Mental health billing and coding guide

ICD-10 coding for mental health: specificity that stops denials.

An unspecified diagnosis code isn't invalid — it's just frequently missing from the covered-diagnosis list a payer's coverage policy actually checks against. F41.9, F32.9, and F43.10 are all valid, billable ICD-10-CM codes, and all three are also a common way to trigger a CO-50 medical-necessity denial when the governing policy names a more specific subtype instead. This guide covers exactly where mental health's four highest-volume diagnosis families — depressive disorders, anxiety disorders, PTSD, and adjustment disorders — need more than the default code, and why the difference decides whether the claim clears.

Key takeaways

  • F32 versus F33 is about episode count, and both carry severity and remission specifiers. A patient with two or more documented depressive episodes belongs in F33, not a permanent F32.9 — and the severity/remission digit is frequently what a coverage policy actually checks.
  • F41.9 is valid but rarely the code a coverage policy names. F41.1 (generalized anxiety disorder) and F41.3 (mixed anxiety disorders) are the subtypes that show up on covered-diagnosis lists far more often than the unspecified default.
  • PTSD's acute-versus-chronic split (F43.11 vs. F43.12) isn't a formality. Duration-based specificity affects testing and higher-intensity psychotherapy authorization review, not just the diagnosis on paper.
  • Adjustment disorder is the most habitually under-coded family in behavioral health. F43.20 (unspecified) gets billed by default when the chart almost always supports F43.21–F43.25 instead — F43.22 (with anxiety) most commonly of all.

Why specificity decides the claim, not just the diagnosis

Every code discussed on this page is a real, HIPAA-valid ICD-10-CM code — verified live against the FY2026 code set for this guide. None of them are wrong to use clinically. The issue is narrower and more mechanical: Medicare LCDs and coverage articles for psychotherapy and psychological testing publish a list of the specific diagnosis codes that support medical necessity for that service, and an unspecified or category-level code frequently isn't on it, even when a more specific code describing the same patient would be. Three coverage documents govern this territory depending on jurisdiction — First Coast Service Options' A57520, "Billing and Coding: Psychiatric Diagnostic Evaluation and Psychotherapy Services" (effective 01/01/2025), First Coast's earlier L33252 LCD of the same name (effective 07/01/2020), and Palmetto GBA's L39853, "Outpatient Psychotherapy" (effective 10/20/2024) — all confirmed live against the CMS Coverage Database for this guide. Find your own MAC's version before treating any one of these as universal; coverage policy in behavioral health is written per contractor, not nationally.

⚠️ This build confirmed the document IDs, contractor names, and effective dates above directly against the CMS Coverage Database. The specific covered-diagnosis code lists inside each article were not independently re-verified here — CMS's article pages returned an access error to automated retrieval during this build — so pull the actual diagnosis list from your governing article before finalizing a coding rule against it. The pattern described below (specific subtype coded, category-level default avoided) reflects standard coverage-policy structure across behavioral health LCDs generally, not a confirmed line-by-line reading of any one article's list.

Depressive disorders: F32 vs. F33, and the specifiers that follow

F32 codes a single depressive episode; F33 codes a recurrent one — the distinction requires the chart to actually establish episode count, not just current symptoms. Both branches carry a severity digit, and F33 adds a separate remission-status branch that F32 splits differently.

Depressive disorder specificity, verified against the FY2026 ICD-10-CM code set. All codes below confirmed billable for HIPAA transactions.
PresentationCodeWhat the chart needs to establish
Single episode, mild / moderate / severe without psychotic featuresF32.0 / F32.1 / F32.2First documented episode; severity per symptom count and functional impact
Single episode, severe with psychotic featuresF32.3Documented psychotic symptoms alongside the depressive episode
Single episode, in partial / full remissionF32.4 / F32.5Symptom reduction from a prior documented episode, not a fresh presentation
Recurrent, mild / moderate / severe without psychotic featuresF33.0 / F33.1 / F33.2Two or more distinct prior episodes documented in the history, not inferred
Recurrent, severe with psychotic symptomsF33.3Recurrent pattern plus current psychotic features
Recurrent, in partial / full remissionF33.41 / F33.42Established recurrent history, current episode resolved or resolving
Single episode / recurrent, unspecifiedF32.9 / F33.9 / F32.AAvoid as a standing default — defensible only when episode count and severity genuinely aren't yet established

The trap here isn't picking the wrong code at intake — it's never revisiting it. A first visit often can't support a severity or episode-count determination, so F32.9 is a legitimate first claim. The problem is the chart that still carries F32.9 eight months and a documented second episode later, when the diagnosis has clearly moved to F33-something and nobody updated the superbill. Recurrent-episode history has to be written into the note explicitly — "second documented episode, prior episode [date], full interim remission" — or a coder has no basis to move off F32 even when the clinician knows the history.

Anxiety disorders: F41.0, F41.1, F41.3 vs. F41.9

The F41 family is smaller than depression's, which makes the unspecified default even harder to justify once the presentation is clear.

Anxiety disorder specificity, verified against the FY2026 ICD-10-CM code set.
CodePresentationDistinguishing feature
F41.0Panic disorderRecurrent, unexpected panic attacks plus persistent worry about future attacks or their consequences
F41.1Generalized anxiety disorderExcessive, hard-to-control worry across multiple life domains for six months or more — the subtype most coverage policies name explicitly
F41.3Other mixed anxiety disordersClinically significant anxiety symptoms that don't cleanly meet full criteria for GAD or panic disorder alone
F41.8Other specified anxiety disordersA named anxiety presentation that doesn't fit F41.0, F41.1, or F41.3 — requires the note to state what the presentation actually is
F41.9Anxiety disorder, unspecifiedValid and billable, but the subtype least likely to appear on a payer's covered-diagnosis list

F41.9 is defensible at an initial visit before diagnostic clarity exists, or genuinely when a patient's presentation doesn't yet sort into a specific subtype. It stops being defensible once a course of psychotherapy is underway and the treatment note keeps describing GAD-consistent symptoms — persistent, hard-to-control worry, months of duration, multiple worry domains — without the diagnosis on the claim ever moving to F41.1. That gap is exactly where an otherwise clean psychotherapy claim denies against a coverage policy that names F41.1 but not F41.9.

