Science for a Healthier Tomorrow
Adding emotion-regulation training did not speed teenagers’ weight loss during treatment — but a year later those families had regained less, but by less than half the study’s own threshold for a meaningful change.
A two-site randomized trial in 166 families found that adding emotion-regulation skills to family-based weight treatment for teenagers changed nothing while the program ran: after six months, BMI z scores differed by 0.03 (95% CI -0.05 to 0.10). A year after treatment ended, the add-on group had regained less — 0.09 BMI z units (95% CI -0.17 to -0.02), under half the trial’s 0.20 threshold for a meaningful change — and reported less emotional eating. The emotion-regulation score the program targeted never differed between arms.
What this means
In 166 families, adding emotion-regulation skills to family-based weight treatment did not produce more weight loss during the six-month program, but a year later the add-on group had regained less and reported less emotional eating. The difference was smaller than the trial’s own threshold for a meaningful change, and the skill the program targeted — measured emotion regulation — did not move, so the trial cannot explain its own result. Both groups regained weight; the comparison is slower regain, not no regain.
It is possible the study failed to effectively instill emotion-regulation skills in the participants.
Three results that have to be read together
FRESH-Teen randomized 166 parent-teenager pairs to family-based behavioral treatment, or that treatment plus emotion-regulation skills adapted from dialectical behavior therapy and emotion-focused therapy, over six months on the same schedule, and followed to 18 months.
The first result is a null. At the end of treatment, BMI z — a teenager’s BMI against others of the same age and sex — differed by 0.03 (95% CI -0.05 to 0.10). The extra skills did not make the program work faster.
The second is a small difference the other way: 12 months after treatment ended, the between-group BMI z difference was -0.09 (95% CI -0.17 to -0.02), favoring the add-on but less than half the trial’s own 0.20 benchmark for a meaningful change in an individual. By that point 22% of the add-on arm (95% CI 8% to 48%) and 4% of the standard arm (95% CI 1% to 12%) had lost at least 0.20 BMI z units — a secondary proportion analysis with wide, overlapping intervals.
The third is the awkward one. The score the program was designed to change did not change: on the Difficulties in Emotion Regulation Scale the arms differed by -0.06 (95% CI -0.25 to 0.50, P = .63) at 12 months, and at no earlier point.
The trial’s own overall model found no group effect
Read alone, the 12-month figure looks like a maintenance win. The paper’s repeated-measures model, run across all 18 months, separated the arms through its time-by-treatment contrast, not through its group effect: the group effect on %BMIp95 — BMI relative to the 95th percentile for age and sex — was -0.403 percentage points (95% CI -2.978 to 2.172, P = .76). That null belongs beside the -0.09. The per-week slope contrasts and responder proportions are secondary analyses, not the registered primaries.
Both groups regained weight
Neither arm held its loss. BMI z fell in both groups during treatment, by 0.003 per week (95% CI -0.006 to -0.002); after treatment ended it rose in both, by 0.004 per week (95% CI 0.002 to 0.006, P < .001). The arms diverged by 0.003 BMI z per week during follow-up (95% CI -0.005 to -0.001, P = .01), in the add-on arm’s favor: a slower climb, not a plateau, and not a stop.
Why the explanation does not follow the result
The trial can describe what happened. It cannot say why. The mechanism it was built around — better emotion regulation — was measured and did not move, so any account in which these teenagers kept weight off by handling feelings better is a hypothesis this trial did not confirm.
One candidate it did measure: emotional eating, self-reported on the Emotional Eating Scale, was lower in the add-on arm at 12 months by 0.39 points (95% CI -0.63 to -0.15, P = .001), with smaller differences at end of treatment (-0.22, 95% CI -0.42 to -0.03) and six months later (-0.31, 95% CI -0.48 to -0.13) — a behavior score moving the expected way, not evidence of a pathway.
