90-day graft survival hazard ratio: A program’s rate of graft failure from date of transplant to 90 days post-transplant, relative to the expected rate based on transplants with similar recipient and donor organ characteristics across the nation.

1-year conditional on 90-day graft survival hazard ratio: A program’s rate of graft failure from day 90 post-transplant to day 365 post-transplant, conditional on the graft surviving for the first 90-days post-transplant, relative to the expected rate based on transplants with similar recipient and donor organ characteristics across the nation.

• Beginning July 14, 2022, the MPSC began to consider program performance on the above metrics for transplants that occurred from January 2019 through June 2021 (excluding the March 13– June 12, 2020, SRTR COVID carve-out).

Offer acceptance: A program’s actual rate of accepting organ offers relative to the expected rate, based on similar offers across the nation.

• Beginning July 27, 2023, the MPSC will begin to consider program performance on the above metric for offers received from January 2022 through December 2022. 

• The MPSC sent notice to transplant programs six months prior to implementation. 
• Download the PDF: Effective Practices to Improve Offer Acceptance (PDF)

• Recorded presentations from the OPTN Offer Acceptance Collaborative's 2023 hybrid kickoff event are available in the OPTN Learning Management System (known as UNOS Connect) and offer access to a range of thought leadership on system performance.

  Search for the OPTN Offer Acceptance Collaborative Playlist – CEPTCs available

  The modules have information about:

  • MPSC transplant program performance monitoring, the offer acceptance rate ratio, and OPTN tools to drive improvement.
  • Effective practices related to offer acceptance and lessons learned from previous collaborative participants.

Pre-transplant mortality: A program’s pre-transplant mortality rate relative to the mortality rate experienced by similar patients across the nation.

•  Beginning in July 2024, the MPSC will consider program performance on mortality for candidates appearing on the waiting list at any time during the period of January 2022 through December 2023.
• The MPSC will send notice to transplant programs six months prior to implementation.
• Download the PDF: Effective Practices to Improve Pre-transplant Outcomes

Without risk adjustment, all programs would be expected to achieve the same observed survival rate, regardless of the acuity of their patients or the quality of the donor organs they transplant. The goal of using risk adjustment is to be able to differentiate between program performance, given the known patient and donor factors.

The MPSC wants to understand if a program’s outcomes differ significantly from what would be expected from that program based on the outcomes of similar patients or organ offers nationally; or, in other words, what is the individual program’s effect on the outcome of interest (graft survival, pre-transplant mortality, offer acceptance).

Some transplants involve greater risk than others. The risk adjustment models do not expect higher-risk transplants to succeed at the same rate as lower-risk transplants. If the higher-risk transplants a program performs are associated with a bad outcome, and the model expects those transplants to have a poorer outcome, then the program can still be considered to be performing as expected.

In this example, assume two liver transplant programs perform the same number of transplants and that most of the donor and recipient characteristics are the same between the programs. One program utilizes a high volume of donor after cardiac death (DCD) livers, while the other program does not use any.

• The model expects that liver transplants using a DCD liver have an approximately 60 percent higher risk of graft failure based on national data.
• Because of risk-adjustment, the model would expect the program that utilizes DCD livers to have more graft failures (a lower “expected” rate) than the program that uses none.
• However, a program that rarely performs higher-risk transplants would be expected to have fewer poor outcomes, and would be identified as performing worse than expected if the program had poor outcomes in average-risk transplants.

If the SRTR models did not use risk adjustment, the MPSC would not be well-equipped to understand if a program’s outcomes differ from what would be expected based on patient and donor risk factors that are beyond the program’s control, such as the age of the recipient or whether a donor is a donor after cardiac death or donor after brain death. Instead, the MPSC would only have the raw, absolute number of events at a program. The MPSC would not be able to determine easily if those events were similar to events experienced by other programs, or were associated with more complex transplants.

The SRTR risk adjustment models are limited to the data currently collected by the OPTN. The MPSC acknowledges that current risk adjustment models cannot adjust for all risks; however, the MPSC does not believe that the lack of perfect risk adjustment supports the complete rejection of risk adjustment. If the MPSC and the transplant community acknowledge that transplant programs have different risk profiles – different patient acuity, different case mix, and different donor availability depending on where they are located and the donor supply - the way to reflect that difference is through risk adjustment.

No. Risk-adjusted models cannot be built for some types of transplants, because there are not enough events or available factors that predict the outcome. For example, there were no predictors identified for pediatric kidney deceased donor 1-year graft survival, so a risk-adjusted model was not built.

In cases such as these, the national failure rate is used to determine an expected rate for a program.

The SRTR is continuously improving its risk adjustment models and encourages any feedback from the community and the MPSC. Every six months, the SRTR analyzes national data to determine if the risk associated with a variable has changed and refits the models to capture the current risk profile.