Most F32 and F33 mix-ups come from the same instinct. When a client's symptoms look worse than last time, the impulse is to reach for F33. But severity isn't what separates these two ICD-10 codes for depression. History is.

Picking the wrong code creates real problems. Denials. Documentation request letters. Audit flags. Whether you're the clinician writing the note or the biller reviewing it before submission, the F32 vs F33 call happens fast and often without much guidance.

Here's a practical way to think about it, what documentation actually supports each code, and the edge cases that trip up even experienced practices.

F32 vs F33: It's About History, Not Severity

Start with the binary split. F32 covers a single episode of major depression. F33 covers recurrent major depressive disorder.^[1][2] The decision is about the client's depressive history, not how bad they feel today.

F32 applies when there's no documented prior episode. The current presentation is the first one on record.

F33 applies when the client has had at least two distinct episodes separated by a stretch of full remission lasting at least two months.

The severity digit, the fourth character in the code, comes after the history question. A 42-year-old client returning after a depressive episode six years ago, currently moderate, codes to F33.1. Not F32.1. The fact that the current episode is mild, moderate, or severe has nothing to do with whether F32 or F33 is correct. That's already decided by the history.

It's an easy instinct to get wrong. A solo therapist coding F32.1 for a returning client because "this episode feels worse than the last one" is making a severity call where a history call belongs. The code looks defensible on the claim, but it doesn't match the chart. And payers notice patterns.

Clean claims start with the right diagnosis. When the diagnosis code doesn't match the documented history, the rest of the claim is built on shaky ground, even when the CPT, modifier, and units are all correct. Tracking your clean claim rate usually surfaces these patterns long before an auditor does.

What "Recurrent" Actually Requires

Recurrent depression has a specific definition, and getting it wrong in either direction causes problems.

Recurrence requires at least two distinct depressive episodes separated by at least two months of full remission.^[2] Not partial remission. Full. A client whose symptoms wax and wane without ever clearing for that two-month stretch stays in the F32 family until a clean recovery-relapse pattern emerges.

The reverse also matters. A prior episode from a decade ago, fully resolved, still triggers F33 on a new presentation. There's no statute of limitations on recurrence. A client whose first depressive episode was in college, untreated and fully resolved, who now presents at 35 with new symptoms, codes to F33. But only if the intake captured that college episode.

This is where intake history becomes a billing question, not just a clinical one. If the intake form doesn't ask directly about prior episodes of depression, F33-eligible clients quietly get coded as F32. The chart has nothing in it to support recurrence, and the default becomes single-episode by absence of evidence.

A group practice that audited a quarter of its returning-client charts found that nearly half of its F32 codes were actually recurrent presentations. The intake form simply didn't ask. One short question about prior depression history shifted the mix and made the documentation match the codes on the claims that followed.

When history is ambiguous, the right move is to query the clinician before coding. Not default to unspecified. Not guess. Ask.

Choosing the Severity Digit

Once history is settled, severity comes next. The fourth digit and any fifth-digit specifiers tell the rest of the story.

.0 mild
.1 moderate
.2 severe without psychotic features
.3 severe with psychotic features
.4 partial remission (F32) or unspecified remission (F33.40)
.5 full remission (F32); F33 splits this into .41 partial and .42 full
.9 unspecified

Each level needs documented support. Mild, moderate, and severe aren't synonyms for "client's mood today." Payers expect the chart to show symptom count, duration, and functional impact in line with DSM-5 criteria.^[3] A note that just says "depression, moderate" without symptom detail gives a payer reason to ask follow-up questions.

A client on maintenance medication, asymptomatic for eight months, codes to F33.42 (full remission), not F33.9. The remission code is more specific, more accurate, and easier to defend on audit. Defaulting to F33.9 because "the client isn't really depressed anymore" looks fine on a single claim and becomes a pattern across many.

F32.9 and F33.9 are valid codes, but they shouldn't be the house default. CMS and most commercial payers expect coding to the highest available specificity the documentation supports.^[4] When unspecified codes start showing up repeatedly across encounters, payers read that as incomplete documentation, not as a recurring clinical ambiguity. The claims still pay, but documentation requests start arriving, and audit risk creeps up.

