Early Intensive Behavioral Intervention (EIBI): A Practitioner's Guide to Dosage, Curriculum, Outcomes, and the Neurodiversity Reckoning

What "EIBI" actually means in 2026 practice

EIBI is one of three principal comprehensive ABA-based treatment models for young children with autism, distinguished from focused (skill-specific) ABA by its breadth of curriculum, intensity of contact hours, and early-life delivery window Gitimoghaddam et al. (2022). The canonical parameters in the implementation literature are 20–30 (and often up to 40) hours per week of one-to-one or small-group programming for children typically under age six, delivered over 1–3 years (Lindgren et al., 2024) Gitimoghaddam et al. (2022). The curriculum is simultaneously broad and deep: imitation, receptive and expressive language, play, social engagement, adaptive and self-care skills, and — when problem behavior emerges — function-based behavior reduction (Lindgren et al., 2024). Inside this envelope, programs blend procedural dialects: traditional discrete trial training (DTT), progressive DTT, naturalistic developmental behavioral intervention (NDBI), embedded teaching, peer-mediated formats, Direct Instruction (DI), and CABAS-style learn-unit accounting (LaMarca et al., 2024) Ferguson et al. (2022) Raulston et al. (2024) Vidovic et al. (2021) Park et al. (2020). The label "EIBI" is therefore better understood as a dose-and-scope envelope around ABA-based early intervention than as a single procedure.

Historical lineage: from Lovaas 1987 to the modern variants

Modern EIBI traces to the UCLA Young Autism Project work published by Lovaas in 1987, which described 40 hours per week of one-to-one ABA-based teaching for preschoolers with autism and reported substantial gains in IQ and educational placement for a meaningful subset of children Gitimoghaddam et al. (2022). Over the following decades, the Lovaas package was extended into program variants that the field still uses: the Early Start Denver Model (ESDM), with its developmentally embedded NDBI curriculum; LEAP, with its inclusive preschool architecture; JASPER, focused on joint attention and symbolic play; CABAS, with its learn-unit-driven verbal-behavior orientation; and progressive DTT, which folded learner-selected targets, varied language, and instructive feedback into the DTT chassis Ferguson et al. (2022) Park et al. (2020) Raulston et al. (2024). The contemporary scoping-review map places EIBI alongside ESDM and other comprehensive ABA-based packages as frontline comprehensive options for young autistic children Gitimoghaddam et al. (2022). The corpus does not adjudicate definitively among these models head-to-head; prudent practice treats them as overlapping rather than competing, and most modern programs draw from multiple lineages within a single curriculum (LaMarca et al., 2024).

Outcomes evidence: what the field can and cannot honestly claim

The most useful 2024 programmatic evidence comes from McKinnon and colleagues' description of an Australian clinic-based EIBI program delivering an average of 27 hours per week of intervention to 154 preschoolers with autism: 84% of children accelerated their rate of learning over an average year of treatment, 52% more than doubled it, and standardized cognitive and adaptive scores improved at the group level, with younger age at entry — but not pre-treatment IQ — predicting larger developmental-trajectory gains McKinnon et al. (2024). The honest reading: at the group level the program looks effective; the heterogeneity is wide, with a meaningful minority not showing acceleration; and the design is single-arm without a control group, so causal language has to stay calibrated McKinnon et al. (2024). Park and colleagues' Korean CABAS case-series across seven preschoolers documented 542,706 learn-units and 7,247 mastered short-term objectives across a year, with above-average developmental gains and remarkably low cost per learn-unit ($0.65) and per objective ($48.90), but again without a comparison group Park et al. (2020). These studies illustrate the basic shape of the EIBI evidence base for the average practitioner: substantial within-program gains, considerable heterogeneity in who benefits how much, and most studies operating at the case-series, single-subject, or single-arm programmatic level rather than randomized comparison McKinnon et al. (2024) Park et al. (2020).

Dosage: 25–40 hours per week is the convention, not a settled prescription

The 25–40 hr/week intensity envelope is anchored more in convention and meta-analytic averages than in head-to-head dose-response trials Gitimoghaddam et al. (2022) (Lindgren et al., 2024). The Australian program landed at a mean of 27 hr/week and reported strong group-level outcomes McKinnon et al. (2024); the Korean CABAS site delivered intensive learn-unit programming and produced above-average developmental gains Park et al. (2020); and current scoping-review work continues to identify 25–40 hr/week as the typical EIBI prescription for young children with ASD Gitimoghaddam et al. (2022) (Lindgren et al., 2024). To formalize dose-planning at intake, Toby and colleagues developed and preliminarily validated the 21-item Patient Outcome Planning Calculator (POP-C) across 63 cases: total POP-C scores were strongly related to insurer-authorized hours, suggesting a defensible objective basis for documenting requested EIBI dosage (Toby et al., 2024). The corpus is honest about what is missing: there is no large prospective trial comparing — for example — 25 hr/week versus 40 hr/week within matched samples on equivalent curricula, and individual-level dose-response prediction remains an open problem the POP-C addresses only partially (Toby et al., 2024) McKinnon et al. (2024).

Comparative outcomes: vs. eclectic, vs. NDBI, vs. focused intervention

The corpus surveyed for this page includes one preschool RCT in the social-communication space and several efficiency comparisons inside EIBI programs, but no head-to-head EIBI-vs-eclectic trial in the extracted set Bourque et al. (2025) Ferguson et al. (2022). Bourque and colleagues' RCT scaling the peer-mediated Stay-Play-Talk package combined with speech-generating devices (SGD) across 45 minimally verbal preschoolers with moderate-to-severe autism produced large, maintained increases in spontaneous intentional communication to peers and generalization to new settings — a piece of evidence that peer-mediated social-communication intervention with SGD can rapidly multiply communicative opportunities even for very young children with the most limited initial repertoires Bourque et al. (2025). Inside EIBI, Ferguson and colleagues' single-subject comparison across four children showed that progressive DTT produced mastery of conditional discriminations several sessions faster than equivalence-based instruction, with higher participant social validity ratings — useful for procedural decisions inside an existing program but not a head-to-head test of comprehensive packages Ferguson et al. (2022). Practically: the corpus supports adding peer-mediated and NDBI-style components inside EIBI rather than rejecting EIBI in favor of NDBI, and within a comprehensive curriculum, progressive DTT formats appear efficient relative to some equivalence-based alternatives for conditional-discrimination teaching Bourque et al. (2025) Ferguson et al. (2022).

