Clearinghouses

A clearinghouse is the intermediary that sits between providers and payers, taking the electronic transactions a provider sends, eg. claims, eligibility checks, status queries, and scrubbing them for errors, translating them into each payer's required format, before routing them to the right plan. Then it carries the payer's answers back.

ℹ️ In plain English

Health insurance is fragmented so a a doctor's office bills hundreds of different plans, and every plan wants the data slightly differently. Rather than build and maintain a dedicated pipe to each one, the provider integrates with a single clearinghouse. The clearinghouse is the "post office" of healthcare billing that connects to thousands of payers. A clearinghouse checks the envelope of data, fixes the address, and makes sure the claim is well-formed before it ever reaches the insurer.

💡 Zahid's take

"Well-formed" sounds like an easy problem to solve, until you learn that defining "well-formed" is the crux of the challenge and only the tip of the iceberg. Fragmentation naturally means multiple standards to adhere to and so with each payer the list of rules evolves just as quickly.

Being able to fingerprint each payer and digitise their standard of "well-formed" would be huge in eliminating some of the tensions in Payer↔Provider communication.

How it works

Everything a provider exchanges with a payer travels as a standardized X12 EDI transaction. This is machine-to-machine billing data in formats mandated under HIPAA. The clearinghouse behaves as the switchboard for that traffic. The core transactions it handles:

837 — the claim. The provider's billing system produces an 837 (professional, institutional, or dental) containing patient, provider, diagnosis, procedure, charge, and coordination-of-benefits data. The clearinghouse validates it, applies payer-specific edits, and forwards it for adjudication.
270/271 — eligibility. The provider sends a 270 asking "is this patient covered, and for what?" The clearinghouse routes it to the payer and returns the 271 with coverage status, copays, deductibles and accumulators, and benefit details. See Eligibility & Enrollment.
276/277 — claim status. A 276 asks "where is my claim?"; the 277 answers received / pending / paid / denied. A 277CA acknowledgment reports claim-level acceptance or rejection.
835 — remittance (ERA). The Electronic Remittance Advice tells the provider what was paid, denied, and adjusted, line by line, for automated cash posting in Revenue Cycle Management (RCM).
278 — prior authorization. The request-and-response transaction for Prior Authorization decisions, increasingly the focus of regulators (see below).
834 / 820 — enrollment and premium. Used between employers, exchanges, and payers to enroll members and remit group premium.

The clearinghouse's real value is the middle step: scrubbing and translation. It runs each transaction through thousands of edits — missing fields, invalid codes, member-ID mismatches, payer-specific quirks — and returns errors back to the provider before the payer sees them. A clean claim adjudicates faster and denies less. It also normalizes formats: every payer has its own companion guide, and the clearinghouse attempts to absorb that complexity so the provider doesn't have to.

Why it matters

Without clearinghouses, every one of ~6,000 hospitals and hundreds of thousands of practices would need a direct EDI connection to every one of thousands of payers — an N×M mesh that is impossible to maintain at the local level. The clearinghouse collapses that into a hub-and-spoke model: connect once, reach everyone. They are the connective tissue of the money flow, and they sit on a chokepoint — if the switch goes down, claims and cash stop.

That chokepoint became national news. The largest US clearinghouse — acquired by a unit of a major vertically integrated payer in a deal that closed in late 2022 after the Department of Justice unsuccessfully sued to block it. The clearinghouse processes roughly one-third of all US medical claims. In February 2024 a ransomware attack forced it offline for weeks. Providers couldn't submit claims or verify eligibility, cash flow froze across the system, and the government stood up emergency advance-payment programs. It was the clearest possible demonstration that a back-office plumbing layer most patients have never heard of is systemically critical infrastructure.

⚠️ Where it breaks

Clearinghouses are a concentrated, low-margin, deeply embedded layer, and that structure defines where the pressure sits. Concentration means a single outage has system-wide reach, as 2024 showed which is why redundancy, multi-connectivity, and graceful failover are becoming design priorities. Embeddedness makes switching costly because a provider's connections, payer enrollments, and ERA mappings are sticky, so incumbents earn durable relationships and can extend that lead by investing in reliability and service. Vertical integration, ie a clearinghouse operating inside a larger payer organization, brings scale and investment capacity. There is also room to add value on transparency: when providers can see why a claim was scrubbed out, fewer legitimate claims get held and reworked. This is the layer with the most operational leverage and the most upside from adding visibility.

💡 Zahid's take: The opportunity for startups

I have a soft spot for clearinghouses. They are utility-grade infrastructure that ended up holding some of the richest data in healthcare. Two things are true about them at once.

