Does IAPB Improve Survival? Uncovering Disease Severity by the Survival Mixture Modeling

• Activity Number: 550
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Chung-Han Ho*+ and Fu-Wen Liang and Shih-Feng Weng and Ya-Wen Hsu and Chin-Chen Chu and Chun-Yen Chiang
• Companies: Chi Mei Medical Center, Department of Medical Research and National Cheng Kung University, College of Medicine Department of Public Health and Chi Mei Medical Center, Department of Medical Research and Chia Nan University of Pharmacy and Science, Department of Hospital & Healthcare Admin. and Chi Mei Medical Center, Department of Anesthesiology and Chi Mei Medical Center, Division of Cardiology, Department of Internal Medicine
• Keywords: Intra-aortic balloon pump ; Survival analysis ; Mixture modeling ; Unobserved heterogeneity ; Latent class

Abstract

Cadiogenic shock (CS) is the leading cause of death in patients with acute myocardial infarction (AMI), accounting for more than 40% of death. Intra-aortic balloon pump (IABP) is currently the most common device applied in these patients. However, the benefit of IABP in infarct related CS is still controversial. In addition, when examining the prognosis of AMI patients with CS, it is problematic to take the disease severity into consideration. The disease severity is a typical unobserved heterogeneity in medical research. A nationwide cohort from the National Health Insurance Research Database in Taiwan was analyzed to examine the survival in the presence of unobserved heterogeneity under a framework of mixture modeling. A mixture model is applied to incorporate the latent survival trajectory along with the observed information. The unobserved survival trajectories are identified through estimating the probability that a subject belongs to a particular class based on observed information. The length of intensive care unit stay, a commonly used alternative to disease severity, is compared across different trajectories to validate the performance of estimated latent classes.