PTSD: F43.10, F43.11, F43.12

PTSD's specificity axis is duration, and it's a short, clean decision once the chart states it.

Prior authorization reviewers for psychological testing and step-up levels of care look at chronicity as a clinical urgency signal, not a coding footnote — a chronic, treatment-resistant PTSD presentation reads differently on utilization review than a fresh acute reaction. Coding to F43.10 on a patient who has been in treatment for a year, with a chart that clearly supports chronic, understates the clinical picture on every claim and testing authorization request that follows.

Adjustment disorder: F43.20–F43.29, especially F43.22

This is the family with the widest gap between what the default code says and what the chart actually supports. Adjustment disorder is, by definition, a reaction to an identifiable stressor with a specified emotional or behavioral presentation — the presentation is almost always documented in the note even when the diagnosis code doesn't reflect it.

Adjustment disorder specificity, verified against the FY2026 ICD-10-CM code set.
CodePresentation
F43.21With depressed mood
F43.22With anxiety
F43.23With mixed anxiety and depressed mood
F43.24With disturbance of conduct
F43.25With mixed disturbance of emotions and conduct
F43.29With other symptoms
F43.20Unspecified — avoid once the note names an emotional or behavioral presentation, which it almost always does

F43.22 gets flagged in the title of this guide specifically because it's the most common miss: a patient presenting with anxiety symptoms tied to an identifiable stressor — a job loss, a diagnosis, a relationship change — documented clearly in the note, billed as F43.20 out of habit instead of F43.22. Because the note already contains the qualifying detail in almost every real adjustment-disorder chart, this is close to a pure documentation-to-code mapping gap, not a diagnostic uncertainty problem the way an early-presentation anxiety or depression case can genuinely be.

Do
  • Code to the specific subtype, episode pattern, and duration the chart actually supports at the diagnostic evaluation.
  • Revisit an unspecified code once diagnostic clarity, episode history, or duration is established — the diagnosis on the claim should track the current chart, not the intake note.
  • Check the covered-diagnosis list in your MAC's specific psychotherapy or testing coverage article before defaulting to unspecified out of habit.
Don't
  • Don't leave F32.9, F41.9, F43.10, or F43.20 as a permanent diagnosis once the chart supports a more specific code.
  • Don't assume "unspecified" means invalid — it doesn't. It means a coverage-risk decision, not a coding error.
  • Don't assume every payer's covered-diagnosis list is identical — behavioral health coverage policy is written per MAC, and commercial payers layer their own lists on top.
Pro tip

Run a quarterly report of every active patient still coded to an unspecified diagnosis (F32.9, F41.9, F43.10, F43.20) past their second or third visit. In our experience this single report recovers more denial-prevention value than any individual appeal, because it catches the gap between chart and code before the claim goes out, not after it comes back.

Coding unspecified diagnoses out of habit?

We'll audit a sample of your recent behavioral health claims against the specific covered-diagnosis lists in your MAC's coverage policy and show what's costing you.

Book a free claims review

Frequently asked questions

Is F41.9 (anxiety disorder, unspecified) an invalid ICD-10 code?

No. F41.9 is a valid, billable ICD-10-CM code, confirmed for HIPAA transactions in the FY2026 code set, and using it is never a coding error by itself. The problem is medical necessity, not validity: coverage policies for psychotherapy and testing often name specific anxiety subtypes, most commonly F41.1 (generalized anxiety disorder), as the covered indication, and F41.9 frequently isn't on that list. A claim can be denied for medical necessity while the diagnosis code on it is completely valid — those are two separate questions, and conflating them is why practices keep defaulting to unspecified codes without realizing the risk.

How specific does a PTSD diagnosis need to be for ongoing psychotherapy or testing authorization?

As specific as the clinical picture actually supports at the time of coding, which for PTSD means choosing between F43.11 (acute, generally under three months of symptoms) and F43.12 (chronic, three months or more) rather than defaulting to F43.10 (unspecified) once that duration is established. Authorization reviewers for testing and higher-intensity psychotherapy frequently weigh chronicity directly, since it affects both clinical urgency and expected treatment course, and a note that never states duration leaves the coder stuck defaulting to F43.10 indefinitely even after the chart clearly supports one of the specific codes.

Can we bill an unspecified diagnosis code at intake and update it once the diagnosis is clearer?

Yes, and this is the correct workflow rather than a shortcut — a first visit often doesn't yet support a specific episode count, severity, or duration, so an unspecified code can be the clinically honest choice for that single claim. The failure mode is never revisiting it: once the diagnostic evaluation or a subsequent session establishes the specific subtype, episode pattern, or duration, the diagnosis on the claim needs to move with it. Practices that leave F32.9 or F41.9 as the permanent diagnosis on every claim for a patient with months of established treatment history are coding to the intake note, not the current chart.

Verify before billing. CPT is a registered trademark of the American Medical Association; codes here are paraphrased, not reproduced from the CPT Professional edition. CPT, HCPCS and ICD-10 codes, coverage policy, and bundling edits change, including annual code-set updates. This page reflects standard industry practice and is provided for general education — it is not a substitute for your own compliance review, your current payer contracts, or the current-year code sets. Confirm requirements against your specific payer mix before submitting claims.

Related resources