Human Factors Methods to Support the Experts' Review of Automatically Detected Adverse Drug Events

2 Abstract. The European project Patient Safety through Intelligent Procedures in Medication (PSIP) aims at semi-automatically identifying and preventing Adverse Drug Events (ADEs). Data mining of the structured hospital data bases provides a list of potential ADEs, along with their frequencies and probabilities. Once a set of ADEs has been detected by data mining techniques, it is necessary to have them validated by human experts. This paper presents the methods used to support the review by clinicians and pharmacologists of these automatically detected ADEs. We use think-aloud methods and portable labs to track and record the experts reasoning and their reviewing cognitive procedures. We present preliminary results obtained with this method, which allows identifying the key data and information used to characterize the ADEs.