Table 1 Cognitive behavior therapy trials for adolescent eating disorders
Authors (year) | Study design | Population | Therapy (n) | Attrition (%) | Tx length | FU | Summary of relevant findings | Quality appraisal | |
|---|---|---|---|---|---|---|---|---|---|
Dx(s) | Age M [range] | ||||||||
CBT (outpatient settings) | |||||||||
Ball and Mitchell (2004) [37] | RCT | AN | 18.5 [13–23] | CBT (13) versus behavioral family therapy (12) | CBT: 30.8 Behavioral family therapy: 25 | 21–25 sessions over 12 mo | 6 mo | No differences between treatment groups at EOT or FU in weight, return of menses, or ED psychopathology. At EOT, 77.8% (n = 7) of each treatment group achieved min. weight gain of 4 kg (8.8 lbs) and within 10% of adolescent’s average body weight, with binge and purge episodes occurring less than once per week. Improvements were maintained at FU. Statistically significant improvement in ED psychopathology and behavior (EDE, EDI-2, ABOS) for both treatment groups at EOT, but all measures remained in the symptomatic range at all FU timepoints. Of patients that completed either treatment (n = 18), 72% achieved min. weight gain of 4 kg (8.8 lbs) and within 10% of adolescent’s average body weight, return of menses, and abstinence from binge and purge episodes at 6mo FU | Faira |
Byford et al. (2007)† [34] | Secondary analysis of RCT data (Gowers et al., 2007; [82]) | AN | 14.9 [12–18] | CBT w/parent counseling, dietary therapy (55) versus multi-disciplinary inpatient treatment (57) versus general community care (55) | CBT: 25.5 Inpatient: 50.9 General community care: 30.9 | CBT: 6 mo Inpatient: 6 wks General community care: 6 mo | 1 yr 2 yr | The outpatient CBT group spent the least amount of time in the hospital compared to the other two treatment groups. There were no statistically significant differences in cost of treatment between the three treatment groups, though the CBT group was least expensive at non-significant level | Faira |
Charpentier et al. (2003) [41] | Cohort study | AN BN | 17.7 [13–22] | Group CBT (26) | 12 | 13 wks (≈ 3 mo) | 3 mo | For patients with BN, statistically significant reduction in ED psychopathology (EDI) and frequency of binges and purging at EOT, which was maintained at 3mo FU. For patients with AN, statistically significant increase in BMI, with no significant changes in other ED symptoms (EDI) | Faira |
Cibich and Wade (2019) [77] | Case study | BN | 16 [N/A] | CBT-T (1) | N/A | 10 sessions | 3 mo | Abstinence from binge eating and purging by second session, and ED psychopathology (EDE-Q) within community norm range by EOT. All changes maintained at 3mo FU | Includeb |
Cowdrey and Davis (2016) [78] | Case study | AN | 15 [N/A] | CBT-E (1) | N/A | 11 wks (≈ 3 mo) | 8 mo | Marked reduction in ED behaviors and "feeling fat" (self-monitoring records), global ED psychopathology (EDE-Q) within community norms, and clinically significant increases in weight at EOT. At 8 mo, progress was partially maintained (global EDE-Q score had increased to just above clinical cut-off) | Includeb |
Craig et al. (2019) [79] | Cohort study | AN, atypical AN, BN, atypical BN | 15.5 [13–18] | CBT-ED (54) | 38.9 | Varied (M = 22 wks; range: [6–46 wks], or M ≈ 5 mo, range: [1–10 mo]) | None | Statistically significant improvement in ED psychopathology (EDE-Q) at EOT (d = 0.82). For patients with AN or atypical AN, statistically significant increases in mean %EBW (AN: d = 0.61; atypical AN: d = 0.36) | Faira |
Dalle Grave et al. (2013) [35] | Cohort study | AN | 15.5 [13–17] | CBT-E (46) | 17.4 | 40 wks (≈ 9 mo) | 60 wk (≈ 14 mo) | Nearly two thirds (63%, n = 29) completed 40 sessions without need for additional treatment. Of treatment completers, 31% (n = 9) reached 95% EBW by EOT. Nearly all treatment responders (97%, n = 28) had global ED psychopathology (EDE-Q) within community norms at EOT. Changes remained stable at 60wk FU | Gooda |
