The Medical Debt Wars: The Fight Over What a Hospital Bill Does to Your Credit

Roughly 100 million Americans carry medical debt, and the question of whether it belongs in a credit file has now produced one of the strangest regulatory sequences in consumer finance: a federal rule that would have erased $49 billion from 15 million reports — killed in court with its own agency's consent; a voluntary purge by the bureaus that quietly removed 70% of medical tradelines and outlived the rule that would have mandated it; fifteen states building their own bans; and a federal preemption offensive now aimed at those states. Underneath the legal war sits an empirical finding almost nobody disputes: medical collections barely predict whether you'll repay anything else. This report is the full anatomy of the fight.

The core thesis

The medical debt fight is a referendum on the first principle of the credit file: is the report a ledger of what you owe, or a prediction of how you'll pay? Those two ideas usually coincide — a defaulted card is both a debt and a signal. Medical collections are where they diverge: the debt is real, but the research (including the regulator's own) keeps finding it a weak predictor of future repayment, because nobody underwrites their appendix. The bill arrives involuntarily, priced opaquely, frequently erroneous, often owed by insurance rather than the patient — and then lands in the file with the same gravity as a walked-away loan. Every combatant in this war is really arguing about which principle governs: the predictive camp (bureaus' own purge, the newer scoring models, the vacated rule, the state bans) says signal-less data doesn't belong; the ledger camp (the court's FCRA reading, collector trade groups) says Congress built a reporting system, not a curation system, and the tool for bad debts is the dispute process, not category deletion.

Our thesis: the predictive camp has already won the substance — 70% of medical tradelines are gone via voluntary bureau action, the newer models discount what remains, and no serious actor defends the old regime's accuracy — while the ledger camp is winning the law, having killed the federal rule and now pressing preemption against the states. That divergence is the story: the market and the models fixed most of the problem before the law decided who was allowed to; what remains contested is the residual — unpaid collections over $500, in non-ban states, scored by legacy models — and the deeper precedent question of who curates the file. For the leverage-holders in our collections economics report, that residual is the point: credit reporting is the collection industry's cheapest pressure instrument, and every removal is a pressure valve closed.

Nobody underwrites their appendix. The whole war is about whether the file should pretend otherwise.

Why medical debt is different — the evidence

The empirical case, briefly, because it drives everything: regulator research found medical collections systematically over-penalize — consumers with medical collections repay other obligations better than their scores predict, unlike consumers with non-medical collections. The mechanism isn't mysterious. Medical debt is involuntary (no application, no ability-to-pay decision); informationally broken (opaque pricing, surprise billing, insurance disputes — a large share of reported medical collections are contested or simply wrong, which is why they dominate complaint volumes); and billing-system-generated (a single episode can spawn multiple collections from providers the patient never chose or met). None of that describes a borrowing decision, which is what the file exists to characterize. This is the same lesson our invisibility and score wars reports keep surfacing from the other direction: the file is a proxy for creditworthiness, and proxy error — counting what shouldn't count, missing what should — is measured in real people's basis points.

The voluntary purge: 70% gone without a rule

The most consequential act of the entire war required no statute: in 2022–2023, the three bureaus jointly announced that paid medical collections would be deleted (previously they lingered seven years, paid or not), unpaid medical collections under $500 would not be reported, and new medical collections would face a one-year waiting period before appearing — time for insurance ping-pong to resolve. Combined effect: roughly 70% of medical-debt tradelines vanished from the reporting system. Read the move through the bureau economics lens and it's rational twice over: the data was predictively weak (furnishing it degraded the product) and reputationally expensive (medical collections drove disputes, complaints, and the political heat that produces rules like the one then brewing). The purge's fragility is equally important: it's policy, not law — reversible by announcement — which is precisely why the states legislated and the CFPB ruled. The voluntary regime is the war's center of gravity: everything still reported is what the purge chose to leave.

