Have questions or an idea to bring to life? Fill out the form, and our experts will get back to you promptly!
Empowering businesses across the United States with secure, scalable, and AI-powered digital solutions.
Share This Post
Every behavioral health practice in the United States bills using two psychotherapy CPT codes more than any others: 90834 for 45-minute sessions and 90837 for 60-minute sessions. The distinction between them seems straightforward — it is based entirely on time. Yet this seemingly simple time threshold is responsible for more revenue leakage, more claim denials, and more audit risk than any other single billing factor in outpatient behavioral health.
The financial stakes are significant. CPT 90837 reimburses 15 to 25 percent more than 90834 — an average difference of $25 to $45 per session depending on the payer. For a multi-location behavioral health group with 50 clinicians each seeing 6 patients per day, even a small percentage of coding errors in either direction compounds into tens of thousands of dollars per month in lost or at-risk revenue.
This guide breaks down everything behavioral health operations leaders need to understand about these two codes in 2026 — the exact time thresholds, the documentation requirements that differ between them, the payer-specific rules that trip up billing teams, the audit triggers that put revenue at risk, and the technology approaches that eliminate coding errors before claims are submitted.
The distinction is entirely time-based, but the specific thresholds are where confusion begins.
CPT 90834 covers individual psychotherapy sessions lasting between 38 and 52 minutes of face-to-face therapeutic contact. CPT 90837 covers individual psychotherapy sessions lasting 53 minutes or longer. There is no rounding. A session that lasts 52 minutes is 90834. A session that lasts 53 minutes is 90837. That single minute is the dividing line between two different reimbursement rates.
The critical detail that causes the most billing errors: only face-to-face psychotherapy time counts toward the threshold. Time spent on charting after the session, scheduling the next appointment, speaking with the patient in the hallway, waiting for the patient to settle in, or reviewing intake forms does not count. If a patient arrives 10 minutes late to a scheduled 60-minute appointment and the clinician provides 50 minutes of actual psychotherapy, the correct code is 90834, not 90837 — regardless of what was scheduled.
This is where multi-location behavioral health groups encounter systemic problems. When an organization has 30, 50, or 100 clinicians across multiple locations, each clinician interprets and documents session time differently. Some document actual face-to-face therapy time precisely. Others document the scheduled appointment length. Others document the total time the patient was in the room, including non-therapeutic activities. Without standardized time documentation practices, coding accuracy becomes inconsistent across the organization — and that inconsistency costs real money in both directions.
The revenue impact of 90834/90837 coding confusion operates in two directions, and both cost money.
This is the more common error and the one most practices never detect. A clinician conducts a 55-minute session — clearly within the 90837 threshold — but documents it as a "45-minute session" out of habit or because their note template defaults to 90834. The claim is submitted as 90834 and paid at the lower rate.
The per-session revenue loss ranges from $25 to $45 depending on the payer. That sounds small until you multiply it across a practice. Consider a 10-location behavioral health group with 40 clinicians. Each clinician sees approximately 6 patients per day, 5 days per week. If even 20 percent of sessions that should be billed as 90837 are incorrectly billed as 90834, the math looks like this: 40 clinicians multiplied by 6 sessions per day equals 240 sessions daily. Twenty percent under-coded means 48 sessions per day billed at the wrong code. At an average $35 revenue difference per session, that equals $1,680 per day in lost revenue. Over a month of 20 working days, that totals $33,600. Annualized, the practice is leaving approximately $403,200 on the table — not from denied claims, not from patient no-shows, but purely from billing the wrong code for sessions that were properly delivered.
Most practices never discover this revenue leak because the claims are paid. There is no denial to investigate. The money simply never arrives because it was never billed for. The only way to detect under-coding is to systematically compare documented session durations against billed CPT codes — something most practices lack the infrastructure to do.
This error is less common but far more dangerous. When a clinician bills 90837 for a session that lasted 48 minutes — within the 90834 range — the claim may be paid initially, but it creates significant audit and compliance risk.
