Bayesian networks for antibiotics prescription

Deals with the use of a Bayesian network as a computer-aided decision system. We have constructed a belief network for antibiotic prescription assistance in the case of urinary infections. The clinical diagnosis must have been made before using the belief network and the expertise is devoted to improving the choice of antibiotics. We compare our results with those given by human experts specialized in antibiotherapy. It proves to be quite efficient: in 29 clinical cases, the concordance kappa coefficient is as high as 0.67. This system can help the non-expert physician with antibiotherapy in non-trivial cases.