Computer assisted diagnosis in renal transplantation a bayesian classification for differential diagnoses

Poor renal allograft function shortly after renal transplantation and a rising serum creatinine concentration on follow up are common occurrences. Various diagnostic possibilities are considered in this situation but the differential diagnostic process may be hampered by the non specificity of the clinical findings and laboratory results. We have developed a Bayesian classification of diagnostic categories in this setting to facilitate diagnosis. This model has shown a 96.3% accuracy when compared to the clinical diagnosis. It also offers great potential in the systematic analysis of various diagnostic elements used in the follow up of renal transplant recipients.