Anesth Pain Med.  2016 Apr;11(2):130-148. 10.17085/apm.2016.11.2.130.

An introduction to propensity score matching methods

• Lee DK

• KMID: 2358475
• DOI: http://doi.org/10.17085/apm.2016.11.2.130

Abstract

Propensity score matching method (PSM) is widely used in observational study to reduce selection bias. Observational study lacks randomization, hence, statistical inferences without bias adjustments usually include observed or unobserved effects of covariates. If a subject with specific characteristics has a higher chance to be selected for a specific treatment, the characteristics have a possible effect on statistical results. PSM is the method for controlling covariate imbalance that produces the selection bias. In this paper, we introduce the basic concepts of PSM and simplified methods of PSM process. However, PSM is a rapidly developing statistical area with many limitations and some disadvantages. These points are described in the concluding section to emphasize the importance of considering the various features of PSM in the study design.

Keyword

Observational study; Propensity score matching method; Selection bias

MeSH Terms

Bias (Epidemiology)
Methods*
Observational Study
Propensity Score*
Random Allocation
Selection Bias

Reference

1. Luellen JK, Shadish WR, Clark MH. Propensity scores: an introduction and experimental test. Eval Rev. 2005; 29:530–58. DOI: 10.1177/0193841X05275596. PMID: 16244051.

2. Day AG. Why the propensity for propensity scores? Crit Care Med. 2015; 43:2024–6. DOI: 10.1097/CCM.0000000000001175. PMID: 26274709.

3. Caliendo M, Kopeinig S. Some practical guidance for the implementation of propensity score matching. J Econ Surv. 2008; 22:31–72. DOI: 10.1111/j.1467-6419.2007.00527.x.

4. Heinrich C, Maffioli A, Vazquez G. A primer for applying propensity score matching. Inter-American Development Bank. 2010. cited 2016 Jan 15. Available from http://idbdocs.iadb.org/wsdocs/getdocument.aspx?docnum=35320229.

5. Hullsiek KH, Louis TA. Propensity score modeling strategies for the causal analysis of observational data. Biostatistics. 2002; 3:179–93. DOI: 10.1093/biostatistics/3.2.179. PMID: 12933612.

6. Rosenbaum PR. Constructing matched sets and strata. Observational studies. 2002. 2nd ed. New York: Springer-Verlag;p. 295–331. DOI: 10.1007/978-1-4757-3692-2.

7. NCSS statistical Software. Chapter 123 Data Matching - Optimal and Greedy [Internet]. cited 2016 Jan 15. Kaysville: NCSS LLC;Available from http://www.ncss.com/wp-content/themes/ncss/pdf/Procedures/NCSS/Data_Matching-Optimal_and_Greedy.pdf.

8. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res. 2011; 46:399–424. DOI: 10.1080/00273171.2011.568786. PMID: 21818162. PMCID: PMC3144483.

9. Austin PC, Mamdani MM. A comparison of propensity score methods: a case-study estimating the effectiveness of post-AMI statin use. Stat Med. 2006; 25:2084–106. DOI: 10.1002/sim.2328. PMID: 16220490.

10. Lu B. Propensity score matching with time-dependent covariates. Biometrics. 2005; 61:721–8. DOI: 10.1111/j.1541-0420.2005.00356.x. PMID: 16135023.

11. King G, Nielsen R. Why propensity scores should not be used for matching [Internet]. 2016. cited 2016 Jan 15. Available from http://gking.harvard.edu/files/gking/files/psnot.pdf.

12. Shah BR, Laupacis A, Hux JE, Austin PC. Propensity score methods gave similar results to traditional regression modeling in observational studies: a systematic review. J Clin Epidemiol. 2005; 58:550–9. DOI: 10.1016/j.jclinepi.2004.10.016. PMID: 15878468.

13. Weitzen S, Lapane KL, Toledano AY, Hume AL, Mor V. Principles for modeling propensity scores in medical research: a systematic literature review. Pharmacoepidemiol Drug Saf. 2004; 13:841–53. DOI: 10.1002/pds.969. PMID: 15386709.

14. Ho DE, Imai K, King G, Stuart EA. Match It: nonparametric preprocessing for parametric causal inference [Internet]. J Stat Soft. 2011. Jun. 42. Available from http://www.jstatsoft.org/article/view/v042i08.

15. Sekhon JS. Multivariate and propensity score matching software with automated balance optimization: The matching package for R [Internet]. J Stat Soft. 2011. May. 42. Available from http://www.jstatsoft.org/article/view/v042i07.

16. Hansen B, Klopfer SO. Optimal full matching and related designs via network flows. J Comput Graph Stat. 2006; 15:609–27. DOI: 10.1198/106186006X137047.

17. Leuven E, Sianesi B. PSMATCH2 Stata module to perform full Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing. Statistical Software Components S432001. 2003. cited 2016 Jan 15. Chestnut Hill, Massachusetts: Boston College Department of Economics;Available from https://ideas.repec.org/c/boc/bocode/s432001.html.

18. Abadie A, Drukker D, Herr JL, Imbens GW. Implementing matching estimators for average treatment effects in Stata. Stata J. 2004; 4:290–311.

19. Becker SO, Ichino A. Estimation of average treatment effects based on propensity scores. Stata J. 2002; 2:358–77.

20. Parsons LS. Reducing bias in a propensity score matched pair sample using greedy matching techniques [Internet]. 2001. Cary NC: SAS Institute Inc., Proceedings of the Twenty-Sixth Annual SAS ÒUsers Group International Conference;p. 214–26. Available from http://www2.sas.com/proceedings/sugi26/p214-26.pdf.

21. Levesque R, Balabanov A. Raynald’s SPSS Tools. Match cases on basis of propensity scores [Internet]. 2001. Dec. 8. cited 2016 Jan 15. Available from http://spsstools.net/en/syntax/375/.