Use of Predictive Models to Determine Transplant Eligibility.

PURPOSE OF REVIEW: This paper summarizes predictive models developed to determine transplant eligibility over the past 5 years, focusing on application of novel data sources and methodologic approaches. RECENT FINDINGS: The contemporary body of research employing predictive models to inform transplant eligibility mainly relies on pre- or post-transplant patient survival. No studies have sought to assimilate all features collected during the transplant evaluation process to produce a composite prediction of post-transplant success or failure. SUMMARY: Predictive modeling is a commonly used statistical technique that uses available data on a subset of a target population to estimate the current health state or the probability of developing a future health outcome among individuals in the target population. Modern analytic techniques allow for transformation of vast amounts of data into actionable information but require curated organized well-defined data to deploy. That data is currently lacking for patients referred for transplant.

Duke Scholars

Author Nrupen Bhavsar Surgery, Surgical Sciences

Author Roland Albert Matsouaka Biostatistics & Bioinformatics, Division of Biostatistics

Author Lisa M McElroy Surgery, Abdominal Transplant Surgery