Hospital Length-of-Stay Modeling by Coxian Phase-Type Regression with Heterogeneity

• Activity Number: 351
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Xiaoqin Tang*+ and Zhehui Luo and Joseph Gardiner
• Companies: Michigan State University and Michigan State University and Michigan State University
• Keywords: Length of stay ; heterogeneity ; phase-type model ; Coxian ; RJMCMC ; cluster

Abstract

Hospital length of stay (LOS) is an important measure of healthcare utilization and is generally positively skewed and heterogeneous. We apply a latent homogeneous Markov model to fit a Coxian phase-type distribution and systematically identify states of hospital LOS. We demonstrate that selecting an appropriate number of phases and a regression model for hazard rates in the Coxian phase-type model can account for some observed heterogeneity in LOS. Reversible Jump Markov Chain Monte Carlo (RJMCMC) enables us to dynamically uncover the hidden stochastic Markov structure. A classification method is used to assign patients to different LOS groups. The methodology is illustrated with application to hospital admissions for acute myocardial infarction in the 2003 Nationwide Inpatient Sample from the Healthcare Utilization Project.