Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity‐score matched samples

Statistics in Medicine - Tập 28 Số 25 - Trang 3083-3107 - 2009
Peter C. Austin1,2,3
1Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada
2Department of Health Policy, Management, and Evaluation, University of Toronto, Canada
3Institute for Clinical Evaluative Sciences, G1 06, 2075 Bayview Avenue, Toronto, Ontario, Canada

Tóm tắt

AbstractThe propensity score is a subject's probability of treatment, conditional on observed baseline covariates. Conditional on the true propensity score, treated and untreated subjects have similar distributions of observed baseline covariates. Propensity‐score matching is a popular method of using the propensity score in the medical literature. Using this approach, matched sets of treated and untreated subjects with similar values of the propensity score are formed. Inferences about treatment effect made using propensity‐score matching are valid only if, in the matched sample, treated and untreated subjects have similar distributions of measured baseline covariates. In this paper we discuss the following methods for assessing whether the propensity score model has been correctly specified: comparing means and prevalences of baseline characteristics using standardized differences; ratios comparing the variance of continuous covariates between treated and untreated subjects; comparison of higher order moments and interactions; five‐number summaries; and graphical methods such as quantile–quantile plots, side‐by‐side boxplots, and non‐parametric density plots for comparing the distribution of baseline covariates between treatment groups. We describe methods to determine the sampling distribution of the standardized difference when the true standardized difference is equal to zero, thereby allowing one to determine the range of standardized differences that are plausible with the propensity score model having been correctly specified. We highlight the limitations of some previously used methods for assessing the adequacy of the specification of the propensity‐score model. In particular, methods based on comparing the distribution of the estimated propensity score between treated and untreated subjects are uninformative. Copyright © 2009 John Wiley & Sons, Ltd.

Từ khóa


Tài liệu tham khảo

10.1093/biomet/70.1.41

10.1080/01621459.1984.10478078

10.1002/sim.2328

10.1002/pds.969

10.1016/j.jclinepi.2004.10.016

10.1016/j.jclinepi.2005.07.004

10.1002/sim.3150

10.1016/j.jtcvs.2007.07.021

10.1161/CIRCOUTCOMES.108.790634

10.1023/A:1020363010465

10.1002/pds.968

10.1093/pan/mpl013

10.1016/j.ahj.2005.06.034

10.1111/j.1467-985X.2005.00380.x

10.1002/sim.2770

Tu JV, 2004, Quality of Cardiac Care in Ontario

10.1002/sim.2580

10.1002/bimj.200810488

Moher D, 2001, The CONSORT statement: revised recommendations for improving the quality of reports of parallel‐group randomized trials, Journal of the American Medical Association, 285, 1787

10.7326/0003-4819-134-8-200104170-00012

10.2307/2684560

10.2307/2683903

10.1016/S0895-4356(00)00321-8

10.1016/j.ahj.2006.06.020

10.1093/eurheartj/ehi890

Cohen J, 1988, Statistical Power Analysis for the Behavioral Sciences

Hedges LV, 1985, Statistical Methods for Meta‐Analysis

10.2202/1557-4679.1146

10.1111/j.1467-985X.2007.00527.x

Rosner B, 1995, Fundamentals of Biostatistics

10.1007/978-1-4757-3462-1

Hoaglin DC, 1983, Understanding Robust and Exploratory Data Analysis

Casella G, 1990, Statistical Inference

10.1002/pds.986

10.1002/sim.4780131703

10.1002/sim.4780080410

10.1002/sim.4780100514

10.1056/NEJM198311243092105

10.1093/biomet/71.3.431

10.1016/j.jclinepi.2009.06.002

Sackett DL, 1996, Down with odds ratios! for publication, Evidence‐Based Medicine, 1, 164

10.1002/sim.2683

10.1046/j.1524-4733.2002.55150.x

10.1136/bmj.310.6977.452

Jaeschke R, 1995, Basis statistics for clinicians 3: assessing the effects of treatment: measures of association, Canadian Medical Association Journal, 152, 351

10.1016/0895-4356(94)90191-0

10.1002/pds.1674

Thông tin
Thông tin xuất bản

Statistics in Medicine

Tập 28 Số 25

3083-3107

Thông tin tác giả