Chiara Cornalba, R. Bellazzi, S. Quaglini, R. G. Bellazzi, and M. Stefanelli
The dependability of health care organizations may be improved limiting clinical risks and adverse events. This paper describes a project, implemented in a haemodialysis department, which is aimed at exploiting the results of a dialysis monitoring system to dynamically extract risk profiles for the patients and the clinical centres. These risk profiles allow to define decision support strategies able to adaptively minimize risks and improve patient safety. We focus on the problem of managing clinical risks, in terms of events which influence the risk of hospitalization and mortality and their expected costs. The developed tool takes advantage of an Incident Reporting System (Hemostat) and exploits a Bayesian network approach to estimate clinical risk and to propose therapeutic and strategic decisions.