RESUMEN
Since 2002, the Model for End-Stage Liver Disease (MELD) has been used to rank liver transplant candidates. However, despite numerous revisions, MELD allocation still does not allow for equitable access to all waitlisted candidates. An optimized prediction of mortality (OPOM) was developed (http://www.opom.online) utilizing machine-learning optimal classification tree models trained to predict a candidate's 3-month waitlist mortality or removal utilizing the Standard Transplant Analysis and Research (STAR) dataset. The Liver Simulated Allocation Model (LSAM) was then used to compare OPOM to MELD-based allocation. Out-of-sample area under the curve (AUC) was also calculated for candidate groups of increasing disease severity. OPOM allocation, when compared to MELD, reduced mortality on average by 417.96 (406.8-428.4) deaths every year in LSAM analysis. Improved survival was noted across all candidate demographics, diagnoses, and geographic regions. OPOM delivered a substantially higher AUC across all disease severity groups. OPOM more accurately and objectively prioritizes candidates for liver transplantation based on disease severity, allowing for more equitable allocation of livers with a resultant significant number of additional lives saved every year. These data demonstrate the potential of machine learning technology to help guide clinical practice, and potentially guide national policy.
Asunto(s)
Hepatopatías/mortalidad , Trasplante de Hígado , Listas de Espera , Femenino , Humanos , Hepatopatías/cirugía , Aprendizaje Automático , Masculino , Modelos EstadísticosRESUMEN
BACKGROUND: When a deceased-donor kidney is offered to a waitlisted candidate, the decision to accept or decline the organ relies primarily upon a practitioner's experience and intuition. Such decisions must achieve a delicate balance between estimating the immediate benefit of transplantation and the potential for future higher-quality offers. However, the current experience-based paradigm lacks scientific rigor and is subject to the inaccuracies that plague anecdotal decision-making. METHODS: A data-driven analytics-based model was developed to predict whether a patient will receive an offer for a deceased-donor kidney at Kidney Donor Profile Index thresholds of 0.2, 0.4, and 0.6, and at timeframes of 3, 6, and 12 months. The model accounted for Organ Procurement Organization, blood group, wait time, DR antigens, and prior offer history to provide accurate and personalized predictions. Performance was evaluated on data sets spanning various lengths of time to understand the adaptability of the method. RESULTS: Using United Network for Organ Sharing match-run data from March 2007 to June 2013, out-of-sample area under the receiver operating characteristic curve was approximately 0.87 for all Kidney Donor Profile Index thresholds and timeframes considered for the 10 most populous Organ Procurement Organizations. As more data becomes available, area under the receiver operating characteristic curve values increase and subsequently level off. CONCLUSIONS: The development of a data-driven analytics-based model may assist transplant practitioners and candidates during the complex decision of whether to accept or forgo a current kidney offer in anticipation of a future high-quality offer. The latter holds promise to facilitate timely transplantation and optimize the efficiency of allocation.