Functional Specifications for Propensity Score Table 1

Project Title Functional Specifications for Propensity Score Table 1
Date Posted
Wednesday, July 11, 2018
Status
Complete
Deliverables
Description

The objectives of this project are to determine the optimal output display(s) for a baseline covariate characteristics table, i.e., Table 1, for propensity score stratified analyses and to write a detailed functional specification to enhance the tool. 

The Sentinel Propensity Score (PS) adjustment tool enables stratified, 1:1 matched, and variable-ratio matched analyses. In contrast to fixed-ratio matching (e.g., 1:1 propensity score matching), stratification and variable-ratio matching require accounting for the different numbers of patients in each stratum or matched set in order to control for confounding. Baseline covariate characteristics tables (i.e., ‘Table 1’) among the matched populations are typically used to assess covariate balance as a proxy for confounding control when performing matched analyses. Currently, no such table exists for stratified analyses to enable capture of balance within and across strata in an aggregated overall description of the stratified cohort.

Workgroup Leader(s)

Josh Gagne ScD; Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA

Workgroup Members

Shirley Wang PhD; Yinzhu Jin MPH; Rishi Desai PhD; Richie Wyss PhD; Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 

Ting-Ying Huang BSPharm, PhD; Zilu Zhang MSc; Andrew Petrone MPH; Margaret Johnson; Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Healthcare Institute, Boston, MA 

Rongmei Zhang PhD; Mingfeng Zhang PhD; David Tyler Coyle MD; Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, MD