Polk & Johnson County: two counties, one Medicaid ledger

A local panel over three snapshots: the state’s county-month Medicaid payments, county budget levies, hospital DSH examinations, and the IRS 990 / federal single-audit records of the nonprofit providers who actually deliver Medicaid-funded care. County hospital (Broadlawns) meets university hospital (UIHC); Des Moines’s provider ecosystem meets Iowa City’s — with Adair County alongside as the state’s per-eligible outlier: the highest Medicaid cost per enrolled member in Iowa, spread over a denominator of only ~1,500 eligibles. Small-county per-eligible extremes are usually facility mix (a single nursing or ICF facility dominates a tiny base); confirming that for Adair specifically needs the facility-level LTC vendor dataset, the natural next snapshot.

chart scale:

Medicaid payments attributed to each county

Rolling 12-month sums from Iowa Data Hub dataset 973. County attribution has weakened over time (a growing share books to a “Central Office” pseudo-county), so levels understate resident spending increasingly; trends within a county remain meaningful.

The county mental-health levy, and its abolition

County debt service (context for fiscal capacity)

Annual budgeted debt-service payments from the same county budget dataset \u2014 the cash cost of county borrowing, shown to normalize the health-spending lines against each county\u2019s fiscal obligations. Outstanding principal (bond balances) is a different number, published in the Iowa DOM Outstanding Obligations reports.

County budgeted MH/DD service-area expenditures (hub dataset 939). SF 619 (2021) phased county mental-health levies out in favor of state funding: statewide county MH budgets fell from ~$134M (FY2020) to zero (FY2026). By FY2023 the statewide line is Polk — the last county budgeting MH — reaching zero in FY2026. The money moved to state appropriations; this chart is the county half of that handoff.

Broadlawns vs UIHC: the DSH examination trail

Where the registered money sits

Circle area \u221d latest IRS-990 total revenue (organization-wide \u2014 not Medicaid-specific), placed at each org's registered (IRS) address, which is a headquarters, not service-delivery sites; multi-site orgs collapse to one point. Gray crosses are city centers for orientation. Geocoding: U.S. Census Bureau geocoder.

The provider nonprofits: 990s and single audits

Why is Adair County such an outlier?

990 revenue growth since 2019 — leads for reading the filings

filter:

Ratios compare the last pre-2020 fiscal year filed to the latest filing. Growth is a lead, not a verdict: this window contains CCBHC prospective payment (2024), HCBS rate increases, ARPA-era grants, and pandemic programs — the filings themselves say which. Every link opens the organization in ProPublica's free Nonprofit Explorer (full 990 PDFs and XML); bulk data: IRS 990 e-file downloads.

Reading this honestly

990s are organization-level. An org’s revenue covers every program and location it operates, not a facility and not a payer split; Medicaid revenue appears only where the org itemizes it. Single-audit “Medicaid” is narrower than Medicaid revenue: the SEFA counts federal awards expended, and ordinary Medicaid patient-care payments are not awards — so $0 in the 93.778 column is expected for providers whose Medicaid money is service revenue, and the nonzero entries (waiver- and grant-shaped flows) are the interesting part. Orgs under $750K/yr federal spend file no single audit at all. University of Iowa has no separate single audit — state universities are audited inside the State of Iowa’s statewide single audit, so its absence from the federal database is structural, not suspicious; UIHC’s Medicaid trail here is the DSH examination schedule (as “Clinics”, a name truncation) plus its Auditor-of-State financial audits. The two DSH measures disagree: the examinations’ “DSH payments” (~$3.2M for Broadlawns) and the hospital’s own audited “DSH payment” figures ($8.5–10.3M/yr) are different measures of different slices; they are shown separately and never summed. Flagged rows (⚑) carry unresolved entity-match questions — treat them as leads, not facts.

Sources: Iowa Data Hub datasets 973 & 939 (CC BY 4.0); Iowa Auditor of State DSH examinations SPRY 2017 & 2021; IRS Form 990 via ProPublica Nonprofit Explorer; Federal Audit Clearinghouse. Refresh scripts and the statewide explorer: iowa-medicaid.html.