Medical research has recently established the high frequency of New-Onset Diabetes After Transplant (NODAT), which refers to the incidence of diabetes in transplanted patients with no prior history of diabetes. The dynamic and complex interactions between immunosuppressive drugs used to ensure organ survival, medications used to prevent NODAT, and the simultaneous risks of NODAT and organ rejection has created a conundrum for physicians, leaving them in an ambiguous state in their post-transplant decisions. To assist physicians, the research will develop mathematical models using techniques from operations research, statistics, and econometrics. The models will consider multiple perspectives including a patient's quality adjusted lifespan, the risk of developing NODAT, the risk of organ rejection, the potential errors in estimating the health transition and observation probabilities, and the sensitivity and specificity of available medical tests. If successful, this collaborative award will help generate new guidelines and a data-driven decision support system that has the potential to increase patient safety and help hospitals reduce NODAT, organ rejection, and patient mortality.
The intellectual merits of the research include new directions for applications of operations research to healthcare. Specifically, the models consider the decision-maker's pessimism/optimism, direct incorporation of time-varying medical risk factors, and empower the decision maker to dynamically optimize with respect to a "cloud" of models (as opposed to a single model), thereby gaining robustness to potential model misspecifications without the need to perform sensitivity analyses. This is in sharp contrast with currently available techniques that solve a single dynamic optimization model (with parameters estimated from data sets), and then attempt to mitigate potential estimation errors via sensitivity analyses. Although motivated by the interactions between immunosuppressive drugs and diabetes medications for NODAT patients, the methodological contributions have other potential uses, such as in advancing the science of medication management for a variety of diseases for which therapeutic interventions have conflicting effects.
