Early experiences from a guideline-based computerized clinical decision support for stroke prevention in atrial fibrillation

Atrial fibrillation (AF) affects 1-2% of the population. Twenty percent of all strokes are caused by AF. In this study, we represented the clinical knowledge in the European Society of Cardiology guideline using Guideline Definition Language (GDL), a format that binds openEHR archetypes, rule languages, and reference terminologies together. The computerized guidelines are applied to electronic health record (EHR) data retrospectively in order to identify possible gaps between current clinical practice and optimal care recommended by the evidence-based guidelines. Treatment compliance is checked in two patient groups: one received treatment from a cardiologist who is responsible for AF treatment in the region (n=514) and the other was a much larger patient group from the whole region (n=8130). The compliance checking shows the cardiologist group has substantially higher percentage of compliant treatment compared with that of the general population group. Based on this important finding, we are now implementing at-point-of-care clinical decision support reusing the same computerized guideline knowledge in GDL format in order to increase the guideline adherence of the treatment.