Core components: what a defensible EIBI curriculum actually contains

A defensible EIBI program is not a single procedure repeated across hours but a structured curriculum stack. LaMarca and LaMarca's ADDIE-based programming framework explicitly walks BCBAs through assembling the dense, individualized curricula EIBI requires — pulling from progressive ABA, NDBI, embedded teaching, and DTT — with documented analyze/design/develop/implement/evaluate steps so the program is reproducible and supervisable rather than improvised (LaMarca et al., 2024). Within that frame, Mason and colleagues' analysis of 118 children's VB-MAPP echoic trajectories shows that listener and visual-perceptual milestones reliably precede rapid echoic gains; programming early language without first establishing those prerequisites tends to underperform (Mason et al., 2025). When echoic shaping stalls despite differential reinforcement, Sun and colleagues' accelerated auditory-matching protocol moved 27- to 36-month-old CABAS-EI students to ≥90% echoic accuracy across a small number of sessions (Sun et al., 2024). For pre-vocal mand training, Frampton and colleagues' tutorial argues that non-vocal indicating responses (proximity, eye gaze, pulling, pointing) embedded in VB-MAPP and EFL protocols should be systematically taught and probed under EOs, not skipped on the assumption that vocal mands will emerge first (Frampton et al., 2024). When DTT plateaus on intraverbals and problem behavior climbs, LaLonde and colleagues showed that a wh-bingo listener-training format produced rapid intraverbal emergence and near-elimination of problem behavior in two preschoolers — a small sample, but a useful demonstration that game-based formats are legitimate EIBI procedures, not optional flair LaLonde et al. (2020). Direct Instruction can be embedded inside EIBI school models, but only with deliberate implementation planning across BCBA, SLP, and OT — Vidovic and colleagues document the EPIS-framework rollout in a Midwestern charter school for students with ASD, including the staffing and transition tradeoffs that derail naive adoption Vidovic et al. (2021).

For mealtime and feeding, Ibañez and colleagues showed that BST plus a manualized checklist enabled three EIBI-setting behavior analysts to deliver a structured mealtime protocol with high procedural integrity — meaning feeding interventions can live inside the EIBI day rather than requiring a separate feeding clinic Ibañez et al. (2023). For peer engagement and community life-skill teaching, Torelli and colleagues' kindergarten Life Skills Program demonstrated that tiered group-format ABA teaching (whole-group → small-group → individual) with BST and least-to-most prompting produced acquisition for five children with IDD without continuous 1:1 staffing, an existence proof that not every EIBI hour has to be solo with an RBT (Torelli et al., 2026). The high-level takeaway is that a 2026 EIBI curriculum draws from a wide procedural shelf and the work of program design is choosing which procedures fit which targets — not picking a single "method" and applying it to every objective.

Measurement: trial counts vs. engagement vs. learning rate

How EIBI tracks progress determines what it learns to optimize. Raulston and colleagues argue that for NDBI-flavored EIBI with autistic toddlers, rigid trial counts should be replaced — or at least supplemented — with developmentally grounded play-diversity counts, frequency of varied pre-symbolic actions, and interval-rated engagement states (person-engaged, object-engaged, person-object), because the construct of "trial" misses what early intervention is trying to build at this developmental level Raulston et al. (2024). Lewon and Ghezzi's retrospective analysis of 15 children in a center-based EIBI program found that early acquisition rates on the Early Learning Measure — particularly vocal imitation and expressive labeling — predicted later adaptive functioning, autism symptom severity, and support needs at discharge more strongly than pre-treatment standardized scores Lewon & Ghezzi (2021). The clinical implication is concrete: rapid early-skill trajectories are the most reliable mid-treatment signal a BCBA has for whether the program is working for that specific child, and stagnation in vocal imitation and labeling within the first ~12 weeks is a cue to intensify or restructure procedures rather than wait for outcome scores to mature Lewon & Ghezzi (2021). Mason and colleagues' echoic-prerequisite work fits the same logic: track milestone sequences, and if listener and visual-perceptual prerequisites have not been mastered, do not blame slow echoic gains on the child (Mason et al., 2025).

The neurodiversity reckoning: from "indistinguishability" to assent and learner-defined outcomes

The Lovaas-era goal of producing children "indistinguishable" from neurotypical peers no longer survives intact in the contemporary literature, and the modern EIBI corpus is reorienting toward learner-defined outcomes, assent-based instruction, and meaningful generalization rather than topographical normalization. Weber and colleagues trained instructors in an early intensive DTT program for preschoolers with ASD to embed continuous monitoring of student assent and assent withdrawal — reliable detection and brief contingent pauses or choices — without degrading instructional pace (Weber et al., 2025). The program-level argument behind this work is that EIBI cannot be defended as ethical practice in 2026 without a documented process for honoring assent withdrawal during high-density teaching, and the empirical demonstration is that doing so is cheap and trainable (Weber et al., 2025). Tryggestad and colleagues' 17-week BST + biweekly coaching package raised APERS-P-SE-measured environmental quality in preschool units, increasing correct evocative situations and decreasing incorrect ones — i.e., the unit got better at offering meaningful learning opportunities, which is a different optimization target than counting trials (Tryggestad et al., 2025). Raulston and colleagues' play and engagement work explicitly frames its measurement reform as following the child's lead and indexing genuine affect rather than item-based response counts Raulston et al. (2024). Read together, the modern EIBI position is: assent, engagement, and learner-defined functional outcomes are inside the program, not constraints on it (Weber et al., 2025) Raulston et al. (2024) (Tryggestad et al., 2025).

