CMS Open Data

Dashboards over the mart layer of the cms-open-data warehouse: Medicare drug spending (Parts B and D), Medicaid drug utilization, and hospital quality, modeled with dbt on DuckDB from CMS public datasets.

Pages

• Drug spending — top drugs, multi-year trends, and a Part B vs Part D comparison for drugs billed in both programs.
• Hospital quality — overall star ratings, healthcare-associated infection performance, HRRP excess readmission ratios by condition and state, and VBP total performance scores with domain decomposition.
• Cost vs quality — Medicare inpatient payment per discharge by star rating, with case mix as the confounder.
• Prescribers — who drives spending on the top Part D drugs: specialties, top-1% concentration, and billed-vs-gross cost.
• Nursing homes — star ratings, ownership, quality measures, and health-inspection enforcement for certified nursing homes.
• Home health and hospice — Care Compare star ratings, ownership, OASIS / claims quality vs national benchmarks for home-health agencies, and CAHPS family-caregiver survey plus Hospice Item Set process measures for hospices.
• Clinicians and industry payments — Open Payments' 2024 general-payment file joined to the clinician roster, Part B utilization, and Part D prescribing: dollar concentration, top payers and specialties, and observational associations between industry payments and prescribing volume.
• Marketplace (QHPs) — PY2026 Federally-Facilitated Marketplace qualified health plans: benchmark silver premiums by state and county, premium spread by metal level, issuer concentration, the CSR deductible cliff, and stand-alone dental.
• Medicaid drug utilization — 2023 State Drug Utilization Data: national totals, the fee-for-service vs managed- care split (including which states carve pharmacy out of managed care), top drugs and labelers by reimbursement, and a program-level comparison against Medicare Part D and Part B.
• Specialty facilities — Medicare-certified dialysis clinics, inpatient rehabilitation facilities, and long- term care hospitals: chain landscape and five-star distribution for dialysis, standardized-ratio and percent-of-patient measures, freestanding-vs-hospital-unit split for IRFs, and LTCH bed capacity.
• Dialysis chains — do DaVita and Fresenius, who run roughly three-quarters of Medicare-certified outpatient dialysis facilities, deliver measurably different quality than the rest of the market? Chain-level star mix, key clinical measures, and the reporting-rate gap that conditions every headline.
• Dataset freshness — how fresh is each CMS extract, when did upstream last publish it, and what changed between weekly warehouse builds. Reads the durable vintage ledger committed to the repo.
• Data quality — dbt build results for the run that produced this deployment: per-model status, slowest builds, test-outcome rollup, and prose write-ups of parked data-model decisions (dim_drug name-only matching, no NDC bridge).
• dbt docs — browsable model/column documentation and lineage for the dbt project, generated by `dbt docs generate --static` and served alongside this site.

Warehouse coverage

The warehouse is rebuilt locally by the Dagster full_refresh_job (raw CMS extracts → dbt staging → the marts above); these pages query the marts only.