Why TMS Therapy billing is not straightforward
TMS is considered by payers as a high-scrutiny service. Most payers want proof of diagnosis and medical necessity, and that the treatment course matches policy limitations. Medicare coverage is governed through Local Coverage Determinations (LCDs) and related billing/coding articles rather than a single national rule, so requirements can vary by MAC/jurisdiction, making things more complex.
On the coding side, the CPT set is small but easy to mis-bill in day-to-day operations. 90867 is used for the initial TMS session (mapping, motor threshold determination, and delivery/management), 90868 is used for subsequent treatment sessions, and 90869 is used when motor threshold re-determination is performed with delivery/management and the chart clearly supports why it was necessary. We've seen examples where 90867 is accidentally billed more than once in the same treatment course or when 90869 is submitted without documentation explaining why re-determination was clinically necessary.
Even when the clinical case is appropriate, denials commonly occur when the delivered course doesn’t match the authorized dates/units. We’ve seen this most often when a patient is authorized for a fixed number of sessions but the course extends slightly due to missed visits or scheduling shifts, causing the last few sessions to fall outside the authorization window.
We'll not talk about how TMS Billing can break across the patient lifecycle. In case you want a quick summary of the best practices, navigate to TL;DR section at the end of the article.
Failure points in TMS Therapy billing workflow
1) Intake
Intake documentation often fails to capture the elements payers expect to see when evaluating medical necessity for TMS. Charts may lack clear diagnosis documentation, sufficient history of prior medication trials, or contraindication screening.
A second frequent issue occurs when the ordering and rendering providers do not align with payer credentialing or supervision requirements. We’ve often seen situations where treatment was delivered appropriately but the claim later failed because the rendering or supervising provider did not match the payer’s credentialing rules.
What to do:
- Build a standardized TMS intake template that captures diagnosis, current episode details, and baseline severity scores when payer policies require them.
- Document prior medication trials clearly (drug, dose, duration, and reason for discontinuation or failure).
- Include contraindication screening and relevant clinical history in the intake packet so it is present before authorization or treatment begins.
- Confirm that the ordering provider and rendering clinician meet payer credentialing, supervision, and site-of-service requirements before scheduling the first session.
2) EBV (Eligibility & Benefits Verification)
Once intake is complete, the risk shifts to Eligibility. The most common pitfall here is failing to check if it’s covered for this specific patient and indication. Teams often miss details like prior authorization requirements, visit limits, payer-specific TMS policies, or network restrictions. We’ve seen teams confirm that TMS is listed as a covered benefit but miss plan-level conditions such as prior authorization requirements, step therapy criteria, or visit caps tied to the diagnosis.
What to do:
- Confirm benefit coverage for TMS for the patient’s diagnosis and indication.
- Verify whether prior authorization is required before treatment begins.
- Check visit or unit limits, including taper or maintenance rules.
- Validate in-network status for both the rendering clinician and treatment location.
- Document EBV results clearly, noting that authorization does not guarantee payment without correct coding and medical necessity documentation.
3) Prior Authorization (PA)
Prior authorization failures in TMS usually stem from incomplete submissions or poor tracking after approval. Requests are often sent without the documentation payers expect. Operational gaps also occur when treatment starts before approval, sessions are billed outside the authorized date span, or the clinic continues treatment after authorized units are exhausted without submitting re-authorization. The authorization may approve a fixed number of sessions.
We've seen scenarios where without a system to monitor remaining units the treatment course can extend beyond what the payer approved, leading to write-offs.
What to do:
- Submit a standardized PA bundle: physician order, psych evaluation, prior treatment documentation, requested CPT codes, total units, date span, and clinical rationale aligned with payer criteria.
- Ensure PA approval is logged before the first session and treatment stays within the approved dates and units.
- Track authorization like inventory: remaining sessions, authorization end date, and when to submit re-authorization.
4) Coding & Charge Capture (Session-by-Session)
Coding errors in TMS typically come from misunderstanding how the CPT codes apply across the treatment course. The most common issues include using 90867 more than once in a course, billing 90869 without documentation supporting a re-determination, or pairing CPT codes with ICD-10 diagnoses that the payer does not recognize as medically necessary for TMS. Operational mistakes such as incorrect place of service, rendering NPI, or taxonomy can also cause otherwise valid claims to reject.
What to do:
- Bill 90867 only once at treatment initiation per course, consistent with payer policy conventions.
- Use 90868 for each subsequent treatment session in the course.
- Bill 90869 only when re-determination criteria are documented in the chart.
- Ensure the ICD-10 diagnosis aligns with payer-covered indications for TMS.
- Validate operational details on the claim, including place of service, rendering NPI, and taxonomy.
Common TMS Claim Denials and Fixes
| Denial Reason | What Causes It | How to Manage |
|---|---|---|
| Authorization issues | Service started before PA. Units exceed authorization. Dates outside approval window. | Add a PA gate before the first session. Run a pre-claim check to ensure units ≤ authorized and DOS falls within the approval span. |
| Medical necessity | Diagnosis not covered by payer policy. Missing prior treatment trials. Incomplete clinical documentation. | Use payer-specific templates. Ensure charts show diagnosis, prior trials, severity, and rationale for TMS. |
| Coding errors | 90867 repeated in a course. 90869 billed without documentation. Session billing inconsistent with payer rules. | Add billing system edits to block duplicate 90867. Require structured documentation when 90869 is used. |
| Diagnosis mismatch | ICD-10 on the claim does not match payer-covered indications for TMS CPTs. | Maintain a payer-specific ICD-10 allowlist mapped to CPT codes and update it from payer policies. |
TL;DR: Best Practices
A. Build a chronological “RCM spine” for every TMS patient
- Intake completeness check (diagnosis, prior treatment history, contraindication screening, baseline documentation).
- EBV checklist (coverage, PA requirement, visit limits, network status).
- PA submission with standardized bundle (requested CPTs/units, date span, clinical proof).
- Session-level charge capture rules (90867 once per course; 90868 per session; 90869 only with supported documentation).
- Pre-bill audit (auth match, coding edits, ICD-10 alignment).
B. Align your ops to the expected course structure
A common clinical schedule for depression-related protocols is 30–36 sessions delivered 5 days/week for 4 to6 weeks (varies by protocol and patient), and some policies describe structured intensive + taper/continuation patterns. If your PA request doesn’t match your actual cadence, you create downstream claim risk.
C. Make denials management part of day-to-day workflow
- Categorize denials into auth, medical necessity, coding, eligibility/network.
- For medical necessity denials, appeal with the payer’s own criteria checklist and the specific missing documentation resolved (LCD/policy alignment is often decisive).
D. Maintain a living “payer policy library”
- Store the top payer policies your clinic bills against and update quarterly (or when you see denial pattern changes). Policies change, and small criteria shifts drive big approval swings.
In case you want to evaluate AI Agents
Over time, clinics that consistently win TMS approvals treat RCM as an engineered workflow, not a set of handoffs. AI agents fit naturally into that model, they can continuously check for missing documentation, keep prior auths and authorized units in sync with the treatment course, and flag coding/diagnosis mismatches before claims go out, so denials become the exception, not the operating mode.
The practical takeaway is, when you standardize and monitor every step, approvals follow.
Comments