Ghassan B Hamra1,2, Catherine R Lesko1, Jessie P Buckley1,2, Elizabeth T Jensen3, Daniel Tancredi4, Bryan Lau1, Irva Hertz-Picciotto5. 1. From the Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD. 2. Department of Environmental Sciences and Engineering, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD. 3. Department of Epidemiology and Prevention, Wake Forest School of Medicine, Winston Salem, NC. 4. Department of Pediatrics, University of California, Davis School of Medicine, Sacramento, CA. 5. Division of Environmental and Occupational Health, University of California, Davis School of Medicine, Davis, CA.
Abstract
BACKGROUND: Collaborative research often combines findings across multiple, independent studies via meta-analysis. Ideally, all study estimates that contribute to the meta-analysis will be equally unbiased. Many meta-analyses require all studies to measure the same covariates. We explored whether differing minimally sufficient sets of confounders identified by a directed acyclic graph (DAG) ensures comparability of individual study estimates. Our analysis applied four statistical estimators to multiple minimally sufficient adjustment sets identified in a single DAG. METHODS: We compared estimates obtained via linear, log-binomial, and logistic regression and inverse probability weighting, and data were simulated based on a previously published DAG. RESULTS: Our results show that linear, log-binomial, and inverse probability weighting estimators generally provide the same estimate of effect for different estimands that are equally sufficient to adjust confounding bias, with modest differences in random error. In contrast, logistic regression often performed poorly, with notable differences in effect estimates obtained from unique minimally sufficient adjustment sets, and larger standard errors than other estimators. CONCLUSIONS: Our findings do not support the reliance of collaborative research on logistic regression results for meta-analyses. Use of DAGs to identify potentially differing minimally sufficient adjustment sets can allow meta-analyses without requiring the exact same covariates.
BACKGROUND: Collaborative research often combines findings across multiple, independent studies via meta-analysis. Ideally, all study estimates that contribute to the meta-analysis will be equally unbiased. Many meta-analyses require all studies to measure the same covariates. We explored whether differing minimally sufficient sets of confounders identified by a directed acyclic graph (DAG) ensures comparability of individual study estimates. Our analysis applied four statistical estimators to multiple minimally sufficient adjustment sets identified in a single DAG. METHODS: We compared estimates obtained via linear, log-binomial, and logistic regression and inverse probability weighting, and data were simulated based on a previously published DAG. RESULTS: Our results show that linear, log-binomial, and inverse probability weighting estimators generally provide the same estimate of effect for different estimands that are equally sufficient to adjust confounding bias, with modest differences in random error. In contrast, logistic regression often performed poorly, with notable differences in effect estimates obtained from unique minimally sufficient adjustment sets, and larger standard errors than other estimators. CONCLUSIONS: Our findings do not support the reliance of collaborative research on logistic regression results for meta-analyses. Use of DAGs to identify potentially differing minimally sufficient adjustment sets can allow meta-analyses without requiring the exact same covariates.
Authors: Catherine R Lesko; Lisa P Jacobson; Keri N Althoff; Alison G Abraham; Stephen J Gange; Richard D Moore; Sharada Modur; Bryan Lau Journal: Int J Epidemiol Date: 2018-04-01 Impact factor: 7.196
Authors: Ellis Voerman; Susana Santos; Bernadeta Patro Golab; Pilar Amiano; Ferran Ballester; Henrique Barros; Anna Bergström; Marie-Aline Charles; Leda Chatzi; Cécile Chevrier; George P Chrousos; Eva Corpeleijn; Nathalie Costet; Sarah Crozier; Graham Devereux; Merete Eggesbø; Sandra Ekström; Maria Pia Fantini; Sara Farchi; Francesco Forastiere; Vagelis Georgiu; Keith M Godfrey; Davide Gori; Veit Grote; Wojciech Hanke; Irva Hertz-Picciotto; Barbara Heude; Daniel Hryhorczuk; Rae-Chi Huang; Hazel Inskip; Nina Iszatt; Anne M Karvonen; Louise C Kenny; Berthold Koletzko; Leanne K Küpers; Hanna Lagström; Irina Lehmann; Per Magnus; Renata Majewska; Johanna Mäkelä; Yannis Manios; Fionnuala M McAuliffe; Sheila W McDonald; John Mehegan; Monique Mommers; Camilla S Morgen; Trevor A Mori; George Moschonis; Deirdre Murray; Carol Ní Chaoimh; Ellen A Nohr; Anne-Marie Nybo Andersen; Emily Oken; Adriëtte J J M Oostvogels; Agnieszka Pac; Eleni Papadopoulou; Juha Pekkanen; Costanza Pizzi; Kinga Polanska; Daniela Porta; Lorenzo Richiardi; Sheryl L Rifas-Shiman; Luca Ronfani; Ana C Santos; Marie Standl; Camilla Stoltenberg; Elisabeth Thiering; Carel Thijs; Maties Torrent; Suzanne C Tough; Tomas Trnovec; Steve Turner; Lenie van Rossem; Andrea von Berg; Martine Vrijheid; Tanja G M Vrijkotte; Jane West; Alet Wijga; John Wright; Oleksandr Zvinchuk; Thorkild I A Sørensen; Debbie A Lawlor; Romy Gaillard; Vincent W V Jaddoe Journal: PLoS Med Date: 2019-02-11 Impact factor: 11.069