Treatment dosage and service intensity

Dosage has several dimensions

A recommendation states:

• component, purpose, frequency, session length, period, and roles
• setting, modality, group size, qualifications, and supervision
• individualized start, review, transition, and ending criteria; contextual assumptions; and benefit, burden, implementation, missed-service, and adverse-effect measures

Codes and units vary by payer and jurisdiction; calculate components separately. The CASP ABA Practice Guidelines Version 3.0 public summary addresses planning, implementation, evaluation, and coordination for ABA treatment of people diagnosed with autism. The full guide requires a license; only the public summary is used here.

Four amounts answer four different questions

A submitted request should preserve each recommended component and period. If payer forms or codes require conversion, keep the clinical recommendation and document the crosswalk. A schedule does not prove delivery. Interpret a gap only after matching the component, code or unit, setting, period, and reason. Record assent or dissent separately from illness, access, staffing, scheduling, and payer barriers.

Individualize the recommendation

Clinical reasoning can include:

1. chosen goals, strengths, baseline, supports, complexity, urgency, and opportunity
2. generalization, maintenance, transition, prior response, and implementation quality
3. communication, health, pain, sensory, mobility, safety, and interdisciplinary context
4. benefit, burden, fatigue, distress, masking risk, and time displaced from relationships, school, play, rest, community, or other care
5. missed service, changing conditions, and feasible safe delivery

The BACB BCBA Test Content Outline, 6th edition is examination content covering assessment, client-informed goals, intervention selection, integrity, data-based modification, collaboration, risk, generalization, and maintenance, rather than a fixed-hours rule. Administrative staff and software may perform arithmetic or flag missing submission fields within documented current capabilities. They do not author dosage, clinical rationale, consent decisions, or payer determinations.

For BCBA and BCaBA certificants and people who completed an application for either credential, the BACB Ethics Code addresses competence, client involvement, applicable consent and assent, preferences, assessment-based intervention, risk, documentation, service agreements, and evaluation. BACB has no separate jurisdiction over organizations or corporations.

Research does not yield one universal hour rule

The studies below answer different questions. Sandbank and colleagues analyzed 144 controlled-group studies with 9,038 children across naturalistic developmental behavioral, behavioral, technology-based, and developmental interventions. Study-level meta-regressions found no significant positive association between daily intensity, duration, or cumulative intensity and intervention effects within intervention type. The null finding has limits: daily and cumulative amount was often unavailable, caregiver and clinician hours could not be examined separately, and study-level models cannot assign an individualized dose.

Eldevik and colleagues obtained participant data from 15 of 17 studies of EIBI lasting at least 12 months for autistic children ages two through six; 341 received EIBI and 280 received comparison care. Weekly hours were a statistically significant predictor of change in adaptive behavior, intellectual functioning, and autism severity. All studies had serious risk of bias from nonrandom assignment. Treatment amount was not randomized, so the association does not show that adding hours caused improvement or identify the best dose for a person.

No universal hour rule follows. Apply relevant population, intervention, outcome, and design; monitor benefit, burden, assent/dissent, implementation, adverse effects, and opportunity costs.

Payer rules define coverage, not clinical truth

For EPSDT-eligible Medicaid members under 21, the May 2026 CMS EPSDT coverage guide says decisions for section 1905(a) services must be individualized, consider long-term need, and cannot apply flat, fixed, hard, or arbitrary limits. State criteria may not be more restrictive than the federal correct-or-ameliorate standard. A soft limit may operate only pending an individualized determination, and services beyond it must be covered when medically necessary. The treating provider recommends or determines need; the state or its contracted plan applies the coverage standard, with the state retaining responsibility. This framework does not govern every commercial product.

As another scoped example, the June 24, 2026 TRICARE T-5 Operations Manual snapshot, Revision C-57, is Chapter 18, Section 3 for the DoD Comprehensive Autism Care Demonstration. It requires recommended CPT codes and units in the treatment plan, contractor clinical-necessity review of goals, targets, progress, and hours before a six-month authorization, required outcome measures, and a reassessment and treatment-plan update for reauthorization every six months. These program requirements do not establish a clinical dose outside the demonstration or guarantee approved units.

When coverage differs, preserve both records. Explain financial/scheduling effects, continuity/safety steps, and applicable review or appeal routes.

A fictional dosage record

Mara is a fictional 12-year-old who uses AAC and chose support for entering two community routines and teaching partners to respond to stop and help messages. For 12 weeks, the clinician recommends three two-hour direct visits weekly: 3 × 2 × 12 = 72 direct hours. Six one-hour partner sessions add 6 × 1 = 6 partner hours. Each component retains its own purpose, code, unit, setting, and period. If each governing code uses 15-minute units, the conversions are 72 × 4 = 288 direct units and 6 × 4 = 24 partner units; verify each code’s actual unit definition before conversion.

The fictional payer authorizes 60 direct hours and 4 partner hours for the same period. The direct gap is 72 − 60 = 12 hours, and authorized divided by recommended is 60/72 = 83.3%. The partner gap is 6 − 4 = 2 hours, and the ratio is 4/6 = 66.7%. Do not combine the components or denominators. If both codes use 15-minute units, the authorized amounts are 240 direct units and 16 partner units. The recommendation remains the clinical record; authorization remains a coverage fact. The team shares the determination, notice, and review route and confirms dates, codes, and locations before scheduling.

During weeks one through four, the practice schedules 12 equal two-hour direct visits, or 24 hours. Ten visits totaling 20 hours occur; one is not furnished because of illness and one because of a staffing failure. Visit completion is 10/12 = 83.3%. Delivered divided by scheduled hours is 20/24 = 83.3%. The percentages match only because every visit had the same scheduled length. Authorization-period utilization is 20/60 = 33.3% at week four, a partial-period measure that does not prove underdelivery. Continuing the same cadence would total 72 hours and exceed the 60 authorized hours, so the schedule must not represent excess hours as covered. The team addresses the gap through applicable review or appeal, continuity planning, and informed financial discussion without changing the clinical recommendation. Mara may pause or decline. Track assent or dissent, AAC access, benefit and burden, outcome data, cancellations, staffing, and adverse effects separately.