Propensity Score Methods in Rhinologic Observational Research: Principles, Applications, Extensions, and Practical Considerations.

Randomized controlled trials (RCTs) remain the gold standard for evaluating treatment efficacy because randomization minimizes measured and unmeasured confounding. In rhinology, however, RCTs are often impractical because of ethical constraints, patient preferences, financial limitations, and logistical challenges related to surgical interventions or long-term follow-up. Consequently, observational studies based on real-world data have become increasingly important for comparative effectiveness research. Despite their practical advantages, observational studies are inherently vulnerable to confounding by indication, selection bias, prevalent-user bias, and immortal time bias. Propensity score (PS) methods were developed to mitigate measured confounding and improve the internal validity of nonrandomized comparisons. Since their introduction, PS approaches have expanded beyond traditional matching to include advanced weighting strategies, such as inverse probability of treatment weighting, overlap weighting, and doubly robust estimation. This review provides a clinician-oriented overview of PS methods, including their conceptual foundations, practical implementation, study design considerations, target population implications, diagnostic tools, strengths, limitations, and modern methodological extensions. The review emphasizes the application of these methods to rhinologic research, in which treatment allocation is frequently influenced by disease severity and patient characteristics.