CUSUMs for Survival Outcomes with Dependent Censoring: Application to Monitoring Waitlist Mortality in Liver Transplant Programs

• Activity Number: 587
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Jie (Rena) Sun*+ and John David Kalbfleisch and Douglas E. Schaubel
• Companies: University of Michigan and University of Michigan and University of Michigan
• Keywords: inverse probability of censoring weighting ; cumulative sum ; dependent censoring ; sequential probability ratio test ; V-mask ; quality control and quality improvement

Abstract

Cumulative Sum (CUSUM) charts are often utilized to monitor processes for quality control and quality improvement purposes. In essence, a CUSUM chart compares the cumulative observed outcomes to the cumulative expected outcomes derived from a null or reference model. In survival analysis settings, existing CUSUM methods assume independent censoring, which may not hold in practice. We propose a CUSUM that allows for dependent censoring by incorporating inverse probability of censoring weighting. We derive large-sample properties of the proposed CUSUM, and evaluate their applicability in finite samples through simulation. We demonstrate our method through evaluating liver transplant programs with respect to waitlist mortality, where dependent censoring arises from liver transplantation. We also implement a graphical approach using a Cumulative Sum (CUSUM) to monitor center outcomes.