Computerized Bayesian Evaluation of Adverse Events

Abstract

The Bayesian approach is one of three main methods that can be used in the differential diagnosis of adverse events. The other two methods are medical diagnosis and standardized decision aids. The Bayesian approach has unique advantages (eg, etiologic balancing, explicitness, flexibility, etc.) that are not seen with the other two methods, making it appropriate for unexpected, serious adverse events and complex cases where controversies in expert opinion may arise. Some benefits can be anticipated from computerization of the Bayesian approach because of the complexity of the calculations. Some preliminary efforts have been made to computerize Bayesian causality assessment and one of these programs is illustrated. Although Bayesian-based computerized diagnostic systems show potential as decision aids, continued development and testing is needed before widespread use will be possible.