The 90834 CPT code covers individual psychotherapy sessions lasting 38 to 52 minutes. It is the middle of the three time-based individual psychotherapy codes and the one most outpatient practices bill most often.
The official CPT descriptor reads "Psychotherapy, 45 minutes with patient," while the billable time range runs from 38 to 52 minutes. The descriptor names the typical session; the range decides the claim.
90834 applies to licensed therapists, psychologists, clinical social workers, counselors and psychiatrists delivering psychotherapy in person or by telehealth. It is the default code for routine weekly therapy, which is exactly why two things go wrong with it: clinicians bill it when a different code was correct, and payers move claims into it that belonged somewhere else.
The 90834 CPT Code Time Range
All three individual psychotherapy codes are timed, and the billable range for each is set by the midpoint rule rather than by the number in the descriptor. A session qualifies for a code once it passes the midpoint between that code and the one below it.
| Code | Documented time | Typical use |
|---|---|---|
| 90832 | 16 to 37 minutes | Brief individual psychotherapy. A session under 16 minutes does not meet the threshold for any psychotherapy code. |
| 90834 | 38 to 52 minutes | Routine outpatient individual psychotherapy. The standard "45 minute hour" and the most frequently billed psychotherapy code. |
| 90837 | 53 minutes or more | Extended individual psychotherapy. No upper limit, and the code payers scrutinise most. |
Two boundaries matter more than the rest. A session that ends at 37 minutes is a 90832, not a 90834. A session that reaches 53 minutes is a 90837, and billing it as 90834 is under-billing your own work. One minute decides the code at either end, which is why the time in the note has to be the time that actually elapsed.
What Counts Toward the 38 to 52 Minutes
Only time spent face to face with the patient in the therapeutic encounter counts. That includes clinical assessment within the session and time spent with a family member when it is part of the patient's treatment and the patient is the identified client.
What does not count is the administrative shell around the session: writing the note afterwards, reviewing records beforehand, coordinating with a school or a prescriber outside the session, waiting for a patient who arrives late, or the portion of a telehealth appointment lost to a failed connection. A 50 minute calendar slot is not a 50 minute session, and payers know the difference.
Record the actual elapsed time or the start and stop times. "45 minute session" written into every note in a schedule of 45 minute slots is the single easiest pattern for a reviewer to spot, because it describes the calendar rather than the encounter.
When 90834 Is the Wrong Code Even at 45 Minutes
This is the most expensive mistake in the code, and it applies to prescribers. If psychotherapy is delivered during the same encounter as an evaluation and management service, the psychotherapy is reported with an add-on code, not with 90834.
The add-on codes carry the same time ranges as the standalone codes:
- 90833 for 16 to 37 minutes of psychotherapy with E/M
- 90836 for 38 to 52 minutes of psychotherapy with E/M
- 90838 for 53 minutes or more of psychotherapy with E/M
So a psychiatrist who spends 45 minutes on psychotherapy alongside medication management bills the appropriate E/M code plus 90836. Billing 90834 instead, or billing 90834 and an E/M code together, is a claim built wrong at the structure level rather than at the time level. The E/M portion is selected on medical decision making, and the psychotherapy time is counted separately from the time supporting the E/M service.
Practices with both therapists and prescribers on staff tend to get this wrong in one direction only: the prescribers' claims. It is worth auditing a sample of prescriber encounters specifically against this rule.
90834 vs 90837: The Decision and the Downcoding
The clinical decision is simple. Document the time; the time selects the code. There is no discretion to exercise and no need to average across a caseload.
The commercial reality is less simple. 90837 reimburses more than 90834, and several payers treat a high proportion of 90837 claims as a utilisation flag. The result is a well known pattern: a clinician documents 55 minutes, bills 90837, and the payer reimburses at the 90834 rate or requests records first.
Two responses to that are wrong. Pre-emptively billing 90834 for sessions that ran 53 minutes or longer is under-billing, and doing it across a caseload is a standing revenue loss you have chosen. Ending sessions early to stay inside 90834 is letting a fee schedule set clinical length.
The correct response is to bill the code the documentation supports and make the documentation strong enough to survive review. A 90837 note that names the modality, the clinical reason the session needed the additional time, and the elapsed time is a note that holds up. Our guide to psychiatric billing covers how the codes sit together across a full practice.
What to Do When a Payer Downcodes 90837 to 90834
A downcode is an adjustment, not a denial, and it is appealable. It also frequently goes unnoticed, because the claim was paid and the remittance shows a payment rather than a rejection.
- Find them first. Compare the procedure code submitted against the code paid on the remittance advice. If your reports only show paid versus denied, downcoding is invisible in your own data.
