Balance diagnostics after propensity score matching

被引:660
作者
Zhang, Zhongheng [1 ]
Kim, Hwa Jung [2 ,3 ]
Lonjon, Guillaume [4 ,5 ,6 ,7 ]
Zhu, Yibing [8 ]
机构
[1] Zhejiang Univ, Sir Run Run Shaw Hosp, Dept Emergency Med, Sch Med, Hangzhou 310016, Zhejiang, Peoples R China
[2] Asan Med Ctr, Dept Clin Epidemiol & Biostat, Seoul, South Korea
[3] Univ Ulsan, Dept Prevent Med, Coll Med, Seoul, South Korea
[4] INSERM, UMR 1153, Ctr Res Epidemiol & Stat Sorbonne Paris Cite CRES, METHODS Team, Paris, France
[5] AP HP, Orthopaed Dept, Paris, France
[6] Hosp Georges Pompidou, Paris, France
[7] Paris Descartes Univ, Med Sch, Sorbonne Paris Cite, Paris, France
[8] Capital Med Univ, Fuxing Hosp, ICU, Beijing 100045, Peoples R China
关键词
Propensity score; standardized mean difference (SMD); balance diagnostics; prognostic score;
D O I
10.21037/atm.2018.12.10
中图分类号
R73 [肿瘤学];
学科分类号
100214 ;
摘要
Propensity score matching (PSM) is a popular method in clinical researches to create a balanced covariate distribution between treated and untreated groups. However, the balance diagnostics are often not appropriately conducted and reported in the literature and therefore the validity of the findings from the PSM analysis is not warranted. The special article aims to outline the methods used for assessing balance in covariates after PSM. Standardized mean difference (SMD) is the most commonly used statistic to examine the balance of covariate distribution between treatment groups. Because SMD is independent of the unit of measurement, it allows comparison between variables with different unit of measurement. SMD can be reported with plot. Variance is the second central moment and should also be compared in the matched sample. Finally, a correct specification of the propensity score model (e.g., linearity and additivity) should be re-assessed if there is evidence of imbalance between treated and untreated. R code for the implementation of balance diagnostics is provided and explained.
引用
收藏
页数:8
相关论文
共 16 条
[1]  
Austin PC, 2008, STAT MED, V27, P2037, DOI 10.1002/sim.3150
[2]   A comparison of 12 algorithms for matching on the propensity score [J].
Austin, Peter C. .
STATISTICS IN MEDICINE, 2014, 33 (06) :1057-1069
[3]   A Tutorial and Case Study in Propensity Score Analysis: An Application to Estimating the Effect of In-Hospital Smoking Cessation Counseling on Mortality [J].
Austin, Peter C. .
MULTIVARIATE BEHAVIORAL RESEARCH, 2011, 46 (01) :119-151
[4]   Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples [J].
Austin, Peter C. .
STATISTICS IN MEDICINE, 2009, 28 (25) :3083-3107
[5]   Report Card on Propensity-Score Matching in the Cardiology Literature From 2004 to 2006 A Systematic Review [J].
Austin, Peter C. .
CIRCULATION-CARDIOVASCULAR QUALITY AND OUTCOMES, 2008, 1 (01) :62-67
[6]   STANDARD DISTANCE IN UNIVARIATE AND MULTIVARIATE-ANALYSIS [J].
FLURY, BK ;
RIEDWYL, H .
AMERICAN STATISTICIAN, 1986, 40 (03) :249-251
[7]   Propensity scores in intensive care and anaesthesiology literature: a systematic review [J].
Gayat, Etienne ;
Pirracchio, Romain ;
Resche-Rigon, Matthieu ;
Mebazaa, Alexandre ;
Mary, Jean-Yves ;
Porcher, Raphael .
INTENSIVE CARE MEDICINE, 2010, 36 (12) :1993-2003
[8]   The prognostic analogue of the propensity score [J].
Hansen, Ben B. .
BIOMETRIKA, 2008, 95 (02) :481-488
[9]  
Ho D. E., 2017, POLIT ANAL, V15, P199
[10]   Misunderstandings between experimentalists and observationalists about causal inference [J].
Imai, Kosuke ;
King, Gary ;
Stuart, Elizabeth A. .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES A-STATISTICS IN SOCIETY, 2008, 171 :481-502