In an eighteen-month prospective observational investigation tracking thirty-nine autistic children receiving Early Intensive Behavioral Intervention, the number of delivered one-to-one therapy hours strongly predicted the rate at which children mastered instructional learning targets. Conducted across natural clinical service delivery and published in the Journal of Autism and Developmental Disorders, the evaluation demonstrated that average weekly treatment intensity accounted for approximately 45.5 percent of the variance in children’s average weekly target acquisition. The investigative team observed that this predictive relationship remained robust and statistically significant even after adjusting for baseline chronological age and intake developmental functioning, indicating that the observed differences in skill acquisition were not merely artefacts of initial developmental competence or age at enrollment.
Within this specific clinical cohort, the researchers sought to resolve an ongoing question in autism intervention science regarding how directly prospectively recorded, actual delivered treatment hours correspond to measurable learning gains. While previous intervention research has often relied on authorized, planned, scheduled, or retrospective cumulative service hours, the investigators focused strictly on delivered one-to-one intervention time logged each week. By recording weekly treatment hours alongside verified mastered instructional objectives over an extended observation period, the study examined both overarching between-child trajectories and granular within-child weekly variations.
The study followed thirty-nine children across eighteen continuous months of intervention, compiling a total dataset that encompassed 1,970 positive-hour child-week observations. In the primary sample, participating children received an average delivered intensity of 24.5 hours per week of one-to-one therapy and successfully mastered an average of 41.4 instructional targets each week over the course of the observation window. The researchers also examined an age-restricted subsample comprising twenty-four children, which exhibited comparable intervention metrics, averaging 24.7 hours per week of delivered one-to-one treatment with a standard deviation of 6.7 hours and a range spanning from 6.8 to 32.7 hours. Within this age-restricted subgroup, children mastered a mean of 42.6 targets weekly, with a standard deviation of 15.4 and observed individual rates ranging from 17.5 to 87.1 targets per week.
Linear regression analyses demonstrated a strong positive association between average weekly intensity and average weekly target mastery, yielding a correlation coefficient of r = .674 and an unstandardized regression slope indicating that each additional hour of average weekly intensity was associated with approximately 1.7 additional mastered targets per week. To illustrate the practical clinical magnitude of this slope, the researchers noted that a difference of 5 delivered hours per week corresponded to roughly 8 additional mastered learning targets each week, while an intensity difference of 10 hours per week corresponded to approximately 17 additional mastered targets weekly. In a hierarchical regression framework, the standardized regression coefficient for average treatment intensity remained substantial, between beta = 0.61 and 0.67, confirming that neither baseline chronological age nor baseline developmental scores explained away the dominant predictive contribution of delivered treatment hours.
Beyond between-child averages, the researchers evaluated weekly target mastery across four distinct weekly-intensity bands after excluding non-treatment weeks where zero delivered hours were recorded. Across the 1,970 positive-hour child-weeks, the mean number of mastered targets increased systematically across bands: children mastered an average of 11.6 targets (SD = 9.2) during weeks with more than 0 to fewer than 10 treatment hours, 23.3 targets (SD = 12.9) during weeks with 10 to fewer than 20 hours, 40.8 targets (SD = 18.1) during weeks with 20 to fewer than 30 hours, and 59.1 targets (SD = 25.5) during weeks with 30 or more hours. Consequently, children acquired an average of 47.5 more instructional targets during weeks in which they received 30 or more hours compared to weeks with fewer than 10 hours of intervention.
A more detailed five-hour banded breakdown further clarified this progression across the observed distribution of intervention hours. Weekly mastery climbed from an average of 6.3 targets in child-weeks with more than 0 to fewer than 5 treatment hours up to 63.5 targets in weeks with 35 to fewer than 40 hours, and 65.0 targets in weeks with 40 to fewer than 45 hours. Although child-weeks logging 45 or more hours displayed the highest absolute average target mastery, the authors cautioned that this specific estimate rested upon only 12 recorded observations from 7 individual children. The overall scatterplot across all 1,970 child-weeks yielded an R-squared of 0.446, which the investigators explicitly presented as a descriptive characterization of the dataset rather than an independent inferential estimate, given the inherent nested clustering of repeated weekly measures within individual children.