Payers, particularly Blue Cross Blue Shield, UnitedHealthcare, and Medicaid managed care organizations, actively monitor 90837 billing patterns. Their audit algorithms flag several specific patterns: practices where more than 70 to 80 percent of sessions are billed as 90837, clinicians who bill 90837 for every single session with no variation, sessions documented at exactly 53 minutes consistently (suggesting manufactured time records rather than genuine clinical variation), and identical progress note language across multiple sessions.
When a payer audit identifies systematic over-coding, the consequences extend beyond the specific claims reviewed. The payer typically demands repayment for all over-coded claims within a look-back period — often 12 to 24 months. For a practice that has been over-coding 30 percent of sessions, a two-year repayment demand can reach six figures. Additionally, the practice may be placed on a corrective action plan, subjected to ongoing heightened review, or in extreme cases, removed from the payer's network entirely.
The complexity multiplies for multi-location and multi-state behavioral health groups because each payer applies slightly different rules to these codes.
Medicare reimburses 90834 at approximately $113 to $120 in non-facility settings in 2026, and 90837 at approximately $154 to $160, depending on geographic locality. The CMS 2026 Physician Fee Schedule includes a conversion factor increase to approximately $33.40 to $33.57, up from $32.35 in 2025. Importantly, CMS finalized a 2.5 percent efficiency adjustment reducing work Relative Value Units for many procedure codes in 2026, but time-based behavioral health codes including 90834 and 90837 are exempt from this reduction.
Medicare requires documentation of the exact start and stop time of therapeutic contact, not just the duration. Progress notes must include the presenting problem, interventions used, patient response, and a treatment plan update. Medicare Administrative Contractors have explicit documentation requirements that, if not met, can result in post-payment recoupment.
LMFTs and LMHCs billing Medicare receive approximately 75 percent of the standard rate — roughly $115 for 90837 instead of $154 — a credential-based differential that affects revenue planning for practices with mixed credential levels.
BCBS plans audit 90837 more frequently than any other psychotherapy code. The higher reimbursement — typically $130 to $155 depending on the state plan — combined with the 53-minute threshold makes it a natural audit target. BCBS audit algorithms specifically flag practices with high 90837-to-90834 ratios, identical start and stop times across sessions, and vague intervention descriptions in progress notes.
Blue Cross and Blue Shield of Kansas explicitly caps unit reporting at one unit for psychotherapy codes including 90834 and 90837 — reporting multiple units on a single date of service triggers automatic denial and recoupment. Blue Cross and Blue Shield of Rhode Island reaffirmed coverage for both codes effective May 2026 but requires add-on codes like interactive complexity (90785) to be billed with an appropriate primary procedure per CPT guidelines.
UnitedHealthcare has historically required prior authorization for 90837 in many plans, though this requirement has been relaxed in some markets. Reimbursement rates tend to fall on the lower end of commercial payers — approximately $110 to $145 for 90837 depending on the plan and provider credentials. Practices billing UHC should verify authorization requirements for each patient's specific plan before scheduling extended sessions.
Aetna reimburses 90837 at approximately $140 to $155 for in-network providers. Aetna's documentation requirements emphasize medical necessity justification for sessions exceeding the standard 45-minute length. Notes must articulate why the clinical situation required the additional time — crisis intervention, complex trauma processing, acute symptom exacerbation, or initial assessment complexity.
Medicaid reimbursement varies dramatically by state and is typically 20 to 30 percent below Medicare rates. Some states have raised behavioral health rates in 2025 and 2026, while others remain stagnant. Most states now use Medicaid managed care organizations for behavioral health, which may negotiate rates slightly above or below the state fee schedule. The key Medicaid-specific consideration is that documentation standards are often more stringent than commercial payers, with state mental health authorities conducting periodic audits that review medical necessity, treatment plan alignment, and time documentation.
Based on patterns observed across multi-location behavioral health organizations, these are the specific errors that cause the most revenue leakage and audit risk.
