More heart transplants through informed donor selection - PROJECT SUMMARY / ABSTRACT The demand for heart transplant (HT) in the United States (US) far exceeds the number of HTs performed, resulting in dire consequences for candidates on the HT waitlist. The applicant’s recent European Heart Journal publication refutes the long-held assumption that US HT volume is limited by a shortage of donors. It showed that the US utilizes only 44% of potential donors for HT and that adopting the more liberal donor acceptance practices of the Eurotransplant (ET) consortium would lead to hundreds more US HTs annually. The barriers to doing so include 1) lack of evidence on which donor risk factors impact HT outcomes and 2) difficulty synthesizing a wide array of clinical datapoints during donor selection. The applicant’s proposed research aims will address both barriers, with the goal of increasing heart donor utilization in the US. The first aim is to investigate the impact of donor risk factors on HT outcomes, focusing on coronary artery disease (CAD), left ventricular dysfunction and hypertrophy (LVH), and older age. Each is a common reason for donor heart non-acceptance. ET registry data will be used to test the hypothesis that mild CAD and LV dysfunction, in the absence of other risk factors, do not confer increased risk. Data from a novel US-wide cohort will be used to test the hypotheses that donor LVH is often transient or spurious and does not impact post-HT outcomes. Finally, the impact of a large HT center's strategy to utilize older donors will be evaluated. The second aim is to develop prediction models to rationalize donor heart assessment and selection. Having already produced an online tool to predict donor acceptance for HT, the applicant will employ machine learning algorithms to develop three additional decision-support tools to inform donor assessment. The first prediction tool will predict the likelihood of donor CAD and thereby inform selection of which donors need screening coronary angiography (which is highly time- and resource- intensive). The other prediction tools will quantify 1) the impact of a donor’s risk factors on predicted post-HT mortality and 2) how long a HT candidate will expect to wait before receiving a better donor offer. The expected impact of this research is to enable a more data-driven and systematic approach to heart donor selection, leading to discard of fewer viable donor hearts and thus more lives saved through HT. The K08 training plan will enable the applicant’s growth as an independent clinician-investigator in the field of HT policy and outcomes research, via expert mentorship and skill development in 1) machine learning and decision analytic modeling; 2) clinical informatics and observational research; 3) leadership and organization; and 4) oral and written communication. This skillset and the above research output will enable the applicant to lead prospective studies under a subsequent R01 grant to demonstrate the impact of increased donor utilization.
