Building a Better Way to Measure Risk in Transplant Patients

Predicting what will happen after an organ transplant is difficult. One major reason is that the most common tools used to measure patient health risks were designed for general hospital patients—not for people receiving transplants.

Faith Mendoza, MB a graduate of the Duke Biostatistics Master’s program, led a multicenter study as her final master’s project to address this problem. Her team tested the Elixhauser Comorbidity Index in kidney and liver transplant patients and then created new, organ-specific versions that better fit these groups.

Using data from 6,404 transplant patients across five U.S. centers, the team found that the updated, reweighted indices performed better than the original tool. These new models more accurately predicted patient risk, showed stronger links to post-transplant mortality, and more clearly separated survival patterns—especially for liver transplant patients, where the older methods often fell short.

Reweighting comorbidities for transplant populations significantly improves our ability to identify which patients are truly at highest risk after transplant.

The organ-specific indices also showed how different health conditions affect transplant patients in unique ways, highlighting why general risk tools don’t always work well for these complex cases. Results improved across centers and demographic groups, with especially notable gains among groups that are usually underrepresented.

These findings suggest that transplant-specific risk tools could help doctors make more informed decisions, support fairer evaluations of transplant centers, and strengthen research and policy related to transplant outcomes.