Common Methods for Missing Data in Marginal Structural Models: What Works and Why.

Common Methods for Missing Data in Marginal Structural Models: What Works and Why. Am J Epidemiol. 2020 Oct 15;: Authors: Leyrat C, Carpenter JR, Bailly S, Williamson EJ Abstract Marginal structural models (MSMs) are commonly used to estimate causal intervention effects in longitudinal non-randomised studies. A common challenge when using MSMs to analyse observational studies is incomplete confounder data, where a poorly informed analysis method will lead to biased intervention effect estimates. Despite a number of approaches described in the literature to handle missing data in MSMs, there is little guidance on what works in practice and why. We reviewed existing missing data methods for MSMs and discussed the plausibility of their underlying assumptions. We also performed realistic simulations to quantify the bias of five methods used in practice: complete case analysis, the last observation carried forward, the missingness pattern approach, multiple imputation and inverse-probability-of-missingness weighting. We considered three mechanisms for non-monotone missing data encountered in electronic health record data research. Further illustration of the strengths and limitations of these analysis methods are provided through an application using a cohort of individuals with sleep apnoea, the research database of the French "Observatoire Sommeil de la Fédération de Pneumologie" (OSFP). We recommend a careful consideration of (i) the...
Source: Am J Epidemiol - Category: Epidemiology Authors: Tags: Am J Epidemiol Source Type: research