Depression With Psychotic Features: ICD-10

Depression with psychotic features codes to F32.3 (single episode) or F33.3 (recurrent) in ICD-10-CM. The codes themselves are easy; what trips up clinicians and coders is the structural mismatch between DSM-5-TR — which lets psychotic features attach to an episode of any severity — and ICD-10-CM, which only offers “severe with psychotic features” codes. This guide covers the two codes as published in the FY2026 tabular list ↗, the severity trap, the differentials, and the documentation that makes these high-acuity claims stick.

This page is part of our ICD-10 directory for behavioral health; for the full mood-disorder block, see the F30–F39 guide.

The codes: F32.3 vs. F33.3

(The titles differ slightly — “features” vs. “symptoms” — an ICD-10-CM quirk with no clinical significance.)

The selection logic is pure episode history: first lifetime major depressive episode → F32.3; any prior episode → F33.3. That makes history-taking a revenue-integrity issue. Psychotic depression typically presents acutely — often at an ED or crisis intake — where longitudinal history is hardest to get. Verify prior episodes through collateral and records before settling the code, and document the basis for “single” vs. “recurrent.”

Mood-congruent vs. mood-incongruent psychotic features both map to the same codes — the tabular inclusion terms cover each explicitly. The distinction still belongs in the documentation: it informs prognosis and the differential with schizoaffective disorder, even though it doesn’t change the ICD-10 code. In DSM-5-TR terms, congruent content tracks depressive themes (guilt, worthlessness, deserved punishment, nihilism); incongruent content does not. (Described here in original words — DSM-5-TR text is not reproduced.)

The DSM/ICD severity mismatch

Here is the trap. DSM-5-TR decoupled psychotic features from severity: under current DSM convention, a clinician may diagnose a major depressive episode of mild or moderate severity with psychotic features, because the psychotic-features specifier is rated separately from the severity rating. ICD-10-CM never made that move. Its only MDD-with-psychosis codes are F32.3 and F33.3, and both embed “severe” in the code title.

The practical consequence: a chart that reads “MDD, moderate, with psychotic features” has no exactly-matching ICD-10-CM code. The coder’s options are F32.3/F33.3 (which assert severe) or F32.1/F33.1 (which drop the psychosis — clinically the most important fact in the chart). Neither is faithful to the note.

How to handle it deliberately:

• Reassess severity in light of the psychosis. In most cases, the presence of delusions or hallucinations within a depressive episode reflects severe illness by functional-impairment standards — many “moderate with psychotic features” notes are under-rated severity, not genuine mismatches. If the clinician concurs on reassessment, document severe and code F32.3/F33.3 with a clean conscience.
• If the clinician genuinely rates it sub-severe, query — don’t silently re-rate. The coder cannot upgrade severity by inference. A documentation query asking the provider to reconcile severity rating with the psychotic-features specifier is the compliant route.
• Write code-supportable documentation in the first place. The clean note states: episode type (single/recurrent), severity with functional anchors, psychotic features present, congruence, and risk findings. That note codes itself.

Differential coding

Psychotic symptoms plus depressed mood has a four-way differential, and each branch changes the code family:

For behavioral health facilities running co-occurring programs, the substance-induced branch is the everyday one: stimulant- or cocaine-related psychosis with depressive features at intake is common, and the 72-hour course often clarifies what the cross-sectional picture cannot. Provisional coding at intake with planned diagnostic review is better practice than committing early to F32.3.

Documentation and medical necessity

F32.3 and F33.3 are among the strongest acuity anchors in the mood-disorder block — psychotic features drive inpatient and PHP-level medical necessity in a way few other depression codes do. To make the claims hold:

• Describe the psychosis, don’t just check it. “MDD severe with psychotic features” as a bare phrase is weak; documented delusional content (in clinical summary form), hallucination modality, the patient’s degree of conviction and concealment, and the functional consequences are strong.
• Document the safety assessment. Psychotic depression carries elevated suicide risk; reviewers expect a risk assessment and safety plan in the record — see our suicidality coding guide for the companion codes (R45.851, Z91.51).
• Show the treatment response. Antipsychotic augmentation alongside antidepressant therapy — or ECT evaluation in refractory cases — is the expected treatment signature. Continued-stay reviews go better when medication rationale and response tracking are explicit in the chart.
• UR language that works ties the code to observable acuity: psychotic symptoms, impaired reality testing, risk findings, failed or infeasible lower levels of care, and 24-hour-structure rationale, updated at each review.

Coding the episode as it resolves

Psychotic depression follows a characteristic course in treatment — psychotic symptoms typically remit before the mood episode fully resolves — and the coding should track it:

Worked example (fictional). A 47-year-old with two prior major depressive episodes is admitted to inpatient care with severe depression, mood-congruent delusions of guilt, and passive suicidal ideation. Inpatient claims carry F33.3 with R45.851 secondary. After three weeks the delusions have resolved; she steps down to PHP, where the psychiatrist documents “MDD, recurrent, severe, psychotic features resolved.” PHP claims carry F33.2, with the psychosis history documented in the continued-stay narrative. At outpatient follow-up four months later in full remission: F33.42.

• Psychosis resolved, depression ongoing. Once psychotic features have remitted but the depressive episode continues, the current-episode code drops to the severity actually present (e.g., F33.2 or F33.1). The psychotic-features history stays in the clinical record and risk formulation even though no current code carries it — which is why discharge summaries and step-down treatment plans should state it narratively.
• Episode in remission. F32.4/F32.5 (single episode, partial/full remission) or F33.41/F33.42 (recurrent) cover the maintenance phase. There is no “in remission, history of psychotic features” combination code; the chart, not the code, preserves that history.
• Relapse vigilance. For a client with prior F33.3 episodes, returning psychotic features mark a severe recurrence — re-code F33.3 and revisit level of care promptly.

Common coding errors

• F32.3/F33.3 without documented psychosis. The code asserts delusions or hallucinations; a checked box without described symptoms is an audit exposure.
• Coding the psychosis separately. Psychotic features within the mood episode are built into F32.3/F33.3 — adding a separate psychotic-disorder code double-reports the same findings.
• Missing bipolarity. Anchoring on F33.3 without screening the history for mania/hypomania; one documented manic episode moves the claim to F31.5.
• Letting F33.3 persist after remission of psychosis. Severity and features coding should move with the clinical course, as above.

Treatment planning

The diagnosis drives a dual-target treatment plan: mood episode resolution and psychosis remission, each with measurable goals, plus relapse-signature monitoring as the psychosis resolves before the mood symptoms (the typical course). For plan structure and goal examples in depressive disorders, see comprehensive depression treatment plans. Initial diagnostic workups at this acuity are billed under 90791; ongoing individual therapy under 90834 / 90837 as clinically indicated.

EHR connects diagnosis, treatment plans, progress notes, and claims in one behavioral-health-native system. Request a Demo: contact us.

Sources: FY2026 ICD-10-CM code set and tabular list ↗ (CMS, effective Oct 1, 2025); ICD-10-CM Official Guidelines for Coding and Reporting, FY2026 ↗; CDC NCHS ICD-10-CM ↗. Accessed June 2026.

Reference tables

Common questions

Official sources

• FY2026 tabular listcms.gov
• ICD-10-CM Official Guidelines for Coding and Reporting, FY2026cms.gov
• CDC NCHS ICD-10-CMcdc.gov