90832 vs 90834 vs 90837: The Therapist's Guide to Psychotherapy CPT Codes in 2026

If you bill psychotherapy, almost every claim you submit uses one of three CPT codes: 90832, 90834, or 90837. They describe the same service — individual, face-to-face psychotherapy — and the only thing that separates them is the documented time spent with the patient. That single detail decides how much you get paid, whether the claim is denied, and whether the code holds up under audit.

Getting it wrong is expensive in both directions. Undercoding (billing 90834 for a 55-minute session) quietly caps your revenue every visit. Overcoding (billing 90837 without the time to support it) invites a recoupment that can claw back months of payments and flag the provider for ongoing review. This guide walks through the exact time thresholds, what each code's documentation must show, the 2026 reimbursement picture, and the mistakes payers look for first.

Key Takeaways

• 90832, 90834, and 90837 are time-based codes: the code you bill must match documented face-to-face psychotherapy time — not the scheduled appointment slot.
• 90832 covers 16–37 minutes, 90834 covers 38–52 minutes, and 90837 covers 53+ minutes. A session of exactly 52 minutes is 90834, not 90837.
• 90834 is the most frequently billed psychotherapy code nationally, accounting for roughly 45–50% of sessions.
• 90837 reimburses more, so payers scrutinize it — flagging clinicians whose 90837 use runs high relative to their own historical pattern.
• Behavioral health claims carry about a 14.2% initial denial rate, above the 11.8% all-specialty average — and time-code mismatches are one of the most common, most preventable causes.
• Every session note should record start/stop time or total minutes; without it, the code is indefensible under review.

What These Codes Are

90832, 90834, and 90837 all report individual psychotherapy — insight-oriented, behavior-modifying, and/or supportive — delivered face-to-face with the patient. They differ only in the typical session length each describes. Because they are time-based codes, CPT uses a midpoint rule: you bill the code whose typical time is closest once you pass its lower threshold.

90832 — Psychotherapy, 30 minutes

Billed when documented face-to-face time is 16–37 minutes. This is the brief session code — useful for check-ins, stabilization visits, or medication-adjunct therapy. Because it's shorter, the note should briefly justify why a shorter format was clinically appropriate.

90834 — Psychotherapy, 45 minutes

Billed when documented time is 38–52 minutes. This is the standard therapy session and the most-billed of the three. The note should clearly state the modality and interventions that support a standard-length session.

90837 — Psychotherapy, 60 minutes

Billed when documented time reaches 53 minutes or more. Because 90837 pays more, it draws more payer scrutiny: some plans flag clinicians whose 90837 use runs above the norm for prepayment review or records requests. The note must explicitly support the longer session — the clinical complexity, interventions, and time that justify going beyond a standard 45-minute visit.

The Time Rules at a Glance

• Under 16 minutes → not separately billable as psychotherapy.
• 16–37 minutes → 90832 (the 30-minute code).
• 38–52 minutes → 90834 (the 45-minute code).
• 53 minutes or more → 90837 (the 60-minute code).
• Exactly 52 minutes → 90834, not 90837.
• Document start/stop time or total minutes for every session.

The threshold that trips people up most is the 90834/90837 line. A session documented as '50 minutes' is 90834. A session documented as '60 minutes' with no start/stop time is a denial risk, because payers apply medical review and the time isn't verifiable. The decision is based on documented face-to-face psychotherapy time only — never the scheduled slot.

2026 Reimbursement: What Therapists Actually Get Paid

Rates vary by payer, contract, and market, but the ranges therapists are seeing across commercial plans in 2026 look roughly like this:

• 90837 — common range ~$125–$165 (roughly $105 low end, $185+ high end).
• 90834 — common range ~$100–$135 (roughly $85 low end, $150+ high end).
• 90832 — common range ~$75–$100 (roughly $65 low end, $115+ high end).

Two things follow from these numbers. First, the per-session gap between 90834 and 90837 is real — a clinician who consistently sees 55-minute sessions but bills 90834 is undercoding every encounter, and across a full caseload that compounds into meaningful annual revenue. Second, because 90837 pays more, the audit risk concentrates there: payers look for providers whose 90837 rate deviates from their own historical pattern, not just from a national average.

Documentation That Holds Up Under Review

Every code shares a core documentation requirement, and longer sessions carry a higher necessity bar. A defensible psychotherapy note includes:

• Start and stop time, or total minutes spent face-to-face — the single most important field, and the one most often missing.
• The modality and type of intervention (insight-oriented, supportive, behavior-modifying, etc.).
• The clinical content of the session — enough to show psychotherapy actually occurred, not just a medication check.
• For 90837 specifically: explicit clinical justification for why a longer session was necessary.
• For 90832: a brief rationale for the shorter format.

