Skip to content

Treatment in adults

TL;DR — Specialist adult psychotherapies improve weight and symptoms on average, but no approach is a reliable first-choice winner. ANTOP randomized 242 adult women (BMI 15.0–18.5 kg/m²) from ten German university hospitals to focal psychodynamic therapy (n=80), CBT-E (n=80) or optimized treatment as usual (n=82); BMI increased in all groups (0.73, 0.93, 0.69 kg/m² at end of treatment; 1.64, 1.30, 1.22 at 12 months) with no significant between-group differences, and 22% were lost by end of treatment and 30% by 12 months (Zipfel 2014, PMID 24131861). The trial's own interpretation is more differentiated than "no winner": focal psychodynamic therapy was advantageous for recovery at 12-month follow-up and CBT-E for speed of weight gain and eating-disorder psychopathology, while optimized treatment as usual — psychotherapy plus structured family-doctor care — was recommended as the solid baseline. MOSAIC randomized 142 broadly defined AN outpatients to MANTRA or SSCM and found no significant BMI or symptom difference at 6 or 12 months, although MANTRA was rated more acceptable/credible (Schmidt 2015, PMID 25984803). Evidence supports offering a credible specialist therapy with medical/nutritional care and measurement-based adjustment.

Named treatments

Treatment Core emphasis Typical evidence limitation
CBT-E/CBT-ED Regular eating, formulation, cognitive/behavioural maintaining processes, relapse prevention Small/moderate trials; therapist expertise and intensity vary
MANTRA Cognitive-interpersonal maintenance model, motivation, thinking style, social context No superiority over SSCM in MOSAIC
SSCM Supportive relationship, psychoeducation, nutrition/weight restoration, patient-selected issues Active comparator, not “no treatment”
Focal psychodynamic therapy Symptom meaning, relationships, self-esteem and therapeutic focus Evidence concentrated in ANTOP
Optimized TAU Structured medical plus community psychotherapy Heterogeneous but clinically relevant comparator

Comparative trials

Study N Main result Interpretation
ANTOP 242 (80 FPT / 80 CBT-E / 82 optimized TAU) BMI rose in all three groups with no primary-outcome difference; FPT advantageous for recovery at 12 months, CBT-E for speed of weight gain and ED psychopathology Screened 727 to randomize 242; attrition 22% at EOT, 30% at 12 months (PMID 24131861)
ANTOP 5-year 242 randomized, 154 (64%) reassessed at mean 5.96 years No significant BMI difference between arms (estimated means 18.64–18.99 kg/m²); on observed data 41% full recovery, 41% partial, 18% full-syndrome AN; one suicide death Naturalistic post-trial care and 36% attrition limit purity; predictors of 5-year BMI were baseline BMI, illness duration and depression (PMID 35294860)
MOSAIC 142 (MANTRA n=72, SSCM n=70), broadly defined AN with BMI ≤18.5; 20–30 weekly sessions by severity Both improved BMI, symptoms, distress and clinical impairment; no significant between-group difference at 6 or 12 months; neurocognitive/social-cognitive gains inconsistent MANTRA rated significantly more acceptable and credible at 12 months; no difference in additional service use; one SSCM patient died (PMID 25984803)
Network meta-analysis 16 RCTs from 14,003 screened reports; 13 in the network, 1,047 patients (97.4% female) None of the interventions outperformed treatment as usual on BMI or eating-disorder psychopathology; the one significant comparison was lower all-cause dropout for CBT than psychodynamic therapies (OR 0.54, 95% CI 0.31–0.93) Confidence in the evidence rated low to very low (PMID 33600749)
Meta-review of (network) meta-analyses Systematic meta-review to December 2020 Interventions involving family outperformed active control in adults as well as adolescents with AN A meta-review inherits the limitations of its inputs; the adult family-therapy evidence base is thin (PMID 36084848)

The three-arm Australian trial and the older comparisons

Two further randomized trials belong in the comparative table above, and one of them is the result that most complicates the field.

