Be Best Practice Guidelines for Propensity Score Methods in Medical Research: Considerations on Theory, Implementation, and Reporting. A Review
Authors: Chen JW, Maldonado DR, Kowalski BL, Miecznikowski KB, Kyin C, Gornbein JA, Domb BG
DOI: 10.1016/j.arthro.2021.06.037
Background
Propensity score matching (PSM) is widely used to reduce confounding in retrospective studies where randomization is not feasible. This review provides guidance on PSM theory, correct application, and transparent reporting.
Methods
The authors present a conceptual overview of PSM, contrast it with traditional regression techniques, and offer step-by-step examples to demonstrate its practical use in orthopaedic and medical research.
Key Findings
- PSM improves the internal validity of observational studies by balancing covariates across treatment groups.
- The review highlights common pitfalls in implementation, such as poor variable selection, inadequate matching ratios, and lack of balance diagnostics.
- Proper reporting, including pre-matching diagnostics and post-matching covariate balance, is essential for study reproducibility.
Conclusions
Effective application of PSM can enhance the credibility of nonrandomized studies. Researchers must understand its theoretical underpinnings and adhere to rigorous implementation standards.
What Does This Mean for Providers?
- For clinicians involved in research, PSM is a powerful tool to control confounding in retrospective cohort studies.
- Proper use requires thoughtful selection of covariates, balance diagnostics, and transparent methodology reporting to ensure scientific validity.
- Familiarity with PSM strengthens your ability to critically appraise and produce methodologically sound outcomes research.