Skip to content

Treatment in adolescents

TL;DR — Eating-disorder-focused family therapy, commonly operationalized as family-based treatment (FBT), has the strongest adolescent psychotherapy evidence. It mobilizes caregivers to interrupt restriction and restore nutrition before returning developmentally appropriate autonomy. In one RCT, FBT outperformed adolescent-focused individual therapy on full remission at 6- and 12-month follow-up — but not at end of treatment, where the two were similarly effective; in another, parent-focused treatment produced higher end-of-treatment remission than conjoint FBT but not statistically different follow-up remission (Lock 2010, PMID 20921118; Le Grange 2016, PMID 27453082). The Cochrane synthesis is more restrained than the individual trials: family approaches beat treatment as usual on low-quality evidence from two small studies, and did not clearly beat other psychological interventions (Fisher 2019, PMID 31041816). “Family-based” should not be mistaken for blaming families or for a single feasible model in every household.

Treatment architecture

Phase/task Core aim Measurement
Medical triage Determine whether outpatient care is safe Vitals, trajectory, laboratories/ECG as indicated
Early nutritional restoration Caregivers temporarily organize meals and interrupt compensatory behaviour Weight/growth, intake, instability
Return of control Transfer eating responsibility as stability permits Sustained eating and developmental readiness
Adolescent development Resume age-appropriate identity, peer and autonomy work Function, school, relationships, symptoms
Relapse prevention Recognize early restriction and act Individualized warning/response plan

Randomized evidence

Trial/review Sample Contrast Quantitative result
Lock 2010 121 adolescents aged 12–18, DSM-IV AN without the amenorrhoea requirement; 24 outpatient hours over 12 months FBT vs adolescent-focused individual therapy No difference in full remission at end of treatment; FBT significantly superior at both 6- and 12-month follow-up. FBT superior for partial remission and BMI percentile at EOT only, with no follow-up difference (PMID 20921118)
Le Grange 2016 107, age 12–18 Parent-focused vs FBT, 18 sessions/6 months EOT remission 43% vs 22%; OR 3.03, 95% CI 1.23–7.46; no significant 12-month difference (PMID 27453082)
Le Grange 2014 79 long-term participants Follow-up of FBT/AFT trial 2/33 previously remitted relapsed; 10/44 newly remitted; selected follow-up (PMID 25440306)
Fisher 2019 Cochrane review, 25 trials (16 adolescent, 8 adult, 1 mixed) Family therapy vs treatment as usual, education, or other psychological interventions Advantage over TAU on post-intervention remission RR 3.50 (95% CI 1.49–8.23) but from only 2 studies and 81 participants, low-quality evidence, not maintained at follow-up; versus other psychological interventions no clear advantage (RR 1.22, 0.89–1.67; 5 studies, n=252) or at long-term follow-up (RR 1.08, 0.91–1.28; 4 studies, n=200). 68% of studies at high risk of selective reporting bias (PMID 31041816)
Austin 2025 Systematic review and meta-analysis; 23 publications across 18 RCTs Eating-disorder-focused family therapy vs other psychotherapies, and FT-ED variants FT-ED produced significantly greater weight gain than individual psychotherapy at end of treatment; parent-only or separated delivery gave preferable end-of-treatment weight outcomes and recovery rates than conjoint FT-ED; no other meta-analysed outcome reached significance at end of treatment or follow-up (PMID 39041682)

Where the evidence started, and what it originally showed

The founding trial is narrower than its reputation. Eighty patients (57 AN, 23 bulimia nervosa) were first admitted for inpatient weight restoration and only then randomized, before discharge, to family therapy or individual supportive therapy for one year. Family therapy was more effective only in patients whose illness was not chronic and had begun before age 19; the more tentative finding was that individual supportive therapy was of greater value in older patients (Russell 1987, PMID 3318754). Two features of that design are routinely lost in summary: everyone was weight-restored as an inpatient before randomization, and the age-and-chronicity restriction was the trial's own conclusion, not a later caveat.

The second foundational question — whether the family must be seen together — was answered by a trial whose five-year follow-up traced all 40 patients and reassessed 38. At five years there was "little to distinguish" conjoint from separated family therapy, with over 75% symptom-free, no deaths, and relapse in only 8% of those who had reached a healthy weight at end of treatment. The single durable difference was a moderator: patients from families with raised maternal criticism did worse if offered conjoint family meetings, and that disadvantage was still visible at five years (Eisler 2007, PMID 17537071). Expressed emotion, not diagnosis, is the variable that selects delivery format in the only long-term randomized data available.

