HEDIS Abstraction

Quality & StarsAsync jobs required

Chart abstraction for HEDIS / Stars measure compliance with verbatim evidence and audit.

HEDISStarsQuality measuresGap closureAbstraction

About

Abstracts clinical charts against a curated HEDIS/Stars measure catalog (MY2026). A map-reduce pipeline reads the note, matches numerator events and exclusions to each measure's value sets, and returns a per-measure determination — compliant, gap, exclusion, escalated, or not-eligible — with verbatim evidence, confidence, and an audit verdict.

Member demographics drive deterministic denominator eligibility, and a claims-side administrative baseline enables documentation-only gap-closure detection (supplemental data).

How it works

  1. 1Member + chart upload (or pasted text) + optional measure scope
  2. 2Deterministic denominator eligibility per measure
  3. 3Map-reduce abstraction: numerator / exclusion matching with evidence
  4. 4Audit pass verifies each determination is evidence-supported

Intended use

  • •Retrospective HEDIS / Stars chart abstraction and gap closure
  • •Supplemental-data capture for documentation-only closures
  • •Quality worklists by measure and gap status

Key outputs

  • ▸determinations[] — measure_id, status, numerator_event / exclusion, evidence, confidence, audit
  • ▸supplemental_data[] — documentation-only closures with codes
  • ▸summary — compliance_rate, gap / escalated measures, audit_pass_rate

Model comparison

F1 on Gwen's healthcare benchmark for this task — the Gwen pipeline vs the prompt-optimized model alone, with the uplift the pipeline adds, per model.

#ModelGwen pipelineModel onlyUplift
1
Claude Opus 4.8Best
0.923
0.905+0.018
2
GPT-5.5
0.910
0.854+0.056
3
Gemini 3.5 Flash
0.890
0.837+0.053

Endpoints

Try each endpoint with your signed-in session — usage counts toward your monthly budget.

Use synthetic data only. Do not submit real patient records or PHI when testing endpoints.

Limitations & caveats

  • –Value sets are core-code approximations of the licensed NCQA VSD — illustrative, refresh annually
  • –Denominator eligibility needs member age/sex; without numerator evidence a measure is a gap
  • –Multi-pass LLM (1–3 minutes); the async /jobs flow is mandatory for uploads