The F32.9 vs F33.9 choice itself still hinges on history. If recurrence is documented, the unspecified code is F33.9. If only a single episode is on record, it's F32.9. Severity unspecified, history specified.

The Edge Cases That Cause Denials

A handful of patterns catch even experienced billers.

F32.A vs F32.9. These look interchangeable. They aren't. F32.A is "depression, unspecified" where major depressive disorder has not been confirmed. F32.9 is MDD confirmed, severity unspecified. Different clinical meaning, different documentation requirements. Using F32.A when the chart actually supports MDD undersells the diagnosis and can affect medical necessity determinations down the line.

Postpartum presentations. Most postpartum depression actually codes to an F32 or F33 code with the postpartum context noted in the documentation, not to a separate puerperal code by default. Get the depression code right first, then add specifiers as the chart supports them.

Comorbid anxiety. When a client meets criteria for both depression and an anxiety disorder, code both. Don't reach for an unspecified depression code as a workaround. Add the second diagnosis, document both, and let the claim show the full clinical picture.

Ambiguous notes. This is the most common source of bad F32 vs F33 calls. The note says "depression" without episode count, severity, or duration. The biller has a claim to submit and a clock ticking on timely filing. The temptation is to default to F32.9 and move on.

A better workflow: query the clinician before submission. A two-line message asking whether this is a first episode or a recurrent one, and what the documented severity is, takes minutes and produces a code that matches the chart. A dedicated billing coordinator who reads the chart with the same care as the clinician makes this exchange routine instead of awkward.

When a depression claim paired with a psychotherapy CPT like 90834 vs 90837 gets denied or downcoded, the underlying cause often traces back to a depression diagnosis that didn't match the documentation. Strong denial tracking catches whether the pattern points to a coding habit or a documentation gap.

A Simple Decision Flow for Your Practice

Three questions, in order, before any depression claim goes out.

Has the client ever had a prior major depressive episode that fully remitted for at least two months? Yes points to the F33 family. No points to F32.
What severity does the documentation actually support? Mild, moderate, severe with or without psychosis, or remission. Pick the most specific code the chart can back up.
Does the note include duration, symptom count, and functional impact? If not, the chart needs more before the claim is clean.

If the answer to any of these is "I'm not sure," that's a query-the-clinician moment, not a code-it-unspecified moment.

The check works in two places. On the intake side, a short prior-episode question catches recurrence before the first claim ever goes out. On the billing side, the same three questions run before claim submission catch documentation gaps while they're still cheap to fix.

A practice that builds this into routine claim review usually sees two things shift. Unspecified depression codes drop noticeably. And documentation requests from payers slow down, because the codes on the claims actually match what's in the charts.

That's the kind of detail a dedicated billing coordinator brings to the relationship: claim review that reads the chart, not just the cover page. It's the difference between submitting what the practice sent and submitting what the chart supports.

Final Thoughts

F32 vs F33 is a history question, not a severity question. Severity comes next. Unspecified codes should be a last resort, not a default. And the chart has to back up whichever code goes on the claim.

When a billing partner reads the chart with the same attention as the clinician who wrote it, ICD-10 codes for depression get chosen correctly the first time. The practice doesn't pay for sloppy coding in denials, downcodes, or audit flags. The clinician doesn't get pulled into documentation request loops weeks after the visit.

If unspecified depression codes are creeping into your claims, or if F32 vs F33 calls feel like guesswork, BreezyBilling is here to help. We work with behavioral health practices on exactly this kind of detail.

Sources

2026 ICD-10-CM Codes F32: Depressive episode. ICD10Data, 2026.
2026 ICD-10-CM Codes F33: Major depressive disorder, recurrent. ICD10Data, 2026.
Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision (DSM-5-TR). American Psychiatric Association, 2022.
ICD-10-CM Official Guidelines for Coding and Reporting. Centers for Medicare & Medicaid Services, 2026.

ICD-10 Codes for Depression: Choosing Between F32 and F33