The first and most prevalent mistake is defaulting to one code regardless of actual session length. Many practices develop a habit of billing 90837 for every session or 90834 for every session, without checking actual time documentation. A clinically healthy billing pattern shows a natural mix — typically 30 to 50 percent of sessions billed as 90837 and 50 to 70 percent as 90834, with variation by clinician, patient acuity, and treatment phase. A practice showing 90 percent of sessions under one code is almost certainly miscoding.
The second mistake is documenting scheduled time instead of actual therapeutic contact time. When a 60-minute appointment results in 48 minutes of actual therapy due to a late arrival, the correct code is 90834. Many clinicians and billing teams default to the scheduled length, creating systematic over-coding.
The third mistake involves inconsistent time documentation formats across clinicians. One clinician writes "session length: 50 minutes." Another writes "met from 2:00 to 2:55 PM." Another writes "45-minute session." Without a standardized format that captures exact start time, exact end time, and total face-to-face therapy minutes, billing teams cannot accurately determine which code applies.
The fourth mistake is failing to document medical necessity for 90837 sessions. Many payers require that 90837 notes specifically justify why the session exceeded the standard 45-minute length. A progress note that reads identically to a 45-minute session note but happens to be billed as 90837 is a red flag for auditors.
The fifth mistake involves copy-forward documentation. When clinicians copy notes from previous sessions and make minimal changes, payer audit systems detect the repetitive language patterns. This triggers heightened review not just of the documentation quality but of the time and coding accuracy for every copied note.
The sixth mistake is overlooking the telehealth modifier requirements. For telehealth sessions billed as either 90834 or 90837, Modifier 95 must be appended and place of service must be coded as 02. The same time documentation rules apply regardless of whether the session is in-person or virtual. Some practices apply the telehealth modifier inconsistently, creating a subset of claims vulnerable to denial.
The seventh mistake occurs when organizations bill 90837 with other time-based codes on the same date of service. The National Correct Coding Initiative sets automatic blocks on certain code combinations. Billing 90837 with 90832 or 90834 on the same day triggers automatic rejection because the codes cover overlapping time ranges.
For a solo practitioner, managing 90834 versus 90837 accuracy is a matter of personal discipline — document time carefully and code accordingly. For a multi-location behavioral health group with dozens of clinicians across multiple states, the challenge is fundamentally different. It becomes a systems problem, not an individual behavior problem.
The organizations that maintain coding accuracy at scale implement three specific capabilities.
Every clinician, at every location, documents session time in the same format: exact start time, exact end time, and total face-to-face therapeutic minutes, with a clear notation of any interruptions or non-therapeutic time. This standardization removes ambiguity from the coding decision. When the billing team sees "Start: 2:03 PM | End: 2:55 PM | Therapeutic contact: 52 minutes," the code selection is mechanical — 90834.
Before any claim leaves the organization, an automated system compares the documented session duration against the billed CPT code. If a clinician documented 48 minutes but the claim shows 90837, the system flags the discrepancy before submission. This catches both under-coding (missed revenue) and over-coding (audit risk) at the point where correction is easiest and cheapest.
Leadership can see coding patterns across all locations in real time: which locations have a healthy 90834/90837 mix, which are over-indexing on one code, which individual clinicians show patterns that need attention, and how coding accuracy correlates with denial rates by payer. This organizational visibility is impossible when each location operates as a documentation island.
The emerging approach in behavioral health documentation uses AI to address the 90834/90837 problem at its source — the moment the clinician creates the note, not after the claim is submitted.
AI-powered documentation systems can automatically track therapeutic contact time based on session data, suggest the appropriate CPT code based on documented duration, validate that the progress note content supports the selected code (for example, flagging when a 90837 note lacks medical necessity language for the extended session), and identify patterns across the organization that suggest systematic coding errors.
The critical distinction is between generic AI documentation tools and systems trained on a specific organization's data. A generic tool applies the same rules to every practice. An organization-specific system learns which payers audit most aggressively, which clinicians tend to under-document time, which locations show coding patterns that need attention, and which note templates produce the highest clean-claim rates for each payer in the organization's specific mix. Over time, the system becomes more accurate for that specific organization than any off-the-shelf tool could be.
