Evidence-Based Psychosocial Treatments for Eating Problems and Eating Disorders

Abstract

Eating disorders represent a significant source of psychological impairment among adolescents. However, most controlled treatment studies have focused on adult populations. This review provides a synthesis of existing data concerning the efficacy of various psychosocial interventions for eating disorders in adolescent samples. Modes of therapy examined in adolescent samples include family therapy, cognitive therapy, behavioral therapy, and cognitive behavioral therapy mostly in patients with anorexia nervosa. At this time, the evidence base is strongest for the Maudsley model of family therapy for anorexia nervosa. Evidence of efficacy for other treatments and other conditions is limited by several methodological factors including the small number of studies, failure to use appropriate control conditions or randomization procedures, and small sample sizes (i.e., fewer than 10 participants per treatment arm). Potential moderators and mediators of treatment effect are reviewed. Finally, results from adolescent studies are contrasted with those from adult studies of eating disorders treatment. Many studies of adult populations comprise late adolescent/young adult participants, suggesting that findings regarding the efficacy of cognitive behavioral therapy for bulimia nervosa in adults likely extend to older adolescent populations.

This work was supported by grants from the National Institute of Mental Health (R01 61836; R01 63758; PI: Pamela K. Keel). We thank Kyle De Young and Elizabeth Damstetter for their assistance with the literature search for this article.

Notes

Note: Superscripts indicate who completed each outcome measure; no superscript indicates that outcome was objectively measured. DSM = Diagnostic and Statistical Manual of Mental Disorders; ICD = International Classification of Diseases; AN = anorexia nervosa; CFT = conjoint family therapy; SFT = separated family therapy; BMI = body mass index; EAT = Eating Attitudes Test; EDI = Eating Disorders Inventory; M-R = Morgan & Russell scales; MOCI = Maudsley Obsessional Compulsive Index; RSE = Rosenberg Self Esteem Inventory; SMFQ = Short Mood and Feeling Questionnaire; wt = weight; IBW = ideal body weight; FT = family therapy; FPE = family group psychoeducation; ind = individual; fam = family; tx = treatment; ABW = average body weight; BSI = Brief Symptom Inventory; CDI = Children's Depression Inventory; FAM-III = Family Assessment Measure; Cauc = Caucasian; Hisp = Hispanic; Afr Am = African American; BN = bulimia nervosa; SPT = supportive psychotherapy; EDE = Eating Disorder Examination; BDI = Beck Depression Inventory; Nat Am = Native American; STFT = short-term family therapy; LTFT = long-term family therapy; YBC-ED = Yale-Brown-Cornell Eating Disorder Scale; CBCL = Child Behavior Checklist; SADS-SAC = Schedule for Affective Disorders and Schizophrenia for School-Age Children; YSR = Youth Self-Report; BFST = behavioral family systems therapy; EOIT = ego-oriented individual therapy; IT = nonspecific individual treatment; COT = cognitive orientation treatment; DSM-SS = DSM Symptomatology Scale for Anorexia and Bulimia; ART = activity restriction therapy; TET = token economy; KPT = Kyoto Prefectural University of Medicine behavior therapy; ED = eating disorder; RT = relaxation training; VR = virtual reality; ASI = Appearance Schemas Inventory; BASS = