Executive Summary: Healthcare practices generate more litigation per dollar of revenue than almost any other borrower type you will fund, and most underwriting teams read that litigation wrong. They count cases instead of classifying them, which means a dermatology group with one closed malpractice claim gets flagged while a multi-site practice with three active payer reimbursement suits and a wage-and-hour class action sails through. This post explains which healthcare case types actually predict repayment risk, how entity structure hides litigation from the name on your application, and where Cobalt's court records coverage fits, which is New York State and Miami-Dade County only, not a nationwide malpractice search.
Why Does Healthcare Carry a Litigation Profile No Other Borrower Type Has?
A restaurant that gets sued is usually being sued for one of two or three things. A medical practice is exposed on four separate fronts at once, each with a different cause, a different counterparty, and a different implication for whether your advance gets repaid.
What Are the Four Litigation Categories a Practice Generates?
Every healthcare borrower sits inside four distinct streams of legal exposure, and they do not correlate with each other:
• Professional liability (malpractice). Patient-initiated claims alleging negligent care. Volume is driven far more by specialty and patient contact hours than by practice quality. Roughly 28.7 percent of physicians reported having been sued at some point in their career as of 2024, down from 34 percent in 2016.[1]
• Payer reimbursement disputes. Suits and arbitrations against commercial insurers over denied, downcoded, or recouped claims. These are fights over money the practice already earned and already spent against.
• Employment claims. Wage-and-hour actions, misclassification suits, discrimination and retaliation claims from clinical and administrative staff. Healthcare is one of the most heavily litigated sectors for overtime and independent contractor misclassification.[2]
• Regulatory and licensing actions. State medical board proceedings, DEA registration matters, Medicaid fraud control unit investigations, and state corporate practice enforcement.
• Commercial and landlord disputes. Equipment leases, EHR vendor contracts, medical office space, and increasingly, funder judgments from prior advances.
Only two of those five categories reliably tell you anything about whether the practice can service new debt. The other three tell you the practice exists and sees patients.
Why Does Case Count Mislead Underwriters in Healthcare?
Because litigation volume in medicine is substantially a function of specialty, not of operator quality. The foundational specialty-risk research found that physicians in low-risk specialties faced roughly a 75 percent chance of facing a claim by age 65, while those in high-risk surgical specialties approached 99 percent.[3] A neurosurgery group with four historical malpractice filings is statistically ordinary. A five-provider family medicine practice with four is not.
If your rule is "three or more open cases triggers a decline," you have built a specialty filter, not a credit filter. You will systematically decline the surgical specialties that carry the highest procedure revenue and approve the primary care practices operating on the thinnest reimbursement margins in the sector.
Which Healthcare Court Cases Actually Predict Repayment Risk?
The useful question is not how many cases exist. It is which counterparty is on the other side, and what the outcome does to next quarter's cash position.
Why Is a Single Malpractice Suit a Weak Credit Signal?
Three reasons, all of which underwriters routinely miss.
First, it is insured. Professional liability coverage is effectively mandatory for hospital privileges and most payer contracts, so the defense costs and the indemnity payment usually sit with the carrier, not the practice's operating account. The practice's real exposure is the deductible and the future premium.
Second, most claims produce nothing. The vast majority of medical liability claims close without any indemnity payment at all, and most are dropped, withdrawn, or dismissed before reaching a verdict.[1] A filed complaint in the docket is an allegation with a low base rate of conversion into a payment.
Third, frequency is falling even as severity rises. Actual claims frequency ran about 4.6 percent in 2025, down from 7.5 percent in 2016, while the average NPDB physician payment climbed to roughly $514,000, about 20 percent higher than 2022.[4] Fewer suits, bigger ones. That trend moves the carrier's premium and therefore the practice's fixed cost base, but a single pending case tells you almost nothing about the next six months of cash flow.
The malpractice signal worth acting on is different: a claim large enough to exceed policy limits, a carrier non-renewal, or a pattern of claims concentrated on one named physician who is also the practice's primary revenue producer. Those are cash flow events. A routine filing is not.
Why Do Payer Disputes Speak Directly to Cash Flow?
Because a payer dispute is, by definition, a fight about receivables the practice has already booked. When a practice sues an insurer, or arbitrates against one, it is telling you that a meaningful share of its accounts receivable is contested rather than merely slow.
