Preview of New, Improved Transplant Rate, and Waitlist Mortality Models
SRTR has updated the transplant rate and waitlist mortality models for the program-specific reports (PSRs) of kidney, liver, lung, and heart programs. The models will be used to derive the expected number of transplants and deaths on the waiting list in the January 2018 PSRs.
There are several important characteristics of the updated models. First, a much wider range of candidate characteristics at listing are considered, including information from the status history and/or justification files. Second, inactive status was removed from the liver/heart models, due to concerns that this would create potential non-clinical incentives to inactive candidates unlikely to undergo transplant. Third, separate models are estimated for pediatric and adult candidates at listing. Additionally, the updated models use a single 2-year cohort rather than two separate 1-year cohorts. Lastly, the models are now estimated with the Least Absolute Shrinkage and Selection Operator (LASSO). The LASSO can select important covariates while improving the predictive performance of the models.
Pediatric candidates currently receive significantly more priority in allocation than adult candidates. Thus, beginning with the January 2018 PSR release, the transplant rate and waitlist mortality sections of the PSR will classify the observed and expected transplants and deaths on the waiting list by pediatric and adult status.
The updated transplant rate and waitlist mortality models can be previewed on our website. The models were built using a similar process to the posttransplant models; read the SRTR publication Developing Statistical Models to Assess Transplant Outcomes Using National Registries: The Process in the United States for more information.
Share your feedback on the transplant rate and waitlist mortality models by contacting us.