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

Peter C. Austin

doi:10.1002/sim.3697

journal article Oct 13, 2009

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

Peter C. Austin

Statistics in Medicine Vol. 28 No. 25 pp. 3083-3107 · Wiley

View at Publisher Save 10.1002/sim.3697

Abstract

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.

Topics

No keywords indexed for this article. Browse by subject →

References

47

[1]

The central role of the propensity score in observational studies for causal effects

Paul R. Rosenbaum, Donald B. Rubin

Biometrika 10.1093/biomet/70.1.41

[2]

10.1080/01621459.1984.10478078

[3]

10.1002/sim.2328

[4]

10.1002/pds.969

[5]

10.1016/j.jclinepi.2004.10.016

[6]

10.1016/j.jclinepi.2005.07.004

[7]

10.1002/sim.3150

[8]

10.1016/j.jtcvs.2007.07.021

[9]

10.1161/circoutcomes.108.790634

[10]

10.1023/a:1020363010465

[11]

10.1002/pds.968

[12]

10.1093/pan/mpl013

[13]

10.1016/j.ahj.2005.06.034

[14]

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

[15]

10.1002/sim.2770

[16]

Tu JV (2004)

[17]

10.1002/sim.2580

[18]

Some Methods of Propensity‐Score Matching had Superior Performance to Others: Results of an Empirical Investigation and Monte Carlo simulations

Peter C. Austin

Biometrical Journal 10.1002/bimj.200810488

[19]

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

[20]

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

[21]

10.2307/2684560

[22]

10.2307/2683903

[23]

10.1016/s0895-4356(00)00321-8

[24]

10.1016/j.ahj.2006.06.020

[25]

10.1093/eurheartj/ehi890

[26]

Cohen J (1988)

[27]

Hedges LV (1985)

[28]

10.2202/1557-4679.1146

[29]

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

[30]

Rosner B (1995)

[31]

10.1007/978-1-4757-3462-1

[32]

Hoaglin DC (1983)

[33]

Casella G (1990)

[34]

10.1002/pds.986

[35]

10.1002/sim.4780131703

[36]

10.1002/sim.4780080410

[37]

10.1002/sim.4780100514

[38]

10.1056/nejm198311243092105

[39]

10.1093/biomet/71.3.431

[40]

10.1016/j.jclinepi.2009.06.002

[41]

Sackett DL "Down with odds ratios! for publication" Evidence‐Based Medicine (1996)

[42]

10.1002/sim.2683

[43]

10.1046/j.1524-4733.2002.55150.x

[44]

The number needed to treat: a clinically useful measure of treatment effect

Richard J Cook, David L Sackett

BMJ 10.1136/bmj.310.6977.452

[45]

Jaeschke R "Basis statistics for clinicians 3: assessing the effects of treatment: measures of association" Canadian Medical Association Journal (1995)

[46]

10.1016/0895-4356(94)90191-0

[47]

Assessing balance in measured baseline covariates when using many‐to‐one matching on the propensity‐score

Peter C. Austin

Pharmacoepidemiology and Drug Safety 10.1002/pds.1674

Cited By

5,504

Glucagon-like peptide-1 receptor agonists and risk of substance use disorders among US veterans with type 2 diabetes: cohort study

Miao Cai, Taeyoung Choi · 2026

BMJ

Clinicopathological and Prognostic Characteristics of Gastric-Type Endocervical Adenocarcinoma: A Nested Case–Control Study

Yang Liu, Yundi Hu · 2026

Cancers

Association of carotid revascularization with epilepsy and seizures in patients with carotid stenosis

Ming‐Tsung Chuang, Yu Chang · 2026

Epilepsia

Bullying victimisation as a mediator in the association between childhood epilepsy and later emotional and behavioural difficulties: Evidence from the British National Child Development Study

Emma Blundell, Vaughan Bell · 2026

Epilepsy & Behavior

Minimal benefit of co-testing over HPV primary screening with cytology triage from resource-limited settings in China

Xinhua Jia, Xi’ao Da · 2026

Communications Medicine

Non-invasive continuous versus intermittent oscillometric arterial pressure monitoring and maternal hypotension during cesarean delivery: a randomized controlled trial

Youngwon Kim, Hansol Kim · 2026

Scientific Reports

Transition to Virtual Diabetes Self-Management Education Delivery in the Setting of Health Care Disruption for Adults With Diabetes and Their Support Persons

Denise J. Deverts, Margaret F. Zupa · 2026

The Science of Diabetes Self-Manage...

Robotic Versus Open Pancreaticoduodenectomy: A Single-Center Analysis of Safety and Efficacy Using Inverse Probability of Treatment Weighting

Mariano Cesare Giglio, Silvia Campanile · 2025

Cancers

Comparative Safety of Biologic and Targeted‐Synthetic DMARDs in Patients With Rheumatoid Arthritis: A Multi‐Database Real‐World Cohort Study

Yinzhu Jin, Jun Liu · 2025

Pharmacoepidemiology and Drug Safet...

Interpretable machine learning model to predict 90-day radiographically confirmed pneumonia after chemotherapy initiation in non-Hodgkin lymphoma: development and internal validation of a single-center cohort

Zhanna Zhang, Manqi Su · 2025

Frontiers in Medicine

Midazolam and Ketamine for Convulsive Status Epilepticus in the Out-of-Hospital Setting

Tony Zitek, Kenneth A. Scheppke · 2025

Annals of Emergency Medicine

Risk of Incident Atrial Fibrillation in Women With a History of Hypertensive Disorders of Pregnancy: A Population-Based Retrospective Cohort Study

Amy Johnston, William Petrcich · 2025

Circulation

Clinical and Economic Burden of Cytomegalovirus (CMV) Infection/Disease Among Hospitalized Adult Allogeneic Hematopoietic Stem Cell Transplant (Allo‐HSCT) Recipients in China

Chenhua Yan, Linghui Xia · 2025

Journal of Medical Virology

Effectiveness of the 2024–2025 KP.2 COVID-19 vaccines in the United States during long-term follow-up

George N. Ioannou, Kristin Berry · 2025

Nature Communications

Apixaban for Extended Treatment of Provoked Venous Thromboembolism

Gregory Piazza, Behnood Bikdeli · 2025

New England Journal of Medicine

Postsurgical Medication Awareness, Recovery, and Tracking using a Phone-Based App (SMART-APP): a randomized clinical trial

Megan L Rolfzen, Karan Shah · 2025

Regional Anesthesia & Pain Medi...

The Influence of Knee Phenotypes Based on Coronal Plane Alignment of the Knee on Intraoperative Soft Tissue Balance and Clinical Outcomes: Comparison between Kinematically and Mechanically Aligned Total Knee Arthroplasty

Shotaro Tachibana, Tomoyuki Matsumoto · 2025

The Journal of Knee Surgery

Health Outcomes of Discontinuing Antipsychotics After Hospitalization in Older Adults

Chun-Ting Yang, James M. Wilkins · 2025

JAMA Psychiatry

The Accelerating Exposure of European Protected Areas to Climate Change

Marta Cimatti, Valerio Mezzanotte · 2025

Global Change Biology

Mapping the effectiveness and risks of GLP-1 receptor agonists

Yan Xie, Taeyoung Choi · 2025

Nature Medicine

Metrics

5,504

Citations

47

References

Details

Published: Oct 13, 2009
Vol/Issue: 28(25)
Pages: 3083-3107
License: View

Authors

P

Peter C. Austin

Cite This Article

Peter C. Austin (2009). Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity‐score matched samples. Statistics in Medicine, 28(25), 3083-3107. https://doi.org/10.1002/sim.3697

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

You May Also Like