg are recoded as 0.01 f/mg, including 5/106 mesothelioma, 35/181 lung cancer and 14/46 pneumothorax samples in which no fibres were counted. Results In men, the average amphibole lung burden fell from 62 f/mg (born 1940–44) to 11 f/mg (born 1960–64) and mesothelioma risk per million to age 50 fell from 184 to 35 (Table 1, Figure 2a). In women, the average lung burden fell from 18 f/mg (born 1940–44) to 9 f/mg (born 1960–64) and their risk per million to age 50 fell from 33 to 16. The dose-specific mesothelioma risk to age 50 estimated from these data is 0.00032% per f/mg [95% confidence interval (CI) 0.00026%, 0.00040%)] for men and 0.00019% per f/mg (95% CI 0.00014%, 0.00024%) for women (P < 0.002). Average lung burdens unadjusted for asbestos-related lung cancer risk for those born 1940–64 are shown in Table 1 footnote a. (Only one lung cancer patient was born after 1964.) The adjustment has a material impact only for men born before 1955. Figure 2 Open in new tabDownload slide (a) National mesothelioma mortality and average amphibole asbestos lung burdens in Britain by year of birth (fibres/mg longer than 5 microns). Subjects born 1940–64 are predominantly resected lung cancer patients, whereas all but one of those born 1965–92 are pneumothorax patients. (b) Average amphibole asbestos lung burdens in occupationally exposed men by year of birth (fibres/mg longer than 5 microns). Data for environmental exposure only include both sexes. Figure 2 Open in new tabDownload slide (a) National mesothelioma mortality and average amphibole asbestos lung burdens in Britain by year of birth (fibres/mg longer than 5 microns). Subjects born 1940–64 are predominantly resected lung cancer patients, whereas all but one of those born 1965–92 are pneumothorax patients. (b) Average amphibole asbestos lung burdens in occupationally exposed men by year of birth (fibres/mg longer than 5 microns). Data for environmental exposure only include both sexes. Table 2 and Figure 2b show lung burdens by year of birth and highest risk occupation. For environmental exposure (those who never worked in hazardous occupations), burdens were much lower and were similar in men and women. In those born 1940–64, the proportion with lung burdens exceeding 200 f/mg was 19% (14/75) among men who worked in the three highest risk categories (carpenters; plumbers, electricians and painters; other high-risk occupations), 2% (3/152) among other men and 1% (1/109) among women. None exceeded 60 f/mg in those born since 1965. Table 3 shows counts for each fibre type and unadjusted lung burdens for all amphiboles and chrysotile by year of birth, sex and occupation. In men, the overall distribution of counted fibres was 75% amosite, 13% crocidolite, 9% other amphiboles and 3% chrysotile, and in women 52% amosite, 11% crocidolite, 23% other amphiboles and 14% chrysotile. Fibre type differed between occupational groups, carpenters having the highest proportion of amosite (90.8%) and the lowest of crocidolite (0.4%). Chrysotile concentrations were uniformly low and showed no consistent relationship with occupation or gender. People born in 1965–74 began work after 1980 when amosite materials were no longer being installed. Their average lung burden was as low in carpenters (1.8 f/mg) as in unexposed men and women (1.7 f/mg) but remained substantially higher among plumbers, painters and electricians (9.1 f/mg: Table 2, Figure 2b). Figure 3 shows that crocidolite burdens fell sharply in men born after 1950, about 5 years earlier than amosite. Figure 3 Open in new tabDownload slide Average asbestos lung burdensa in Britain by year of birth (fibres/mg longer than 5 microns). Upper graph: crocidolite and amosite by sex. Lower graph: other amphiboles and chrysotile (both sexes). aAverage lung burdens unadjusted for the effect of asbestos on lung cancer (see Statistical Methods). bExcluding a chrysotile concentration of 72 f/mg based on 24 fibres in a woman who reported no asbestos exposure. Her inclusion increases the chrysotile average for those born 1960–64 from 2.0 to 26.0 f/mg. Figure 3 Open in new tabDownload slide Average asbestos lung burdensa in Britain by year of birth (fibres/mg longer than 5 microns). Upper graph: crocidolite and amosite by sex. Lower graph: other amphiboles and chrysotile (both sexes). aAverage lung burdens unadjusted for the effect of asbestos on lung cancer (see Statistical Methods). bExcluding a chrysotile concentration of 72 f/mg based on 24 fibres in a woman who reported no asbestos exposure. Her inclusion increases the chrysotile average for those born 1960–64 from 2.0 to 26.0 f/mg. Discussion Trends in lung burden and dose-specific risk in those born before 1965 Average lung burdens in men born 1940–54 (Table 2) reflect the ranking of occupational and environmental relative risks seen in our case-control study2 (154 f/mg in carpenters, 88 f/mg in plumbers, electricians and decorators, 60 f/mg in other high-risk occupations (including shipbuilding and lagging), 49 f/mg in medium-risk (mainly factory) work