Date on Master's Thesis/Doctoral Dissertation

5-2012

Document Type

Master's Thesis

Degree Name

M.S.

Department

Bioinformatics and Biostatistics

Committee Chair

Brock, Guy

Author's Keywords

Propensity score

Subject

Statistical matching

Abstract

Because of limitations in randomized controlled trials, medical researchers are often forced to rely upon studies of observational data. Confounding is a major difficulty encountered in such studies that can create considerable bias in estimates of treatment effects. Propensity score analysis was developed by Rosenbaum & Rubin in 1983 to overcome these difficulties. In essence, a propensity score allows balance to be achieved on confounding covariates in treatment and control groups, thus creating a 'quasi-randomized' trial from observational data. In this study, I illustrate the use of propensity matching to demonstrate that African American race is a significant risk factor for receiving a lower quality donor kidney using a national database on transplantation. I then use propensity matching to demonstrate the benefits of laparoscopic resection for hepatic colorectal metastases. In doing so, the great value of propensity matching in reducing bias in observational studies is demonstrated.

Download

Share

COinS