Caregiver-implemented EIBI components and parent training

Caregiver-implemented EIBI is not just a workforce convenience — it is the layer that determines whether intervention generalizes outside the teaching environment. Tomlinson and colleagues' systematic review of telehealth-delivered ABA training documents that caregivers and frontline staff can be taught to implement EIBI procedures with young children with ASD remotely, with single-case demonstrations across the included studies producing high procedural fidelity and comparable child skill gains relative to in-person training Tomlinson et al. (2018). The review also notes that most telehealth EIBI evidence remains single-case and few studies directly compare telehealth with in-person training or examine long-term child outcomes, so caregiver-implemented telehealth EIBI should be deployed with clear fidelity checks and planned generalization probes rather than assumed to be equivalent across all conditions Tomlinson et al. (2018). The practical operating model the review supports: brief BST plus real-time remote coaching (e.g., bug-in-ear) can move caregivers above 80% procedural integrity using a standard webcam and a HIPAA-compliant platform, meeting BACB supervision standards in geographically underserved or waitlist contexts Tomlinson et al. (2018).

Telehealth and hybrid delivery of EIBI itself

Direct EIBI service delivery via telehealth — not just training — is now an empirically supported option, narrowly. Lindgren and colleagues' alternating-treatments single-subject design across three preschoolers already enrolled in EIBI directly compared telehealth-delivered DTT with in-person DTT for object labeling and found rates of skill acquisition closely matched across modalities, with no clinically meaningful differences for any participant (Lindgren et al., 2024). The authors' caveat is the right one to carry into practice: the children in the study had prior DTT exposure, the data do not yet speak to learners completely naive to DTT, and any telehealth EIBI session needs explicit generalization tactics built in to compensate for the loss of in-person context (Lindgren et al., 2024). Combined with Tomlinson and colleagues' systematic review, the operational picture is that hybrid EIBI — in-person plus targeted telehealth components for caregivers, generalization probes, and direct teaching of selected targets — is defensible 2026 practice for many programs, but blanket telehealth-only EIBI for new learners is not yet supported by the published corpus Tomlinson et al. (2018) (Lindgren et al., 2024).

Staff training, supervision pyramid, and program-level fidelity

EIBI's effects depend on the quality of teaching at the moment of delivery, and Tryggestad and colleagues' 17-week staff-training package — a 4-hour BST workshop plus biweekly on-site coaching — raised APERS-P-SE-measured environmental quality in two of three inclusive preschool units, increasing correct evocative situations and decreasing incorrect ones (Tryggestad et al., 2025). The implication is that staff training should not be a single workshop; coaching has to follow into the room, on the schedule, with feedback. Ibañez and colleagues' BST package for structured mealtime protocols moved three EIBI-setting behavior analysts to high procedural integrity quickly, replicating across all participants — meaning that complex protocols can be installed inside an EIBI program without standing up parallel clinics, but only with manualized checklists and rehearsal Ibañez et al. (2023). Weber and colleagues' assent-monitoring training similarly reached fidelity within two sessions using live + video feedback, suggesting that the specific competencies modern EIBI requires (assent detection, learner-defined goal alignment, generalization programming) are trainable inside the supervision pyramid rather than treated as background dispositions (Weber et al., 2025).

Insurance, access, and dosage authorization

Even when a child clinically warrants 30+ hr/week, what gets delivered is often constrained by what insurers authorize and what families can sustain. Toby and colleagues' POP-C — 21 items mapped to age, adaptive level, risk factors, and other variables that predict authorized service hours — was developed precisely to give clinicians and families an objective intake document that aligns clinical need with payer logic before EIBI begins, and was strongly related to authorized hours in the validation sample (Toby et al., 2024). The corpus does not solve the access problem, but the practitioner moves it supports are concrete: complete a structured dosage rationale at intake, document re-evaluation at six-month reviews, and treat dose as a managed variable that changes with progress rather than a fixed prescription (Toby et al., 2024). Where geography or transportation barriers limit access, the telehealth literature offers a partial bridge for caregiver-implemented components and selected direct-teaching targets, but not a wholesale replacement for in-person early teaching for newly-diagnosed learners Tomlinson et al. (2018) (Lindgren et al., 2024).

Maintenance and generalization

Comprehensive early teaching is only worth the hours if skills survive outside the teaching context. The corpus is more programmatic than experimental on maintenance and generalization in EIBI, but several anchors apply. Lindgren and colleagues explicitly note that telehealth DTT sessions need planned generalization strategies bolted on to offset the loss of in-person contextual variability (Lindgren et al., 2024). Bourque and colleagues' peer-mediated SGD intervention produced not only spontaneous communication gains but generalization to new settings — supporting the strategy of using peer-mediated and inclusive components specifically because they program generalization in from the start Bourque et al. (2025). Torelli and colleagues' kindergarten Life Skills program used brief booster sessions plus naturalistic practice with progressive inter-trial intervals to maintain newly taught skills inside existing classroom schedules — a generalizable maintenance pattern for EIBI life-skills programming (Torelli et al., 2026). The implication for program design is to write generalization into the original objective, not append it after mastery — including peers, varied environments, and naturalistic contingencies as design parameters rather than transition tasks Bourque et al. (2025) (Torelli et al., 2026).

A foundation page on EIBI has to be honest about where the evidence sits across tiers — and the corpus surveyed for this page reflects the broader EIBI literature's structural skew toward case-series and single-subject designs with relatively few large RCTs McKinnon et al. (2024) Gitimoghaddam et al. (2022).

Scoping and systematic reviews. Gitimoghaddam and colleagues' scoping review identifies EIBI as one of the three principal comprehensive ABA-based treatment models for young children with ASD and anchors the canonical 25–40 hr/week intensity envelope Gitimoghaddam et al. (2022). Tomlinson and colleagues' systematic review of telehealth-delivered ABA training documents that telehealth coaching reliably produces high procedural fidelity on EIBI procedures and yields comparable child skill gains, while noting that most evidence is single-case and few studies directly compared modalities or measured long-term outcomes Tomlinson et al. (2018).

Randomized controlled trials. Bourque and colleagues' preschool RCT scaling the peer-mediated Stay-Play-Talk + SGD package across 45 minimally verbal preschoolers with moderate-to-severe autism is the strongest randomized evidence in the extracted corpus, showing large, maintained increases in spontaneous communication and generalization to new settings — a component-level RCT rather than a test of comprehensive EIBI as a whole Bourque et al. (2025).

Programmatic case-series. McKinnon and colleagues' description of an Australian clinic-based EIBI program with 154 preschoolers averaging 27 hr/week is the corpus's most useful real-world outcomes anchor: 84% of children accelerated their rate of learning, 52% more than doubled it, group-level cognitive and adaptive scores improved, younger age predicted larger gains, but the design is single-arm without a control group McKinnon et al. (2024). Park and colleagues' Korean CABAS case-series across seven preschoolers documents dense learn-unit accounting, above-average developmental gains, and very low cost per learn-unit and per objective, again without a comparison group Park et al. (2020). Mason and colleagues' analysis of VB-MAPP echoic trajectories across 118 EIBI-receiving children with ASD identifies listener and visual-perceptual prerequisites as predictors of subsequent vocal imitation (Mason et al., 2025).

Single-subject experimental designs. Most procedural EIBI evidence is single-subject (Lindgren et al., 2024). Lindgren and colleagues' alternating-treatments comparison of telehealth vs. in-person DTT across three EIBI preschoolers shows comparable acquisition rates (Lindgren et al., 2024). Sun and colleagues' multiple-baseline study of an accelerated auditory-matching protocol for echoic accuracy across five 27- to 36-month-olds in a CABAS-EI classroom demonstrates rapid mastery (Sun et al., 2024). Lewon and Ghezzi's retrospective single-subject analysis of 15 children's Early Learning Measure trajectories links early-skill rates to later outcomes Lewon & Ghezzi (2021). Tryggestad and colleagues' delayed multiple-probe staff-training study across three preschool units demonstrates raised environmental quality after BST + coaching (Tryggestad et al., 2025). Ibañez and colleagues' multiple-baseline BST study installs structured mealtime protocols across three EIBI behavior analysts Ibañez et al. (2023). Weber and colleagues' single-subject training study for assent detection and assent-withdrawal response across three instructors documents rapid fidelity gains (Weber et al., 2025). Ferguson and colleagues' single-subject comparison of progressive DTT vs. equivalence-based instruction across four children in EIBI shows faster mastery and higher social validity for progressive DTT on conditional discriminations Ferguson et al. (2022). LaLonde and colleagues' game-based intraverbal study (n=2) demonstrates a procedural alternative when DTT plateaus LaLonde et al. (2020). Torelli and colleagues' kindergarten Life Skills evaluation across five children with IDD uses tiered teaching with BST and least-to-most prompting (Torelli et al., 2026).

Methodology and decision-tool papers. Toby and colleagues' POP-C (n=63 validation) provides a structured intake tool for documenting and authorizing EIBI dosage (Toby et al., 2024). LaMarca and LaMarca's ADDIE-based programming framework offers a reproducible structure for assembling EIBI curricula (LaMarca et al., 2024). Raulston and colleagues' methodology paper reframes play and engagement measurement for NDBI-style EIBI Raulston et al. (2024). Frampton and colleagues' tutorial on indicating responses in mand training organizes pre-vocal mand assessment within EI-BI protocols (Frampton et al., 2024).

Theoretical and case-study papers. Vidovic and colleagues' EPIS-framework case study of Direct Instruction implementation in an EIBI school illustrates phased rollout and barrier patterns but reports no student outcomes Vidovic et al. (2021).

Bottom line. The convergent picture across these tiers supports the operational claims this page makes — that EIBI as a 25–40 hr/week comprehensive ABA-based package can accelerate learning for the majority of young autistic children at the program level, that early-skill trajectories are the most reliable mid-treatment signal, that telehealth and caregiver-implemented components are now defensible with appropriate fidelity safeguards, and that assent and engagement measurement can be installed without degrading teaching density McKinnon et al. (2024) Lewon & Ghezzi (2021) (Lindgren et al., 2024) (Weber et al., 2025). The literature is weaker for any precise dose-response prescription within the 25–40 hr/week band, for head-to-head comparisons of comprehensive packages, and for long-term maintenance evidence beyond program discharge — and this page treats those gaps explicitly rather than smoothing them over McKinnon et al. (2024) (Toby et al., 2024).

Practical EIBI decisions are dose, format, and curriculum decisions — not "EIBI or no EIBI" decisions in most cases. A defensible logic, drawn from the corpus:

1. Newly diagnosed preschooler, family available, no medical complications. Document a structured dosage rationale at intake using the POP-C or an equivalent structured tool; align requested hours with payer criteria before authorization rather than after (Toby et al., 2024). Default to the upper portion of the 25–40 hr/week band when access permits and family capacity supports it; younger age at entry predicts larger trajectory gains in real-world programmatic data McKinnon et al. (2024) Gitimoghaddam et al. (2022).
2. Limited initial language and minimal vocal-imitation repertoire. Sequence the curriculum so listener and visual-perceptual VB-MAPP milestones are addressed before pushing echoic targets — the prerequisite structure predicts whether echoic gains will be rapid or stalled (Mason et al., 2025). Probe non-vocal indicating responses (proximity, eye gaze, pulling, pointing) under EOs as initial mand topographies, not as fallbacks (Frampton et al., 2024). If echoic shaping stalls under differential reinforcement, add an accelerated auditory-matching protocol (Sun et al., 2024).
3. DTT plateau on intraverbals or escape-maintained problem behavior during instruction. Switch the format — e.g., embed a wh-bingo listener game — rather than escalate trial counts; corpus demonstrations show rapid intraverbal emergence with reduced problem behavior under game-based formats LaLonde et al. (2020). For conditional-discrimination teaching specifically, progressive DTT formats produced faster mastery than equivalence-based instruction in the available comparison Ferguson et al. (2022).
4. Toddler-aged learner, NDBI-aligned program. Replace rigid trial counts with frequency counts of varied pre-symbolic actions and interval-rated engagement states; write play goals that index diversity and affect, not item-level response counts Raulston et al. (2024). Use the child's instructional play level to set targets, follow the child's lead, and record naturally occurring play variations Raulston et al. (2024).
5. Geographic, transportation, or workforce barriers to in-person delivery. Use telehealth for caregiver training and caregiver-implemented EIBI components first, with brief BST + real-time remote coaching to reach ≥80% procedural integrity Tomlinson et al. (2018). For direct teaching via telehealth, restrict to learners with prior DTT exposure and embed explicit generalization tactics; do not assume telehealth-only EIBI is equivalent for naive learners (Lindgren et al., 2024).
6. No measurable acceleration in vocal imitation or expressive labeling within ~12 weeks. Treat the early-skill stagnation as a program-level signal: intensify or restructure procedures, recheck prerequisite-milestone status, and verify assent and engagement quality — do not wait for pre-treatment standardized scores to play out Lewon & Ghezzi (2021) (Mason et al., 2025).
7. Feeding or mealtime concern in the EIBI client's daily routine. Embed a structured mealtime protocol inside the existing EIBI schedule using BST + manualized checklists rather than referring to a separate feeding clinic; corpus evidence supports installation without parallel infrastructure Ibañez et al. (2023).
8. Inclusive preschool placement available. Use peer-mediated and tiered group-format components — Stay-Play-Talk + SGD for minimally verbal learners, tiered BST + least-to-most prompting for life-skills targets — rather than defaulting to continuous 1:1 staffing Bourque et al. (2025) (Torelli et al., 2026).
9. Six-month review. Re-run the structured dosage tool, re-examine early-skill trajectories, and re-document assent and engagement quality before authorizing the next dose increment; treat dose as a managed variable, not a fixed prescription (Toby et al., 2024) Lewon & Ghezzi (2021).
10. Persistent learner distress, withdrawal, or loss of assent during high-density teaching. Stop, run an assent-and-engagement audit, and retrain instructional staff on assent-withdrawal detection and response with live + video feedback before resuming density (Weber et al., 2025).

Center- and clinic-based EIBI

Center-based delivery is where most of the programmatic outcomes evidence lives McKinnon et al. (2024) Park et al. (2020). McKinnon and colleagues' Australian clinic-based program at 27 hr/week produced acceleration in 84% of 154 preschoolers across an average year of treatment, with younger entry predicting larger gains and standardized cognitive and adaptive scores improving at the group level McKinnon et al. (2024). Park and colleagues' Korean CABAS case-series documents dense learn-unit accounting (~542,706 learn-units across seven children) at low cost per unit, illustrating how structured center-based programs can sustain throughput Park et al. (2020). The corpus supports center-based delivery as the modal setting for high-intensity EIBI and as the natural site for embedding feeding, mealtime, and other complex protocols inside the schedule with appropriate BST Ibañez et al. (2023). Within centers, Tryggestad and colleagues' BST + biweekly coaching package raised environmental quality on the APERS-P-SE in inclusive preschool units, increasing correct evocative situations — meaning that center-level fidelity is a managed variable, not an emergent property of staff motivation (Tryggestad et al., 2025).

Inclusive preschool and school-based EIBI

Inclusive preschool placement is a viable EIBI setting when peer-mediated and tiered group-format components are designed in. Bourque and colleagues' RCT showed that minimally verbal preschoolers paired with trained peers using Stay-Play-Talk scripts plus SGD produced large, maintained gains in spontaneous communication and cross-setting generalization, leveraging the inclusive setting rather than treating it as a generalization afterthought Bourque et al. (2025). Torelli and colleagues' kindergarten Life Skills program demonstrated that whole-group → small-group → individual tiered teaching with BST and least-to-most prompting can build social-communication and adaptive skills without continuous 1:1 staffing (Torelli et al., 2026). For schools attempting Direct Instruction within an EIBI model, Vidovic and colleagues document the EPIS-framework rollout — preparation, active implementation, sustainment — with explicit attention to staffing trust, transition management, and interdisciplinary leadership; the case-study evidence is implementation-process rather than student-outcome data, and the practical lesson is that DI rollout in EIBI schools fails without deliberate planning Vidovic et al. (2021).

Home-based and caregiver-implemented EIBI

Home-based delivery and caregiver-implemented components have a clearly supported role, especially for generalization, low-rate behaviors that don't appear in clinic, and underserved geographies. Tomlinson and colleagues' systematic review of telehealth-delivered ABA training shows caregivers and frontline staff can be coached to high procedural fidelity on EIBI procedures via secure video platforms, narrowing geographic gaps in service access Tomlinson et al. (2018). Brief BST plus real-time remote coaching can move caregivers above 80% procedural integrity using a standard webcam and a HIPAA-compliant platform, meeting BACB supervision standards Tomlinson et al. (2018). The caveat is that most of this evidence is single-case and few studies directly compared telehealth with in-person training, so caregiver-implemented components should be deployed with explicit fidelity probes and planned generalization checks Tomlinson et al. (2018).

Telehealth and hybrid delivery

Direct telehealth EIBI for children already in EIBI is supported by Lindgren and colleagues' alternating-treatments comparison of telehealth vs. in-person DTT across three preschoolers, with comparable acquisition rates for object labeling (Lindgren et al., 2024). The defensible operating model in 2026 is hybrid: in-person foundation teaching plus targeted telehealth components for caregiver training, generalization probes, and selected direct-teaching targets in learners with prior DTT exposure Tomlinson et al. (2018) (Lindgren et al., 2024). Blanket telehealth-only EIBI for newly-diagnosed naive learners is not yet supported by the published corpus and should not be marketed as equivalent (Lindgren et al., 2024).