and 30 f/mg in general construction). Lung burdens in those born 1940–54 with environmental exposure only were similar in men and women (average 17 f/mg). Occupationally exposed women had a similar level (14 f/mg, 95% CI 8, 31). Occupational and environmental lung burdens were substantially lower in those born 1955–64 but show a similar pattern. Regression analysis of the parallel decline in mesothelioma mortality and average amphibole lung burden in male birth cohorts from 1940–44 to 1960–64 (Figure 2a, Table 1) gives a cumulative risk by age 50 in men of 0.00032% per f/mg. Multiplying by 51.8 (see Statistical Methods) gives a lifelong mortality of 0.017% per f/mg, close to the lifetime incidence of 0.020% per f/mg estimated from case-control analysis of lung burdens in male mesothelioma patients.3 However, the male data are dominated by a heavily exposed minority. The estimated increase in lung cancer RR from our case-control study (0.00255 per f/mg) is very imprecise,3 and adjusting for it substantially reduced the estimated average lung burdens of men born before 1955 (see Table 1 footnote a). Lung burdens in women are much lower and are hardly altered by the adjustment. Therefore we believe that the female estimate of the risk per f/mg (0.00019% by age 50, lifetime risk 0.010%) provides a more reliable indication of future mesothelioma rates in both sexes from recent exposure, which is predominantly environmental. This predicts a lifetime mesothelioma risk of the order of 1 in 10 000 at the average lung burden of ∼1 f/mg due to environmental exposure in men and women born since 1965 (Table 2). Asbestos exposure since 1980 By 1980, when those born in 1965 were starting work, traditional high-risk occupations such as lagging and shipbuilding had disappeared and carpenters no longer cut amosite board. The only occupational groups born since 1965 with substantially higher lung burdens than the general population are the 43% (6/14) of plumbers, electricians and decorators, 17% (3/18) of other construction workers and 14% (3/22) in medium-risk occupations in whom two or more amphibole fibres were counted in approximately 0.3 mg of lung tissue. The mean lung burden in these 12 cases (11 f/mg) implies a lifetime risk of ∼1 in 1000. Potentially remediable work practices seem likely to underlie this continuing occupational hazard. The distribution among the other 48 men and women in jobs classed as occupationally exposed in whom fewer than two fibres were counted, including the remaining eight plumbers, electricians and decorators, was 35 with no fibres and 13 with one fibre, similar to that among those with environmental exposure only. The reduction in the asbestos-exposed workforce and their declining lung burdens are reflected in the converging trends in male and female mesothelioma rates (Figure 2a). The majority of mesotheliomas in people born since 1965 will be caused by environmental exposure, presumably mainly in buildings. Numbers of amphibole fibres counted in 105 men and women born since 1965 with environmental exposure only (77 with none, 22 with one, four with two, one with three and one with four fibres) suggest fairly uniform environmental exposure across the UK, with a minority having higher (probably unsuspected) exposure. For example, these fibre counts are consistent with about 10% having a mean lung burden of ∼6 f/mg (lifetime risk ∼1 in 2000), with the remaining 90% having a lung burden an order of magnitude lower (∼0.6 f/mg; lifetime risk ∼1 in 20 000). The steep decline in mean lung burden in men and women with environmental exposure only from 17 f/mg born 1940–54 to 1 f/mg born 1975–84 (P < 0.001) indicates that environmental as well as occupational exposure levels fell abruptly around 1980 when use of amphibole products had ended. This suggests that until the 1970s, most asbestos entered the environment during or soon after installation of new asbestos materials. Current environmental releases may also occur mainly during construction or demolition work on asbestos-containing buildings.8 (Our sample included no asbestos removal workers, but removal and demolition may contribute substantially to both occupational and environmental exposure.) However, airborne asbestos fibres released by weathering and everyday occupation of buildings may also be an important source of environmental exposure. Identifying asbestos in buildings that warrants containment or removal should continue to be a regulatory priority, but unnecessary asbestos removal may increase the number of fibres released to the environment. The trend in average lung burden for men and women born before 1965 with only environmental exposure suggests an annual increment in eventual lung burden of ∼1 f/mg per year in adults until about 1980, when it fell sharply. The crucial question is whether environmental exposures, particularly in childhood, have remained fairly constant since 1980. In men and women born since 1965 with only environmental exposure, the average lung burden declines from 1.7 