Dalle Grave et al. (2015) [48] | Cohort study | BN, BED, or other specified ED | 16.5 [13–19] | CBT-E (68) | 25 | 20 wks (≈ 5 mo) | None | Statistically significant improvements in ED psychopathology (EDE-Q) at EOT (d = 1.03). More than half (67.6%, n = 46) had minimal residual ED psychopathology (global EDE-Q scores below 1 SD above the community mean) at EOT. Of participants with binge eating or purging at BL, 50% (n = 25) were abstinent at EOT. Of treatment completers, 81.2% (n = 42) achieved minimal residual ED psychopathology at EOT, and 76.5% (n = 26) of those with binge eating or purging at BL were abstinent at EOT | Gooda |
Dalle Grave et al. (2019) [31] | Cohort study | AN | 15.5 [11–18] | CBT-E (49) | 28.6 | 40 wks (≈ 9 mo) | 20 wk (≈ 4 mo) | The majority of treatment completers (62.9%, n = 22) achieved both good weight outcome (corresponding to an adult BMI ≥ 18.5 kg/m2) and global ED psychopathology (EDE-Q) within community norms at EOT. Nearly half of patients that completed the 20wk FU (48.3%, n = 14) maintained this outcome | Gooda |
DeBar et al. (2013) [30] | RCT | BED, BN | 15.12 [12–18] | CBT (13) versus TAU (12) | 23 (CBT) 30 (TAU) | 8 wks + 4 optional wks (≈ 2–3 mo) | 3 mo 6 mo | At FU, CBT held advantage over TAU, with statistically significant higher rates of abstinence from binge eating (at 6mo FU, d = 1.47), and more improvements in eating, shape and weight concerns (at 6mo FU, EDE Eating Concerns d = 0.80, EDE Shape Concerns d = 1.04, EDE Weight Concerns d = 0.64). 100% of CBT participants were abstinent from binge eating at 6mo FU. Participants in the CBT group reported high post-treatment satisfaction (Client Satisfaction Survey) | Faira |
Gorrell et al. (2019)† [80] | Secondary analysis of RCT data (Le Grange et al., 2015; [44]) | BN, BN-type EDNOS (DSM-IV) | 15.8 [12–18] | FBT-BN (52) versus CBT-A (58) | 10 | Varied (FBT-BN, M = 13.6 wks; CBT-A, M = 14.7 wks) (both ≈ 3 mo) | none | Across both treatments, participants with a higher level of motivation to change in ED-related preoccupations and rituals (Motivation for Change subscale of YBC-EDS) at BL were more likely to have reduced ED psychopathology (EDE global score) at EOT. Motivation to change ED-related preoccupations and rituals at BL had no effect on abstinence from bingeing and purging at EOT in either treatment | Gooda |
Gowers and Smyth (2004) [81] | Cohort study | AN | 16.1 [12–20] | CBT + parental counseling + dietary therapy (42) | 21.5 | 12 sessions CBT, 4 sessions parental counseling, 3 sessions dietary counseling | 6 wk | At 6wk FU, significant improvements in ED psychopathology (EDI-2) and weight gain. Those with higher motivation (assessed using survey designed by research team) were more likely to complete treatment. Motivational status did not predict self-rated outcome (including EDI-2) at FU, though higher level of motivation at BL predicted more weight gain at 6wk FU | Faira |
Gowers et al. (2007) [82] | RCT | AN | 14.9 [12–18] | CBT w/parent counseling, dietary therapy (55) versus multi-disciplinary inpatient treatment (57) versus general community care (55) | CBT: 25.5 Inpatient: 50.9 General community care: 30.9 | CBT: 6 mo Inpatient: 6 wks General community care: 6 mo | 1 yr 2 yr | At 1 yr FU, mean improvement in weight and ED psychopathology (EDI-2, MRAOS) across treatment groups, with no statistically significant differences between them. At 2 yr FU, further improvement in outcome for all groups, with no statistically significant differences between them | Faira |
Gowers et al. (2010) † [32] | Secondary analysis of RCT data (Gowers et al., 2007; [82]) | AN | 14.9 [12–18] | CBT w/parent counseling, dietary therapy (55) versus multi-disciplinary inpatient treatment (57) versus general community care (55) | CBT: 25.5 Inpatient: 50.9 General community care: 30.9 | CBT: 6 mo Inpatient: 6 wks General community care: 6 mo | 1 yr 2 yr 5 yr | At 5 yr FU, trend of improvement across treatment groups in ED outcome (MRAOS), and no statistically significant differences between groups in % weight for height, diagnostic outcome, or ED psychopathology (EDI-2, MRAOS). No significant differences in total cost of hospital use across treatment groups at 5 yr FU. No statistically significant differences between groups in parental expectations of treatment (at 1 yr FU, self-reported rating of prior expectation of treatment received on 7-point Likert scale), but adolescents’ expectations of general outpatient care were lower than of CBT outpatient care. At 1 yr FU, parents reported satisfaction across treatments (rated on 7-point Likert scale), but were significantly more satisfied with CBT outpatient care compared to general community care. Adolescents had significantly lower satisfaction with care across treatments compared to parents, though a non-significant trend favored adolescent satisfaction with CBT outpatient care compared to general community care | Faira |
Hilbert et al. (2020) [47] | RCT | BED | 15.3 [12–20] | CBT (32) vs WLC (36) | 32 | 20 sessions over 4 mo | 6 mo 12 mo 24 mo | Compared to WLC group, CBT group had significantly higher rates of abstinence from binge eating (51% vs. 33%) and remission from BED (57% vs. 33%) and significantly less ED psychopathology at EOT (EDE). CBT group maintained its advantage relative to WLC in in ED behaviors and psychopathology at all FU timepoints | Gooda |
Hurst et al. (2017) [83] | Case study | BN | 15 [15–15] | FBT-BN (1) versus FBT-BN + CBT (1) | 0 | Not reported | None | Participants in both treatments achieved full remission at EOT (EDE global score within 1 SD of community norms). Participant 1 reduced binge and purge episodes from an average of 12 per week at BL, to 100% abstinence at EOT. Participant 2 reduced binge and purge episodes from an average of ten per week at BL, to 100% abstinence at EOT | Includeb |
Hurst and Zimmer-Gembeck (2019) [36] | Cohort study | AN | 14.9 [12–17] | FBT + CBT-P (21) | 9 | Varied (M = 32 wks; 23 FBT sessions and 9 CBT sessions) (≈ 7 mo) | None | Over half (57%; n = 12) of adolescents achieved full remission by EOT, defined as a minimum of 95% of EBW and global EDE-Q score within community norms. Two participants demonstrated increased ED symptoms at EOT | Gooda |
Jaite et al. (2018) [33] | RCT | AN | 16.9 [12–21] | CBT (24) vs DBT (26) | Not reported | 25 wks (wkly individual and group sessions) (≈6 mo) | None | ED symptoms (EDI-2) decreased and BMI increased significantly from BL to EOT in both treatments (CBT: EDI-2 d = -0.61, BMI d = 1.04; DBT: EDI-2 d = -0.55, BMI d = 0.7). Treatment assignment (CBT vs DBT) did not predict treatment satisfaction. Significant correlation between therapist and parent level of satisfaction of treatment (Questionnaire for the Evaluation of Treatment). Agreement was not observed between adolescent and therapist, nor adolescent and parent level of satisfaction in treatment | Faira |
Le Grange et al. (2015) [44] | RCT | BN, BN-type EDNOS (DSM-IV) | 15.8 [12–18] | FBT-BN (52) versus CBT-A (58) | 10 | Varied (FBT-BN M = 13.6 wks; CBT-A M = 14.7 wks) (both ≈ 3 mo) | 6 mo 12 mo | Abstinence from binge eating and purging was significantly higher in FBT-BN than CBT-A at EOT and 6mo FU. Abstinence from these behaviors was no longer significantly different between the groups at 12mo FU. No significant differences between groups in %EBW, EDE global score, YBC-EDS total score | Gooda |
Le Grange et al. (2020) [49] | Cohort study | Any DSM-5 ED diagnosis, excluding ARFID | 14.6 [12–19] | FBT (51) versus CBT-E (46) | FBT: 35 CBT: 37 | FBT: 20 sessions over 6 mo CBT: 40 sessions over 9–12 mo for lower weight pts; 20 sessions over 6 mo for higher weight pts | 6 mo 12 mo | Rate of weight gain was faster in FBT than CBT by EOT, though differences in weight gain were no longer significant at 6 and 12mo FUs. There were no differences in ED psychopathology (EDE global score) between the treatments at any time point | Gooda |
Lock (2005) [43] | Cohort study | BN | 15.8 [12–18] | CBT (34) | 18 | Varied (M = 15.8 sessions) (≈ 4mo) | None | Majority of patients (82%, n = 28) completed at least 10 weeks of treatment. The mean rate of binge eating and purging reduced from 15.8 episodes/wk (range: [2–21]) at BL to 3.4 episodes/wk (range: [0–21]) at EOT. Over half of participants (56%, n = 19) were abstinent from binge eating and purging at EOT | Faira |
