Revenue Cycle Distress at Independent U.S. Hospitals

Prepared by Diego Armas Morales, Founder & Director of Research, Synergize AI.

Abstract

Objective. Characterize the financial state of independent (non-system) PPS hospitals in the United States as of FY2024, with attention to geographic and Medicare-Advantage-penetration gradients.

Methods. Dual-source-verified universe of 518 independent PPS hospitals (CMS PECOS + AHRQ Compendium consensus, non-CAH/non-REH, balanced FY2021–FY2024 HCRIS coverage; 50 states + DC). Per-hospital metrics computed from public CMS HCRIS filings. Geographic enrichment via Census Geocoder, USDA Rural-Urban Continuum Codes 2023, HRSA HPSA and AHRF 2024–2025. Medicare Advantage penetration computed at the county level from CMS Medicare Monthly Enrollment (January 2026 snapshot). IRS Form 990 Schedule H Part III bad-debt cross-regime check on the strict-aligned subset (n=32).

Results. FY2024 universe-median paid-claim ratio 0.369; Days in AR 101.6 (~2.5× HFMA well-managed-practice band). 77% of the universe operated at a net loss. 45.2% of hospitals with both metrics populated (151 of 334) entered FY2024 in compound distress (negative operating margin AND DAR > 90). Metro/nonmetro DAR gradient: 82.4 days metro median vs 133.0 days nonmetro median (+51 days). MA-penetration quartile gradient: 89.4 days bottom-quartile vs 105.0 days top-quartile median DAR (+15.6 days). IRS Schedule H bad debt is 1.84× larger than HCRIS S-10 line 30 bad debt at the median hospital across the strict-aligned subset.

Conclusions. Independent PPS hospital revenue-cycle distress is structural, geographically concentrated outside metro counties, and measurably tracks MA penetration. The two reporting regimes (HCRIS S-10 vs. IRS Schedule H) are quantifying related but non-equivalent constructs.

The headline figure for FY2024 is not 101.6. It is 133.0.

Among 518 dual-source-verified independent (non-system) PPS hospitals analyzed from CMS HCRIS filings for FY2021–FY2024, the universe-median Days in AR was 101.6 days, already roughly 2.5× the HFMA well-managed-practice band. Outside metro counties, the median independent PPS hospital is waiting 133.0 days to collect cash already earned, 51 days longer than its metro peers, and 55% of nonmetro hospitals enter FY2024 in simultaneous margin loss and AR-aging distress.

The revenue cycle crisis at independent hospitals

Independent PPS hospitals (not large systems, not critical access facilities) have absorbed four consecutive years of falling collection rates, rising bad debt, and slowing cash velocity. The FY2024 picture is not a cyclical trough. It is a structural baseline. The median hospital in this analysis has never, across the four years of public filings examined here, collected more than 40 cents per gross charge dollar. The distribution is skewing worse outside metro counties.

Key takeaways

1. Collection rates have compressed structurally, not cyclically. The universe-median independent PPS hospital collected 36.9 cents per dollar of gross charges in FY2024, down from 39.7 cents in FY2021. The most distressed cohort, the 102 hospitals where the paid-claim ratio has collapsed below 0.25, collected a median of 19.8 cents. This level of compression reflects payer mix deterioration, denial accumulation, and contractual adjustment rates that have outpaced charge growth over the four-year window. No operational efficiency program recovers that gap. The issue is structural.

2. Cash velocity is the diagnostic metric most CFOs are not watching closely enough. Universe-median Days in AR rose from 104.5 days in FY2021 to 101.6 days in FY2024, with a peak of 107.0 days in FY2023. The HFMA-cited benchmark for well-managed systems is 35–50 days. The universe-median runs at roughly 2.5× that benchmark; the 75th percentile sits at 183.6 days in FY2024, six months of services delivered but uncollected. Outside metro counties, the median hospital is waiting 133.0 days, 51 days longer than its metro peers.

3. Compound distress, defined as simultaneous margin loss and slow AR, applies to nearly half the universe and a clear majority of nonmetro hospitals. Of the 334 hospitals with both Days in AR and operating margin populated for FY2024, 151 (45.2%) entered FY2024 with a negative operating margin AND Days in AR exceeding 90 simultaneously. In nonmetro counties (RUCC tiers 4–9), that share climbs to 55%; in metro counties (RUCC 1–3), 38%. This is not a cash-flow management challenge. It is a structural failure in which the cash needed to fund operations cannot be collected at the rate operations consume it. The hospitals in this category are not trending toward distress; they are in it, and the geography of where they sit is itself a finding.

4. Medicare Advantage penetration tracks measurably with AR aging. Hospitals in counties where Medicare Advantage holds the top quartile of beneficiary share (median 64.9% MA penetration) carry a median 105.0 days in AR, 15.6 days longer than hospitals in the bottom MA-penetration quartile (median 37.0% MA, 89.4 days in AR). That gradient is consistent with the published 17.3% MA inpatient denial rate documented in Health Affairs (June 2025).

What are hospitals actually collecting per dollar billed?

The paid-claim (PC) ratio measures what a hospital actually collects against what it charges. A ratio of 0.37 means 63 cents of every gross charge dollar is absorbed by contractual adjustments, denials, bad debt, and uncompensated care. It is the single most compressed summary of revenue cycle performance available in public cost report data.

The FY2021–FY2024 decline of 2.8 percentage points represents structural compression, not a single-year anomaly. The progression is monotonic; each year is worse than the last on the universe median.

The Collection-Compressed cohort is the most diagnostically significant grouping in the dataset. A median PC ratio of 0.198 means the median Collection-Compressed hospital is collecting 19.8 cents per dollar of gross charges. At that level, no operational efficiency program recovers the gap. The issue is structural: payer mix deterioration, denial accumulation, and contractual adjustment rates that outpace charge growth.

PC ratio trajectory (FY2021→FY2024, balanced-panel hospitals):

• 38% of hospitals declined more than 2 percentage points
• 27% improved more than 2 percentage points
• 34% held within a 2-point range

About 1.4× as many hospitals deteriorated as improved. The minority that improved did so against the macro trend, a finding that warrants its own analysis.

How long are hospitals waiting to collect?

The HFMA benchmark exists. The distance from it is the finding.

Days in AR measures how long it takes a hospital to convert a billed service into collected cash. The HFMA-cited benchmark for well-managed systems is 35–50 days. The hospitals in this analysis do not operate near that benchmark, and the spread by geography is wider than the spread by clinical cohort.

The universe-median hospital carries AR roughly 2.5× the well-managed benchmark. At the 75th percentile of 183.6 days in FY2024, a hospital is waiting six months to collect on services already delivered.

The Collection-Compressed cohort’s DAR median of 162.3 days indicates that hospitals with collapsed paid-claim ratios are simultaneously carrying AR beyond 5 months. The Persistent-Loss cohort’s 126.7 days reflects the same dynamic at a less-collapsed PC level. These are not cash-flow problems. They are structural revenue failures.

DAR trajectory (FY2021→FY2024):

• 50% of hospitals worsened by more than 5 days
• 36% improved by more than 5 days
• 14% held within 5 days

The DAR trajectory across these four years is asymmetric. Half the universe deteriorated by more than five days; just over a third improved. A hospital that held steady year over year is, by the math of the trajectory distribution, performing above the median. The mathematical center of gravity in this dataset is downward; not by a single shock and recovery, but by continuous compression that the universe-median absorbs slowly enough to obscure year to year and visibly enough to read across four.

How widespread is operating at a loss?

The FY2022 margin compression, from −9.9% to −11.8%, reflects the post-pandemic cost normalization documented in national hospital finance data: labor costs rising, payer mix shifting toward public payers, and Medicare Advantage denial rates accelerating. The FY2024 median of −9.5% represents marginal improvement from FY2022–2023, but 77% of the universe is still in the red. Of these three pressures, the MA denial-rate acceleration is the least visible and the hardest to reverse: it does not show up as a cost line; it accumulates as AR that does not convert.

The Persistent-Loss cohort reached 100% operating at a loss by FY2024, with a median margin of −18.5%. Every hospital in this group is spending more than it collects.

How much revenue is permanently written off each year?

Bad debt at a PPS hospital is partially recoverable. Medicare reimburses 65% of allowable bad debt from cost reports, but only on qualifying Medicare deductible and co-insurance amounts, and only after proper collection efforts are documented and the debt is deemed uncollectible. The delta between what a hospital writes off and what it recovers through Medicare bad debt reimbursement represents a permanent revenue gap.

Per-hospital median bad debt grew 14.2% from FY2021 to FY2024 in the comparable subset; aggregate grew 17.8% from $750M to $883M. Both figures are smaller than v1’s previously reported 28% growth. The difference reflects the worksheet redefinition between FY2022 and FY2023, the subset restriction, and a corrected line-30 extraction (see the audit-finding subsection in the methodology).

Counter-intuitively, the Bad-Debt-Exposed cohort shows the lowest median bad-debt magnitude. The cohort definition keys on bad-debt-as-a-share-of-revenue, not absolute dollars. Smaller rural hospitals with the highest relative bad-debt exposure carry smaller absolute dollar amounts.

Which hospitals face multiple simultaneous failures?

Of the 334 hospitals in this universe with both metrics populated for FY2024, 151 (45.2%) entered the year carrying both a negative operating margin and Days in AR exceeding 90 days simultaneously.

This is the compound-distress threshold. A hospital below zero on margin and above 90 days on AR is not managing a single constraint; it is managing a system failure. The cash needed to stabilize operations is locked in AR that cannot be collected fast enough to fund the operations generating it.

Most denial root causes originate at the front end of the revenue cycle (scheduling, registration, prior authorization), not in the billing and collections team that inherits them. The back end sees the symptom. The front end generated it. Hospitals with denial rates above 15% and no upstream feedback loop from patient financial services to registration will not solve the problem by adding denial-management headcount.

The economics of denials compound this. Industry estimates place the cost of reworking a denied claim at roughly $25, about four times the cost of filing it initially, and a meaningful share of denied claims are never resubmitted at all. A hospital carrying 100+ days in AR with a denial rate above 15% is not slow to collect. It is permanently writing off a predictable share of billable revenue every operating period.

Where is the distress concentrated?

Outside metro counties, the median independent PPS hospital is 51 days slower to collect, 6.6 percentage points further below break-even, and 17 percentage points more likely to be in compound distress. The pattern is a consequence of the fixed-cost mathematics of sparse markets, compressed against a payer mix that has hardened toward public-payer concentration over the four years examined here.

The nonmetro DAR median of 133.0 days is the central operational number in this report. A hospital running 130+ days of AR is collecting on care delivered four months ago and funding present-tense operations from cash already tied up in the rear-view mirror. Within nonmetro, the worst-performing tier is RUCC 6, peri-urban towns of 5,000–20,000 population adjacent to a metro area: 75% compound-distress prevalence, median 191.8 days in AR, median operating margin −17.3% (n=32). These hospitals appear to face the disadvantages of nonmetro economics without the smaller-cost-base offset of deeper rural operations.

Medicare Advantage penetration tracks with AR aging

Hospitals in the highest MA-penetration quartile carry 15.6 more days in AR than hospitals in the lowest. The mechanism is documented: MA inpatient denial rates run at 17.3% of initial claim submissions versus approximately 8% for traditional fee-for-service Medicare (Soto, Hicks, Chernew, Health Affairs, June 2025; doi:10.1377/hlthaff.2024.01485). The MA denial-rate gradient does not stay at the claim-submission stage. It accumulates downstream as AR that does not convert.

The Q1 → Q4 spread on Days in AR is 15.6 days. The shape is non-monotonic at Q4 (Q3 peaks at 111.9 days, Q4 dips to 105.0 days), possibly reflecting saturation effects or that the largest MA plans negotiate better payment-timing terms with hospitals in their core markets. The directional gradient is the headline.

The geography of distress and the geography of payer mix are not separable. Nonmetro counties carry both more rural-economy disadvantages (sparser volumes, thinner administrative capacity, weaker negotiating leverage) and, on average, higher MA penetration in the counties where the most distressed hospitals sit. Each pressure compounds the other. A hospital reading these benchmarks against its own cost report should not ask which gradient is primary. It should ask how many of these gradients its specific service-area position currently exposes it to simultaneously.

How do distress patterns differ by hospital cohort?

Bad-Debt-Exposed cohort (n=127, 24.5% of universe)

The largest cohort. Defined by PC ratio ≥ 0.40 paired with either DAR > 75 days or operating margin < −5%. Median PC ratio of 0.531 is the highest of the four primary cohorts; these hospitals are collecting at a comparatively healthy rate, but bad-debt-to-revenue exposure is what drives the cohort assignment. Mostly small rural and frontier-rural hospitals (median bed count below the universe median). Medicare bad debt reimbursement mechanics (the 65% recovery rule, S-10 exhibit compliance, and the distinction between traditional, crossover, and indigent bad debt) are the primary cost-report optimization levers for this group.

Collection-Compressed cohort (n=102, 19.7% of universe)

Defined by PC ratio < 0.25. Median PC ratio of 0.198 in FY2024. These hospitals are collecting less than 20 cents per gross charge dollar, a level that reflects either severe payer mix deterioration (high Medicare Advantage penetration with elevated denial rates), historical charge master inflation without corresponding collection rate improvement, or both. The DAR median of 162.3 days indicates severe AR aging on top of the collection compression. The Health Affairs (June 2025) Medicare Advantage denial-rate finding applies directly: each percentage point of denial rate increase amplifies the collection shortfall in ways that operational efficiency programs cannot recover.

Persistent-Loss cohort (n=89, 17.2% of universe)

Defined by PC ratio in [0.25, 0.40) AND operating margin < −5%. 100% of this cohort operated at a loss in FY2024. Median margin of −18.5%, the deepest of any cohort. This group is not recovering. The trajectory is straight deterioration across every metric, and the PC ratio band itself (collecting between 25 and 40 cents per gross charge dollar) points the diagnosis at contract structure, not collection operations. There is no billing-operations path out of −18.5% margin under that PC range. The operational lever is contract renegotiation or payer exit.

Below distress thresholds (residual; n=114, 22.0% of universe)

Hospitals in the universe that meet the dual-source-verified independence and balanced-panel criteria but do not trip any of the three primary distress thresholds. Median PC ratio of 0.440, median DAR of 67.6 days, median operating margin +2.2%, only 34% at operating loss. This is the comparison cohort. It documents the financial profile of an independent PPS hospital outside observable revenue-cycle distress. That this cohort is barely larger than the Bad-Debt-Exposed cohort alone, and roughly a third the size of the three distressed cohorts combined, is itself a finding.

Insufficient data (n=86, 16.6% of universe)

Hospitals with HCRIS coverage that does not populate the threshold metrics cleanly enough to assign to one of the three primary distress cohorts or to “below distress thresholds.” Reported separately for transparency.

Considerations for hospital leadership

These findings are drawn from public filings. They are not projections. They describe a population at a point in time. What they reveal is a small set of structural conclusions that follow from the data regardless of where any single hospital sits within the universe.

A PC ratio below 0.35 is a contract problem, not a collection problem. The compression documented in this universe is predominantly contractual (contract terms, denial rates, charge master architecture) rather than back-office execution. Adding AR follow-up staff to a structural contract shortfall is a recurring cost with no structural resolution. The diagnostic question is whether the contracts the back office is working under are collectible at the universe-median PC level, or below it.

AR above 90 days is a front-end problem before it is a back-end problem. The denial root causes sit at scheduling, registration, and prior authorization, not in the collections workflow that inherits them. A hospital carrying AR above 90 days with denial rates above 15% and no upstream feedback loop from patient financial services to registration will not reduce AR by adding denial-management headcount. The diagnostic question is whether any signal travels back from the AR team to the registration desk that generated the denial.

Compound distress (negative margin and AR above 90 simultaneously) has a fixed priority sequence: AR recovery before operational cost cutting. A hospital cutting costs while AR remains above 90 days is reducing the numerator of its margin calculation while the denominator, uncollected billings, compounds. Cash recovery from existing AR is the single fastest path to operational stability for a hospital in this category.

Nonmetro independent PPS hospitals do not face a Medicare Advantage problem, a denial-rate problem, and a payer-mix problem in isolation. They face all three simultaneously, against a smaller revenue base, with thinner administrative capacity. The cost-report-mechanic levers (Medicare bad debt reimbursement, wage index reclassification, DSH qualification, 340B status review) are disproportionately under-pursued in this segment relative to the dollar value at stake, and the dollar value at stake is documented in the universe-median benchmarks above.

Methodology

What follows reproduces every figure in this report deterministically from a frozen set of public analytical inputs. Cost-report mechanics are stated as facts where they are facts. Regime differences and inference limits are flagged where they exist.

Relationship to prior work

This report sits adjacent to three contemporary priors. Williams, Davlyatov, Bowblis, and Braun (Health Affairs Scholar, November 2025; doi:10.1093/haschl/qxaf220) ask what would happen to U.S. acute-care hospital margins if the Medicare bad debt reimbursement program (currently 65% of allowable bad debt) were eliminated. Their counterfactual is modeled on FY2022 HCRIS Worksheet S-10 (Row 27 Col 1) across 4,106 short-term general acute hospitals, finding that elimination would compress total margin by 0.30 percentage points among CAHs, 0.25 among other rural hospitals, and 0.20 among urban; 42 hospitals would flip unprofitable. We share the data layer (HCRIS S-10) and inherit their scope-discipline precedent on the bad-debt extraction, but we ask a different question: how distressed independent PPS hospitals are right now under the current reimbursement regime, on a different cohort of 518 dual-source-verified independent PPS hospitals.

Cataife and Liu (Health Economics Review, 2025; doi:10.1186/s13561-025-00599-7) model the relationship between county-level Medicare Advantage penetration and rural hospital Medicare inpatient days and revenue. We measure a different downstream consequence of the same MA mechanism: the cross-tab between county MA penetration and hospital Days in AR. Same mechanism layer, different outcome (revenue cycle aging rather than utilization).

Soto, Hicks, and Chernew (Health Affairs, June 2025; doi:10.1377/hlthaff.2024.01485) document MA’s 17.3% initial inpatient denial rate. We trace the downstream consequence (hospital AR aging) at the population level for independent PPS hospitals. Their paper supplies the explanatory mechanism for our MA-quartile cross-tab; our paper supplies population-level evidence that the mechanism clears as a measurable revenue-cycle gradient.

Our four first-publisher contributions extend this literature into territory none of the three has reached: the metro/nonmetro DAR gradient (+51 days), the MA-penetration quartile × hospital AR cross-tab (+15.6 days), the IRS Form 990 Schedule H × HCRIS S-10 cross-regime audit (1.84× per-hospital median), and the empirical Transmittal 18 worksheet-rename hospital-level crosswalk applied to the universe.

Population definition and dual-source independence consensus

The research universe is 518 hospitals that satisfy three criteria simultaneously:

• Independent on dual-source consensus, determined via cross-check against two regulatory regimes: CMS Provider Enrollment, Chain & Ownership System (PECOS) Hospital Enrollments (February 2026 release) joined to PECOS Hospital All Owners (March 2026 release); cross-checked against the AHRQ Compendium of U.S. Health Systems, 2023 release. A hospital is tagged Independent only when both sources concur on the absence of multi-hospital affiliation. Hospitals appearing in only one source, or for which the sources disagree, are excluded.
• Non-Critical-Access (CAH = False) and non-Rural-Emergency-Hospital (REH = False) per PECOS provider type.
• Balanced four-year HCRIS coverage: FY2021, FY2022, FY2023, and FY2024 all present in the panel.

The universe spans 50 states and the District of Columbia, a deliberate scope choice that makes universe-median benchmarks defensible against population-bias arguments.

Cohort definitions

Cohorts are defined by HCRIS thresholds applied to the universe:

• Bad-Debt-Exposed: PC ratio ≥ 0.40 AND (DAR > 75 OR operating margin < −5%)
• Collection-Compressed: PC ratio < 0.25
• Persistent-Loss: PC ratio in [0.25, 0.40) AND operating margin < −5%
• Below distress thresholds (residual): all remaining hospitals

Metric definitions

• Paid-claim (PC) ratio: net patient revenue ÷ gross patient charges, calculated from Worksheet G-3 data. Values > 1.0 excluded as data anomalies.
• Days in AR: (AR balance ÷ net patient revenue) × 365, derived from Worksheet A and G-3. Values > 365 excluded as anomalies.
• Operating margin: operating income ÷ net patient revenue, from Worksheet G-3. Values outside [−1.0, +1.0] excluded.
• Bad debt: total bad debt expense from CMS Worksheet S-10 line 30 col 1 (entire facility, all payers).

Bridging CMS Transmittal 18

CMS Transmittal 18 reorganized Worksheet S-10 between FY2022 and FY2023, replacing the pre-T18 worksheet code S100000 with the post-T18 code S100001. Within both worksheet codes, the line numbers for the bad-debt-relevant lines are identical: line 30 col 1 (“Total bad debts entire facility, all payers”) and line 31 col 1 (“Total cost of charity care AND bad debt to the entire facility”). The bridge applied here uses line 30 col 1 throughout the four-year window with the worksheet-code crosswalk. The comparable subset is restricted to 251 hospitals with bad-debt populated in all four FYs to insulate the time series from coverage-expansion artifacts. Per-hospital median is the lead metric; aggregate is secondary. Source: CMS Provider Reimbursement Manual; Transmittal 18 documentation at cms.gov/files/document/r18p240.pdf. Williams et al. (Health Affairs Scholar, November 2025) work the same data plane on FY2022 filings (Worksheet S-10 Row 27 Col 1, the reimbursable Medicare bad debt allowance) and provide the immediate scope-discipline precedent for the line-30-vs-line-31 audit finding documented below and the T18 worksheet-rename crosswalk applied here.

Audit finding (line 30 vs. line 31)

During the 2026-05-07 rebuild, the prior extraction’s bad-debt column was found to be sourced from line 31 col 1, which CMS labels “Total cost of charity care AND bad debt to the entire facility” (i.e., total uncompensated care). The conceptually correct bad-debt measure is line 30 col 1 (“Total bad debts entire facility, all payers”). This rebuild uses line 30 col 1 throughout for all bad-debt figures and discloses the prior extraction error transparently. Bad-debt magnitudes in this report are not directly comparable to any earlier Synergize AI publication that referenced the previous column.

IRS 990 Schedule H cross-regime check

A first-of-its-kind cross-regime check was performed: for the 99 universe hospitals identified as 501(c)(3) nonprofits per HCRIS Worksheet S-2 Type-of-Control with EINs successfully matched against the IRS Exempt Organizations Business Master File, FY24-aligned Form 990 filings were extracted from IRS bulk XML archives and Schedule H Part III bad-debt-expense fields parsed. Across the 32-hospital strict fiscal-year-aligned subset, the median per-hospital ratio of IRS Schedule H bad-debt to CMS S-10 line 30 bad-debt is 1.84×, with an aggregate ratio of 2.23× and Spearman rank correlation of 0.52. The two regimes are reporting on different accounting bases (most likely cost-report-style cost basis vs. financial-statement-style charges basis, with material per-hospital practice variance). Future cross-source bad-debt research should treat the two as related but non-equivalent measures.

Geographic methodology (RUCC, HPSA, AHRF)

Each universe hospital was geocoded via the U.S. Census Geographies endpoint using the HCRIS-reported street address; hospitals where the address did not resolve fell back to the Census ZCTA-to-County 2020 relationship file via ZIP code. 509 of 518 hospitals (98.3%) resolved to a 5-digit county FIPS code. The county FIPS was joined to:

• USDA Rural-Urban Continuum Codes, 2023 release. The 2023 release changed the urban-area population threshold from 2,500 to 5,000, so RUCC 2023 figures are not directly comparable to RUCC 2013 figures.
• HRSA Primary Care HPSA designations, snapshot 2026-05-07. Designation status is time-varying; we use HPSA as a point-in-time enrichment signal, not a longitudinal one.
• HRSA Area Health Resources Files (AHRF), 2024–2025 release, for county population (2020 Census), median family income (ACS 2019–2023), share of families below poverty (ACS 2019–2023), and share of under-65 uninsured (Census SAHIE 2022).

Medicare Advantage penetration methodology

MA penetration is computed at the county level from the CMS Medicare Monthly Enrollment dataset (January 2026 snapshot, the most recent available at time of analysis) as the share of total Medicare beneficiaries enrolled in MA or other Part C plans. The 348-hospital MA × DAR cross-tab uses the universe-internal quartile boundaries documented in Exhibit 5.2. The Health Affairs paper cited as the explanatory framework (Soto, Hicks, Chernew, June 2025; doi:10.1377/hlthaff.2024.01485) is referenced for the denial-rate gradient mechanism, not as the source of the MA penetration data.

Reproducibility

All numbers in this report reproduce deterministically from a frozen, byte-verified set of nine pinned public input files. The recompute pipeline, the analytical inputs, and the cross-regime validation scripts are maintained internally and available on request to qualified researchers via press@synergizeai.io.

A note on universe scope

This rebuild materializes a research universe defined by structural criteria: independence, non-CAH, balanced four-year coverage. It is not a distress-pre-selected cohort. The benchmarks reported here describe the universe-median independent PPS hospital and the geographic and payer-mix gradients within it. They do not describe a national average across all hospital types or a distress-sample artifact. A CFO reviewing these figures should compare them to their own cost report data at the universe-median level and in the metro/nonmetro stratum that matches their own location.

Limitations and scope

The following limitations bound the interpretation of these findings. A reader applying these benchmarks to a specific hospital should treat each limitation as a hedge on transferability.

• Universe scope. The 518-hospital universe represents independent (non-system), non-CAH, non-REH PPS hospitals with balanced four-year HCRIS coverage. It is approximately 8.3% of the full HCRIS panel. Findings do not transfer to system-affiliated hospitals, critical access hospitals, rural emergency hospitals, or hospitals with gaps in their four-year filings. CFOs at hospitals outside the universe should treat the figures as adjacent context, not as a direct benchmark.
• Survivor bias from balanced-panel selection. Hospitals that closed, merged, or filed irregularly between FY2021 and FY2024 are excluded by construction. The most distressed segment of the independent-PPS population is therefore underrepresented; the 77% net-loss share and 45% compound-distress share are conservative under any sample where closures are present.
• Geographic enrichment is at county, not service-area, resolution. A single-county FIPS code obscures the catchment area that a hospital actually serves; metro-classified hospitals at the county boundary may serve nonmetro patients. The +51-day metro/nonmetro DAR gradient is therefore a lower bound on the true service-area gradient.
• Cross-regime difference is documented, not resolved. The IRS Schedule H Part III cross-regime check (n=32 strict-aligned hospitals) shows IRS bad-debt magnitudes are systematically larger than HCRIS S-10 line 30 magnitudes (median 1.84×, aggregate ratio 2.23×, Spearman ρ=0.52). This documents a real difference but does not resolve which figure to use for any specific decision; cost-report-style cost basis versus financial-statement-style charge basis is the most plausible explanatory mechanism but is not directly tested here.
• Cross-sectional design precludes causal inference. The metro/nonmetro gradient and the MA-quartile × DAR association are patterns observed in FY2024 data, not causal estimates. We make no claim that increasing MA penetration causes AR aging at the hospital level. The Health Affairs framework (Soto, Hicks, and Chernew, 2025) is the explanatory mechanism we cite; the gradient we report is consistent with that framework but does not test it.