The rule that died twice

The January 2025 rule would have finished the job: barring medical debt from consumer reports and its use in lending decisions — ~15 million people, ~$49 billion, an average score lift the agency projected at twenty points. It never took effect. Industry plaintiffs sued in the Eastern District of Texas; the change of administration flipped the defendant; and in July 2025 the court vacated the rule with the CFPB's own consent — the agency joining the request to strike down its own regulation, a sequence with few precedents in consumer finance. The holding matters more than the theater: the court found the rule exceeded the agency's FCRA authority — the statute permits coded medical information in reports, so a rule prohibiting it contradicts the text — and the vacatur forecloses a future CFPB from simply re-issuing it. Whatever one thinks of the outcome, the reasoning relocated the fight: if federal regulators can't curate the file's categories, the action moves to whoever can — the bureaus (who already had, voluntarily) and the states (who were already moving). Which is exactly where the war went next.

The state counterattack — and the preemption war

As of early 2026, at least fifteen states restrict or ban medical debt credit reporting — Colorado's outright ban, New York, New Jersey, Illinois, Minnesota, Maryland, Connecticut, Washington, Oregon, New Mexico, Nevada, Delaware, and more — most effective in 2025 or 2026, with variations spanning full bans, higher dollar thresholds, longer waiting periods, and use restrictions in specific lending. The counterstrike followed the vacatur's own logic: the Texas court, beyond killing the federal rule, opined that the FCRA expressly preempts state laws restricting what may be furnished to consumer reports — and the CFPB then issued an interpretive rule embracing that preemption theory, with collector trade groups now suing states to enforce it. Note the jurisprudential switchback: the same statute read narrowly to stop federal curation is being read broadly to stop state curation — a combination whose endpoint would be that nobody elected may decide what's in the file except Congress, leaving the bureaus' voluntary policy as the only operative regulator. Consumer advocates dispute the preemption reading (the vacatur itself, they note, didn't directly reach state laws), and the litigation map is live. For planning purposes: the state protections are real today and contested tomorrow.

What's actually on reports right now

That last row deserves italics in every consumer's mind: paying a hospital with a card — including medical financing cards — converts protected medical debt into ordinary revolving debt with none of the medical carve-outs. The consumer playbook for what does appear runs through the collections removal guide: dispute inaccuracies (medical collections have the sector's highest error rates), demand validation, resolve with the provider (payment deletes by policy), and check your state's law before assuming the national baseline.

What the scoring models did on their own

The third front barely made headlines: the scoring models quietly defected from the ledger camp years ago. FICO 8 reduced small-collection impact; FICO 9, FICO 10/10T, and VantageScore 3.0/4.0 ignore paid collections entirely and discount medical collections relative to other derogatory items — the model architects reading the same predictive research and repricing accordingly. The catch is deployment lag: vast swaths of lending still run legacy models where a medical collection bites at full force — which connects this report to the score wars directly, since the mortgage gate's migration to newer models is, among everything else, a partial de-weighting of medical debt at the largest checkpoint in consumer credit. The composite trajectory across all three fronts — purge, models, states — points one direction: medical debt's grip on credit access is loosening structurally, with the legal war deciding the speed and the residue rather than the direction.

Scenarios and what we're watching

What we're watching: the preemption cases (the war's decisive theater); whether the voluntary purge gets codified, reversed, or simply endures as the strangest form of financial regulation — bureau press release; the reclassification channel (medical credit cards converting protected debt at scale is the loophole with growth ambitions); and the residual's size in the broader consumer-strain picture. The deepest question outlasts every docket: a credit file that curates away noise is more accurate and more humane — and a file curated by unaccountable policy is one announcement away from curating in either direction. The medical debt wars are the system deciding, expensively and in public, who holds that pen.

Frequently asked questions

This report is for general information only and does not constitute legal or financial advice. Litigation and state laws in this area are changing rapidly; figures are drawn from publicly reported sources including court records, CFPB materials, and bureau announcements.