• Treating dosage as the goal rather than a means to outcomes. The 25–40 hr/week envelope is a typical prescription, not a definition of clinical success; rapid early-skill acquisition rates predict outcomes more strongly than pretreatment scores, so progress at the skill level — not adherence to a target hour count — is the right ongoing signal Lewon & Ghezzi (2021) McKinnon et al. (2024).
• Drift from analysis to procedure-by-rote. EIBI programs that abandon the analyze/design/evaluate cycle drift toward procedure-without-function; an ADDIE-style framework keeps programming decisions explicit and supervisable rather than improvised (LaMarca et al., 2024).
• Pushing echoic targets before listener and visual-perceptual prerequisites are mastered. Mason and colleagues' analysis across 118 children shows rapid echoic gains follow — not precede — those milestones; programming out of sequence wastes teaching time and frustrates the learner (Mason et al., 2025).
• Ignoring assent and engagement. EIBI delivered without continuous monitoring of assent and assent withdrawal is no longer defensible 2026 practice, and the empirical work shows assent monitoring can be installed within two sessions of training without degrading instructional pace (Weber et al., 2025).
• Trial counts as the only progress metric in NDBI-style EIBI. For toddler-aged learners on NDBI-flavored curricula, rigid trial counts miss the play diversity, engagement state, and pre-symbolic action variation the program is supposed to build; flexible frequency counts and interval engagement ratings index the actual target Raulston et al. (2024).
• Parent-implementer fidelity gaps treated as training items rather than maintenance items. Caregiver-implemented EIBI components require ongoing fidelity probes — not just initial BST — to remain effective; the systematic review documents that ongoing remote coaching is what sustains parent procedural integrity above mastery Tomlinson et al. (2018).
• No outcome accountability beyond hours billed. Programs that document hours but not learn-units mastered, environmental quality, or early-skill trajectories cannot diagnose underperformance until it is too expensive to fix Park et al. (2020) (Tryggestad et al., 2025) Lewon & Ghezzi (2021).
• Defaulting to continuous 1:1 staffing for every objective. Tiered group-format teaching with BST and least-to-most prompting can build life-skills targets without 1:1 escalation, so 1:1 should be a programming choice, not an automatic baseline (Torelli et al., 2026).
• Telehealth EIBI sold as identical to in-person for naive learners. The published evidence supports telehealth direct teaching for children with prior DTT exposure and embedded generalization tactics — not as a wholesale substitute for in-person foundation teaching (Lindgren et al., 2024) Tomlinson et al. (2018).

• Suspected medical or biological substrate driving the behavior pattern. Sleep, GI, seizure, pain, or other medical contributors warrant evaluation before high-density behavioral teaching is intensified; the dose decision changes when a non-behavioral variable is plausibly in play.
• Persistent feeding refusal or aspiration risk that cannot be addressed within the EIBI mealtime protocol. Refer to a multidisciplinary feeding team while maintaining the structured mealtime protocol where safe to do so Ibañez et al. (2023).
• Severe or escalating problem behavior requiring an experimental functional analysis the local team cannot deliver safely. Refer to a specialist behavioral team rather than running an underpowered FA inside the EIBI program.
• No measurable acceleration in any early-skill domain across two consecutive 12-week reviews despite curriculum revision and assent/engagement audits. Refer for external clinical consultation — including consideration of whether a different comprehensive model (ESDM, JASPER, LEAP) or a focused intervention is a better match — rather than continuing the same dose without diagnosis Lewon & Ghezzi (2021) McKinnon et al. (2024).
• Workforce or fidelity ceiling. When BST + coaching cannot bring staff to ≥80% procedural integrity on core EIBI procedures across two training cycles, refer the case to a regional consultation team rather than running an underpowered program in-house Tomlinson et al. (2018) (Tryggestad et al., 2025).
• Family preference for a non-ABA framework after informed discussion of options. Refer to the chosen model rather than persisting against family values; assent extends to the family-level decision about modality.

The honest read of the corpus is that EIBI's operational claims — that 25–40 hr/week comprehensive ABA-based programming accelerates learning for the majority of young autistic children at the program level, that early-skill trajectories predict outcomes, that telehealth and caregiver-implemented components can be deployed safely with appropriate fidelity — sit on solid programmatic and single-subject evidence, while the comparative-effectiveness layer remains thin McKinnon et al. (2024) Lewon & Ghezzi (2021) Tomlinson et al. (2018). Several gaps are tractable.

The dose-response question — whether 25 hr/week, 30 hr/week, or 40 hr/week produces materially different outcomes for matched children on equivalent curricula — has not been answered with prospective head-to-head trials in the extracted corpus, and the POP-C is a useful authorization aid rather than a dose-response prediction model (Toby et al., 2024). Prospective trials randomizing dose within plausible bands would clarify both clinical practice and payer policy more than any further single-arm program description McKinnon et al. (2024). Comparative-effectiveness work between comprehensive packages — EIBI, ESDM, LEAP, JASPER, progressive ABA, and high-quality eclectic intervention — paired with shared outcome measures would tell us when one model is clearly preferable to another and when clinical equipoise should drive family choice Gitimoghaddam et al. (2022).

The neurodiversity reorientation also requires its own outcomes science. Weber and colleagues' assent-monitoring training is an existence proof that the competency is trainable; what we still need is a longitudinal study linking assent-quality, engagement-quality, and learner-defined goal attainment to long-term well-being outcomes, not just immediate procedural fidelity (Weber et al., 2025). Raulston and colleagues' play-and-engagement measurement reform similarly invites a measurement-validation step: does interval-rated engagement during NDBI-style EIBI predict downstream developmental outcomes better than trial-count metrics? — a question the field can now answer with existing program data Raulston et al. (2024).