Beyond single-family therapy

Trial Design Result
Eisler 2016, PMID 27881106 Pragmatic multicentre RCT, 169 adolescents, multifamily therapy (MFT-AN) vs single-family therapy (FT-AN), assessed to 18 months Good or intermediate Morgan-Russell outcome in just under 60% (FT-AN) vs over 75% (MFT-AN) at end of treatment — OR 2.55, 95% CI 1.17–5.52, p = 0.019. The advantage was no longer statistically significant at 18-month follow-up. About half of FT-AN and nearly 60% of MFT-AN patients had resumed menstruating
Lock 2024, PMID 38142046 Adaptive RCT (NCT03097874, 130 enrolled, completed); 107 families recruited, the 69 who failed to gain 2.4 kg by session 4 randomized to continue standard FBT or add 3 sessions of intensive parental coaching Adding parental coaching did not improve remission overall. It helped only where baseline parental self-efficacy was low (moderator). The 2.4 kg-by-session-4 early-response predictor was replicated

The MFT-AN between-group difference converged by 18 months, the same pattern seen in conjoint-versus-parent-focused delivery. These trials are consistent with a difference in speed rather than final outcome, but neither was designed to test that interpretation directly.

Early response is the most replicated prognostic signal

Across a two-site trial of 121 adolescents randomized to FBT or adolescent-focused therapy, receiver-operating-characteristic analysis identified a gain of 5.8 lb (2.65 kg) by session 3 in FBT (AUC 0.670, p = .043) and 7.1 lb (3.20 kg) by session 4 in AFT (AUC 0.754, p = .014) as the earliest predictors of end-of-treatment remission. Early weight gain did not predict remission at follow-up in either arm, and weight trajectory was only marginally superior in FBT (Wald χ² = 3.692, p = .055) (Le Grange 2014, PMID 24190844). A separate trial replicated the same signal — early gain in FBT predicting greater end-of-treatment weight gain and remission, and remission at 12 months (Madden 2015, PMID 26488111) — and the 2.4 kg-by-session-4 threshold was replicated again a decade later (Lock 2024, PMID 38142046).

The awkward part is what to do about it. The one randomized attempt to act on early non-response by intensifying parental coaching produced no overall benefit (Lock 2024, PMID 38142046). Early response is therefore a well-replicated prognostic marker whose prescriptive value has been tested once and largely failed.

Setting: how much hospital, and what kind

Two randomized trials address the setting question directly, and both point away from longer inpatient stays.

  • 82 medically unstable adolescents aged 12–18 admitted to two Australian paediatric units were randomized to shorter hospitalization for medical stabilization or longer hospitalization for weight restoration to 90% expected body weight, each followed by 20 sessions of manualized FBT. There was no significant difference in hospital days after the initial admission; the weight-restoration group used significantly more total hospital days and more post-protocol FBT sessions. Participants with higher eating psychopathology and compulsive features did better in the medical-stabilization arm (Madden 2015, PMID 25017941).
  • In the ANDI non-inferiority trial, 172 female patients aged 11–18 at first admission across six German centres were randomized after three weeks of inpatient care to continued inpatient treatment or day-patient treatment, with identical programmes and intensity. Day patient was non-inferior on 12-month BMI (mean difference 0.46 kg/m² in favour of day patient, 95% CI −0.11 to 1.02; p_non-inferiority < 0.0001), with a similar count of treatment-related serious adverse events (8 inpatient vs 7 day patient) (Herpertz-Dahlmann 2014, PMID 24439238).

Neither trial licenses discharging an unstable patient. Both were built on a period of inpatient medical care and a defined step-down; what they tested was the duration of that care, not its necessity. See service models and setting.

Parent-focused treatment

In PFT the therapist works with parents while a nurse separately monitors the adolescent. Its end-of-treatment advantage over conjoint FBT in the 107-person trial was not maintained as a statistically significant difference at 6 or 12 months: at 12 months remission was 37% versus 29% (OR 1.39, 95% CI 0.60–3.21) (Le Grange 2016, PMID 27453082). This suggests a delivery option, not universal superiority.

Delivery at a distance, and at lower therapist cost

Two questions have moved since the manualized in-person trials: whether family treatment survives videoconference delivery, and how little therapist time it needs.

A two-site pilot randomized 40 adolescents aged 12–18 with DSM-5 AN and their families to twelve 20-minute guided sessions of an online guided self-help version of FBT for parents (GSH-FBT), or fifteen 60-minute videoconference FBT sessions for the whole family. Recruitment and retention were similar to in-person trials, acceptability was similar between arms, and both produced medium-to-large effect sizes for weight, EDE score, parental self-efficacy and remission, maintained at 3-month follow-up. Between-group clinical differences were small — but efficiency (outcome per unit of therapist time) showed a large effect size favouring GSH-FBT, which uses about one fifth of the therapist contact (Lock 2021, PMID 34553395). A confirmatory trial of online guided self-help FBT is currently recruiting 200 participants (NCT05563649, live ClinicalTrials.gov query 2026-09-02).

