More Contact Hours, Bigger Gains: What 78 Trials Reveal About Treating Childhood Obesity

When a child is diagnosed with obesity, clinicians and families face a deceptively simple question: how much treatment is enough? Behavioral interventions — structured programs of counseling, goal-setting, and lifestyle coaching — remain the cornerstone of pediatric obesity management, yet the field has been haunted by an unresolved debate. Do more sessions, longer programs, or more total hours of contact actually translate into greater weight-related benefit? A new study-level meta-regression published in BMC Medicine by Yakun Liu, Xiaoyan Shi, and colleagues at Children’s Hospital of Soochow University offers one of the most systematic answers to date, drawing on 78 randomized clinical trials involving 16,514 children and adolescents aged 2 to 18.

The research team followed PRISMA guidelines, searching MEDLINE, PsycINFO, and the Cochrane Library from their inception through December 2024, with an updated search in April 2026. To be included, trials had to test behavioral interventions in children with study-defined overweight or obesity, meaning eligibility thresholds at or above the 75th body mass index percentile for age and sex. Two reviewers independently screened studies, extracted data, and assessed risk of bias, a safeguard against the selective judgments that can quietly distort meta-analytic conclusions.

Technically, the analysis was built on a hierarchical effect-size strategy. The authors prespecified an ordering of weight-related outcomes, prioritizing the BMI z score — a measure that adjusts body mass index for age and sex — followed by raw BMI, BMI percentile, percent overweight, and finally body weight. Effect sizes were expressed as Hedges’ g, a standardized mean difference corrected for small-sample bias, and pooled with random-effects meta-analysis. Across 91 comparisons from the 78 trials, the pooled effect size was −0.28 (95% confidence interval, −0.35 to −0.21), a small but statistically robust benefit of behavioral interventions on weight-related outcomes.

That headline number, however, conceals a turbulent landscape. Heterogeneity across studies was substantial, with an I² of 73.4%, meaning nearly three-quarters of the variance in observed effects reflected differences between trials rather than sampling noise. This is precisely where the study’s central methodological move — meta-regression — comes in. Instead of asking simply whether behavioral interventions work, the researchers modeled how specific dose parameters, quantified at the study level, tracked with effect sizes across the entire evidence base.

The dose parameter that emerged most consistently was total contact hours. In the adjusted meta-regression model, each additional hour of intervention contact was associated with a larger treatment effect, with a regression coefficient of −0.016 (95% CI, −0.027 to −0.004). Because more negative values indicate greater weight reduction, the sign matters: programs that accumulated more face-to-face or structured contact time with participants tended to produce measurably larger benefits. Notably, two other intuitive dose metrics — the overall duration of the intervention in weeks and the sheer number of sessions — were not consistently associated with effect sizes once the model accounted for other study characteristics.

This dissociation is one of the study’s most thought-provoking findings. A program stretched over many months with sparse sessions may accumulate little therapeutic momentum, while a denser schedule concentrating meaningful contact hours into a shorter window may deliver more. The implication is that what counts is not the calendar footprint of an intervention but the accumulated time of actual engagement — a distinction that could reshape how clinical guidelines and trial designers conceptualize intensity. The U.S. Preventive Services Task Force, whose recommendations anchor much of pediatric screening practice, has long wrestled with defining adequate intensity; this analysis provides a quantitative anchor for that conversation.

The second major signal concerned who sits in the room. Interventions involving both children and their caregivers were associated with larger effects than child-only interventions, with a coefficient of −0.271 (95% CI, −0.532 to −0.010). The confidence interval barely excludes zero, so the association is statistically defensible but modest, and it aligns with a broad mechanistic rationale: obesity-related behaviors — eating patterns, physical activity, screen time — are embedded in household environments that children rarely control alone. Caregiver participation may reinforce behavioral targets, reshape food availability at home, and sustain motivation between sessions.

The authors are careful to temper these conclusions, and their caution is warranted. Meta-regression operates at the study level, not the individual level, which exposes it to ecological bias: a relationship observed across trials need not hold for any single patient. Residual study-level confounding is also possible — trials with more contact hours may differ systematically in intervention content, population, or setting in ways the model cannot fully capture. Sensitivity analyses yielded directionally consistent findings, strengthening confidence, but the substantial between-study heterogeneity remains a reminder that behavioral interventions for pediatric obesity are far from a monolithic entity.

For clinicians, the practical reading is that total contact hours offer the most reliable dose-response signal available, and that involving caregivers is likely worthwhile. For policymakers and insurers, the findings argue against reimbursing token programs that spread minimal engagement across long periods. And for researchers, the study highlights the value of prespecified, hierarchical outcome selection and transparent dose quantification. As childhood obesity continues to rise globally with long-term consequences, this meta-regression of 16,514 young participants provides a rare quantitative map of the relationship between how much behavioral treatment is delivered and how much benefit it produces — a map that will likely guide the next generation of pediatric weight-management programs.

Subject of Research: Associations between behavioral intervention dose and outcomes in children and adolescents with obesity: meta-regression of randomized clinical trials

Article Title: Associations between behavioral intervention dose and outcomes in children and adolescents with obesity: meta-regression of randomized clinical trials

Article References: Liu, Y., Shi, X., Chen, T., Kong, H., Chen, L., Wang, J., & Huang, S. (2026). Associations between behavioral intervention dose and outcomes in children and adolescents with obesity: meta-regression of randomized clinical trials. BMC Medicine. https://doi.org/10.1186/s12916-026-05323-8

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05323-8

Keywords: Associations, between, behavioral, intervention, dose, outcomes, children, adolescents, obesity, meta-regression, randomized, clinical