Telehealth EIBI for naive learners, longitudinal maintenance evidence beyond program discharge, and the relative contributions of caregiver-implemented vs. clinician-implemented components within the same comprehensive program are all open empirical questions (Lindgren et al., 2024) Tomlinson et al. (2018). Each is tractable; none requires the field to invent new methods.

1. Treat 25–40 hr/week as a typical prescription, not a definition of success. Dose is a managed variable that should be re-justified at intake and at six-month reviews using a structured tool such as the POP-C; the underlying signal of program quality is acceleration in early-skill trajectories, not adherence to a target hour count (Toby et al., 2024) McKinnon et al. (2024) Lewon & Ghezzi (2021).
2. Document a structured dosage rationale before authorization. Align requested EIBI hours with payer logic at intake using the POP-C or an equivalent structured tool; this turns dose conversations into evidence-anchored decisions rather than negotiation (Toby et al., 2024).
3. Track early-skill trajectories — especially vocal imitation and expressive labeling — as the most reliable mid-treatment signal. Stagnation across the first ~12 weeks is a cue to intensify or restructure procedures, not to wait for outcome scores to mature Lewon & Ghezzi (2021).
4. Sequence the language curriculum to match VB-MAPP prerequisite structure. Address listener and visual-perceptual milestones before pushing echoic targets; rapid echoic gains follow those prerequisites (Mason et al., 2025). When echoic shaping stalls under differential reinforcement, add an accelerated auditory-matching protocol (Sun et al., 2024).
5. Teach and probe non-vocal indicating responses as initial mands. Proximity, eye gaze, pulling, and pointing can serve as functional mand operants and bridge to later vocal speech; treat them as primary curriculum, not fallback (Frampton et al., 2024).
6. Use an ADDIE-style framework to script comprehensive curriculum decisions. EIBI programs that document analyze/design/develop/implement/evaluate cycles produce reproducible, supervisable curricula and avoid drift to procedure-without-analysis (LaMarca et al., 2024).
7. Prefer progressive DTT formats for conditional-discrimination teaching. In the available comparison, progressive DTT produced faster mastery and higher social validity than equivalence-based instruction within EIBI Ferguson et al. (2022).
8. Switch the format when DTT plateaus on intraverbals or escape behavior climbs during instruction. Game-based formats (e.g., wh-bingo) produced rapid intraverbal emergence and reduced problem behavior in the available demonstrations — escalating trial counts is not the answer LaLonde et al. (2020).
9. For NDBI-flavored toddler EIBI, replace rigid trial counts with engagement and play-diversity metrics. Frequency counts of varied pre-symbolic actions and interval-rated engagement states index what the program is actually trying to build Raulston et al. (2024).
10. Embed peer-mediated and tiered group-format components in inclusive placements. Stay-Play-Talk + SGD for minimally verbal preschoolers and tiered BST + least-to-most prompting for life-skills targets reduce reliance on continuous 1:1 staffing and program generalization in from the start Bourque et al. (2025) (Torelli et al., 2026).
11. Install assent and assent-withdrawal monitoring across all instructional staff. Live + video feedback can bring instructors to fidelity within two sessions, and the practice is now a defensibility floor, not an extra (Weber et al., 2025).
12. Build feeding and mealtime protocols inside the EIBI day, not in a parallel clinic. BST + manualized checklists move EIBI behavior analysts to high procedural integrity on structured mealtime protocols quickly, with fidelity maintained inside the natural EIBI environment Ibañez et al. (2023).
13. Use telehealth for caregiver-implemented components first; restrict direct telehealth teaching to learners with prior DTT exposure. Hybrid delivery is defensible; blanket telehealth-only EIBI for naive learners is not yet supported by the published corpus Tomlinson et al. (2018) (Lindgren et al., 2024).
14. Treat staff training as ongoing coaching, not a single workshop. A 4-hour BST workshop followed by biweekly on-site coaching raised environmental quality on the APERS-P-SE; one-shot training does not (Tryggestad et al., 2025).
15. Document outcomes at the learn-unit and environmental-quality level, not just hours billed. Programs that track learn-units, mastered objectives, and APERS-style environmental ratings can diagnose underperformance early; programs that only track hours cannot Park et al. (2020) (Tryggestad et al., 2025).

Primary research synthesized in this guide. DOIs link to the original source.

