Agreement between risk and priority for heart transplant: Effects of the geographic allocation rule and status...

Allocation of donor hearts in the United States is accomplished by an algorithm based on status, time waited, and geographic boundaries. Although not designed to always transplant the highest acuity candidates, the ability of current United Network for Organ Sharing policies to prioritize highest acuity candidates is unknown.

We analyzed 32,866 adult match runs generated from 2007 to 2014. Each candidate's sequence number within a match run was compared with the candidate's risk of mortality using Kendall's tau-b-a measure of rank correlation. Two primary methods of evaluating risk of mortality were used: status designation-based risk (i.e., status 1A risk > status 1B > status 2) and status justification-based risk (e.g., status 1A justified by presence of a complication).

Median sequence number for transplanted candidates was 3 (interquartile range [IQR]: 1, 9). Median correlation among match runs for status-based risk was 0.57 (IQR: 0.47, 0.66) and for justification-based risk was 0.51 (IQR: 0.39, 0.60). Sensitivity to status 2 candidates was evident when status 2 candidates were removed from the sample (status-based tau-b = 0.31, justification-based tau-b = 0.1) and with restriction of the data set to only the first 20 candidates (status-based tau-b = 0.35, justification-based tau-b = 0.15).

There is only mild correlation between status and priority under the current allocation algorithm and poor concordance when more detailed risks are considered. The geographic allocation rule is responsible for most of the measured discordance.