MedRecon — AI-Assisted Medication Reconciliation

At patient discharge, pharmacists must manually cross-check three separate medication documents — inpatient orders, discharge summary, and prescription — while simultaneously counseling the patient. In cardiovascular care, patients routinely carry 8–15 medications. The process was time-intensive, unstandardized, and produced no structured data for quality review.

MedRecon automates the comparison of the three document sources and surfaces discrepancies for pharmacist review. The system classifies each drug entry across nine discrepancy categories, handles cardiovascular-specific edge cases (IV-to-oral transitions, therapeutic class switches, ICU drug discontinuation), and produces a verified, auditable record of every reconciliation.

Pharmacists remain the decision-makers. The AI handles pattern recognition; every discrepancy requires explicit pharmacist approval or a documented override before the case is marked complete.

• check_circle20 pharmacists, 13 wards in active pilot
• check_circleEstimated 5–10 min saved per discharge review
• check_circleFull audit trail + ward-level discrepancy dashboard for monthly quality reporting
• check_circleStandardized documentation replacing ad-hoc manual process

Built on Google Workspace infrastructure for zero-friction hospital deployment. Chose a lightweight AI provider for cost efficiency at pilot scale. Kept pharmacists in the loop by design — the tool augments clinical judgment, it does not automate it.