CPT code 97154

Editorial approval scope: The team checked current source fidelity, scope boundaries, dates, arithmetic, reader usefulness, practical workflow, and general-information limitations.

97154 identifies technician-delivered group treatment

The American Medical Association CPT overview identifies CPT as the code set for reporting medical services and procedures. Exact language and reporting instructions come from the licensed code set and its guidelines. The AMA's behavioral health coding guide places 97154 within the adaptive behavior services family.

In the 2026 family, 97154 concerns timed, face-to-face treatment with multiple patients, delivered by a technician following a protocol under the direction of the professional identified in the licensed guidance. It is reported for each patient whose service satisfies the applicable requirements. One shared group clock cannot prove that every participant had the same eligible time, authorization, provider configuration, or supported service.

The ABA Coding Coalition FAQ is stakeholder guidance, not a substitute for the licensed CPT code set or payer policy. It says not to report 97154 and 97158 concurrently for the same group time and describes a group for either code as two to eight patients. It also says 97158 is for a qualified health care professional (QHP)-led group and may apply even when the QHP concludes that no protocol change is needed.

Group, family, and protocol-modification services differ

Choose the service from what occurred, who participated, and who rendered it:

• 97154 is protocol-based treatment with at least two patients, administered by a technician under QHP direction; report the applicable group code separately for each patient
• 97158 is QHP-administered treatment with multiple patients; its label addresses protocol modification, while the Coalition says a QHP-led session can qualify even if no change results
• 97157 is QHP-led guidance for multiple caregiver sets without patients present; the Coalition says to report it once for each caregiver set associated with a patient
• staff supervision, team meetings, training without the required participant, travel, scheduling, and administrative work require their own analysis

A social-skills group name does not select the code. A group program may include introductions, breaks, caregiver handoff, staff discussion, or another activity outside the reported service. Review each participant and interval against the licensed definition and payer policy.

Payers can define group size and provider rules

As checked August 14, 2026, the current Texas Medicaid Children's Services Handbook provides one program example. It describes 97154 as direct group ABA under a treatment-plan protocol for two to eight children or youth, at a complexity appropriate for a behavior technician (BT). Texas permits delivery by a licensed behavior analyst (LBA) or a supervised licensed assistant behavior analyst (LaBA) or BT; LaBAs and BTs do not enroll separately.

For 97154, Texas requires the treatment note to retain the child's name, date, start and stop times, goals and progress when applicable, covered activities or interventions with time, direct-observation data, and the rendering provider's full signature and credentials. Supervisory signatures follow licensure and professional standards. Texas also bars multiple ABA providers in one session except for separate child and family services when the child is absent from the family session.

These are program rules, not universal CPT requirements. Another payer may impose a narrower group cap or different authorization, provider, modifier, place-of-service, telehealth, unit, or documentation conditions; do not assume payer policy broadens licensed CPT rules.

Use one evidence record per participant

The exact required fields come from licensed and payer guidance. As an internal control, connect the group event to a participant-level record containing:

• client, member, payer, product, authorization, service date, location, and modality
• attendance start and stop events plus each eligible interval
• group census and qualified provider configuration during that interval
• individualized treatment-plan goal or protocol addressed and service delivered
• participant response, data, progress, relevant supports, and safety events
• technician, directing professional, signatures, credentials, enrollment, and roster evidence
• code-set year, payer-rule version, unit calculation, modifiers, and exclusions
• claim, acknowledgment, adjudication, remittance, payment, correction, and appeal state

Keep group-level facts once, then link them to every participant. Shared facts should stay consistent; individualized service and time should remain attributable to the correct record. A group note cannot replace any participant-specific evidence required by the payer or clinical standard.

A fictional group uses participant-interval denominators

A fictional practice schedules four clients for six 15-minute group intervals: 24 planned participant-intervals and 360 planned participant-minutes. Two clients attend all six intervals, one arrives after two and attends four, and one attends five, for 21 attended intervals and 315 attended minutes. At least two eligible participants are present throughout.

Assume each attended interval is exactly 15 minutes and otherwise satisfies the fictional payer rule. The five-interval participant's first interval is only administrative check-in, not treatment under the fictional rule, so eligible minutes by participant are 90, 90, 60, and 60. That converts to 6, 6, 4, and 4 units, or 20 of 21 attended intervals (95.2%) supported. This illustration does not replace the payer's time-aggregation and rounding method; Texas Medicaid, for example, totals each client's billable minutes for the calendar day and applies its published 8-minute rule.

Before release, each record must also pass its own authorization, provider, protocol, documentation, duplicate, group-census, and payer-rule gates. The 95.2% measure cannot establish clinical quality, medical necessity, payer intake, adjudication, or payment.

Coding edits answer a narrow question

The CMS Medicaid NCCI FAQ says NCCI edits concern correct coding rather than medical necessity or prior authorization. CMS also explains that private insurers control how they implement voluntarily adopted edits.

An edit result does not prove that the group was clinically appropriate, a participant received the service, the technician was recognized, the units matched authorization, or the payer covers the code. Record the exact edit file or payer rule, effective date, adjudication unit, modifier logic, and appeal route.

Prepare service-date rules for 2027

The AMA CPT coding resources identify the 2026 code set as effective January 1, 2026. The AMA's September 2025 panel-action summary lists accepted revisions to the adaptive behavior services family for January 2027.

The summary is transition evidence rather than a billable descriptor. Verify the licensed 2027 materials, payer adoption, authorization language, contracts, templates, claims configuration, and staff training before 2027 services. Preserve the 2026 rules for open claims and corrections from 2026 dates.

Measure access, support, and claim results separately

Useful measures include attended participant-intervals divided by planned participant-intervals; supported intervals divided by attended intervals reviewed; authorization-aligned intervals divided by supported intervals due for release; eligible participant records passing every release gate divided by eligible records due; claims accepted into adjudication divided by submitted claims with a final acknowledgment; and denied claims divided by mature adjudicated claims. Report remittance status and payment separately.

Define the group, participant, interval, census rule, exclusion, payer, product, code-set version, claim maturity, and final state. Pair billing measures with participant choice, clinical fit, safety, access, treatment integrity, and outcomes.