f/mg (95% CI 0.9, 3.4) born 1965–74, to 0.7 f/mg (95% CI 0.3, 1.4) born 1985–92 (P = 0.04), but the data are too sparse for the separate contributions of exposure in infancy, during school age and in adults to be estimated. Table 2 shows predicted lung burdens under two scenarios that are both consistent with these data but imply very different regulatory priorities: continuing exposure from age 5 to 16 with negligible environmental exposure after age 16 since 1980 (scenario A), and environmental exposure being negligible in childhood and beginning at age 16 (scenario B). Domestic exposure in infancy could be included without greatly altering these predicted lung burdens. The excess over these environmental levels in the average lung burdens of men with any occupational exposure increases for each year after age 16 by about 2 f/mg per year from 1955 to 1980, and after 1980 by about 0.1 f/mg per year in plumbers, electricians and painters, almost ceasing in other occupations (Table 2). UK amphibole imports up to 1980 show a similar pattern,3 changing little from 1960 to the late 1970s when amosite imports ended abruptly. Crocidolite use ended in 1970,3 and this is reflected in the earlier decline of crocidolite lung burdens in both sexes (Figure 3). If asbestos levels have not fallen since the 1980s, our results suggest an average lung burden from current environmental amphibole exposure of about 1 f/mg by age 30. Lifetime mesothelioma risk is largely determined by asbestos exposure before age 30,2,9 and most of the amphibole fibres still present in the lungs of those born 1940–64, on whom our linear dose-response is based, were inhaled before age 30. However, the only direct evidence of recent environmental exposure is the average lung burden (0.7 f/mg) in 43 unexposed men and women born 1985–92 (Table 2), which is very imprecise and only includes fibres inhaled up to about age 19, the median age when their lung samples were taken. The 14 fibres counted in these 43 subjects comprised five amosite, one tremolite, one anthophyllite, two actinolite and five chrysotile. Study limitations The consistency of the lung burden patterns in Table 2 with known occupational and environmental risks and national trends in mesothelioma mortality is reassuring. However, prediction of future risk from lung burdens in young adults may be affected by several factors. These include the proportion of environmental exposure that occurs in childhood, differences in amphibole fibre type and dimension between past occupational and current environmental exposure, and the opposite effects of fibre clearance and future accumulation on the lung burdens of those born since 1980 who were aged under 30 when samples were taken. Amosite has a particularly long half-life,3 but it is not known whether most fibres still present 20 years after inhalation remain in the lung forever, or whether carcinogenicity and clearance of tremolite, anthophyllite and actinolite are similar. Our main findings are unaffected by information bias, as the average lung burdens in Table 1 were based on the unselected sample irrespective of reported occupation. Any systematic differences between pneumothorax patients and the general population should have little effect on our prediction of future mesothelioma rates, if the dose-response in those born before 1965 and the lung burdens of younger people had both been based solely on pneumothorax patients. However, 78% of subjects born before 1965 were lung cancers from our previous study.3 The high cost of sample preparation and TEM precluded replacing them with pneumothorax patients, but differences between lung cancer and pneumothorax patients might lead to error in our prediction of future mesothelioma rates even if lung burdens in young pneumothorax patients were known precisely. Mean lung burdens in pneumothorax patients born before 1965 show no consistent difference from the overall estimates but vary irregularly across birth cohorts due to small numbers (Table 1 footnote a). The primary risk factor for both lung cancer and pneumothorax is smoking10,11 (among our participants 94% of lung cancers and 75% of pneumothorax patients had ever smoked), so marked differential bias related to the populations studied seems unlikely, but the lung sample was apical in almost all pneumothorax patients and from various sites in resected lung cancers. To avoid these uncertainties, future studies should use lung samples only from pneumothorax patients. This would also simplify the statistical analysis and might eliminate the difference between the results in men and women. Further studies and international comparisons Lung burden studies on larger numbers of young people would determine whether environmental exposures have fallen since the 1980s and whether they occur predominantly in childhood or after beginning work. Analysis of larger amounts of tissue to increase sensitivity would identify individuals with higher levels that might be linked to specific buildings or other sources of environmental