The scale of the payer friction is now large enough that disputes are routine rather than exceptional. More than 41 percent of providers reported denial rates above 10 percent in a September 2025 survey, and initial claim denials reached 11.8 percent in 2024 with projections into the 12 to 15 percent range.[5] Average denied claim amounts rose 12 percent for inpatient and 14 percent for outpatient from 2024, while external payer audit exposure jumped 30 percent year over year per customer.[6] The formal arbitration channel shows the same picture: roughly 4.8 million No Surprises Act independent dispute resolution cases were filed through the end of 2025, with about 1.2 million initiated in the first half of 2025 alone, more than double the same period in 2024.[7] CMS publishes the underlying counts and outcomes on a rolling basis.[8]
For an underwriter, a payer suit in the docket is a proxy for three things: contested receivables, a payer relationship under strain, and legal spend that competes with your repayment. If the defendant insurer is one of the practice's top two payers by volume, treat it as a material adverse signal regardless of the dollar amount pleaded.
What Does a Pattern of Employment Claims Tell You?
One employment claim is noise. A pattern is an operations diagnosis. Repeated wage-and-hour or misclassification actions usually mean the practice has been managing labor cost by stretching scheduling, classification, or overtime practices, which is what operators do when margin is already compressed.
The exposure is not trivial. The Fourth Circuit upheld a $9 million judgment in July 2025 against a medical staffing firm for misclassifying more than 1,100 travel nurses as independent contractors,[9] and a $13.5 million home health nursing wage settlement was approved in August 2025.[10] Class-wide employment liability lands as a lump sum against a practice with no reserve for it.
Weight employment claims by whether they are individual or collective, and by whether they repeat. Two individual claims across five years from a 40-person practice is background. Three claims in eighteen months, or any claim pleaded as a collective action under the FLSA, is a staffing-cost problem that predates your advance and will outlive it.
How Should You Weight Cases by Type Instead of Count?
Replace the count threshold with a weighted rubric. The rubric does not need to be sophisticated to beat counting, it just needs to encode the fact that different plaintiffs mean different things.
What Does a Type-Weighted Scoring Rubric Look Like?
A workable starting structure, tuned for a 6 to 18 month advance:
• Payer or insurer as counterparty: highest weight. Directly contests booked receivables. Escalate if the payer is top-two by practice volume.
• Collective or class employment action: high weight. Unreserved lump-sum exposure plus signal of margin compression.
• Prior funder judgment or confession of judgment: highest weight. Existing claim on the same revenue you are advancing against.
• Landlord or equipment lessor action: medium weight. Short-term liquidity failure on a fixed obligation, which is the most direct read on current cash position.
• Regulatory or licensing action against the entity: high weight. Threatens the payer contracts and therefore the revenue itself.
• Individual employment claim: low weight. Common at any practice above roughly 20 employees.
• Single malpractice claim within specialty norms: lowest weight. Insured, low conversion to payment, and largely specialty-determined.
Score the file, do not tally it. Two payer suits and a funder judgment should outrank six routine malpractice filings, and under a counting rule it never will. If you want to formalize this, the mechanics of building a weighted model against docket data are covered in how to build litigation risk scoring with court data.
Which Case Types Justify a Decline Versus a Structure Change?
Most healthcare litigation should adjust terms rather than kill a deal. A prior funder judgment or an active regulatory action against the entity's license is a decline conversation. A payer dispute or a collective employment action is a structure conversation: shorter term, lower advance amount, tighter holdback, or a reserve sized to the pleaded amount. A malpractice filing inside specialty norms should not change the offer at all unless it exceeds policy limits or names the practice's primary producer.
Why Doesn't the Borrower Entity Match the Entity Named in the Litigation?
This is the failure mode that makes healthcare court searches return clean when the practice is not. You searched the right county and the right court. You searched the wrong name.
What Is the PC/PLLC and Management Company Split?
Most states restrict the corporate practice of medicine, which means the clinical entity has to be owned by licensed physicians. The standard workaround is a two-entity structure: a physician-owned professional corporation or PLLC that holds the licenses, the payer contracts, and the malpractice exposure, paired with a management services organization that holds the lease, the staff, the equipment, the billing function, and the outside capital.
The prevalence is not marginal. Nearly 80 percent of physicians are now employed by or affiliated with hospitals, health systems, or corporate entities, and non-hospital corporate owners including private equity have surpassed hospitals in practice ownership at 30.1 percent versus 28.4 percent.[11] States are tightening disclosure around these structures, with New York, Massachusetts, Oregon, and California all moving on ownership transparency and transaction review.[12]
For an underwriter, the practical consequence is a name mismatch. The application arrives from the management company, because that is the entity with the bank account and the revenue you are advancing against. The malpractice suits, payer disputes, and licensing actions are captioned against the PC, which shares officers but not a name. Search the applicant only and you get a clean docket for an entity that has been operating for fourteen months and never treated a patient.