- Read the adjustment reason. A reduction applied without a records request is different from one applied after review, and the two need different appeals.
- Appeal with the note, not with an argument. The elapsed time, the modality and the clinical rationale in the original documentation are the case. Nothing added afterwards carries the same weight.
- Track the rate by payer. One payer downcoding most of your 90837 claims is a contract conversation. Several payers doing it occasionally is normal claim friction.
Downcoding sits in the same operational blind spot as underpayments, and the fix is the same: reconcile what was paid against what was billed rather than against what was expected. Our notes on denial management go further into the workflow.
90832 vs 90834
The difference is time and nothing else. 90832 covers 16 to 37 minutes, 90834 covers 38 to 52. Both are individual psychotherapy, both use the same documentation standard, and both are billed by the same clinician types.
The practical trap is at the bottom of 90832 rather than at the boundary with 90834: a session of 15 minutes or less does not support any psychotherapy code. Brief check-ins, medication-only touchpoints and short crisis calls need to be coded for what they actually were, and for prescribers that often means an E/M service instead.
Which Providers Can Bill 90834, and Where Payers Differ
90834 is not restricted to one licence type. Psychologists, licensed clinical social workers, licensed professional counselors, marriage and family therapists, psychiatric nurse practitioners and psychiatrists can all bill it when psychotherapy is within their scope of practice and they are credentialed with the payer for that service.
Where practices get caught is that scope of practice and payer recognition are two different things. A licence that permits psychotherapy in your state does not guarantee that every plan credentials that licence type for it, and Medicaid programmes in particular vary state by state on which independent practitioners may bill directly rather than under supervision. Associate-level and pre-licensure clinicians are the usual flashpoint: some payers accept them under a supervising provider with a specific modifier, others will not reimburse the session at all.
The practical check is per payer, per licence type, before the clinician starts seeing that plan's patients rather than after the first denial. Where a supervising provider is required, confirm which NPI belongs on the claim, because a correctly coded 90834 still denies if it is submitted under the wrong rendering provider.
Undercoding and Overcoding Cost You in Different Ways
Both directions of error are worth separating, because they carry different consequences and neither is fixed by being cautious.
Undercoding is the quiet one. Billing 90834 for sessions that documented 53 minutes or more is money you earned and did not claim, and because the claim pays cleanly, nothing in your reporting flags it. It compounds silently across a caseload, and it is invisible in any report that only tracks denials.
Overcoding is the loud one. Billing 90837 for sessions the note cannot support raises audit exposure, and a high proportion of 90837 claims relative to peers can prompt a records request on its own. The exposure is not just the disputed claims: an audit that finds a pattern can extend backwards across a period.
The way out is not to pick a safer code. It is to make the documented time reliable enough that the code follows it automatically, so neither error has room to form. If you want to know what undercoding has cost you specifically, take your own contracted rates for 90834 and 90837, take the difference, and multiply by the number of sessions per month that genuinely ran past 53 minutes. That figure is yours rather than an industry average.
Add-On Codes and Modifiers That Travel With 90834
Interactive complexity (90785) is an add-on for sessions complicated by specific communication factors, such as involving a third party, managing a caregiver conflict that interferes with treatment, or using an interpreter or other device to communicate. It is reported alongside 90834 when those circumstances are documented, and it is not a general marker for a difficult session.
Telehealth is reported with 90834 plus the place of service reflecting where the patient was, commonly 02 for telehealth outside the home and 10 for telehealth in the home, and modifier 95 where the payer requires it. Requirements have been revised more than once and differ between Medicare and commercial plans, so confirm the current expectation with each payer rather than carrying forward last year's configuration.
There is no prolonged-service add-on for psychotherapy. This is worth stating plainly, because the advice persists in older material and in some practice templates. CPT deleted the prolonged service codes 99354 and 99355 effective 1 January 2023, and they were not replaced for psychotherapy. The current prolonged-service code 99417 is an add-on for office and outpatient evaluation and management services and cannot be appended to 90834 or 90837. A session that runs past 52 minutes moves to 90837; there is nothing to add on top of it. If a coding reference or internal cheat sheet in your practice still pairs psychotherapy with 99354, it is describing a rule that ended.
What 90834 Pays and How to Look Up Your Rate
90834 reimburses less than 90837 and more than 90832, but there is no single national figure worth quoting. Medicare rates are set by the Physician Fee Schedule and adjusted by geographic locality, so the same code pays differently in different metropolitan areas, and commercial rates are set by your individual contracts. Any article giving you one number for 90834 is giving you an average that does not describe your practice.
