Abstract
Introduction: Access to kidney transplantation is often unequal due to structural bias across regions, states, sexes, and ethnic groups. In Australia, this disparity is evident as patients in more populous states may experience longer waiting times and therefore a reduced likelihood of successful treatment. As such, a comprehensive methodology to reduce bias in existing kidney allocation algorithms before implementation is required.
Methods: To this end, we developed the Kidney Allocation Process Evaluation System (KAPES), a comprehensive framework to systematically evaluate allocation along four key dimensions: matching effectiveness, individual outcomes, cost and social/ethical equity. KAPES leverages the simKAP simulation platform along with other statistical and Agent-Based Model (ABM) procedures to assess outcomes of patients within each of these four dimensions. To illustrate the capabilities of KAPES, we evaluate the impact of changes to the Australian allocation system before and after May 2021, as well as the potential impact of the proposed system. We further assess the impact of removing the current two-tier system and examine how changes in donor distribution influence outcomes.
Results: Firstly, we evaluated the current allocation framework before and after May 2021 strategies to assess their impact on patients. We found that compared to the previous allocation algorithm, the current allocation algorithm had mixed effects on jurisdictional bias: while overall wait times for many states remain unchanged, we observed that selected states, such as NSW/ACT were reduced by 23 months (p < 0.001) and SA/NT reduced by 15 months (p = 0.01). In addition, SA/NT and TAS/VIC saw marginally better survival probability (0.09 and 0.03). Next, we considered a series of hypothetical allocation algorithms. We found removing the multi-tiered systems made no significant changes to wait time or survival probability (p > 0.05). Furthermore, we demonstrated via simulations that wait times are largely influenced by the underlying donor pool rather than the allocation algorithm. For example, minority groups in non-diverse states will have to wait significantly longer (2 years, p < 0.001) more than average compared to the more diverse states, while non-minority groups found minimal difference between diverse and non-diverse donor pools.
Conclusion: Fundamentally, KAPES enables policymakers to quantitatively test allocation strategies. Our results suggest that improving clinical outcomes may increase jurisdictional bias and that achieving truly equitable transplant outcomes is unlikely through algorithmic adjustments alone. Addressing persistent bias may require targeted supply-side interventions, including efforts to expand the diversity and volume of the donor pool. Public policy approaches that encourage greater donor diversity could help improve donor–recipient matchability and support more equitable access.