• Lindgren, N. A., Higbee, T. S., Osos, J. A., Nichols, B., & Campbell, V. E. (2024). Comparing the effectiveness of discrete trial training delivered via telehealth and in‑person on skill acquisition. Behavior Analysis in Practice, 17, 783–795. https://doi.org/10.1007/s40617-023-00855-4 https://doi.org/10.1007/s40617-023-00855-4
• McKinnon, K., Cihon, J. H., & McEachin, J. (2024). A programmatic description of an early, intensive behavioral intervention program in Australia. Behavioral Interventions, 39(4). https://doi.org/10.1002/bin.2060 https://doi.org/10.1002/bin.2060
• Tryggestad, H., Eldevik, S., Eikeseth, S., Strømgren, B., & Kingsdorf, S. (2025). Staff training to improve the preschool educational environment for children with special needs. Behavior Analysis in Practice. https://doi.org/10.1007/s40617-025-01123-3 https://doi.org/10.1007/s40617-025-01123-3
• Ibañez, V. F., Peters, K. P., Bacotti, J. K., Lloveras, L. A., Van Arsdale, A., & Vollmer, T. R. (2023). Teaching behavior analysts in an early intensive behavioral intervention setting to conduct a structured mealtime protocol. Behavioral Interventions, 38(1), 159-177. https://doi.org/10.1002/bin.1915 https://doi.org/10.1002/bin.1915
• Sun, Y., Sun, T., Farrell, C., & Nuzzolo, R. (2024). The effects of an accelerated auditory matching protocol for early intervention students. Behavior Analysis in Practice, 17, 553–564. https://doi.org/10.1007/s40617-023-00882-1 https://doi.org/10.1007/s40617-023-00882-1
• Lewon, A. B. & Ghezzi, P. M. (2021). An evaluation of the Early Learning Measure as a predictor of outcomes in early intensive behavioral intervention. Behavioral Interventions, 36(2), 388-406. https://doi.org/10.1002/bin.1768 https://doi.org/10.1002/bin.1768
• Park, H. L., Du, L., & Choi, J. (2020). Education for Children With ASD in South Korea: A Case Study. Behavior Analysis in Practice, 13(4), 838-848. https://doi.org/10.1007/s40617-020-00453-8 https://doi.org/10.1007/s40617-020-00453-8
• Raulston, T. J., Ousley, C. L., Hinton, E. M., & Ramirez, A. M. (2024). Beyond Trial Counts: Considerations for Measuring Play and Engagement During Early Intervention for Autistic Children. Behavior Analysis in Practice, 17(4), 1216-1227. https://doi.org/10.1007/s40617-024-01002-3 https://doi.org/10.1007/s40617-024-01002-3
• Toby, L. M., Hustyi, K. M., Hartley, B. K., Dubuque, M. L., Outlaw, E. E., & Logue, J. J. (2024). Development and preliminary validation of the Patient Outcome Planning Calculator (POP‑C): A tool for determining treatment dosage in applied behavior analysis. Behavior Analysis in Practice, 17, 601–614. https://doi.org/10.1007/s40617-023-00861-6 https://doi.org/10.1007/s40617-023-00861-6
• Tomlinson, S. R. L., Gore, N., & McGill, P. (2018). Training Individuals to Implement Applied Behavior Analytic Procedures via Telehealth: A Systematic Review of the Literature. Journal of Behavioral Education, 27(2), 172-222. https://doi.org/10.1007/s10864-018-9292-0 https://doi.org/10.1007/s10864-018-9292-0
• Gitimoghaddam, M., Chichkine, N., McArthur, L., Sangha, S. S., & Symington, V. (2022). Applied Behavior Analysis in Children and Youth with Autism Spectrum Disorders: A Scoping Review. Perspectives on Behavior Science, 45(3), 521-557. https://doi.org/10.1007/s40614-022-00338-x https://doi.org/10.1007/s40614-022-00338-x
• Mason, L., Bolds, A., Gavagan, M., & Ninness, C. (2025). Understanding echoics: identifying predictive indicators of vocal imitation. The Analysis of Verbal Behavior, 41, 84–100. https://doi.org/10.1007/s40616-024-00213-7 https://doi.org/10.1007/s40616-024-00213-7
• Bourque, K., Buzhardt, J., Kamps, D., Heitzman-Powell, L., & Zhang, E. (2025). Interventions Benefitting Young Autistic Children. Education and Treatment of Children, 48(3), 245-258. https://doi.org/10.1007/s43494-025-00158-z https://doi.org/10.1007/s43494-025-00158-z
• LaMarca, V. J., & LaMarca, J. M. (2024). Using the ADDIE model of instructional design to create programming for comprehensive ABA treatment. Behavior Analysis in Practice, 17, 371–388. https://doi.org/10.1007/s40617-024-00908-2 https://doi.org/10.1007/s40617-024-00908-2
• Torelli, J. N., Snyder, S. K., Griffin, M. L., Hellemn, M. P., Cagliani, R. R., & Morgan, G. A. (2026). Life skills evaluation in a kindergarten special education classroom. Journal of Applied Behavior Analysis, 59(1), e70041. https://doi.org/10.1002/jaba.70041 https://doi.org/10.1002/jaba.70041
• Weber, J., Weiss, M. J., & Ferris, K. (2025). Training instructors to support assent and assent withdrawal during instruction for students with disabilities. Behavior Analysis in Practice. https://doi.org/10.1007/s40617-025-01122-4 https://doi.org/10.1007/s40617-025-01122-4
• Vidovic, J. L., Cornell, M. C., Frampton, S. E., & Shillingsburg, M. A. (2021). Adventures in Direct Instruction Implementation: The Devil Is in the Details. Behavior Analysis in Practice, 14(3), 839-855. https://doi.org/10.1007/s40617-021-00616-1 https://doi.org/10.1007/s40617-021-00616-1
• LaLonde, K. B., Jones, S., West, L., & Santman, C. (2020). An Evaluation of a Game-Based Treatment Package on Intraverbals in Young Children with Autism. Behavior Analysis in Practice, 13(1), 152-157. https://doi.org/10.1007/s40617-019-00397-8 https://doi.org/10.1007/s40617-019-00397-8
• Frampton, S. E., Axe, J. B., Davis, C. R., Meleshkevich, O., & Li, M.-H. (2024). A tutorial on indicating responses and their importance in mand training. Behavior Analysis in Practice, 17, 1238–1249. https://doi.org/10.1007/s40617-024-00965-7 https://doi.org/10.1007/s40617-024-00965-7
• Jessel, J., Fruchtman, T., Raghunauth‑Zaman, N., Leyman, A., Lemos, F. M., Costa Val, H., Howard, M., & Hanley, G. P. (2024). A two step validation of the performance‑based IISCA: A trauma‑ informed functional analysis model. Behavior Analysis in Practice, 17, 727–745. https://doi.org/10.1007/s40617-023-00792-2 https://doi.org/10.1007/s40617-023-00792-2
• Ferguson, J. L., Cihon, J. H., Majeski, M. J., Milne, C. M., Leaf, J. B., McEachin, J., & Leaf, R. (2022). Toward Efficiency and Effectiveness: Comparing Equivalence-based Instruction to Progressive Discrete Trial Teaching. Behavior Analysis in Practice, 15(4), 1296-1313. https://doi.org/10.1007/s40617-022-00687-8 https://doi.org/10.1007/s40617-022-00687-8