The cost case is separately modelled. In an Australian Markov model over six years, FBT was less costly than adolescent-focused individual therapy and cost AUD 5,089 per DALY averted against no intervention (95% UI dominant to 16,659), against AUD 51,897 for adolescent-focused therapy (21,591 to 1,712,491); at a AUD 50,000/DALY threshold FBT was 100% likely to be cost-effective and adolescent-focused therapy 45% likely (Le 2017, PMID 29044637). A Swiss non-randomized comparison adding home treatment to FBT (40 vs 21 patients) found greater three-month weight restoration (70% vs 52% at >85% expected body weight) at roughly a third of the cost per patient meeting the criterion, driven almost entirely by reduced inpatient use (Mayr 2025, PMID 39776084).

Both economic results carry the same caveat as the clinical ones: the modelled analysis inherits its effectiveness estimates from the same small trials, and the home-treatment comparison was not randomized.

The pathway trial that questioned inpatient care

TOuCAN randomized 167 young people (mean age 14 years 11 months) across 35 English services to inpatient care in an experienced unit, a specialist outpatient programme, or general community CAMHS. There was significant improvement in every arm and no difference in effectiveness between inpatient and outpatient, or between specialist and generalist, once baseline characteristics were accounted for; good outcome reached 19% at 1 year, 33% at 2 years and 64% of those still followed at 5 years. Adherence to allocation was 49% for inpatient care against 71% and 77% for the outpatient arms (p = 0.013), and the specialist outpatient programme was the dominant strategy economically (Gowers 2010, PMID 20334748). The authors state that the trial gives "little support to a stepped-care approach in which inpatient care is offered to outpatient non-responders" — a conclusion that sits against routine practice in many systems and has not been overturned. See service models and setting.

Applicability and adaptation

Context Adaptation question
Single caregiver/work constraints Can meal support be delivered without impossible attendance demands?
Family conflict or safety concerns Is conjoint work safe; what external supports are available?
Older adolescent/emerging adult How is autonomy preserved while mobilizing supports?
Atypical AN Are medical risk and target trajectory recognized without low-BMI gate?
Neurodivergence/cultural context Are communication, sensory and meal expectations adapted?

Adolescent treatment must run in parallel with medical management. The Society for Adolescent Health and Medicine position paper explicitly covers restrictive eating disorders in adolescents and young adults, including those not underweight (PMID 36058805). A review of adolescent treatments concludes that family-based treatment has the best evidence of both efficacy and cost-effectiveness among current options (Lock 2019, PMID 31443871) — the cost dimension matters because it is one of the few respects in which the adolescent evidence base is clearly ahead of the adult one.

Open questions

  • Which component — parental empowerment, meal exposure, externalization, rapid nutrition or treatment intensity — mediates FBT outcome? As of September 2026 no dismantling or component trial of family-based treatment for AN has reported. The available evidence contrasts whole delivery formats (conjoint vs parent-focused, single-family vs multifamily) or adds a component rather than removing one, and neither design isolates an active ingredient (Le Grange 2016, PMID 27453082; Eisler 2016, PMID 27881106; Lock 2024, PMID 38142046; Austin 2025, PMID 39041682).
  • Why do format advantages converge? Parent-focused over conjoint FBT, and multifamily over single-family therapy, both produce significant end-of-treatment differences that are no longer significant at 12–18 months — a pattern consistent with faster rather than better recovery, and never explicitly tested (Le Grange 2016, PMID 27453082; Eisler 2016, PMID 27881106).
  • Can early non-response be acted on? The 2.4 kg-by-session-4 predictor is replicated three times, but the single randomized attempt to intensify treatment for early non-responders helped only families with low baseline parental self-efficacy (Le Grange 2014, PMID 24190844; Madden 2015, PMID 26488111; Lock 2024, PMID 38142046).
  • Should expressed emotion determine delivery format? Raised maternal criticism predicted worse outcome under conjoint family meetings at end of treatment and still at five years, in the only long-term randomized comparison; it is not routinely measured in practice (Eisler 2007, PMID 17537071).
  • Who benefits most from parent-focused rather than conjoint delivery? The 107-person trial identified several moderators of primary outcome, and the 2025 meta-analysis found separated/parent-focused delivery preferable on end-of-treatment weight and recovery — but neither yields a rule for allocating an individual (Le Grange 2016, PMID 27453082; Austin 2025, PMID 39041682).
  • How should family models be adapted for young people without available or safe family support?
  • Can guided self-help FBT match full FBT at a fifth of the therapist time? A 40-family pilot found small between-group clinical differences and a large efficiency advantage; a 200-participant confirmatory trial is recruiting (Lock 2021, PMID 34553395; NCT05563649, retrieved 2026-09-02).
  • Is stepped care to inpatient treatment for outpatient non-responders justified? The only pragmatic pathway trial found no effectiveness difference and explicitly declined to support that model (Gowers 2010, PMID 20334748).
  • Does the cost-effectiveness advantage of family-based treatment survive real-world implementation rather than modelling (Le 2017, PMID 29044637; Mayr 2025, PMID 39776084)?

