The 837: how claims are told to payers

The 837 is the X12 transaction that carries a healthcare claim from a provider to a payer. It says who is billing, who rendered the service, who the patient is, what was done, why it was done, and what is being charged. Three variants exist, 837P for professional, 837I for institutional, and 837D for dental.

​

The three variants

​

What an 837 carries

Six categories of information:

​

1. Submitter and receiver

Who transmitted the file and who it’s for. Set by your clearinghouse configuration — this is where a misconfigured multi-entity setup sends claims under the wrong PC.

​

2. Billing provider (Loop 2000A / 2010AA)

Your professional entity. Legal name, NPI (Type 2), Tax ID, address, and taxonomy. Everything here must match what the payer holds from enrollment exactly. The legal name must match your IRS EIN letter and your W-9. The NPI must be the one you enrolled. The taxonomy must match the taxonomy on your contract. Name and taxonomy mismatches are the leading cause of first-claim rejections, and they are enrollment problems wearing a billing costume.

​

3. Rendering, referring, and supervising providers (Loops 2310A–D)

The rendering provider is the individual clinician who performed the service, identified by Type 1 NPI. Where a referring provider is required, common for diagnostics and specialty services, omitting it is a denial. Supervising provider appears in incident-to and supervision scenarios.

​

4. Subscriber and patient (Loops 2000B / 2000C)

The subscriber holds the policy; the patient may be the subscriber or a dependent. Member ID, name, date of birth, gender, address, and relationship code. Precision matters more than it feels like it should: transposed member ID digits, a missing alpha prefix on a Blues ID, a nickname instead of the legal name, or a date of birth that differs from the payer’s record all produce rejections.

​

5. Claim level (Loop 2300)

Total charge, place of service, claim frequency code (original, corrected, void), prior authorization number, and up to twelve diagnosis codes in the HI segment.

​

6. Service lines (Loop 2400)

One per procedure. Each carries the procedure code and up to four modifiers, the charge, units, the service date, and diagnosis pointers linking that line to the diagnoses that justify it.

​

How EHR fields map in

Practically, most 837 errors are data-entry or configuration errors upstream:

​

For CMS-1500 veterans

If you learned billing on paper, the mental translation: The important shift: the paper form has a fixed grid with a hard cap on service lines. The 837 has loops that repeat, so a claim can carry far more detail, and correspondingly more ways to be wrong.

​

Where errors originate

Ranked by how often they cause first-submission failures:

1. Enrollment data mismatches, name, NPI, Tax ID, taxonomy not matching the payer’s records
2. Provider not credentialed or not linked to the group contract
3. Subscriber ID errors, the single most common patient-data failure
4. Missing prior authorization number, obtained but not captured
5. Place of service errors, especially telehealth 02 vs 10
6. Missing or wrong modifiers, 25 and 59 lead the list
7. Diagnosis pointer errors, line pointing at a diagnosis that doesn’t support it
8. Missing referring provider where required
9. Date of service before the provider’s effective date
10. Duplicate submission instead of a corrected claim

Note that items 1, 2, and 9 are enrollment problems and item 4 is a workflow problem. Only 5–7 are properly billing errors.

​

Corrected claims

To fix a claim already adjudicated, resubmit with the appropriate claim frequency code (7 for replacement, 8 for void) and the original claim reference number.