Study N Design Result
SWAN (Byrne 2017, PMID 28552083) 120 outpatients, three Australian cities, 25–40 sessions over 10 months, blinded assessment, intention-to-treat SSCM vs MANTRA vs CBT-E No significant differences between treatments on any continuous outcome; all three produced clinically significant, maintained improvement. At 12-month follow-up, healthy weight was achieved by a mean of 50% and remission by a mean of 28.3%, with no between-treatment difference. Treatment was completed by 60%, and only 52.5% of the total sample completed 12-month follow-up (ACTRN12611000725965)
McIntosh 2005, PMID 15800147 56 women, 20 sessions over ≥20 weeks CBT vs interpersonal psychotherapy vs non-specific supportive clinical management (intended as the control) The control condition was superior to interpersonal psychotherapy on the primary global outcome, with CBT intermediate; among completers the control was superior to both specialist therapies. The authors state the result "was opposite to the primary hypothesis and challenges assumptions about the effective ingredients"
Dare 2001, PMID 11230031 84 adults, 1 year Focal psychoanalytic psychotherapy vs cognitive-analytic therapy vs family therapy vs low-contact routine treatment Psychoanalytic psychotherapy and family therapy were significantly superior to the routine-treatment control; CAT trended toward benefit. Improvement across the whole sample was modest, with several patients still significantly undernourished at follow-up

The SWAN remission figure is the number to hold: 28.3% at 12 months, in a trial of three well-delivered manualized specialist therapies, with 40% not completing treatment. That is the realistic ceiling of current adult outpatient care, and it is the reason "no clear winner" is a statement about a low common ceiling rather than about three equally good options.

A secondary analysis of the SWAN data addressed one candidate moderator directly. Set-shifting improved from baseline to end of treatment; central coherence results were less clear; and, contrary to the assumption that detail-focused thinking impedes therapy, participants with low baseline central coherence showed more rapid reduction in eating-disorder psychopathology and clinical impairment. Baseline executive functioning predicted neither early change nor remission (Keegan 2022, PMID 35715854). The neuropsychological profile that motivates cognitive remediation did not, in this trial, select who responds to standard therapy.

McIntosh 2005 is the counter-result the field has never fully absorbed. It is the origin of SSCM's odd status as a control that became a treatment, and its finding — that the deliberately non-specific comparator beat the specialist therapies among completers — is consistent with the later transcript analysis showing that outcome did not track the ratio of clinical-management to supportive-psychotherapy content (McIntosh 2022, PMID 35258113). Two independent lines of evidence therefore point away from technique-specific mechanisms in adult AN psychotherapy. Dare 2001 points the other way, finding specific therapies superior to low-contact routine care. Both are small; neither has been overturned.

After the hospital: relapse prevention

The largest effect size in adult AN psychotherapy is not in initial treatment but in the post-hospitalization period. Thirty-three weight-restored adults randomized after discharge to one year of outpatient CBT or nutritional counselling showed relapse in 22% vs 53%, treatment failure (relapse plus dropout) in 22% vs 73%, and good outcome in 44% vs 7% (Pike 2003, PMID 14594754). The sample is small and the comparator weak — nutritional counselling alone is not a credible active control — but the separation is far larger than anything obtained in first-line outpatient trials, and the design targets the highest-risk window. It has not been replicated at scale in over two decades.

Severe and enduring illness: the only dedicated trial

The severe/enduring literature is not entirely uncontrolled. Sixty-three adults with at least a seven-year illness history were randomized across multiple sites to CBT-AN or SSCM, both modified for SE-AN, delivered over 30 outpatient visits across 8 months, with quality of life, mood and social adjustment — not weight — as the main outcomes. Retention was 85%. Both groups improved significantly; there were no between-group differences at end of treatment. At 6 months CBT-AN participants had higher social-adjustment scores (p = 0.038), and at 12 months lower EDE global scores (p = 0.004) and higher readiness for recovery (p = 0.013) (Touyz 2013, PMID 23642330). The design decision that matters most here is the choice of primary outcome: this is the trial that demonstrated a long-duration cohort can be recruited, retained and improved on quality-of-life endpoints. See severe and enduring illness.

Setting for adults

The Cochrane review of treatment setting found five eligible trials, four in AN totalling 511 participants. For AN there may be little or no difference between specialist inpatient care and active outpatient or combined brief-hospital-plus-outpatient care in weight gain at 12 months (SMD −0.22, 95% CI −0.49 to 0.05; 2 trials, 232 participants; low-quality evidence). People may be more likely to complete treatment when randomized to outpatient settings, but that finding is rated very uncertain (RR 0.75, 95% CI 0.64 to 0.88; 3 trials, 319 participants; very-low-quality evidence). The reviewers' conclusion is that there was insufficient evidence to declare any setting superior for moderately severe or less severe AN (Hay 2019, PMID 30663033).

