OHIP diagnostic codes, and why claims fail without them

What the code is

A diagnostic code tells the ministry what condition you treated. It is short — something like 7159 for osteo-degenerative arthritis, or 300 for anxiety and depression — and it comes from a list published in the Schedule of Benefits.

Two things about that list trip people up:

• It is not ICD-10 and not ICD-9. The codes look familiar because they derive from an older ICD revision, but the OHIP list is its own thing, is much smaller, and does not accept codes from a hospital coding system. A diagnosis pulled from an Epic problem list is a clinical diagnosis, not an OHIP diagnostic code.
• It is coarse. A few hundred codes cover all of medicine, so most of them are broader than the clinical picture. Picking the closest available category is correct and expected, not a compromise.

Which fee codes require one

The requirement is attached to the fee code, not to the visit. The large majority of codes in the Schedule require a diagnostic code; a minority do not take one at all. Three practical consequences:

1. You cannot guess from the kind of service

   Assessments almost always require one. Premiums are mixed — some require a code and some do not. Procedural codes vary. The only reliable answer is what the Schedule says for that specific code, which is why a billing system that knows the requirement per code is more useful than a rule of thumb.

2. Sending one where it is not expected is also a problem

   It is not usually fatal, but on some codes a diagnosis field that should be empty causes a refusal, so "always fill it in" is not a safe default either.

3. Some combinations are constrained

   A handful of fee codes are only payable with particular diagnoses. The Schedule states these where they apply. If a claim is refused and the fee code and diagnosis both look individually valid, a combination rule is the thing to check.

The leading-zero problem

This is worth its own section because it fails silently and it is not obvious.

Some diagnostic codes are three characters and some are four. A three-character code sent in a four-character field has to be padded the way the ministry expects. If a code is stored as a number somewhere in the chain rather than as text, the leading zero disappears — 009 becomes 9 — and the code no longer matches anything in the list.

What you see when this happens is not an error about padding. You see a claim refused as though no diagnosis had been given, or a premium that should have been payable quietly not paid. Nothing in the rejection says "leading zero".

When the code changes the money

For a routine assessment, the diagnostic code does not change the fee. Where it does matter is on premiums and special-case codes whose eligibility depends on the condition treated.

The pattern to understand: the visit code is correct, the claim is accepted, the payment is lower than you expected, and nothing on the Remittance Advice says why in plain language. The premium simply was not applied. That is a diagnosis-dependent premium that did not qualify, and the usual cause is a diagnostic code that was defensible clinically but not the one the premium requires.

Two examples of the shape, without quoting rules that change:

• Chronic-disease and complexity premiums in several specialties are payable only for a defined set of conditions. The code you submit is what the adjudication sees.
• Some consultation and procedure combinations require a specific diagnosis, so a near-miss code produces a refusal rather than a reduction.

Both are checkable against the Schedule in advance, and neither is visible after the fact without comparing what you expected to what was paid. Reading the Remittance Advice covers how to spot the gap.

Choosing codes in practice

A few habits that reduce the error rate:

• Pick the code for the reason you billed the visit, not the patient's most serious diagnosis. A diabetic patient seen for a laceration is a laceration visit.
• Build a short list per specialty. Most physicians use a dozen or two codes for the overwhelming majority of claims. Having those at hand beats searching a list of several hundred each time.
• Set a sensible default where the work is repetitive. A clinic or roster where nearly every patient has the same presenting problem is better served by a default that is occasionally overridden than by an empty field that is occasionally forgotten.
• Treat a refused claim as information about the code. If one diagnosis-and-fee-code pair fails repeatedly, the pair is wrong; fixing it once fixes every future claim of that shape.

Frequently asked questions

Sources

• OHIP Schedule of Benefits and Fees — the diagnostic code list and per-code requirements.

Related reading