A note that says '60-minute session, therapy' with no start/stop time and no clinical detail will not survive a payer review. The same session documented with times, modality, and a sentence of intervention detail will.

The Most Common — and Most Preventable — Coding Mistakes

1. 1Billing by the scheduled slot instead of documented time. A '60-minute appointment' that ran 45 minutes is 90834, not 90837. This is the single most common source of time-code errors.
2. 2Defaulting to one code out of habit. Practices that bill 90834 for every encounter regardless of actual time are either undercoding long sessions or overcoding short ones — both are systemic errors.
3. 3Missing start/stop times. Without documented time, the code is indefensible. Payers increasingly use automated review to flag notes that lack it.
4. 4Overusing 90837. Because it pays more, a 90837 rate that runs high relative to the provider's own pattern draws prepayment review and records requests.
5. 5Confusing psychotherapy with an E/M visit. A medication management visit is not psychotherapy and should not be billed as one; when both occur, use the appropriate E/M code with a psychotherapy add-on code.

Every one of these is a preventable, catchable error — not a clinical judgment call. They are exactly the kind of mismatch a pre-submission coding scrub is designed to catch before the claim ever reaches the payer.

Add-On Codes: When Psychotherapy Happens With an E/M Visit

When a psychiatrist or other prescriber does both evaluation and management (E/M) and psychotherapy in the same visit, you bill the E/M code plus a psychotherapy add-on code — 90833 (30 min, add-on to 90832), 90836 (45 min, add-on to 90834), or 90838 (60 min, add-on to 90837). Done correctly, this recovers meaningful additional revenue per psychiatrist visit; done carelessly, it triggers denials for lack of supporting documentation that the psychotherapy was distinct from the E/M.

Telehealth: Where Time Codes Meet Modifier and POS Errors

Behavioral health moved heavily to telehealth, and that added a second layer of denial risk on top of the time-code rules. Telehealth modifier and place-of-service errors cause roughly 22% of behavioral health denials. The key points for 2026:

• Use Place of Service 10 (patient's home) for most outpatient telehealth therapy visits — POS 02 now triggers automatic denials from many commercial payers.
• Append Modifier 95 for synchronous audio-video telehealth; omitting it causes denial rates as high as 30–34% on telehealth claims.
• Use Modifier 93 for audio-only telehealth — it is not interchangeable with Modifier 95.
• All psychotherapy codes remain on the Medicare telehealth-approved list, and audio-only mental health sessions are permanently covered — but only when billed with the correct POS and modifier.

Getting the CPT code right is only half the battle on a telehealth claim; the POS/modifier combination is the other half, and a clean time-coded session can still be denied if those fields are wrong.

How Pre-Check Catches These Errors Before Submission

The pattern across all of these mistakes is the same: the error is embedded before the claim is ever submitted, and the denial comes back weeks later. That's the problem a pre-submission coding scrub is built to solve. Before a claim leaves the practice, the right validation can:

• Confirm the billed code matches the documented face-to-face time, not the scheduled slot.
• Flag missing start/stop times before the claim is filed.
• Check the 90837 rate against the provider's own historical pattern to surface audit risk early.
• Verify the telehealth POS and modifier against the specific payer's current telehealth policy.
• Confirm that an add-on psychotherapy code is supported by documentation distinct from the E/M service.

This is exactly what AI Medi Coder's Gate 1 coding scrub does — auditing each claim against time rules, CPT/ICD consistency, modifiers, and payer-specific policy before it's submitted, so the time-code mismatch never reaches the payer in the first place. When a denial does come back, the same process runs in reverse to determine whether it was a coding, billing, or policy issue and get a corrected claim back out fast.

The code that gets audited most isn't the one billed incorrectly most often — it's the one billed at the highest rate compared to the provider's own pattern. Consistency matters as much as accuracy.

The Bottom Line

90832, 90834, and 90837 are simple in principle — bill the code that matches your documented time — and treacherous in practice, because the documentation, the slot, and the payer's review logic don't always line up. The practices that lose the least revenue are the ones that document time every session, pick the code from the note rather than the schedule, and catch mismatches before submission instead of appealing them after. That discipline, applied claim by claim, is the difference between a clean revenue cycle and one that leaks quietly all year.

Sources

• CPT Codes 90832–90837 Guide 2026: Psychotherapy Billing & Time Rules — MedCloud MD
• Psychotherapy CPT Codes 2026: 90832 vs 90834 vs 90837 Guide — Sirius Solutions Global
• Behavioral Health Billing: The Complete 2026 Guide — Go Medical Billing
• 90834 vs 90837: Time Rules, 2026 Rates & Denials — MedPrecision Billing
• Insurance Reimbursement Rates for Therapists in 2026 — Upstate Healthcare Admin
• Mental Health Telehealth Billing: POS 10, Modifier 95 & 2026 Rules — Rapid Growth Trend