To rigorously test whether weekly variations reflected a true intra-individual association rather than the passage of time or stable individual characteristics, the authors conducted a child fixed-effects regression. When time in treatment was entered alone, treatment week did not significantly predict target mastery (B = 0.097, p = .264). When weekly treatment intensity was incorporated into the fixed-effects model, treatment week remained nonsignificant (B = -0.053, p = .408), whereas weekly intensity proved highly significant (B = 1.56, SE = 0.12, 95% CI [1.33, 1.80], p < .001). This demonstrates that within the exact same child, each additional delivered treatment hour was associated with approximately 1.56 additional mastered targets during that week, after controlling for linear time in treatment and accounting for stable between-child traits.
The investigation also tracked whether treatment hours resulted in diminishing learning returns by examining instructional efficiency, defined as targets mastered per delivered treatment hour. Across positive treatment weeks, children mastered an average of 1.9 targets per hour during weeks with more than 0 to fewer than 10 hours of intervention, compared with 1.5 targets per hour in the 10 to fewer than 20 hour band, 1.6 targets per hour in the 20 to fewer than 30 hour band, and 1.7 targets per hour in the 30 or more hours band.
In addition to proximal target acquisition, the researchers evaluated broader developmental outcomes using the Learning Accomplishment Profile-Third Edition, or LAP-3, developmental index across twelve months. In the full sample with available twelve-month assessment data (n = 37), average treatment intensity significantly predicted change in the LAP-3 developmental index (R-squared = 0.137, B = 0.560, beta = 0.371, p = .024). In the age-restricted sample (n = 24), this relationship was even more pronounced, accounting for 30.1 percent of the variance in developmental index gains (R-squared = 0.301, B = 0.96, beta = 0.55, p = .006). Similarly, average weekly target mastery was itself a robust predictor of twelve-month LAP-3 developmental change, accounting for 20.7 percent of the variance in the full sample (R-squared = 0.207, beta = 0.46, p = .005) and 31.6 percent of the variance in the age-restricted group (R-squared = 0.316, beta = 0.56, p = .004), demonstrating that proximal weekly learning directly reflects broader developmental progression.
Despite these clear associations, the authors emphasized critical methodological limitations, noting that intervention intensity was naturally occurring rather than experimentally assigned through randomized controlled trial designs. Consequently, the findings demonstrate statistical prediction and association rather than definitive causal dose-response mechanisms. Factors varying from week to week, such as child illness, staffing constraints, behavioral challenges, or curriculum difficulty, could have influenced both delivered hours and target mastery. Furthermore, the researchers noted that higher intensity will not automatically benefit every child uniformly, and that target mastery metrics depend on target selection, sequence, and scoring integrity. The team concluded that clinical hour determinations must remain individualized and responsive to ongoing child progress rather than guided by a universal prescribed dose.
Subject of Research: Association between delivered treatment intensity and developmental outcomes in Early Intensive Behavioral Intervention
Article Title: Treatment Intensity Predicts Target Mastery and Developmental Change in Early Intensive Behavioral Intervention
Article References: Eikeseth, S., Larsson, E. V., Farcas, A., Strømgren, B., & Eldevik, S. (2026). Treatment Intensity Predicts Target Mastery and Developmental Change in Early Intensive Behavioral Intervention. Journal of Autism and Developmental Disorders. https://doi.org/10.1007/s10803-026-07592-x
Image Credits: AI Generated
DOI: 10.1007/s10803-026-07592-x
Keywords: Early Intensive Behavioral Intervention, EIBI, autism intervention, treatment intensity, target mastery, delivered treatment hours, LAP-3 developmental index, developmental outcomes, child fixed-effects regression, learning efficiency, instructional pacing, prospectively recorded hours, autistic children, behavioral intervention, dosage effects, pediatric autism research
News Source: Ophelia Keating. (November 12, 2026). Delivered Treatment Hours Predict Skill Mastery in Early Intensive Behavioral Intervention. Scienmag.