The attempt to fix that for adults failed instructively. DAISIES, a two-arm non-inferiority trial of inpatient treatment-as-usual versus stepped-care day-patient treatment in adults with BMI ≤ 16 kg/m², intended 386 participants and randomized 15 over 16 months before terminating. Among 53 patients approached, the most common reason for non-participation was strong treatment preference, compounded by COVID-19 service disruption. Both arms gained BMI similarly at 12 months, but no conclusion about effectiveness is possible; day-patient care was perceived as more acceptable by patients and carers (İnce 2025, PMID 39943786). Patient preference is not a nuisance variable in this field — it is the reason the setting question for adults remains unanswerable by conventional randomization.

What “no clear winner” means

It does not mean therapies are identical or ineffective. It means direct and indirect comparisons do not estimate stable superiority with adequate certainty. The network meta-analysis is blunt about the scale of the problem: 14,003 reports screened yielded 16 randomized trials, of which 13 contributed 1,047 patients to the network, and confidence in the evidence was rated low to very low throughout (Solmi 2021, PMID 33600749). Common factors, nutritional management, patient preference, therapist competence and rescue care can reduce between-arm separation. Small samples and dropout widen uncertainty — and dropout is the one dimension on which the network did separate the treatments, with lower all-cause dropout for CBT than psychodynamic therapies (OR 0.54, 95% CI 0.31–0.93).

One qualification cuts the other way. A systematic meta-review of meta-analyses and network meta-analyses found that interventions involving the family outperformed active control in adults as well as adolescents with AN (Monteleone 2022, PMID 36084848). This is a weaker design than a primary network meta-analysis and rests on a small adult family-therapy literature, but it is the clearest published counterweight to reading "no winner" as "nothing to choose between".

SSCM: the control that became a treatment

SSCM was designed as a control condition and has been retained as a first-line option, which makes it an unusual object in this evidence base. Six clinical trials have found it effective, and a secondary meta-analysis endorses it as promising; it is characterized by a person-centred, non-prescriptive philosophy, a defined but flexible dietetic stance, a directional commitment to reversing starvation, and deliberate absence of specific mandates (Kiely 2022, PMID 35255984). Content analysis of all 178 transcribed sessions from the ten completing therapist-patient dyads in the original SSCM trial found that over three quarters of session content fell within the clinical-management theme — normalizing eating first, then weight, the mechanics of SSCM and encouragement of self-care — and about 20% within supportive psychotherapy, half of that concerning relationships. Crucially, those achieving good outcome did not have a lower ratio of clinical management to supportive psychotherapy content (McIntosh 2022, PMID 35258113). If the ingredient hypothesis were right, that ratio should have mattered; it did not.

Who drops out, and what predicts outcome

Dropout is the dimension on which the adult network meta-analysis did separate treatments, so what predicts it matters. A systematic review and meta-analysis of 27 studies of baseline predictors in individual and family therapy for AN across adolescents and adults found that lower motivation, lower BMI and the binge-eating/purging subtype predicted drop-out, while greater eating-disorder pathology and poorer motivation predicted poorer outcome (Gregertsen 2019, PMID 30551081). The evidence base was described as sparse and inconsistent, and the authors call for carefully designed multi-site studies.

Two things follow. Motivation appears on both lists, which is why it is measured at assessment in most specialist services and why SSCM's explicitly non-prescriptive stance is not obviously a weakness. And the predictors of dropout are largely the predictors of severity, so "dropout" in adult AN trials is not a neutral loss to follow-up — it removes the sicker patients, biasing completer analyses toward optimism. That is worth holding next to SWAN's 40% non-completion and 47.5% loss by 12-month follow-up (Byrne 2017, PMID 28552083).

Measurement-based care

Domain Track Reason to change plan
Medical Vitals, trajectory, labs/ECG as indicated Instability or continued loss
Nutritional Delivered intake, weight trend, behaviours No progress despite credible dose
Psychopathology Restriction, fear, rituals, exercise, EDE-type measures Worsening or static core symptoms
Function Work/study, relationships, quality of life Weight-only improvement
Engagement Attendance, alliance, treatment credibility Dropout risk or treatment mismatch

Severe and enduring illness

The severe/enduring label lacks a single validated threshold and can become prognostically self-fulfilling. Reviews find small, heterogeneous studies and emphasize quality of life and collaborative goal-setting alongside continued recovery orientation (Dalle Grave 2020, PMID 32400903; Zhu 2020, PMID 32265758). It should not be equated automatically with futility.

