Most longitudinal studies do not address potential selection biases due to selective attrition. Using empirical data and simulating additional attrition, we investigated the effectiveness of common approaches to handle missing outcome data from attrition in the association between individual education level and change in body mass index (BMI).
Using data from the two waves of the French RECORD Cohort Study (N = 7,172), we first examined how inverse probability weighting (IPW) and multiple imputation handled missing outcome data from attrition in the observed data (stage 1). Second, simulating additional missing data in BMI at follow-up under various missing-at-random scenarios, we quantified the impact of attrition and assessed how multiple imputation performed compared to complete case analysis and to a perfectly specified IPW model as a gold standard (stage 2).
With the observed data in stage 1, we found an inverse association between individual education and change in BMI, with complete case analysis, as well as with IPW and multiple imputation. When we simulated additional attrition under a missing-at-random pattern (stage 2), the bias increased with the magnitude of selective attrition, and multiple imputation was useless to address it.
Our simulations revealed that selective attrition in the outcome heavily biased the association of interest. The present article contributes to raising awareness that for missing outcome data, multiple imputation does not do better than complete case analysis. More effort is thus needed during the design phase to understand attrition mechanisms by collecting information on the reasons for dropout.
From the aSorbonne Universités, UPMC Univ Paris 06, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Paris, France; bInserm, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Paris, France; cEHESP School of Public Health, Rennes, France; dDepartment of Family Medicine and Public Health & Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA; and eCentre d’Investigations Préventives et Cliniques, Paris, France.
Submitted January 15, 2015; accepted September 14, 2017.
Supported by a doctoral grant of Région Île-de-France attributed to Antoine Lewin. The RECORD study is funded by the Institute for Public Health Research (IReSP, Institut de Recherche en Santé Publique); the National Institute for Prevention and Health Education (INPES, Institut National de Prévention et d’Education pour la Santé) (Prevention Program 2007; 2010–2011 financial support; 2011–2013 financial support; 2012–2014 financial support); the National Institute of Public Health Surveillance (InVS, Institut de Veille Sanitaire) (Territory and Health Program); the French Ministries of Research and Health (Epidemiologic Cohorts Grant 2008); the National Health Insurance Office for Salaried Workers (CNAM-TS, Caisse Nationale d’Assurance Maladie des Travailleurs Salariés); the Ile-de-France Regional Health Agency (ARS, Agence Régionale de Santé); the Ile-de-France Regional Council (Conseil Régional d’Île-de-France, DIM SEnT and CODDIM); the National Research Agency (ANR, Agence Nationale de la Recherche) (Health–Environment Program 2005); the City of Paris (Ville de Paris); and the Ile-de-France Youth, Sports, and Social Cohesion Regional Direction (DRJSCS, Direction Régionale de la Jeunesse, des Sports et de la Cohésion Sociale).
The authors report no conflicts of interest.
Supplemental digital content is available through direct URL citations in the HTML and PDF versions of this article (www.epidem.com).
Correspondence: Antoine Lewin, Inserm U1136, Faculté de Médecine Saint-Antoine, 27 rue Chaligny, 75012 Paris, France. E-mail: firstname.lastname@example.org.