exposure. The mesothelioma risk from chrysotile is low6 but cannot be estimated from our results,12 and an international study of average TEM asbestos lung burdens is needed to show whether or not mesothelioma mortality in different birth cohorts can be explained by historical amphibole exposure even in countries where almost all asbestos was chrysotile. The risk per fibre for different amphiboles might also be estimated. Lower amphibole imports account for the much lower mesothelioma rate in the USA than in Britain and Australia,2,3 despite similar overall asbestos consumption per head. There is no consistent international correlation between overall asbestos consumption and mesothelioma risk, but crocidolite, amosite and chrysotile consumption were not recorded separately for most countries. Lin et al.13 reported a strong international correlation between the logarithm of recent mesothelioma mortality and historical asbestos consumption, which was predominantly chrysotile even in Britain. The exponential dose-response this would imply is interpreted as evidence of the mesothelioma risk from chrysotile,14 but the apparent correlation merely reflects two separate clusters of countries. There is little correlation either among the countries of North America, Australasia, Western Europe and Japan (the only outlier being Portugal) or in Eastern Europe, South America and the rest of Asia, where registered mesothelioma death rates and asbestos imports in the 1960s also varied widely but were much lower.13 This is confirmed in an updated analysis restricted to European countries.15 Replacement of chrysotile by safer substitutes is justified by the lung cancer and asbestosis risks, and the likelihood of some mesothelioma risk strengthens the case; but population-based data on amphibole lung burdens as well as total asbestos imports will be needed to identify any countries in which a large proportion of mesotheliomas were caused by chrysotile. Conclusion The British mesothelioma death rate will decline from the current peak (0.75% of male deaths and 0.13% of female deaths in 2015) until about 2055, when those born before 1965 will be aged over 90.16 If the average lung burden by age 30 from environmental asbestos exposure is now ∼1 f/mg and remains at that level, there will be a continuing lifetime mesothelioma risk of the order of 1 in 10 000, averaged across the whole population. With projected population growth and ageing over the next 40 years, this would imply almost 100 mesotheliomas per year caused by asbestos, and there may be a similar number unrelated to asbestos.17 The risk is an order of magnitude higher in a subgroup of plumbers, electricians, decorators and presumably asbestos removal workers who do not take adequate precautions and probably in a minority of the general population with unusually high environmental exposure. Further samples from young people are needed to estimate current average lung burdens at each age more precisely. This would indicate whether the environmental hazard is declining and whether exposure is predominantly before or after school-leaving age. Our results suggest that a minority of the general population may have unusually high environmental exposure, but more sensitive fibre counting will be needed to confirm this. We are now recruiting further young pneumothorax cases, to identify those with high lung burdens so that their schools and homes can be studied. Funding This work was supported by the British Lung Foundation (grant no. APG11-3), Health and Safety Executive and Cancer Research UK. A.G.N. was supported by the National Institute of Health Research Respiratory Disease Biomedical Research Unit at the Royal Brompton and Harefield NHS Foundation Trust and Imperial College London. Acknowledgements We thank Jayne Simpson and Maria Kelly for the majority of the telephone interviews; Jane Hatch, Samantha Warnakulasuriya, Laura Newmark, Bavithra Vijayakumar, Chris Heaton, Qichen Liu, Katie Field and Emma Afify for administrative work on the study; and Richard Houlihan, Catherine Taylor, Kirsty Dewberry and James Staff at the Health and Safety Laboratory for TEM counting. Most importantly, we thank the mesothelioma, lung cancer and pneumothorax patients who agreed to participate. Author Contributions C.G. performed the majority of the statistical analyses and contributed to the design of the study, interpretation of data and writing the report. C.R. contributed to the design and supervised the conduct of the study. J.H. fitted the maximum likelihood models and contributed to interpretation of data and writing the report. A.D. contributed to interpretation of data and writing the report. G.B. supervised the TEM and contributed to study conduct and writing the report. J.P.W. and M.S. contributed to the conduct of the study and data collection. A.N., L.D. and M.S. provided advice on lung pathology and contributed to study design and conduct. T.T. contributed to study design and conduct. J.P. designed and supervised the study and the