Open questions

  • Are there replicable moderators that select CBT-E, MANTRA, SSCM or focal psychodynamic therapy for an individual? The network meta-analysis authors explicitly call for participant-level data to be released precisely because aggregate data cannot answer this, and as of September 2026 no individual-participant-data meta-analysis of adult AN psychotherapy has been published (Solmi 2021, PMID 33600749).
  • Which early response rule should trigger augmentation or a setting change?
  • Can family/support involvement effective in adolescents be adapted for adults without undermining autonomy?
  • Does the non-specific comparator really outperform specialist therapy, or was McIntosh 2005 an underpowered anomaly? Two decades on, the finding has neither been replicated nor refuted, and the SSCM transcript analysis is consistent with it (McIntosh 2005, PMID 15800147; McIntosh 2022, PMID 35258113; Dare 2001, PMID 11230031).
  • Why has the post-hospitalization relapse-prevention result with CBT (relapse 22% vs 53%) never been replicated at scale (Pike 2003, PMID 14594754)?
  • Is the ~28% 12-month remission rate a ceiling of the therapies or of the trials' populations and retention, given 40% non-completion in SWAN (Byrne 2017, PMID 28552083)?
  • Can the adult setting question be answered at all by randomization, given that strong treatment preference was the leading cause of DAISIES' failure (İnce 2025, PMID 39943786; Hay 2019, PMID 30663033)?

References

  1. Zipfel S, et al. Focal psychodynamic therapy, CBT and optimized treatment as usual in ANTOP. Lancet. 2014. PMID 24131861.
  2. Herzog W, et al. ANTOP 5-year follow-up. Lancet Psychiatry. 2022. PMID 35294860.
  3. Schmidt U, et al. MOSAIC: MANTRA compared with SSCM. J Consult Clin Psychol. 2015. PMID 25984803.
  4. Solmi M, et al. Comparative efficacy and acceptability of psychological interventions for adult outpatients with anorexia nervosa. Lancet Psychiatry. 2021. PMID 33600749.
  5. Kiely L, et al. Conceptualising specialist supportive clinical management. J Eat Disord. 2022. PMID 35255984.
  6. McIntosh VVW, et al. SSCM: therapy content and relation to outcome. Int J Eat Disord. 2022. PMID 35258113.
  7. Dalle Grave R, et al. Severe and enduring anorexia nervosa: no easy solutions. Int J Eat Disord. 2020. PMID 32400903.
  8. Zhu J, et al. Psychological treatments for people with severe and enduring anorexia nervosa. Front Psychiatry. 2020. PMID 32265758.
  9. Monteleone AM, et al. Treatment of eating disorders: a systematic meta-review of meta-analyses and network meta-analyses. Neurosci Biobehav Rev. 2022. PMID 36084848.
  10. Byrne S, et al. A randomised controlled trial of three psychological treatments for anorexia nervosa. Psychol Med. 2017;47:2823-2833. PMID 28552083. (ACTRN12611000725965)
  11. McIntosh VVW, et al. Three psychotherapies for anorexia nervosa: a randomized, controlled trial. Am J Psychiatry. 2005;162:741-747. PMID 15800147.
  12. Dare C, et al. Psychological therapies for adults with anorexia nervosa: randomised controlled trial of out-patient treatments. Br J Psychiatry. 2001;178:216-221. PMID 11230031.
  13. Pike KM, et al. Cognitive behavior therapy in the posthospitalization treatment of anorexia nervosa. Am J Psychiatry. 2003;160:2046-2049. PMID 14594754.
  14. Touyz S, et al. Treating severe and enduring anorexia nervosa: a randomized controlled trial. Psychol Med. 2013;43:2501-2511. PMID 23642330.
  15. Hay PJ, et al. Inpatient versus outpatient care, partial hospitalisation and waiting list for people with eating disorders. Cochrane Database Syst Rev. 2019;1:CD010827. PMID 30663033.
  16. İnce B, et al. Stepping into day treatment approach versus inpatient treatment for adults with anorexia nervosa: the DAISIES RCT. Health Technol Assess. 2025;29:1-37. PMID 39943786.
  17. Keegan E, et al. An exploratory examination of executive functioning as an outcome, moderator, and predictor in outpatient treatment for adults with anorexia nervosa. J Eat Disord. 2022;10:83. PMID 35715854.
  18. Gregertsen EC, et al. Pre-treatment patient characteristics as predictors of drop-out and treatment outcome in individual and family therapy for adolescents and adults with anorexia nervosa: a systematic review and meta-analysis. Psychiatry Res. 2019;271:484-501. PMID 30551081.