Matheson et al. (2020)† [84] | Secondary analysis of RCT data (Le Grange et al., 2015; [44]) | BN, BN-type EDNOS (DSM-IV) | 15.7 [12–18] | CBT-A (26) vs FBT-BN (30) vs SPT (15) | See Le Grange et al. (2015) | Varied (M = 14 sessions) (≈ 3 mo) | 6 mo 12 mo | Reducing purge episodes by ≥ 96.8% by session 2 and reducing binge eating episodes by ≥ 96.4% by session 4 predicted abstinence in these behaviors at EOT, regardless of treatment type. Reducing binge eating episodes by ≥ 96.4% by session 8, and purge episodes by ≥ 94.4% by session 9 predicted abstinence at 6mo FU, and reductions in binge eating by ≥ 96.4% at session 9 predicted abstinence at 12mo FU, regardless of treatment type | Gooda |
Ohmann et al. (2013) [85] | Cohort study | AN | 14.3 [13–17] | Group CBT (29) | 28 | 40 wkly group sessions + monthly family sessions (≈ 9 mo) | 1 yr | Over half of patients (55%, n = 16) achieved good outcome, defined as reaching the 25th age-related BMI percentile, and absence of restriction or bulimic behaviors. At EOT, weight and BMI improved significantly in patients with good outcome and remained stable at 1 yr FU. There was little to no improvement in weight in patients with poor outcome or who dropped out of treatment (44.8%, n = 13). Patients who achieved good outcome also demonstrated statistically significant improvement in ED psychopathology, per EDE dietary restraint, eating concern and weight concern subscales, though improvements were not observed in the EDE shape concern subscale | Faira |
Puls et al. (2019)† [46] | Secondary analysis of RCT data (Hilbert et al., 2020; [47]) | BED | 14.17 [12–20] | CBT (64) | 32 | 20 wkly sessions (≈ 5 mo) | None | High levels of therapist treatment adherence (ACF) and therapeutic alliance (WAI-OS) observed across all sessions. Decreased adherence was associated with higher patient treatment expectation (VAS rating of expectation of treatment success). No association was observed between treatment adherence and ED outcomes (# OBE and SBE episodes, EDE global score). Alliance was negatively associated with # OBE and SBE episodes, and positively associated with treatment adherence. There was no association between alliance and general ED psychopathology (EDE) or patient treatment expectation | Gooda |
Schapman-Williams et al. (2006) [42] | Cohort study | BN, BN-type EDNOS | 16.3 [range not reported] | CBT (7) | 0 | Varied (M = 15.3 sessions over 5.4 mo; range: [10–20 sessions over 4–8 mo]) | None | At EOT, significant reduction in ED psychopathology (EDE), and mean # binge episodes/wk. Reduction in mean # purge episodes/wk approached statistical significance at EOT, and 57% (n = 4) were abstinent from both binge eating and purging | Faira |
Schapman-Williams and Lock (2007) [86] | Case study | BN | 16 | CBT (1) | N/A | 20 sessions (≈ 5 mo) | None | Consistent normalization of eating patterns by session six (assessed via daily food records). Abstinence from binge and purge episodes maintained throughout treatment, aside from one episode in week 13 (a reduction from > 20 binge and purge episodes/wk pre-treatment). Distress while eating reduced from 60% at session one to 0% by session nine. ED psychopathology (EDE) reduced substantially by mid-treatment, which was maintained at EOT | Includeb |
Stefini et al. (2017) [39] | RCT | BN, BN-type EDNOS (DSM-IV) | 18.7 [14–20] | CBT (39) versus PDT (42) | 38.5 (CBT) 21.4 (PDT) | Varied (M = 36.6 sessions or ≈ 8 mo; up to 60 sessions over 1 yr) | 12 mo | No significant differences in attrition, diagnostic changes or reduction in ED psychopathology at EOT or 12mo FU between groups. Significant reductions in binge and purge frequency, and ED symptoms (EDE, EDE-Q) in both groups at EOT, which were maintained at 12mo FU. CBT showed a small advantage over PDT in reduced frequency of binge eating (d = 0.23) and purging (d = 0.26) at EOT (EDE-Q), while PDT showed a small advantage over CBT in eating concern (d = -0.35) at EOT (EDE) | Faira |
