Modeling multi-level survival data in multi-center epidemiological cohort studies: Applications from the ELAPSE project.

de Hoogh K

Samoli E(1), Rodopoulou S(1), Hvidtfeldt UA(2), Wolf K(3), Stafoggia M(4), Brunekreef B(5), Strak M(6), Chen J(5), Andersen ZJ(7), Atkinson R(8), Bauwelinck M(9), Bellander T(10), Brandt J(11), Cesaroni G(12), Forastiere F(13), Fecht D(14), Gulliver J(15), Hertel O(11), Hoffmann B(16), de Hoogh K(17), Janssen NAH(5), Ketzel M(18), Klompmaker JO(6), Liu S(7), Ljungman P(19), Nagel G(20), Oftedal B(21), Pershagen G(10), Peters A(3), Raaschou-Nielsen O(22), Renzi M(12), Kristoffersen DT(23), Severi G(12), Sigsgaard T(24), Vienneau D(17), Weinmayr G(20), Hoek G(5), Katsouyanni K(25).
Author information:
(1)Dept. of Hygiene, Epidemiology and Medical Statistics, Medical School, National and Kapodistrian University of Athens, 75 Mikras Asias Str, 115 27 Athens, Greece.
(2)Danish Cancer Society Research Centre, Copenhagen, Denmark.
(3)Institute of Epidemiology, Helmholtz Zentrum München, Neuherberg, Germany.
(4)Department of Epidemiology, Lazio Region Health Service ASL Roma 1, Rome, Italy; Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden.
(5)Institute for Risk Assessment Sciences
(IRAS), Utrecht University, Postbus 80125, 3508 TC Utrecht, the Netherlands.
(6)Institute for Risk Assessment Sciences
(IRAS), Utrecht University, Postbus 80125, 3508 TC Utrecht, the Netherlands; National Institute for Public Health and the Environment
(RIVM), Antonie van Leeuwenhoeklaan 9, 3721 MA Bilthoven, the Netherlands.
(7)University of Copenhagen, Department of Public Health, Section of Environmental Health, Øster Farimagsgade 5, 1014, Copenhagen, Denmark.
(8)Population Health Research Institute, St George's, University of London, Cranmer Terrace, London SW17 0RE, UK.
(9)Interface Demography, Department of Sociology, Vrije Universiteit Brussel, Brussels, Belgium.
(10)Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden; Centre for Occupational and Environmental Medicine, Region Stockholm, Stockholm, Sweden.
(11)Department of Environmental Science, Aarhus University, Frederiksborgvej 399, Roskilde, Denmark.
(12)Department of Epidemiology, Lazio Region Health Service ASL Roma 1, Rome, Italy.
(13)NIHR HPRU Health Impact of Environmental Hazards, Environmental Research Group, Analytical, Environmental & Forensic Sciences, King's College London, UK.
(14)Small Area Health Statistics Unit, MRC Centre for Environment and Health, School of Public Health, Imperial College London, Norfolk Place, London W2 1PG, UK.
(15)Centre for Environmental Health and Sustainability & School of Geography, Geology and the Environment, University of Leicester, Leicester, UK.
(16)Institute for Occupational, Social and Environmental Medicine, Medical Faculty, Heinrich-Heine-University of Düsseldorf, Germany.
(17)Swiss Tropical and Public Health Institute, Basel, Switzerland; University of Basel, Basel, Switzerland.
(18)Department of Environmental Science, Aarhus University, Frederiksborgvej 399, Roskilde, Denmark; Global Centre for Clean Air Research
(GCARE), University of Surrey, Guildford GU2 7XH, UK.
(19)Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden; Department of Cardiology, Danderyd University Hospital, Stockholm, Sweden.
(20)Institute of Epidemiology and Medical Biometry, Ulm University, Ulm, Germany.
(21)Department of Environmental Health, Norwegian Institute of Public Health, Oslo, Norway.
(22)Danish Cancer Society Research Centre, Copenhagen, Denmark; Department of Environmental Science, Aarhus University, Frederiksborgvej 399, Roskilde, Denmark.
(23)Cluster for Health Services Research, Norwegian Institute of Public Health, Oslo, Norway.
(24)Department of Public Health, Environment Occupation and Health, Danish Ramazzini Centre, Aarhus University, Aarhus, Denmark.
(25)Dept. of Hygiene, Epidemiology and Medical Statistics, Medical School, National and Kapodistrian University of Athens, 75 Mikras Asias Str, 115 27 Athens, Greece; NIHR HPRU Health Impact of Environmental Hazards, Environmental Research Group, Analytical, Environmental & Forensic Sciences, King's College London, UK.

https://dx.doi.org/10.1016/j.envint.2020.106371

BACKGROUND: We evaluated methods for the analysis of multi-level survival data using a pooled dataset of 14 cohorts participating in the ELAPSE project investigating associations between residential exposure to low levels of air pollution (PM2.5 and NO2) and health (natural-cause mortality and cerebrovascular, coronary and lung cancer incidence). METHODS: We applied five approaches in a multivariable Cox model to account for the first level of clustering corresponding to cohort specification: (1) not accounting for the cohort or using (2) indicator variables, (3) strata, (4) a frailty term in frailty Cox models, (5) a random intercept under a mixed Cox, for cohort identification. We accounted for the second level of clustering due to common characteristics in the residential area by (1) a random intercept per small area or (2) applying variance correction. We assessed the stratified, frailty and mixed Cox approach through simulations under different scenarios for heterogeneity in the underlying hazards and the air pollution effects. RESULTS: Effect estimates were stable under approaches used to adjust for cohort but substantially differed when no adjustment was applied. Further adjustment for the small area grouping increased the effect estimates' standard errors. Simulations confirmed identical results between the stratified and frailty models. In ELAPSE we selected a stratified multivariable Cox model to account for between-cohort heterogeneity without adjustment for small area level, due to the small number of subjects and events in the latter. CONCLUSIONS: Our study supports the need to account for between-cohort heterogeneity in multi-center collaborations using pooled individual level data.