EUVIMEDSkopri EUVIMED
Open access · CC BY 4.0

Propensity score matching after multiple imputation when a confounder has missing data

Corentin Ségalas iD, Clémence Leyrat iD, James R Carpenter iD, Elizabeth Williamson iD

Statistics in Medicine · 2023-03-30 · 42 (7) · 1082-1095

DOI: 10.1002/sim.9658 ↗

At a glance

Journal article in Statistics in Medicine, published online on 2023-01-25. Written by Corentin Ségalas and 3 co-authors. Institution: Department of Medical Statistics, London School of Hygiene and Tropical Medicine, London, UK. Major MEDLINE topics: Propensity Score. Cited 14 times so far (Crossref); 45 references. The full text is freely available (open access, CC BY 4.0). Funded by: Medical Research Council, Medical Research Council Canada.

AstrattSubjectsDettalji biblijografiċiCiteRelated studies

Abstract (original language)

One of the main challenges when using observational data for causal inference is the presence of confounding. A classic approach to account for confounding is the use of propensity score techniques that provide consistent estimators of the causal treatment effect under four common identifiability assumptions for causal effects, including that of no unmeasured confounding. Propensity score matching is a very popular approach which, in its simplest form, involves matching each treated patient to an untreated patient with a similar estimated propensity score, that is, probability of receiving the treatment. The treatment effect can then be estimated by comparing treated and untreated patients within the matched dataset. When missing data arises, a popular approach is to apply multiple imputation to handle the missingness. The combination of propensity score matching and multiple imputation is increasingly applied in practice. However, in this article we demonstrate that combining multiple imputation and propensity score matching can lead to over‐coverage of the confidence interval for the treatment effect estimate. We explore the cause of this over‐coverage and we evaluate, in this context, the performance of a correction to Rubin's rules for multiple imputation proposed by finding that this correction removes the over‐coverage.

Subjects

MeSH terms (MEDLINE)

Propensity ScoreHumansData Interpretation, StatisticalCausality

Author keywords

confoundingmissing datamultiple imputationpropensity score matching

Authors and institutions

  • Department of Medical Statistics, London School of Hygiene and Tropical Medicine, London, UK.
  • Université Paris-Cité, Centre of Epidemiology and Statistics (CRESS) Inserm, Paris, France.
  • MRC Clinical Trials Unit at UCL, UCL, London, UK.

ORCID profiles: Corentin Ségalas, Clémence Leyrat, James R Carpenter, Elizabeth Williamson

Funding

  • Medical Research Council
  • Medical Research Council Canada

Cite

APA

Ségalas, C., Leyrat, C., Carpenter, J. R., & Williamson, E. (2023). Propensity score matching after multiple imputation when a confounder has missing data. Statistics in Medicine, 42(7), 1082-1095. https://doi.org/10.1002/sim.9658

Vancouver

Ségalas C, Leyrat C, Carpenter JR, Williamson E. Propensity score matching after multiple imputation when a confounder has missing data. Statistics in Medicine. 2023;42(7):1082-1095. doi:10.1002/sim.9658

BibTeX
@article{Sgalas2023,
  author = {Ségalas, Corentin and Leyrat, Clémence and Carpenter, James R and Williamson, Elizabeth},
  title = {Propensity score matching after multiple imputation when a confounder has missing data},
  journal = {Statistics in Medicine},
  year = {2023},
  volume = {42},
  number = {7},
  pages = {1082--1095},
  doi = {10.1002/sim.9658}
}
Start for free →