Standalone interactive web app for an ACO executive team reviewing value-based care performance from BCDA-style Medicare claims data.
Open index.html directly in a browser for the generated medium-ACO demo, or run the local BCDA proxy:
npm startThen visit http://localhost:5174.
BCDA publishes sandbox credentials for synthetic model entities on its bearer-token documentation page. Use the advanced 10,000-enrollee synthetic dataset for a medium-sized ACO-style demo, or substitute any other sandbox/production credentials.
export BCDA_CLIENT_ID="your-bcda-client-id"
export BCDA_CLIENT_SECRET="your-bcda-client-secret"
export BCDA_BASE_URL="https://sandbox.bcda.cms.gov"
export BCDA_MAX_PATIENTS="1200"
npm startThe server keeps credentials out of browser code, requests a bearer token, starts a Group/all/$export job for ExplanationOfBenefit, Patient, and Coverage, polls until the job is ready, downloads the NDJSON files, and normalizes the FHIR resources into the dashboard model.
Without credentials, the browser app still runs against a deterministic 850-beneficiary synthetic ACO dataset.
- Executive KPIs for attributed lives, total cost, cost PMPM, ED utilization, and quality.
- Patient-level worklist with cost, HCC RAF, inpatient visits, ED visits, quality score, risk tier, and drill-down detail.
- Patient detail with quality/HCC actions and recent BCDA
ExplanationOfBenefitactivity. - Provider-level trend index comparing cost, quality, and HCC capture.
- Filters for cohort, provider, service line, and patient search.
The app uses local synthetic data, shaped around the BCDA resources an ACO would export:
Patient: demographics, identifiers, attribution.Coverage: Medicare coverage context.ExplanationOfBenefit: cost, diagnoses, utilization, provider, service dates, and claim activity.
Source references: