Safety net hospitals lose DSH funding to bad data
Safety net hospitals lose an estimated $5.7 billion annually in disproportionate share hospital payments they cannot fully justify because the underlying data fails on provenance, recency, and concordance. The Medicaid fraction and low-income utilization rate calculations that drive DSH allotments depend on data elements that are systematically incomplete at the institutions that need the funding most.
Safety net hospitals serve a disproportionate share of Medicaid and uninsured patients. They also operate on the thinnest margins in American healthcare. The federal Disproportionate Share Hospital program exists to offset those costs, distributing approximately $24.4 billion in combined federal and state payments in fiscal year 2023. But the data that determines who gets what share of that money is riddled with gaps that systematically disadvantage the hospitals the program was designed to protect.
How DSH payment calculations depend on data
DSH eligibility and payment amounts hinge on two metrics: the Medicaid inpatient utilization rate and the low-income utilization rate. Both require accurate, timely data on patient insurance status, Medicaid enrollment verification, charges, and revenue attributed to low-income patients.
The Medicaid fraction numerator counts inpatient days for patients who were eligible for Medicaid but not entitled to Medicare Part A. The denominator is total inpatient days. A single misclassified day, a Medicaid eligibility lag, or a dual-eligible patient coded incorrectly shifts the fraction and the resulting payment.
The low-income utilization rate adds another layer. It includes the ratio of Medicaid revenue (including state and local subsidies) to total revenue, plus the ratio of charges for charity care to total charges. Every element in this formula carries its own data quality risk.
Where safety net hospital data quality breaks
| | Value | |---|---| | Medicaid eligibility verification | 20 | | Dual-eligible misattribution | 12.8 | | Charity care vs bad debt classification | 15 | | Cost report recency (18-24 month lag) | 24 | | Uncompensated care cap calculation | 18 |
Safety net hospitals face data quality challenges that other facilities do not. The population they serve is more likely to have unstable insurance coverage, incomplete demographic records, and social complexity that standard intake workflows fail to capture.
Medicaid eligibility verification gaps
Medicaid eligibility is notoriously hard to verify in real time. A 2022 report from the HHS Office of Inspector General found that states varied widely in their ability to accurately determine Medicaid enrollment status, with some states carrying error rates above 20% in eligibility determinations. For safety net hospitals, this means the Medicaid fraction numerator is built on a denominator of uncertainty. A patient who was Medicaid-eligible on the day of admission but whose enrollment record shows a coverage gap due to processing delays gets counted as uninsured, not Medicaid. That lowers the Medicaid fraction and reduces the DSH payment.
The Medicaid unwinding following the end of the COVID-19 public health emergency made this worse. The KFF reported in 2024 that over 25 million people were disenrolled from Medicaid between April 2023 and mid-2024, many for procedural reasons rather than actual ineligibility. Safety net hospitals treated many of these patients during gaps in coverage that were artifacts of administrative processing, not changes in actual eligibility.
Charity care classification inconsistencies
Charity care plays a direct role in the low-income utilization rate, but there is no uniform national standard for what counts as charity care versus bad debt. The American Hospital Association has documented this inconsistency for years. A patient who never applies for financial assistance and whose bill goes to collections appears as bad debt on the balance sheet, even if they would have qualified for full charity care. That charge disappears from the charity care numerator, and the hospital's DSH calculation drops.
Safety net hospitals with limited financial counseling staff are most likely to misclassify care this way. The data problem is not clinical. It is operational.
Dual-eligible patient misattribution
Patients who are dually eligible for Medicare and Medicaid represent a significant portion of safety net hospital volume. The MedPAC March 2023 report noted that roughly 12.8 million people were dually eligible in 2021. For DSH purposes, dual-eligible patients are excluded from the Medicaid fraction numerator if they have Medicare Part A entitlement. Correctly identifying Part A status requires cross-referencing CMS enrollment files with hospital admission records. When that cross-reference fails, patients get misclassified, and the Medicaid fraction moves in either direction.
The data systems at many safety net hospitals were not built for this kind of cross-source validation. Legacy billing platforms often carry Medicare and Medicaid identifiers in separate fields with no automated reconciliation.
The audit trail problem in DSH reporting
CMS and state Medicaid agencies audit DSH payments. The Medicaid and CHIP Payment and Access Commission (MACPAC) has repeatedly flagged that DSH audits reveal hospitals cannot always document the data trail from patient encounter to payment claim. The audit process, governed by Section 1923 of the Social Security Act, requires hospitals to demonstrate that their reported Medicaid days and uncompensated care costs are accurate and verifiable.
When auditors find discrepancies, hospitals must repay overpayments. For safety net facilities already operating on margins below 2%, a recoupment of even a few hundred thousand dollars can force service cuts. The irony is structural: the hospitals least able to invest in data infrastructure are the ones most likely to fail the audits that protect their funding.