Why Does Officer Resolution via SOS Have to Come First?
Because the officers are the join key. The Secretary of State filing for the management company gives you the registered agent, the officers, and the formation date. The same officer names appear on the PC's filing. Once you have both entity names, you can run the court search against both, plus any DBA the practice bills under.
The order matters and it is not optional:
• Pull the SOS record for the applicant entity first. Confirm active status, formation date, and officer names. Formation dates under two years on an established-looking practice are a structure signal, not a fraud signal, but they change what you search.
• Search SOS by officer name to surface sibling entities. Same officers, different entity, usually the PC or an affiliated site LLC.
• Check UCC filings on both entities. Prior funder liens frequently sit against the management company while the clinical revenue sits in the PC.
• Run court records against every resolved entity name, not just the applicant. Including former names and assumed names.
• Verify the professional license against the individual physician, not the entity. The entity does not hold the license; a named physician does.
This is the same officer-resolution discipline described in the pre-funding litigation checks workflow guide, and healthcare is the borrower category where skipping it costs you the most.
What Does the Financial Distress Data Say About Healthcare Practices in 2026?
The reason this matters more this year than last is that the physician practice segment has moved from stable to visibly stressed, and the distress is showing up in bankruptcy court.
Why Are Physician Practice Bankruptcies Rising?
Healthcare bankruptcies actually fell in 2025, with 45 filings, down 21 percent from 2024, and clinics and physician practices accounted for only six of those.[13] That reversed sharply. Filings jumped 33 percent in the first quarter of 2026, with clinics and physician practices and senior care each accounting for four of the twelve filings above $10 million in liabilities.[14] Through the first half of 2026, 26 healthcare companies above that threshold filed Chapter 11, and medical practices made up nearly 30 percent of all healthcare filings, on pace for 14 practice bankruptcies by year end against six in all of 2025.[15] The drivers named in that reporting are Medicaid cuts, expiring enhanced ACA subsidies, and rising uninsured volumes.
The reimbursement side reinforces it. CMS projected physician practice costs to grow 2.7 percent in 2026 as measured by the Medicare Economic Index, against a 2.5 percent statutory update, so the payment increase lands just short of covering the practice's own cost inflation.[16]
Where Does Merchant Cash Advance Debt Show Up?
In the docket, before it shows up in the bank statements. Healthcare now represents roughly 15 percent of the global merchant cash advance market, and MCA obligations are appearing directly in provider bankruptcy filings.[17] That reporting documents providers describing weekly remittances that suffocated operating income, and at least one case where an MCA lender diverted roughly $500,000 in Medicaid payments, which collides with federal anti-assignment rules on government reimbursement.
If you are funding healthcare, the single most predictive court record is not a malpractice case. It is a judgment or a UCC filing from another funder against the same revenue stream. Stacking detection in healthcare is harder than in retail precisely because of the entity split covered above, and it is the reason detecting repeat defendants in lending applications is worth running against every resolved entity name rather than just the applicant.
"If courts are cheap enough, then it's worth it to run on every application automatically." Yehudah Aron, Cucumber Capital
That instinct is right, and the economics of it are why teams end up automating the check rather than reserving it for deep dives.
Evaluating whether a court records check belongs in your healthcare underwriting flow? See what the New York and Miami-Dade data returns against your own application volume. Book a demo.
What Are Your Options for Pulling Healthcare Court Records?
Before getting into how any single API works, it is worth being direct about the market, because the honest answer is that no single source solves healthcare litigation nationally and anyone who tells you otherwise is selling aggregation with gaps in it.
How Do Unicourt, LexisNexis, PACER, and Manual Search Compare?
Four practical routes, each with a real tradeoff:
• PACER. Federal courts only. Excellent for bankruptcy and federal question cases, which for healthcare means some ERISA and False Claims Act matters. Malpractice and most payer disputes are state court, so PACER misses the bulk of what you want.
• Unicourt. Broad state court aggregation with an API. Coverage and refresh cadence vary considerably by county, and normalization across jurisdictions is imperfect.
• LexisNexis and Westlaw. Deep and authoritative, priced and structured for legal research workflows rather than per-application underwriting decisions at volume.
• CSC and Wolters Kluwer. Strong on UCC and corporate filings, with court search generally handled as a service order rather than a real-time API call.
• Manual courthouse or portal search. Free or near-free per lookup, accurate, and completely unworkable at 500 files a day.
The teams running real volume already blend these. One MCA operator described the current state plainly: "we run New York court separately and then we run Unicourt, and then we run UCC searches," at roughly 500 files per day with growth expected toward 600 to 700 submissions daily. Three tools, three integrations, three reconciliation problems.
