Diagnostica--a Bayesian decision-aid system--applied to hypertension diagnosis.

Abstract

"Diagnostica" is a Bayesian statistical tool designed to collect and store the patient's data, suggest a diagnosis, and explain the decision in terms of density distributions. The program is written in C language on MacIntosh support. It is described using the case study of differential diagnosis between essential and secondary hypertensions (i.e., fibrodysplastic renal artery stenosis, atheromatous renal artery stenosis, Conn's syndrome, nephropathy and pheochromocytoma). Seventeen experimental parameters were taken into consideration, all of them available during the first medical examination. The density distributions of all items were established from the ARTEMIS experimental database. Both a priori probabilities of different types of hypertension and loss coefficients are taken into account in the calculations. Diagnostica can be used in a "make diagnosis" mode or in an "edition" mode. In the first case it can serve a physician in everyday practice; in the second it becomes a tool for medical research.