Prescription Deduplication

Overview

Prescription transfers can leave clinically equivalent records active at more than one pharmacy. That creates compliance risk, duplicate operational work, and the possibility of processing the same therapy twice.

What I Did

I led the project through three iterations and paired with another engineer across design and implementation. The work expanded from exact prescription matching to transfer-back scenarios and, later, therapeutic duplicates across different medications.

• Implemented matching rules across prescription, patient, prescriber, drug, quantity, and transfer context.
• Added API and React workflows for warnings, review, deactivation, and permission-controlled actions.
• Published events for downstream visibility and built monitoring around detection outcomes.
• Introduced explicit record links and attestations so users could resolve ambiguous cases without discarding audit context.
• Rolled out each capability as a bounded iteration instead of combining every compliance rule into one release.

Engineering Decisions

Deduplication in a regulated domain is not just a matching algorithm. The design had to expose why a record matched, preserve human review for ambiguous cases, and record the decision path for later investigation.

Outcome

The system identified more than 21,627 duplicate prescriptions in the latest measured three-month window. The product gave pharmacy teams a consistent way to detect, review, and resolve duplicate risk before it became downstream operational work.