Clinical Model for Predicting Kidney Failure

This technology includes a model for providing a patient-specific diagnosis of disease using clinical data. Specifically, the present invention relates to a fully unsupervised, machine-learned, cross-validated, and dynamic Bayesian Belief Network model that utilizes clinical parameters for determining a patient-specific probability of transplant glomerulopathy. Kidney failure is a growing problem worldwide, in part related to the increase incidence of diabetes and hypertension. Renal replacement therapy includes dialysis or renal transplantation. The average lifespan of a kidney transplant is about 10 years and graft loss may be due to both patient death as well as primary graft failure mediated by an entity known as transplant glomerulopathy ("chronic rejection"). Understanding the determinants of this disease would lead to new treatments and biomarkers of disease. This invention provides a method to predict the diagnosis based on clinical parameters. Thus, more accurate diagnosis and prediction of disease will help in patient management.