Who was in the trial, and what that bounds
These were 166 treatment-seeking families who enrolled at UC San Diego and University of Minnesota clinics between August 2019 and January 2022, with follow-up visits running through January 2024. Teenagers were 13 to 16 (mean 14.3) with BMI from the 85th to the 99.9th percentile, averaging a BMI z of 2.15; 57% were female, 49% White, 31% Hispanic.
Completion odds did not differ by arm (OR 0.70, 95% CI 0.31 to 1.62, P = .41); main analyses used intention-to-treat with multiple imputation, and per-arm completion counts sit in the trial’s flow diagram. Treatment moved to video conference after the trial’s first weeks, anthropometric measurements were collected remotely during portions of the study, and the authors state the pandemic’s influence is unknown.
So the finding is bounded to teenagers 13 to 16 in treatment-seeking families at two US clinics, below the top of the BMI range — not younger children, not a population sample, not an eating-disorder sample.
Where this sits
Family-based behavioral treatment is the best-supported behavioral option for children, and its effects are understood to shrink in adolescence — the premise this trial was built on, from a review whose identity we verified but did not read. Regain after treatment stops is the ordinary course, which is why maintenance matters more than another loss comparison. The same question is being pursued in adults who stop GLP-1 drugs — in a recent pilot, with an app and medically tailored meals rather than drugs; the behavioral version, in teenagers, has very little trial evidence behind it. Related coverage: How to keep weight off after GLP-1 ends? and steps-weight-maintenance-meta-analysis.
Terms explained
- BMI z score: a teenager’s BMI relative to others of the same age and sex, in standard-deviation units. Zero is average; higher values sit further above. Scaled to age, it is the usual pediatric outcome.
- %BMIp95: BMI as a percentage of the 95th percentile for the same age and sex. A value of 100 sits at that percentile; teenagers here averaged 120.4.
- Family-based behavioral treatment: a program treating child and parent together, combining nutrition and physical activity education with behavior-change strategies and parenting skills, over six months on a fixed schedule.
- Emotion regulation and DBT: emotion regulation is the set of skills for noticing and managing emotional states. The add-on adapted skills from dialectical behavior therapy (DBT) and emotion-focused therapy, measured with the Difficulties in Emotion Regulation Scale (DERS).
- Intention-to-treat: analyzing every participant in the arm they were assigned to, including dropouts, with missing values filled in by multiple imputation.
Sources
- Primary source: Boutelle KN, Manzano MA, Pasquale EK, Strong DR, Bernard RS, Rhee KE, Eichen DM, Miller A, Peterson CB. “Emotion-Focused Behavioral Weight Loss Treatment for Adolescents: A Randomized Clinical Trial.” JAMA Network Open 2026;9(10):e2639867. DOI: 10.1001/jamanetworkopen.2026.39867 · PMID: 42911592. Peer-reviewed two-site parallel-group randomized clinical trial (FRESH-Teen), registered as NCT03674944. Full text opened 2026-10-24: abstract, results, Tables 1-2, figure captions, limitations and article information were read. Supplementary eTables were not opened in the drafting run (Supplement 2 was read during re-verification on 2026-10-24 and confirms the Results-text values); DERS and Emotional Eating Scale time-point values are reported here as described in the results text. Full text is also live at PMC: PMC13691656.
- Landscape — the same question in adults, with a drug: “Preventing Weight Regain After Glucagon-Like Peptide-1 Therapy Discontinuation: A Pilot Randomized Controlled Trial.” Obes Sci Pract 2026. DOI: 10.1002/osp4.70194 (abstract-level access; pilot trial in adults, cited for the shared maintenance question, not for any result reported here).
- Landscape — why adolescence is the hard age: Altman M, Wilfley DE. “Evidence Update on the Treatment of Overweight and Obesity in Children and Adolescents.” J Clin Child Adolesc Psychol 2015;44(4):521-537. DOI: 10.1080/15374416.2014.963854 (context only: bibliographic identity verified through Crossref; no abstract available and full text not examined in this run).