If you lead operations at a multi-location behavioral health group, three immediate actions will protect your revenue and reduce your audit risk.
First, pull your coding distribution report. Look at your organization's 90834 versus 90837 billing ratio for the last 6 months. If more than 75 percent of sessions are billed under either code, you almost certainly have a coding accuracy problem. A healthy distribution typically shows 30 to 50 percent 90837 and 50 to 70 percent 90834, with variation by clinician and patient population.
Second, standardize time documentation across all locations. Implement a single, consistent format for documenting session start time, end time, and total therapeutic contact minutes. Ensure every clinician and every note template uses this format. This single change eliminates the ambiguity that causes most coding errors.
Third, calculate your potential revenue exposure. Take the number of sessions billed as 90834 last month, estimate what percentage might have actually met the 90837 threshold based on typical session lengths at your practice, and multiply by the average reimbursement difference. For most multi-location groups, this exercise reveals $20,000 to $50,000 or more in monthly revenue that could be recovered through more accurate coding — without changing anything about how care is delivered.
The 90834/90837 distinction is the single most impactful billing variable in outpatient behavioral health. Getting it right — consistently, across every clinician and every location — is not a billing department issue. It is an organizational operations issue that directly affects your revenue, your audit risk, and your clinicians' daily experience with documentation.
What is the time threshold for billing 90837 instead of 90834?
CPT 90837 requires 53 minutes or more of face-to-face psychotherapy time. Sessions between 38 and 52 minutes should be billed under 90834. Only direct therapeutic contact counts — charting, scheduling, and non-clinical conversation do not contribute to the time threshold.
How much more does 90837 reimburse compared to 90834?
CPT 90837 typically reimburses 15 to 25 percent more than 90834. Under Medicare in 2026, 90837 pays approximately $154 to $160 compared to $113 to $120 for 90834. Commercial payers vary, but the difference generally ranges from $25 to $45 per session.
Can I bill 90837 for a telehealth session?
Yes. The same time documentation requirements apply for telehealth sessions. Append Modifier 95 to indicate the session was conducted via telehealth and use place of service code 02. Documentation must still reflect exact therapeutic contact time of 53 minutes or more.
What triggers a payer audit on 90837 billing?
Common audit triggers include billing 90837 for more than 70 to 80 percent of all sessions, identical start and stop times across multiple sessions, repetitive progress note language (copy-forward), lack of medical necessity documentation for extended sessions, and identical billing patterns across all clinicians at a location.
Can I bill both 90834 and 90837 on the same day for the same patient?
No. The National Correct Coding Initiative blocks this combination because the codes cover overlapping time ranges. Only one psychotherapy code per patient per date of service is permitted.
What documentation is required specifically for 90837 that differs from 90834?
Both codes require standard progress note elements — presenting problem, interventions, patient response, and treatment plan. However, 90837 notes should additionally document why the extended session length was clinically necessary. Medical necessity language explaining the need for a session exceeding 52 minutes strengthens the claim's defensibility in an audit.
What is a healthy 90834/90837 billing ratio for a behavioral health practice?
A clinically realistic distribution typically shows 30 to 50 percent of sessions billed as 90837 and 50 to 70 percent as 90834. Practices where one code accounts for more than 80 percent of sessions should review their coding accuracy and documentation practices.
Metricoid Technology Solutions builds custom AI-powered clinical workflow automation for behavioral health organizations. Our systems standardize documentation, validate CPT coding before submission, and learn from your organization's specific payer mix and denial patterns. To learn how we can help your practice eliminate coding errors and recover lost revenue, visit metricoidtech.com/healthcare or email hello@metricoidtech.com.
Share This Post
Subscribe to receive the latest blog posts to your inbox every week.
By subscribing you agree to with our Privacy Policy.
Stay updated with the latest trends and insights in technology, business strategies, and industry innovations through our blog.