References

  1. Lock J, et al. Randomized clinical trial comparing family-based treatment with adolescent-focused individual therapy. Arch Gen Psychiatry. 2010. PMID 20921118.
  2. Le Grange D, et al. Randomized clinical trial of parent-focused treatment and family-based treatment. J Am Acad Child Adolesc Psychiatry. 2016. PMID 27453082.
  3. Le Grange D, et al. Relapse from remission at two- to four-year follow-up. J Am Acad Child Adolesc Psychiatry. 2014. PMID 25440306.
  4. Fisher CA, et al. Family therapy approaches for anorexia nervosa. Cochrane Database Syst Rev. 2019. PMID 31041816.
  5. Austin A, et al. Efficacy of eating-disorder-focused family therapy for adolescents with anorexia nervosa. Int J Eat Disord. 2025. PMID 39041682.
  6. Lock J. Updates on treatments for adolescent anorexia nervosa. Child Adolesc Psychiatr Clin N Am. 2019. PMID 31443871.
  7. Society for Adolescent Health and Medicine. Medical management of restrictive eating disorders. J Adolesc Health. 2022. PMID 36058805.
  8. Russell GFM, et al. An evaluation of family therapy in anorexia nervosa and bulimia nervosa. Arch Gen Psychiatry. 1987;44:1047-1056. PMID 3318754.
  9. Eisler I, et al. A randomised controlled treatment trial of two forms of family therapy in adolescent anorexia nervosa: a five-year follow-up. J Child Psychol Psychiatry. 2007;48:552-560. PMID 17537071.
  10. Eisler I, et al. A pragmatic randomised multi-centre trial of multifamily and single family therapy for adolescent anorexia nervosa. BMC Psychiatry. 2016;16:422. PMID 27881106.
  11. Le Grange D, et al. Early weight gain predicts outcome in two treatments for adolescent anorexia nervosa. Int J Eat Disord. 2014;47:124-129. PMID 24190844.
  12. Madden S, et al. Early weight gain in family-based treatment predicts greater weight gain and remission at the end of treatment and remission at 12-month follow-up in adolescent anorexia nervosa. Int J Eat Disord. 2015;48:919-922. PMID 26488111.
  13. Lock JD, et al. Who responds to an adaptive intervention for adolescents with anorexia nervosa being treated with family-based treatment? Outcomes from a randomized clinical trial. J Am Acad Child Adolesc Psychiatry. 2024;63:605-614. PMID 38142046. (NCT03097874)
  14. Madden S, et al. A randomized controlled trial of in-patient treatment for anorexia nervosa in medically unstable adolescents. Psychol Med. 2015;45:415-427. PMID 25017941.
  15. Herpertz-Dahlmann B, et al. Day-patient treatment after short inpatient care versus continued inpatient treatment in adolescents with anorexia nervosa (ANDI): a multicentre, randomised, open-label, non-inferiority trial. Lancet. 2014;383:1222-1229. PMID 24439238.
  16. Lock J, et al. Feasibility of conducting a randomized controlled trial comparing family-based treatment via videoconferencing and online guided self-help family-based treatment for adolescent anorexia nervosa. Int J Eat Disord. 2021;54:1998-2008. PMID 34553395.
  17. Le LK, et al. The modeled cost-effectiveness of family-based and adolescent-focused treatment for anorexia nervosa. Int J Eat Disord. 2017;50:1356-1366. PMID 29044637.
  18. Mayr LM, et al. Cost-effectiveness analysis of family-based treatment with additional home treatment for adolescent anorexia nervosa. Eur Eat Disord Rev. 2025;33:608-615. PMID 39776084.
  19. Gowers SG, et al. A randomised controlled multicentre trial of treatments for adolescent anorexia nervosa including assessment of cost-effectiveness and patient acceptability — the TOuCAN trial. Health Technol Assess. 2010;14:1-98. PMID 20334748.