Uncompensated care data and the Section 1923 requirements
The Affordable Care Act added an uncompensated care cap to DSH payments. Hospitals cannot receive DSH payments exceeding the cost of uncompensated care they provide. Calculating that cost requires accurate data on what care was delivered to uninsured and underinsured patients, what charges were associated with that care, what payments were received, and what the hospital's cost-to-charge ratio is.
The 2023 MACPAC annual report on DSH documented that states use different methodologies to calculate hospital-specific DSH limits, and some methodologies rely on cost report data that is 18 to 24 months old by the time it informs a payment. That lag means DSH payments are calculated against a version of reality that no longer exists.
Recency is one of the eight dimensions in health data trust scoring, and DSH reporting is a textbook case of what happens when the recency dimension collapses. A cost report from fiscal year 2021 informing a 2023 payment allocation cannot reflect the patient mix changes that the Medicaid unwinding caused in 2023.
What disproportionate share data intelligence actually requires
| | Value (%) | |---|---| | Provenance | 25 | | Consent | 20 | | Recency | 15 | | Quality | 10 | | Concordance | 10 | | Validation | 10 | | Breadth | 5 | | Stability | 5 |
Fixing DSH reporting is not a matter of better software. It is a matter of treating the data inputs to DSH calculations as scored, validated records rather than raw administrative artifacts.
Provenance for every Medicaid day
Every inpatient day counted in the Medicaid fraction needs a documented chain showing the patient's Medicaid enrollment status on the date of service, the source system that confirmed it, and the timestamp of verification. Without provenance, a Medicaid day is an assertion, not a fact. When CMS auditors ask how a hospital knows a patient was Medicaid-eligible on April 3rd, the answer cannot be "our billing system says so." It needs to trace back to the state Medicaid eligibility file, with a date-stamped match.
Concordance across state and federal systems
Dual-eligible identification requires concordance between Medicare enrollment data, Medicaid enrollment data, and hospital admission records. When those three sources disagree, the DSH calculation is wrong. Concordance scoring flags records where Medicare Part A status in the CMS file does not match what the hospital's billing system shows. That flag needs to fire before the cost report is filed, not after the auditor finds it.
Recency enforcement for cost reports
DSH payments calculated on cost report data that is two years old are inherently unreliable. Real disproportionate share data intelligence requires cost data that reflects the current fiscal year, or at minimum the most recently closed quarter. If that is not possible, the payment methodology should discount older data or apply an adjustment factor. The current system treats stale data as if it were fresh, and safety net hospitals pay the price.
Validation of charity care versus bad debt
Every uncompensated care charge that enters the DSH calculation should carry a validated classification: charity care with documented financial screening, or bad debt with documented collection activity. The gap between the two is not ambiguous in policy. It is ambiguous in practice because the screening workflows at safety net hospitals are under-resourced. Data validation at the point of financial screening, not after the fact in a cost report, is what the calculation needs.
The downstream cost of unscored DSH data
When DSH data is unscored, three things happen. First, hospitals that serve the most vulnerable patients receive less than they are owed because their data understates their Medicaid volume and charity care burden. Second, state Medicaid agencies allocate DSH pools based on hospital-reported data that may not reflect actual need, creating misallocation across the safety net. Third, federal oversight agencies spend audit resources on post-hoc corrections rather than pre-submission validation.
The Government Accountability Office reported in 2020 that CMS had not finalized DSH audit rules for certain years, leaving billions in payments unverified. That audit backlog is itself a symptom of the data quality problem. If the underlying data were scored and validated before submission, audits would be confirmatory rather than investigatory.
Why this matters beyond DSH
Safety net hospital data quality affects more than DSH payments. The same data feeds 340B program eligibility determinations, graduate medical education calculations, and state-level supplemental payment programs. A hospital that cannot document its Medicaid patient volume accurately for DSH purposes is also vulnerable to 340B audit findings and GME payment errors.
The safety net hospital data quality problem is not isolated. It is a structural feature of how the U.S. healthcare system asks its most under-resourced institutions to produce the most complex data. Disproportionate share data intelligence means treating these data flows with the same rigor applied to clinical trial data or financial audit records, because the consequences of getting them wrong fall on patients who have no alternative.
The DTI Engine scores every record 0 to 100 across eight dimensions before your AI model sees it. For safety net hospitals and the state agencies that rely on their data, that score is the difference between a DSH payment that reflects reality and one that reflects whatever the billing system happened to capture. If your team is evaluating Medicaid data for reporting, compliance, or payment integrity, talk to the SuperTruth commercial team. Schedule a conversation or call (215) 918-4140.
Further reading:
Sources
Jason Alan Snyder
Co-founder of SuperTruth and Artists & Robots, and an inventor on the Data Trust Index patents. Twenty-plus years building technology inside Interpublic Group. He writes here nearly every day on data trust, provenance, and what AI should be allowed to act on, and publishes essays on his Substack.
About SuperTruth · LinkedIn · Substack · jasonalansnyder.com
See it in practice
EHR data scored before any AI model sees it.
DTI integrates with Epic, Oracle Health, and all major EHR systems.