For the Medicare figure that applies to you, use the CMS Physician Fee Schedule Look-Up Tool, entering 90834 with your locality and the current year. For commercial rates, the number that matters is the one in your fee schedule, and the number worth watching is what you were actually paid against it. The official code definition is maintained by the AMA, and APA Services maintains practical guidance on the psychotherapy code set.
Documentation Checklist for a 90834 Session
A 90834 note that holds up under review records the following:
- Elapsed time, or start and stop times, reflecting the actual encounter
- The psychotherapeutic modality used
- The presenting problem and current status
- The intervention delivered in this session, specific enough to distinguish it from the last one
- The patient's response and any change in clinical picture
- The plan, including the reason for the next session
- The service modality, in person or telehealth, and the patient's location if telehealth
- A signature and credentials matching the rendering provider on the claim
Notes that fail review usually fail on two of these: time that mirrors the appointment slot rather than the session, and interventions described in language identical to every other note for that patient. Both are products of writing notes hours later from memory. Our note structure guide covers the format side, and behavioral health billing covers where these fields land on the claim.
Where the Coding Decision Should Live
90834 is not a difficult code. It becomes an expensive one at scale, because the decision depends on a fact, the elapsed time, that is accurate at the moment the session ends and steadily less accurate after that. A practice billing 90834 correctly is a practice where session time is captured while it is still a fact rather than reconstructed at the end of a clinic day.
That is a workflow question rather than a coding question, and it is where billing software built for behavioral health earns its place: the time, the modality and the code arrive on the claim from the session itself, and the clinician confirms rather than remembers.
Streamline Your Practice
Accurate 90834 billing comes down to whether session time and clinical detail were captured while they were still facts. Emma handles the documentation and surfaces the coding so that answer is yes by default, and your clinicians review rather than reconstruct. To see it inside a live practice, book a demo with the mdhub team.
90834 reports individual psychotherapy delivered face to face, in person or by telehealth, for a documented 38 to 52 minutes. The CPT descriptor is "Psychotherapy, 45 minutes with patient," and it is the standard code for routine outpatient therapy sessions. It is billed by therapists, psychologists, clinical social workers, counselors and psychiatrists. When psychotherapy is delivered in the same encounter as an evaluation and management service, the add-on code 90836 is used instead of 90834 for the same time range.
38 to 52 minutes of face-to-face therapeutic time. The descriptor says 45 minutes, but the billable range is set by the midpoint rule: below 38 minutes the session is a 90832, and at 53 minutes it becomes a 90837. Only time in the encounter itself counts, so note writing, record review and time spent waiting for a late patient are excluded. Document the actual elapsed time or start and stop times rather than the length of the appointment slot.
Documented time, and nothing else. 90834 covers 38 to 52 minutes and 90837 covers 53 minutes or more, with no upper limit. Both are individual psychotherapy with the same documentation standard. 90837 reimburses more, and some payers flag practices with a high proportion of 90837 claims or reduce those claims to the 90834 rate. The correct response is to bill the code the documentation supports and appeal reductions with the original note, not to pre-emptively bill 90834 for longer sessions, which is under-billing your own work.
Time only. 90832 covers 16 to 37 minutes of individual psychotherapy and 90834 covers 38 to 52 minutes. The clinician types, documentation requirements and modality options are identical. The boundary that catches practices out is the floor rather than the line between the two codes: a session of 15 minutes or less does not support any psychotherapy code, so brief check-ins and medication-only touchpoints need to be coded for what they were, which for prescribers is often an evaluation and management service.
Both endpoints belong to 90834. The billable range is inclusive, so a session documented at exactly 38 minutes qualifies for 90834 rather than 90832, and a session documented at exactly 52 minutes is still 90834 rather than 90837. The switch to 90837 happens at 53 minutes. Because a single minute decides the code at either boundary, record the actual elapsed time or start and stop times rather than the scheduled length, and avoid rounding to the appointment slot. A note reading 45 minutes in a diary of 45 minute slots is the pattern a reviewer looks for first.
Start with a scoped self-audit rather than a bulk resubmission. Pull a sample of dates, compare the documented elapsed time in each note against the code submitted on the claim, and establish whether the pattern is occasional or systematic. Where the documentation clearly supports 90837, corrected claims can be submitted within each payer's timely filing and adjustment window, which differs by plan. Two cautions: only refile where the original note already supports the longer code, because amending notes after the fact to justify a higher code is the thing audits look for, and refile in measured batches rather than sending a large volume of upward corrections at once. If the sample shows a systematic pattern, the durable fix is upstream in how session time is captured, not in the resubmissions.