statistical analysis. J.P. and C.G. drafted the initial manuscript with major input from J.H., A.D. and G.B. All authors have contributed to revision of the manuscript and approved the final version. The TIPS collaborators provided access to study participants and their lung samples. TIPS collaborators: Andrew Duncan, Blackpool Victoria Hospital, Blackpool, Michael Dusmet, Royal Brompton Hospital, London, John G Edwards, Northern General Hospital, Sheffield, Eric Lim, Harefield Hospital, Middlesex, Richard Milton, St James’s University Hospital, Leeds, Ian Morgan, Royal Wolverhampton Hospitals NHS Trust, Wolverhampton, P O’Keefe, University Hospital of Wales, Cardiff, Danielle Power, Charing Cross Hospital, London, P B Rajesh and Sridhar Rathinam, Birmingham Heartlands Hospital, Birmingham, Doris M Rassl, Papworth Hospital NHS Foundation Trust, Cambridge, T Routledge, Guy’s Hospital, London, Michael Shackcloth, Liverpool Heart and Chest Hospital NHS Trust, Liverpool and Anthony De Soyza, Freeman Hospital, Newcastle upon Tyne, UK. Conflict of interest: None declared. References 1 Health and Safety Executive . Tables MESO02 and MESO03: Death Certificates Mentioning Mesothelioma, 1968-2015. 2017 . http://www.hse.gov.uk/statistics/tables/ (26 July 2017, date last accessed). 2 Rake C , Gilham C, Hatch J, Darnton A, Hodgson J, Peto J. Occupational, domestic and environmental mesothelioma risks in the British population: a case-control study . Br J Cancer 2009 ; 100 : 1175 – 83 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Gilham C , Rake C, Burdett G et al. Pleural mesothelioma and lung cancer risks in relation to occupational history and asbestos lung burden . Occup Environ Med 2016 ; 73 : 290 – 99 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Hodgson JT , Darnton A. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure . Ann Occup Hyg 2000 ; 44 : 56 – 601 . Google Scholar Crossref Search ADS WorldCat 5 Hodgson JT , Darnton A. Mesothelioma risk from chrysotile . Occup Environ Med 2010 ; 67 : 432 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Berman DW , Crump KS. A meta-analysis of asbestos-related cancer risk that addresses fiber size and mineral type . Crit Rev Toxicol 2008 ; 38(Suppl 1) : 49 – 73 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Peto J , Rake C, Gilham C, Hatch J. Occupational, Domestic and Environmental Mesothelioma Risks in Britain: A Case-Control Study. 2009 . http://www.hse.gov.uk/research/rrhtm/rr696.htm (5 February 2018, date last accessed). 8 Burdett GJ , Jaffrey SA, Rood AP. Airborne asbestos fibre levels in buildings: a summary of UK measurements . IARC Scientific Publications 1989 ; 90: 277 – 90 . Google Scholar OpenURL Placeholder Text WorldCat 9 Peto J , Seidman H, Selikoff IJ. Mesothelioma mortality in asbestos workers: implications for models of carcinogenesis and risk assessment . Br J Cancer 1982 ; 45 : 124 – 35 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Bense L , Eklund G, Wiman LG. Smoking and the increased risk of contracting spontaneous pneumothorax . Chest 1987 ; 92 : 1009 – 12 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Bintcliffe O , Maskell N. Spontaneous pneumothorax . BMJ 2014 ; 348 : g2928 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Peto J , Gilham C, Rake C, Darnton A, Hodgson J, Burdett G. Authors' reply to letters from Egilman et al and Oliver et al . Occup Environ Med 2016 ; 73 : 710 – 11 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Lin RT , Takahashi K, Karjalainen A et al. Ecological association between asbestos-related diseases and historical asbestos consumption: an international analysis . Lancet 2007 ; 369 : 844 – 49 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Wagner GR. The fallout from asbestos . Lancet 2007 ; 369 : 973 – 74 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Darnton A , Gilham C, Peto J. Epidemiology of mesothelioma in Europe. In: Mineo T (ed). Malignant Pleural Mesothelioma: Present Status and Future Directions . Sharjah, United Arab Emirates : Bentham Science , 2016 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 16 Hodgson JT , McElvenny DM, Darnton AJ, Price MJ, Peto J. The expected burden of mesothelioma mortality in Great Britain from 2002 to 2050 . Br J Cancer 2005 ; 92 : 587 – 93 . Google Scholar Crossref Search ADS PubMed WorldCat 17 McDonald JC. Health implications of environmental exposure to asbestos . Environ Health Perspect 1985 ; 62 : 319 – 28 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Pharoah PD , Antoniou A, Bobrow M, Zimmern RL, Easton DF, Ponder BA. Polygenic susceptibility to breast cancer and implications for prevention . Nat Genet 2002 ; 31 : 33 – 36 . Google Scholar Crossref Search ADS PubMed WorldCat Statistical Appendix Estimation of mean population lung burden adjusted for the effect of asbestos on lung cancer risk Pharoah et al.18 considered a lognormal risk factor x where log(x) ∼ N(µ,σ2) in the general population and exposure-response is linear. They were modelling susceptibility to breast