Emmanuel Caruana1, Sylvie Chevret1, Matthieu Resche-Rigon1, Romain Pirracchio2. 1. Service de Biostatistique et Information Médicale, Hôpital Saint-Louis, AP-HP, Inserm 1153 ECSTRA Team, Université Paris 7 Diderot, 1 rue Claude Vellefaux, Paris 75010, France. 2. Service de Biostatistique et Information Médicale, Hôpital Saint-Louis, AP-HP, Inserm 1153 ECSTRA Team, Université Paris 7 Diderot, 1 rue Claude Vellefaux, Paris 75010, France; Service d'Anesthésie-Réanimation, Hôpital Européen Georges Pompidou, Université Paris V Descartes-Sorbonne Paris Cité, 20 rue Leblanc, Paris 75015, France. Electronic address: romainpirracchio@yahoo.fr.
Abstract
OBJECTIVES: The propensity score (PS) is a balancing score. Following PS matching, balance checking usually relies on estimating separately the standardized absolute mean difference for each baseline characteristic. The average standardized absolute mean difference and the Mahalanobis distances have been proposed to summarize the information across the covariates. However, they might be minimized when nondesirable variables such as instrumental variables (IV) are included in the PS model. We propose a new weighted summary balance measure that takes into account, for each covariate, its strength of association with the outcome. STUDY DESIGN AND SETTING: This new measure was evaluated using a simulation study to assess whether minimization of the measure coincided with minimally biased estimates. All measures were then applied to a real data set from an observational cohort study. RESULTS: Contrarily to the other measures, our proposal was minimized when including the confounders, which coincided with minimal bias and mean squared error, but increased when including an IV in the PS model. Similar findings were observed in the real data set. CONCLUSION: A balance measure taking into account the strength of association between the covariates and the outcome may be helpful to identify the most parsimonious PS model.
OBJECTIVES: The propensity score (PS) is a balancing score. Following PS matching, balance checking usually relies on estimating separately the standardized absolute mean difference for each baseline characteristic. The average standardized absolute mean difference and the Mahalanobis distances have been proposed to summarize the information across the covariates. However, they might be minimized when nondesirable variables such as instrumental variables (IV) are included in the PS model. We propose a new weighted summary balance measure that takes into account, for each covariate, its strength of association with the outcome. STUDY DESIGN AND SETTING: This new measure was evaluated using a simulation study to assess whether minimization of the measure coincided with minimally biased estimates. All measures were then applied to a real data set from an observational cohort study. RESULTS: Contrarily to the other measures, our proposal was minimized when including the confounders, which coincided with minimal bias and mean squared error, but increased when including an IV in the PS model. Similar findings were observed in the real data set. CONCLUSION: A balance measure taking into account the strength of association between the covariates and the outcome may be helpful to identify the most parsimonious PS model.
Authors: Richard Wyss; Ben B Hansen; Alan R Ellis; Joshua J Gagne; Rishi J Desai; Robert J Glynn; Til Stürmer Journal: Am J Epidemiol Date: 2017-05-01 Impact factor: 4.897
Authors: Tri-Long Nguyen; Gary S Collins; Jessica Spence; Jean-Pierre Daurès; P J Devereaux; Paul Landais; Yannick Le Manach Journal: BMC Med Res Methodol Date: 2017-04-28 Impact factor: 4.615
Authors: Anne Burden; Nicolas Roche; Cristiana Miglio; Elizabeth V Hillyer; Dirkje S Postma; Ron Mc Herings; Jetty A Overbeek; Javaria Mona Khalid; Daniela van Eickels; David B Price Journal: Pragmat Obs Res Date: 2017-03-22