Medicaid Provider Spending Analysis

A project by Scholar Fund / Unify. We built this to understand how people have been navigating Medicaid — where the system is working, where it's falling short, and what can improve so that people have real access to the care they need.

Analysis of 1.08 trillion dollars in Medicaid provider spending data from HHS, covering 615,000+ providers and 10,800+ procedure codes from January 2018 through October 2024. This dashboard explores where the money goes, who it reaches, and where the system shows signs of strain.

Provenance

No Results

Methodological Guardrails

No Results

Total Paid ($B)

1,076.5

Total Rows

223,928,763

Billing Providers

615,655

HCPCS Codes

10,870

What the Data Tells Us

Three headline findings from 227 million rows of Medicaid claims:

  • [Descriptive, High confidence] Home health spending grew ~86% since pre-COVID. This is the dominant story in the data — the shift from institutional to community-based care is real, massive, and accelerating.
  • [Descriptive, High confidence] ER visits account for ~42% of combined ER + primary care spending. When people use emergency rooms instead of doctors, the system is failing at the most basic level. Every ER visit for a condition a primary care doctor could handle costs 3-10x more and delivers worse outcomes.
  • [Descriptive, High confidence] Every service category is higher post-COVID than pre-COVID — in both spending and beneficiary counts. The pandemic didn't shrink Medicaid; it permanently expanded what the system is asked to do.

Citation: headline_home_health (source medicaid.era_comparison), headline_er (source medicaid.er_vs_primary_care), and by_category (source medicaid.monthly_by_category).

Monthly Spending Trend

Loading...
View SQL (`monthly`)
SELECT * FROM medicaid.monthly_totals
WHERE month <= '2024-10'

The April 2020 dip is COVID shutdowns. The recovery was fast — by mid-2021, spending exceeded pre-COVID levels and has kept climbing. Monthly spending grew from roughly 9B in early 2018 to 17B+ in 2024, nearly doubling in six years.

Loading...

Spending by Service Category

Loading...
View SQL (`by_category`)
SELECT * FROM medicaid.monthly_by_category
WHERE month <= '2024-10'

The "Other" category dominates because most HCPCS codes don't fall neatly into our clinical groupings — it includes DME, pharmacy, transportation, home and community-based waiver services, and hundreds of other service types. The named categories represent the analytically interesting slices.

Top 25 Procedure Codes by Total Spend

Loading...

T1019 (personal care / home health aide, billed per 15-minute unit) towers over everything else at 120B+ — more than the next three codes combined. The top of this list is almost entirely home and community-based services, with office visits (99213, 99214) and ER visits (99283, 99284, 99285) the main non-HCBS entries.

No Results
View SQL (`top_codes`)
SELECT * FROM medicaid.top_hcpcs


Explore the Analysis

Geographic Analysis

State-level maps of spending, cost-per-beneficiary, provider capacity, and ER reliance. See where the money goes and where the system is stretched thinnest.

How Medicaid utilization shifted before, during, and after COVID. Era-by-era comparisons, category-level breakdowns, and the growth stories that matter.

Access Signals

Is the system actually reaching people? ER reliance vs primary care, provider capacity trends, and the gap between enrollment growth and provider supply.

Care Gaps

Where the system falls short on treatment and follow-through. Preventive vs reactive spending, mental health treatment depth, substance use service trends, and care continuity metrics.

Provider Concentration

Where is Medicaid access fragile? Market concentration by service category, top providers by spending, and what happens when a small number of providers dominate.

Concerns & Caveats

What this data can and cannot tell us. Data quality notes, known limitations, missing demographics, and the interpretive guardrails you should keep in mind.

Policy Impact Scenarios

State, county, and entity-level stress tests for coverage-loss planning. Includes risk index monitoring, operational readiness KPIs, and role-specific action playbooks.

Methods, Citations & Reproducibility

See the claim-to-metric citation map, SQL source file references, and step-by-step commands to reproduce the outputs from scratch.

Reproduce This Page

cd dashboard
export EVIDENCE_SOURCE__medicaid__token="<your_motherduck_token>"
export EVIDENCE_SOURCE__medicaid__database="medicaid"
npm run sources
npm run build
npm run preview
# then open http://localhost:3000/