Sysko and Hildebrandt (2011) [87] | Case study | BN-type EDNOS | 16 [N/A] | CBT-E (1) | N/A | 29 sessions over 9 mo | None | Clinically significant decrease in SBEs and purging in first 4 weeks. Abstinence from SBEs and purging achieved by session 22. Residual concerns around shape and weight still present at EOT | Includeb |
Thompson-Brenner et al. (2010) [50] | Observational cross-sectional study | AN, BN, EDNOS | 16.4 [15–18] | CBT, dynamic therapy, family intervention, emotion regulation, trauma therapy or conjoint therapy (120; n per treatment type not reported | N/A | Varied (M = 8 mo, range: [6 h–1 yr]) | None | Dynamic therapy was most strongly associated with better global outcome for the entire sample, though CBT was most strongly associated with better outcome for a subsample of participants with poor relational/personality functioning (assessed via clinician rated psychotherapy effectiveness form) | Gooda |
Valenzuela et al. (2018) † [45] | Secondary analysis of RCT data (Le Grange et al., 2015; [44]) | BN, BN-type EDNOS (DSM-IV) | 15.8 [12–18] | FBT-BN (52) versus CBT-A (58) | 10 | Varied (FBT-BN M = 13.6 wks; CBT-A M = 14.7 wks) or (both ≈ 3 mo) | 6 mo 12 mo | Statistically significant and clinically meaningful reduction in symptoms of depression (BDI) and self-esteem (RSE) at FU in both treatments, with no statistically significant differences between them | Gooda |
CBT guided self-help | |||||||||
Perkins et al.(2005)† [54] | Secondary analysis of RCT data (Schmidt et al., 2007; [51]) | BN, BN-type EDNOS | 17.9 [13–20] | CBT guided self-care or Family therapy adapted from FBT (85) | Not reported | Not reported | None | Both treatment groups included opportunities for parental involvement; for present analyses, total sample grouped into “No Parent Involvement” (n = 23) vs “Parent Involvement” (n = 62). Most common reasons participants reported choosing to exclude parents from treatment were lack of comfort discussing personal issues in their presence, feeling that the ED was “their problem” and parents didn’t need to be involved, or another personal factor unique to the adolescent (e.g., parents not thinking the adolescent had a problem). The most common reasons adolescents reported wanting parents to be involved in treatment included that the parent was viewed as supportive, had time for the patient, was interested in the patient and wanted to be involved, and the patient wanted them to learn more about the ED | Faira |
Pretorius et al. (2009) [52] | Cohort study | BN, BN-type EDNOS | 18.8 [13–20] | Online CBT guided self-care (101) | 17 | 8 web-based sessions, encouraged to complete wkly | 3 mo 6 mo | Significant improvements in bulimic behaviors and cognitions at 3mo and 6mo FUs (EDE, EDE-Q) although most remained symptomatic; of treatment completers, 10% (n = 5) were abstinent from bulimic behaviors at 3mo FU, and 17% (n = 11) were abstinent from bulimic behaviors at 6mo FU. High patient satisfaction with treatment convenience, acceptability, and confidentiality (assessed via questionnaire designed by research team) | Gooda |
Pretorius et al. (2010) † [53] | Secondary analysis of cohort study data (Pretorius et al., 2009; [52]) | BN, BN-type EDNOS | M not reported; 16–19: n = 44; 20 years: N n = 7 [16–20] | Online CBT guided self-care (11) | N/A | 8 web-based sessions. At time of interview, # sessions completed ranged from one to eight | None | Results included qualitative data regarding liked and disliked components of the web-based, guided self-care program. Participants found the layout and presentation of the program components to be user-friendly, and reported their choice of using a web-based program was influenced by accessibility, flexibility, sense of control over treatment, anonymity, non-judgmental nature, privacy, and convenience. Most participants felt that the program offered sufficient support, felt motivation to change at the start of treatment, motivation difficulties toward the end, and described improvement in ED symptoms that they attributed to the program | Gooda |
