Public Release Type:
Journal
Publication Year: 2022
Affiliation: 1Quantitative Sciences Unit, Division of Biomedical Informatics Research, Department of Medicine, Stanford University School of Medicine, 1701 Page Mill Road, Palo Alto, CA 94304
2Division of Nephrology, Department of Medicine, Stanford University School of Medicine, 1070 Arastradero Road, Palo Alto, CA 94304
Authors:
Montez-Rath ME,
Ling AY,
Kapphahn K,
Desai M
Request IDs:
21913
Studies:
Frequent Hemodialysis Network Daily Trial
When the distribution of treatment effect modifiers differs between the trial sample and target population, inverse probability weighting (IPSW) can be applied to achieve an unbiased estimate of the population average treatment effect in the target population. The statistical validity of IPSW is threatened when there are missing data in the target population as well as in the trial sample. However, missing data methods have not been adequately discussed in the current literature. We conducted a set of simulation studies to determine how to apply multiple imputation (MI) in the context of IPSW. We specifically addressed questions such as which variables to include in the imputation model and whether they should come from trial or non-trial portion of the target population. Based on our findings, we recommend including in the imputation model as main effects all potential effect modifiers and trial indicator for both trial and non-trial population, as well as treatment and outcome variables from trial sample. Additionally, we have illustrated ideas by transporting findings from the Frequent Hemodialysis Network (FHN) Daily Trial to the United States prevalent hemodialysis population in 2017 defined using the United States Renal Data System.