cancer, but as the log of asbestos lung burden is approximately normally distributed in the general population (see Figure 1) and we assume that the increase in lung cancer relative risk is proportional to lung burden,3 we can apply their results to our data. The distribution of x (i.e. lung burden) among cancers caused by asbestos will also be lognormal with log(x) ∼ N(µ+σ2, σ2). The arithmetic mean lung burden d in the general population equals exp(µ+σ2/2), so µ equals log(d)-σ2/2. Among the proportion p of lung cancers that are caused by asbestos log(x) ∼ N(µ+σ2,σ2), and the mean lung burden is d.exp(σ2). We assume that log(x) ∼ N(µ,σ2) among pneumothorax patients and among the proportion (1-p) of lung cancers that are not caused by asbestos. The lung cancer relative risk is 1+k.d. The estimate of k from our case-control analysis (k = 0.00255 per f/mg) was used.3 The lognormal variance σ2 may vary between birth cohorts, sexes and occupational groups, but is poorly determined in smaller individual cells. Accordingly, the cells in Tables 1 and 2 (and corresponding points in Figures 2a and b) were grouped, and the cells in each group modelled jointly with common group variance σ2 but different means µ(i), where i indexes the cells in the group. The groups were chosen such that the change in overall fit between fitting a different σ in each cell and fitting a common σ across cells was comfortably non-significant (P > 0.3). This resulted in three groups for Table 1 (men born 1940–44; men born later; and women) and five groups for Table 2 (other construction workers 1940–54; all other male occupations 1940–54; all male occupations combined 1940–92; environmental exposure, men and women combined; and all other cells). In the ith cell in a group: log(lung burden) ∼ N(µ(i),σ2) in the general population and in pneumothorax patients, d(i) = exp(µ(i)+σ2/2) = average lung burden in the general population and in pneumothorax patients, p(i) = k.d(i)/(1+k.d(i)) = proportion of lung cancers due to asbestos, and D(i) = p(i). exp(σ2).d(i) + (1-p(i)).d(i) = average lung burden in lung cancers. For the jth individual in cell i, the true lung burden is X(i,j) fibre/mg and n(i,j) fibres are counted in w(i,j) mg of lung tissue, so n(i,j) ∼ P(n(i,j),[w(i,j).X(i,j)]) where P(n,λ) is the Poisson probability of observing n events with expected number λ. Thus likelihood = ∏∏L(i,j), where for each lung cancer in cell i: L(i,j)=integral from x=0 to infinity of P(n(i,j),w(i,j).x).[p(i).g(x,µ(i)+σ2,σ2)+(1−p(i)).g(x,µ(i),σ2)].x−1.d(x)[Eqn 1] and for each pneumothorax patient in cell i: L(i,j)=integral from x=0 to infinity of P(n(i,j), w(i,j).x).g(x,µ(i),σ2).x−1.d(x)[Eqn 2] where g(x,µ,σ2) is the lognormal function (σ√2π)−1 exp-[(log(x)-µ)2/2σ2]. Replacing µ(i) by log(d(i))-σ2/2 in the likelihood gives maximum likelihood (ML) estimates of the population mean lung burden d(i) and its confidence interval in cell i. The ‘mle2’ function in the statistical package R was used to derive ML estimates, using package ‘poilog’ to provide the Poisson-lognormal likelihood. Confidence intervals were derived from the likelihood profile for each estimate. The likelihood shown in Equation 2 was used for lung cancers as well as pneumothorax patients, to calculate the unadjusted mean lung burdens in footnote a of Table 1, Figure 3 and Table 3. In Table 3, average lung burdens were calculated by fitting µ and σ separately in each cell. Estimating the relationship between national mesothelioma mortality and population mean lung burden The slope b in the relationship M(i) = b.d(i) between average lung burden d(i) and cumulative mesothelioma risk by age 50 M(i) was estimated for each sex by maximizing the Poisson likelihood of the m(i) observed deaths in Britain in birth cohort i, given the population Pop(i) = m(i)/M(i), and the distribution of adjusted estimates of lung burden d(i) implied by the likelihood profiles for the five birth cohorts from 1940–44 to 1960–64 (Table 1). For each sex, each of 5000 replicate estimates of b was derived by drawing values of d(i) for each birth cohort at random from the corresponding likelihood profile and fitting a Poisson regression with offset log(d(i).Pop(i)) to estimate the intercept log(b), so b is estimated by exp(intercept). The mean and 2.5% and 97.5% quantiles of these 5000 replicate estimates give (for each sex) the central estimate and 95% confidence limits for the risk coefficient b linking mesothelioma with average lung burden. The ratio of predicted lifetime risk (the actuarial probability of dying of mesothelioma by age 90) to the observed cumulative mortality to age 50 was estimated by simple age and birth cohort analysis of British male mesothelioma death rates from 1990 to 2009, assuming current British mortality rates for all other causes of death. The statistical programming code is available on request. © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. © The Author(s) 2018. Published by Oxford University Press on behalf of the International Epidemiological Association.