First, they are systemically important and anyone building anything that touches claims or eligibility sits downstream of this layer. 2024 was a reminder to design for resilience where multi-connectivity, redundancy, and graceful degradation have become tablestakes.

Second, the vantage point is remarkable. A clearinghouse sees both sides of every transaction. The 837 claim request sent out and the 835 remittance sent back. This makes up one of the richest denial-and-payment datasets anywhere.

A common mistake I have witnessed is founders assume the opportunity is the pipe. It usually isn't. Moving the bytes is capital-intensive and carries HIPAA-grade reliability demands. The more interesting opening is the intelligence the pipe underuses. For example, predicting which claims will deny before they're sent and why; turning terse scrub rejections into clear, actionable explanations; and closing the loop from 271 eligibility and 278 auth into a clean 837. The general lesson from years spent building in adjacent layers is that the cleaner the upstream transaction, the less downstream review and rework everyone has to do.

The opportunity around clearinghouses is helping reduce the number of requests that bounce, and providing the the sender of the request a feedback loop they can use to learn and improve.

Who is involved

Providers and RCM teams / billing vendors — the senders. Many practices reach a clearinghouse through their practice-management or EHR system rather than directly.
Payers, TPAs, and their core claims engines — the receivers, who often prefer transactions to arrive pre-scrubbed.
Clearinghouses and "switches" — some sell direct to providers; others operate wholesale, connecting other clearinghouses and vendors to payers.
Regulators — HIPAA sets the transaction standards; the CMS Interoperability & Prior Authorization Final Rule (CMS-0057-F) pushes API-based prior authorization that could route around traditional EDI over time. See CMS (Centers for Medicare & Medicaid Services).

Numbers that matter

~95% of provider claims are now submitted electronically as X12 837 transactions, not on paper.
The largest clearinghouse alone processes on the order of billions of transactions per year and touches a large share of all US claims.
During the 2024 outage, surveys reported the overwhelming majority of affected practices saw a >20% revenue impact, underscoring how thin provider cash buffers are when the pipe stops.

🔎 In practice

A patient arrives for an MRI. The front desk fires a 270 through the clearinghouse and gets back a 271 confirming coverage and a $400 deductible still owed. After the scan, the radiology group's billing system builds an 837 and sends it to the clearinghouse, which scrubs it, catches a transposed member ID, and the biller fixes it in seconds before the payer ever sees a malformed claim. Two weeks later an 835 lands: paid at the contracted rate, $400 applied to the patient's deductible. The clearinghouse touched every step and the provider never built a single direct connection to the payer.

💡 Zahid's take: AI Digital Twins

I think the race to market for consumer-focused digital twins probably diluted the true value of the concept and technology. If my thesis of modern healthcare systems evolving from a System of Record (SoR) to a System of Intelligence (SoI) to a System of Agency (SoA) is true, then Digital Twins trained on specific roles and given the mandate to solve for communication may one day have significant role to play here.

For instance, I don't think it is too ambitious to see a world where Healthplans decide to tackle their own administrative backlogs by shipping an "AI Digital Twin" to providers to embed in a "harness" that can help package claims requests to avoid duplicate work on the Payer side. The digital twin would be responsible for ensuring the envelope contains the necessary information required to make a prompt, smooth and accurate claims determination and altogether reduce the amount of denials or claims sent back. Providers can submit requests with more certainty, Payers can spend less resource on invalid requests, and ultimately patients are not left waiting on a request that was never going to make it in the first place.

Sources

CMS, Administrative Simplification — Transaction standards (HIPAA-mandated X12 transactions): https://www.cms.gov/priorities/key-initiatives/burden-reduction/administrative-simplification/transactions
X12, Health Care Transaction Flow (270/271, 276/277, 278, 837, 835): https://x12.org/flow/health-care
HFMA, "Cyberattack on Change Healthcare brings turmoil to healthcare operations nationwide": https://www.hfma.org/technology/cybersecurity/cyberattack-on-change-healthcare-brings-turmoil-to-healthcare-operations-nationwide/
Office of Financial Research Brief, "The Cyberattack on Change Healthcare": https://www.financialresearch.gov/briefs/files/OFRBrief-24-05-change-healthcare-cyberattack.pdf
Holland & Knight, "Change Healthcare Cybersecurity Incident: Financial Impact and Resulting Litigation": https://www.hklaw.com/en/insights/publications/2024/06/change-healthcare-cybersecurity-incident-financial-impact
The Source on Healthcare Price and Competition, "DOJ's Challenge of the UnitedHealth and Change Healthcare Merger": https://sourceonhealthcare.org/dojs-challenge-of-the-unitedhealth-and-change-healthcare-merger-what-went-wrong-and-what-does-it-mean/