Heart Transplantation: An In-Depth Survival Analysis

• Activity Number: 54 - Methods and Modeling for Medical Device and Clinical Studies
• Type: Contributed
• Date/Time: Monday, August 3, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Medical Devices and Diagnostics
• Author(s): Gang Xu* and Eileen Michelle Hsich and Eugene Blackstone and Lucy Thuita and Dennis McNamara and Joseph Rogers and Clyde Yancy and Lee Goldberg and Maryam Valapour and Hemant Ishwaran
• Companies: University of Miami and Cleveland Clinic and Cleveland Clinic and Cleveland Clinic and University of Pittsburgh Medical Center and Duke University and Northwestern University Medical Center and Hospital of the University of Pennsylvania and Cleveland Clinic and University of Miami
• Keywords: heart transplantation; mortality; variable importance; interaction; random forest; survival analysis

Abstract

Heart transplant allocation is primarily focused on preventing waitlist mortality. To prevent organ wastage, future allocation will need to balance risk of waitlist mortality with post-transplant mortality. To understand the complex factors affecting post-heart transplant survival and to determine the importance of possible sex-specific risk factors, we studied data from all adults in the Scientific Registry of Transplant Recipients (SRTR) database who underwent heart transplant between January 1, 2004 and July 1, 2018. We identified three phases affecting survival: early post-operative, constant, and a late phase, and found that the factors and interactions predicting post-transplant mortality varied with the three phases. Machine learning methods and conventional statistical methods were both important in this study and provided complementary information to better understand post-transplant mortality.