That fragmentation, not data availability, is the actual bottleneck. Which is where a purpose-built lending API earns its place.
Where Does Cobalt's Court Records API Fit, and What Does It Not Cover?
Cobalt's court records coverage is New York State and Miami-Dade County, Florida. That is the complete list. It is not nationwide, it is not federal, it is not PACER, and it is not a nationwide malpractice search. If your healthcare book is concentrated in Texas or California, this product does not serve that book today, and one prospect said exactly that: "Most of our clients are not in New York."
The coverage is demand-driven rather than exhaustive. Roughly 80 percent of Cobalt's funder customers file judgments in those two places. New York is the center of alternative lending, and Miami-Dade became the second hub after a large migration of funders to South Florida. Broader jurisdictions are on the roadmap, including a planned human-assisted queue for unsupported jurisdictions that would run slower, around an hour, and has not shipped.
Both happen to be dense healthcare markets. New York's health care workforce exceeds one million people, the largest single employer group in the state,[18] and Florida's annual physician workforce report documents the concentration in the southeast counties.[19] New York carries one additional property worth knowing: it is the only state where Cobalt returns SOS status, UCC filings, contractor licensing, and court records against the same entity. For a New York healthcare file, that is the full stack from one vendor. Everywhere else it is a partial stack, and you should plan for that rather than discover it in production.
How Do You Wire a Healthcare Court Check Into an Underwriting Pipeline?
The integration is straightforward. The design decisions around it are where teams get it wrong.
What Does the /courtCases Call Look Like?
The endpoint is asynchronous. You submit a business name and a jurisdiction with a callback URL, and results post back to you, typically in 30 to 120 seconds. There is no synchronous mode.
curl -G "https://apigateway.cobaltintelligence.com/courtCases" \
-H "x-api-key: $COBALT_API_KEY" \
--data-urlencode "businessName=Riverside Family Medicine PC" \
--data-urlencode "jurisdiction=newYork" \
--data-urlencode "callbackUrl=https://underwriting.example.com/hooks/court-cases"Valid jurisdiction values are `newYork` and `miamiDade`, plus `testNewYork` and `testMiamiDade`, which run the full flow without consuming credits. Build your callback handler and your entity-name fan-out against the test jurisdictions before you spend anything. Responses return judgment details, case number, case type and division, filing dates, parties, and amounts where the record includes them, which is not always. Data is pulled live from the court site at request time rather than served from a cached index. Pricing is one credit per lookup, the same as an SOS lookup.
The design decision that matters: because the call is per entity name, and healthcare practices resolve to two or three entity names, budget two to three court lookups per healthcare application, not one. That is the real unit cost of this check on a medical file. One funder framed the threshold this way: "where I would pay $4 a pull is when you have the state index on court search."
How Do You Handle Jurisdictions Cobalt Does Not Cover?
Write the fallback into the policy rather than leaving it to the analyst. A defensible pattern:
• Route by practice location, not by applicant mailing address. The PC's registered address governs which court has the malpractice and payer cases.
• Run the automated check where covered, and record a documented exception where not. An unchecked jurisdiction should be a recorded state in the file, never a silent gap.
• Keep a manual escalation path for the top three uncovered states in your book. Portal searches are slow but they are not unavailable.
• Do not treat a null court result as a clean result. Null in an uncovered jurisdiction means unknown, and your decision memo should say so.
• Re-run at renewal rather than only at origination. Payer disputes and employment classes surface mid-term.
Cobalt is a data source, not a decisioning engine. It returns records. The weighting, the thresholds, and the adverse action logic are yours, and that separation is what makes the output usable inside a policy you can defend to an examiner.
How Do You Prove the Check Is Actually Working?
Write the rubric down, then test it. Tag every funded healthcare file with the case types found at origination and look back at 12 months against actual default and early-payoff behavior. You are testing one hypothesis: that payer disputes, collective employment actions, and prior funder judgments separate defaults, and that routine malpractice filings do not. If your own book disagrees, your book wins. That is the real argument for type weighting. It produces a model you can test, where counting produces a rule nobody can evaluate.
Name the specialty adjustment explicitly in the written policy. A rule that says "three or more open cases requires senior review" without a specialty qualifier will produce a decline pattern concentrated in surgical specialties, which is both a bad credit outcome and a fair lending question you do not want to answer unprepared.
If your healthcare book runs through New York or Miami-Dade, the court layer is one API call per entity name and one credit per lookup, and you can build the whole integration against the test jurisdictions before spending anything. Book a demo to see what the data returns against your own files.
.jpg){kind=small}
.png)