Analysis of Observational Health Care Data Using SAS

Chapter 3

Propensity Score Matching for Estimating Treatment Effects

Peter C. AustinMaria ChiuDennis T. KoRon GoereeJack V. Tu

Abstract

3.1 Introduction.

3.2 Estimating the Propensity Score

3.3 Forming Propensity Score Matched Sets

3.4 Assessing Balance in Baseline Characteristics

3.5 Estimating the Treatment Effect

3.6 Sensitivity Analyses for Propensity Score Matching

3.7 Propensity Score Matching Compared with Other Propensity Score Methods

3.8 Case Study

3.9 Summary

Acknowledgments

References

 

Abstract

Propensity score matching entails forming matched sets of treated and untreated subjects who have a similar propensity score value. The most common implementation of propensity score matching is 1:1 or pair matching, in which matched pairs ...