Comparing multiple imputation methods for systematically missing subject‐level data

• Published in: Research synthesis methods Vol. 8; no. 2; pp. 136 - 148
• Main Authors: Kline, David, Andridge, Rebecca, Kaizar, Eloise
• Format: Journal Article
• Language: English
• Published: England Wiley-Blackwell 01.06.2017; Wiley Subscription Services, Inc
• Subjects: Attenuation; Coefficients; Collection; Comparative Analysis; Correlation analysis; Covariance; Data Collection; Data Interpretation, Statistical; Efficiency; Estimates; longitudinal data; Longitudinal Studies; Mathematical models; Medical Research; Missing data; multiple imputation; Observation; Outcomes of Education; Outcomes of Treatment; Regression (Statistics); Regression analysis; Research Design; Research Reports; research synthesis; Sequential Approach; systematic missing data; Variance; multiple imputation; systematic missing data; research synthesis; longitudinal data
• ISSN: 1759-2879; 1759-2887
• DOI: 10.1002/jrsm.1192

When conducting research synthesis, the collection of studies that will be combined often do not measure the same set of variables, which creates missing data. When the studies to combine are longitudinal, missing data can occur on the observation‐level (time‐varying) or the subject‐level (non‐time‐varying). Traditionally, the focus of missing data methods for longitudinal data has been on missing observation‐level variables. In this paper, we focus on missing subject‐level variables and compare two multiple imputation approaches: a joint modeling approach and a sequential conditional modeling approach. We find the joint modeling approach to be preferable to the sequential conditional approach, except when the covariance structure of the repeated outcome for each individual has homogenous variance and exchangeable correlation. Specifically, the regression coefficient estimates from an analysis incorporating imputed values based on the sequential conditional method are attenuated and less efficient than those from the joint method. Remarkably, the estimates from the sequential conditional method are often less efficient than a complete case analysis, which, in the context of research synthesis, implies that we lose efficiency by combining studies. Copyright © 2015 John Wiley & Sons, Ltd.