יובל שחר

Clinical guidelines (GLs) standardize care but are complex to implement. Most clinical decision support systems (CDSSs) assume continuous use, which does not reflect real-world episodic workflows. We developed and evaluated e-Picard, a CDSS providing GL-based recommendations for episodic, on-demand consultations, enabling prospective decision support informed by retrospective quality assessment. At runtime, e-Picard analyzes offline patient data, computes fuzzy-logic-based compliance, identifies missed actions and generates context-specific recommendations. The system was applied to longitudinal geriatric data managed under pressure ulcers (PU) and diabetes (DM) GLs. Manual technical validation using 3,110 PU and 12,538 DM data instances (43 PU and 82 DM patients) achieved ≥99% correctness and up to 98% completeness. Retrospective simulation on 1,000 patients per domain (57,860 PU and 100,940 DM data instances) demonstrated potential adherence improvements from 68%-69% to 89%-97% for PU and from 14%-15% to 60%-87% for DM, with higher consultation frequencies increasing compliance and reducing care variability. These results demonstrate that episodic CDSSs can deliver accurate, context-aware support even under intermittent use.