Schmidt et al. (2007) [51] | RCT | BN, BN-type EDNOS | 17.7 [13–20] | CBT guided self-care (44) vs Family therapy adapted from FBT (41) | 30 | CBT: 22 sessions, Family therapy: 24 sessions (both ≈ 5 mo) | 6 mo 12 mo | Both interventions demonstrated significant improvements in binge eating and purging behaviors (EATATE interview) across time. A significantly higher proportion of participants were abstinent from binge eating at 6mo FU in the CBT guided self-care group compared to the family therapy group. These differences were no longer significant at 12mo FU. There were no significant differences in change in purging behaviors between the two groups at 6mo or 12mo FU. Mean cost of treatment significantly lower for CBT guided self-care than family therapy (Client Service Receipt Inventory) | Faira |
Wagner et al. (2013) [55] | RCT | BN | 19.3 [16–21] | Online CBT-based guided self-help (18) versus guided self-help biblio-therapy (11) | 31 | Varied (M not reported; range: [4–7 mo]) | 4 mo 7 mo 18 mo | No differences in outcome between treatment groups; outcomes reported combines data from the two groups. Significant improvement over time in monthly binge eating, vomiting and fasting. At 7mo and 18mo FU, about half of adolescents (44%, n = 8 and 55%, n = 11, respectively) were abstinent from ED behaviors or achieved remission, with outcomes comparable to an adult sample receiving the same treatments. There were no significant improvements in the EDI-2 perfectionism scale, and asceticism EDI-2 subscale scores increased slightly between 4- and 7mo FU. All other EDI-2 subscales demonstrated improvement over time, with the highest decrease in scores within the first 4 mo of therapy | Poora |
CBT-based inpatient care | |||||||||
Calugi and Dalle Grave (2019) [56] | Cohort study | AN | 16.4 [13–18] | CBT-E (62) | 9.7 | 20 wks (≈ 5 mo) | 6 mo 12 mo | Significant mean reduction in body image concerns (EDE), and nearly all treatment completers (96.4%) reached a normative BMI percentile at EOT (corresponding to adult BMI ≥ 18.5 kg/m2). At 6mo and 12mo FU, 78.7% and 80.4% of treatment completers maintained a normative BMI, respectively | Gooda |
Dalle Grave et al. (2014) [57] | Cohort study | AN | 16.0 [13–17] | CBT-E (27) | 3.7 | 20 wks (≈ 5 mo) | 6 mo 12 mo | Treatment was well-accepted by patients, demonstrated by low attrition. Nearly all treatment completers (96.2%, n = 25) reached a normative BMI (corresponding to a BMI percentile ≥ 18.5) by EOT. Statistically significant decrease in ED psychopathology (EDE) by EOT, and 38.5% (n = 10) had minimal residual ED psychopathology (EDE global score below 1 SD above community norms). Changes were maintained at 6mo and 12mo FU | Gooda |
Fennig et al. (2017) [58] | Cohort study | AN | 14.8 [11–18] | Multimodal IP program, with CBT component (44) | 29.5 | Varied (M ≈ 4 mo, range [12 days–≈ 9 mo]) | None | BMI significantly increased from admission to discharge, with 70% achieving 100% IBW at discharge, and 25% achieving 90% IBW. Significant improvement in general ED psychopathology severity (EDI-2 total score) and restraint and eating concern subscales, but no significant change in body dissatisfaction, weight concern, drive for thinness, or shape concern subscales | Faira |
Naab et al. (2013) [59] | Cohort study | AN | 16.4 [13–17] | Multimodal IP program, with CBT component (177) | 30.5 | Varied (M = 77.9 days) (≈ 3 mo | None | Significant mean increases in BMI and decrease in global ED psychopathology (SIAB-S global score) from admission to discharge. Compared to a sample of adults receiving the same treatment, differences in outcome were only found in the bulimic symptoms and atypical binge subscales of the SIAB-S, with adults showing greater severity at admission, and greater improvement upon discharge | Faira |
Schlegl et al. (2016) [60] | Cohort study | AN | 15.7 [13–17] | Multimodal IP program, with CBT component (238) | 19.8 | Varied (M = 81.9 days) (≈ 3 mo) | None | Significant increases in BMI from admission to discharge (d = 2.1), and improvement in ED psychopathology (EDI-2) at discharge (d = 0.8), with 44.7% demonstrating a clinically significant improvement, though 3.7% deteriorated | Gooda |