MA
Portfolio

Area EClinical & Commercial Ops

Radiologist Peer-Review Analysis Dashboard

Quality program analytics

A comparable quality score for peer review — with the full distribution always beside it.

analysis views
5
analysis views
ratings → count-weighted GPA
1–5
ratings → count-weighted GPA

01/Problem

Peer-review data sat in operational tables that couldn't answer what a quality program asks: averages per radiologist, facility comparisons, who is under-reviewed, and scores received versus given.

02/Approach

A five-tab Streamlit surface over the peer-review schema with a standardized count-weighted GPA metric per radiologist and facility group. Dual-perspective analysis reconciles radiologists who appear as both reviewer and reviewed; review counts are joined to exam volumes so a handful of reviews isn't over-interpreted.

03/Outcome

  • Standardized quality view at radiologist, facility-group and modality level, with benchmarking.
  • Presentation-ready output for QA review meetings.

04/My role

  • Approved one comparable score as the headline metric while requiring the full distribution stay visible, so a single number can't be used out of context against a clinician.

Stack

StreamlitPostgreSQLpandas