Almost 9% of deceased donors in the United States are classified by the CDC as """"""""high risk for transmission of HIV"""""""" based on a set of behavioral criteria introduced in 1985 and formalized in 1994. When these criteria were originally developed, they were based on estimates of prevalent HIV disease, and the CDC recommended that organs not be used from Infectious Risk Donors (IRDs) except in extenuating circumstances. However, with significant advances in viral testing, the risks are much lower, the diseases of concern have changed, and the original behavioral criteria are less relevant to the predictions required for clinical decision-making than they were in 1985 and 1994. As a result, discard rates for IRD kidneys are significantly higher than their non-IRD counterparts despite good outcomes in those who do receive them. The central problem is that selecting a recipient that will benefit from an IRD kidney is difficult. First, no systematic estimates of the risk of undetected HIV or hepatitis C (HCV) with various IRD behaviors exist, so clinical decision-making is based more on intuition and anecdote than on evidence. Second, no studies exist that compare, in a given patient, the risk of death while waiting for a better kidney offer with the risk of undetected viral infection from IRDs. We hypothesize that subgroups of patients exist for whom the risks of dialysis while waiting for a better kidney offer far exceed the risks of HIV or HCV transmission from IRDs, and that defining this subgroup will increase comfort with and utilization of IRDs. This seems to be a concept intuitive to transplant surgeons, as identification of a target recipient profile was associated with significantly higher likelihood of using kidneys from IRDs in a national survey. In an effort to inform and improve utilization of IRD kidneys, we propose to systematically review the literature on incidence of HIV and HCV seroconversion in various behavioral risk groups, estimate the predicted probabilities of HIV and HCV transmission in IRDs, and design a Markov decision process model for identifying the recipients who are likely to benefit significantly from IRD kidneys compared with waiting on dialysis. The research described in this proposal will directly address, through a novel mathematical approach, a critical clinical need. A successful Markov decision process model will be immediately useable clinically throughout the country, and will with high likelihood increase provider comfort with kidneys from a currently underutilized subgroup of deceased donors.
Although kidney transplantation offers potentially significant benefit to the tens of thousands of patients on the waiting list, there is reluctance to use kidneys from almost 10% of donors who, based on studies of behavioral risk from over 20 years ago, are flagged as having increased (although still very low) risk of carrying HIV. Because the risk of death on the waiting list is now so high, we suspect that, for some patients, the risk of transmitting an undetected infection from these donors is much lower than the risk of dying while waiting for a better kidney. The goal of this project is to update the antiquated behavioral risk flags, quantify the true risk associated with these donors, and use novel mathematical methods to identify patients who will benefit from kidney transplants from this underutilized supply of donors.