Commentary: Past and current asbestos exposure and future mesothelioma risks in Britain

Terracini,, Benedetto

2018 International Journal of Epidemiology

doi: 10.1093/ije/dyy175pmid: 30169798

The recent paper by Gilham et al.1 quantifies the effectiveness of the asbestos regulations implemented in the UK in 1969 and which were followed shortly afterwards by the abandonment of the use of crocidolite and, more than 15 years later, by an official ban on crocidolite and amosite in that country. Both the reduction in mesothelioma incidence and mortality up to age 50 and the fall in the amphibole burden in the lungs in a representative sample of the British population are impressive. Had other countries adopted similar regulations 50 years ago, many asbestos-related deaths would have been avoided worldwide. For instance, this is significant for Italy, where the lack of adequate preventive measures (at the time) led to a delay of more than a decade in the decline in asbestos consumption compared with the UK.2 Given the relatively short half-life of chrysotile in the lung, measuring the asbestos burden in the lung parenchyma means that chrysotile is not included in any estimate of past exposures and current consequences. Such limitations are acknowledged by Gilham et al.1 as well as the fact that, given the method used, fibres shorter than five microns were beyond the limits of detection. Worldwide, past and present chrysotile consumption has been enormous. Over the decades preceding the total ban in 1999, a total of more than 1 million tons of chrysotile entered the UK.3 Most of it was used in industry and construction work. Nowadays, chrysotile is the major (and possibly exclusive) type of asbestos in use worldwide: its mining, processing and trade are still permitted in countries where a total ban has not yet been implemented. Russia, Kazakistan, China, India, Indonesia and Brazil contribute the major part of production and/or consumption of the 1.5 million tons of chrysotile mined yearly worldwide. Gilham et al.1 address the issue of chrysotile in the discussion of their findings. Notably, they use different terms to describe the credibility of the association of chrysotile with lung cancer and asbestosis on one hand and with mesotheliomas on the other. In their words, the risk is only ‘likely’ for mesothelioma. By saying this, they are disagreeing with the International Agency for Research on Cancer (IARC), which, several years ago,4 included the serous membranes among the target organs for which evidence of the carcinogenicity of chrysotile in humans is convincing. This evidence has been strengthened by additional occupational cohort studies in different countries, published in recent years.5,6 The recurrent assertion by investigators associated with the industry,7 that the cause of mesothelioma in workers exposed to chrysotile is not chrysotile itself but amphibole contaminants, is not based on convincing evidence. In terms of carcinogenic potency (i.e. risk from unit of intensity and duration of exposure), the risk of mesothelioma is lower for chrysotile than for amphiboles, but still increased. Potency differences between different types of asbestos for lung cancer are more difficult to estimate.4 Whereas the observations by Gilham et al. regarding the UK1 are most interesting, the possible international impact of their paper is of concern. In the discussion, the authors warn against the identification of ‘any country in which a large proportion of mesotheliomas were caused by chrysotile’ in the absence of population-based data on amphibole lung burdens as well as total asbestos imports. In countries where chrysotile is still used and the debate about a possible ban is still on or has not even started, these words, together with the authors’ scientific prestige, might be interpreted as the request for locally produced additional scientific findings before any decision about banning chrysotile is taken. This would result in continued and widespread use of a recognized carcinogen, perpetuating a disturbing international double standard in the prevention of environmentally induced disease. As for sophisticated ‘local’ studies, it should be kept in mind that in many countries where chrysotile is still