This scope is deliberate. The intent of this report is to characterize a population at a point in time using public, reproducible data. Causal estimation, longitudinal trajectory modeling, and operating-room-level cost analysis are out of scope and would require different data infrastructure.

Acknowledgments

The author thanks the maintainers of CMS HCRIS, CMS PECOS, USDA ERS, HRSA, the U.S. Census Bureau, and IRS Statistics of Income for keeping the source data publicly accessible. The author thanks Soto, Hicks, and Chernew (Health Affairs, June 2025) for the Medicare Advantage denial-rate framework that organizes the MA × Days-in-AR analysis, and Williams, Davlyatov, Bowblis, and Braun (Health Affairs Scholar, November 2025) for the bad-debt scope-discipline precedent that shaped the methodology. The Healthcare Financial Management Association’s well-managed-practice benchmarks underpin the 35–50-day Days-in-AR target referenced throughout.

Conflicts of interest and funding

This research was self-funded by Synergize AI, Inc. The author declares no conflicts of interest. Synergize AI does not consult to, own equity in, or otherwise hold a financial relationship with any hospital, hospital system, payer, or vendor named or analyzed in this report.

Data availability

All analytical inputs are public. CMS HCRIS cost-report filings, PECOS Hospital Enrollments and All Owners files, CMS Medicare Monthly Enrollment, IRS Pub78 and EO BMF and Form 990 bulk XML, USDA RUCC 2023, HRSA Primary Care HPSA, HRSA AHRF 2024–2025, U.S. Census Geocoder, and U.S. Census ZCTA-to-County 2020 relationship file are downloadable from their respective government sources.

The byte-verified set of analytical inputs, the recompute pipeline, and the cross-regime validation scripts are maintained internally and available on request to qualified researchers via press@synergizeai.io. Re-running the scripts against the frozen inputs reproduces every figure in this report deterministically.

References

Primary cited works

• Williams, A., Davlyatov, G., Bowblis, J., Braun, R. (2025). Effects of eliminating the Medicare bad debt reimbursement program. Health Affairs Scholar, 3(11):qxaf220. doi:10.1093/haschl/qxaf220.
• Soto, P., Hicks, D., Chernew, M. (2025). Medicare Advantage inpatient initial denial rates. Health Affairs. doi:10.1377/hlthaff.2024.01485.
• Cataife, G., Liu, X. (2025). County-level Medicare Advantage penetration and rural hospital Medicare inpatient days and revenue. Health Economics Review. doi:10.1186/s13561-025-00599-7.
• Malone, T. et al. (2025). Rural hospital distress modeling via cash-flow, equity, and closure pathways. Journal of Rural Health. doi:10.1111/jrh.12882.

Macro context

• Kaufman Hall (2026). National Hospital Flash Report, January 2026.
• Chartis Group (2026). 2026 State of Rural Health.
• Aptarro (2025). Denial Rate Survey: 15.7% average MA initial denial rate.
• Moss Adams and Dallas-Fort Worth Hospital Council (2024). Denial Management Framework.
• Industry estimates on claim rework cost, MGMA and HFMA attribution (2024).
• Baker Tilly and Moss Adams (2026). Wage Index Webinar, March 2026.