Competing Risk Survival Analysis of Bariatric Surgery Data Using Propensity Score Methods

• Activity Number: 569
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Diqiong Xie*+ and Michael P. Jones and Mason E. Edward
• Companies: University of Iowa and University of Iowa and University of Iowa
• Keywords: propensity score ; matching ; survival analysis ; competing risks ; bariatric surgery

Abstract

Bariatric surgery has been recognized as an effective therapy for morbid obesity. There are currently two major categories of surgical procedures, depending on whether gastric bypass is included or not. 'Simple' operations are purely restrictive, with no bypass of digestive tract, while 'Complex' operations bypass the duodenum and/or jejunum, in addition to decreasing the size of the stomach. There have been several observational studies investigating the difference in outcome between these two types of procedures, focusing on the efficiency of weight loss, post-operative complication and overall mortality. However, the question remains unanswered about how selection bias was adjusted in the above mentioned observational studies, in which the treatments, i.e. operation types, were not randomly assigned, but were determined by the surgeon and patient. In this study, we show that the choice of treatment depends heavily on patient characteristics. Therefore, we adjust for selection bias by using propensity scores, and analyze the effect of operation types on overall mortality and cause-specific mortality using pair-matched patients.