used, current statistics, including mortality statistics, when made accessible are unsatisfactory8 and cancer registration limited. For example, in Latin America, few estimates of the hazards of asbestos have been produced.8 In Brazil, a recent comparison between mortality statistics and hospital registers has shown that the former miss a substantial number of mesothelioma cases.9 In a commentary on a previous paper by the same group, the authors suggested that measuring lung asbestos burden should be a ‘new standard’ in mesothelioma epidemiology.10 Although the results of such an exercise might be of interest, this requirement could be excessive. At an individual level, it is commonly accepted that a history of significant occupational and/or environmental exposure to asbestos is regarded as sufficient evidence to attribute it as a cause of mesothelioma. Furthermore, in population-based mesothelioma registries such as the Italian ReNaM, the systematic evaluation of sources of exposure has been used in order to estimate proportional attributable risks in the population. In countries where the use of asbestos types other than chrysotile has been negligible, such as Brazil, there are no reasons for rejecting the idea that population-based attributable risks could be more efficiently (and at less cost) estimated through recall of exposure, provided this is information is collected systematically. In conclusion, population-based estimates of asbestos burden in the lung parenchyma provide valuable data. Although expensive to carry out, they provide corroborating estimates of the effectiveness of preventive measures in countries where amphiboles were largely used in the past. However, they are not an essential prerequisite for estimating the impact of chrysotile in countries where amphibole use has been negligible. Conflict of interest: B.T.has acted as an expert witness for the Public Prosecution or for the Judge in Italian criminal trials involving the liability of industrial managers for asbestos-related cancers among their employees. B.T. was also an expert witness for Region Piedmont (plaintiff) in the Eternit criminal trial held in Turin in 2012. References 1 Gilham C , Rake C , Hodgson J et al. Past and current asbestos exposure and future mesothelioma risks in Britain: The Inhaled Particles Study (TIPS) . Int J Epidemiol 2018 ; 47 : 1745 – 56 . 2 Marinaccio A , Binazzi A , Marzio DD et al. Pleural malignant mesothelioma epidemic: incidence, modalities of asbestos exposure and occupations involved from the Italian National Register . Int J Cancer 2012 ; 130 : 2146 – 54 . Google Scholar Crossref Search ADS PubMed 3 Virta RL. Worldwide Asbestos Supply and Consumption Trends from 1900 to 2000 . Denver, CO : U.S. Department of the Interior, U.S. Geological Survey , 2003 . 4 International Agency for Research on Cancer . IARC Monographs on the Evaluation of Carcinogenic Risks of Chemicals to Man . Vol. 100C. Lyon, France : International Agency for Research on Cancer , 2012 . 5 Jiang Z , Chen T , Chen J et al. Hand-spinning chrysotile exposure and risk of malignant mesothelioma: A case-control study in Southeastern China . Int J Cancer 2018 ; 142 : 514 – 23 . Google Scholar Crossref Search ADS PubMed 6 Pira E , Romano C , Donato F et al. Mortality from cancer and other causes among Italian chrysotile asbestos miners . Occup Environ Med 2017 ; 74 : 558 – 63 . Google Scholar Crossref Search ADS PubMed 7 Attanoos RL , Churg A , Galateau-Salle F , Gibbs AR , Roggli VL. Malignant mesothelioma and its non-asbestos causes . Arch Pathol Lab Med 2018 ; 142 : 753 – 60 . Google Scholar Crossref Search ADS PubMed 8 Marsili D , Comba P , Pasetto R , Terracini B. International scientific cooperation on asbestos-related disease prevention in Latin America . Ann Glob Health 2014 ; 80 : 247 – 50 . Google Scholar Crossref Search ADS PubMed 9 Santana VS , Algranti E , Campos F et al. Recovering missing mesothelioma deaths in death certificates using hospital records . Am J Ind Med 2018 ; 61 : 547 – 65 . Google Scholar Crossref Search ADS PubMed 10 Boffetta P , La Vecchia C. Setting new standards for epidemiological research on mesothelioma . Occup Environ Med 2016 ; 73 : 289 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)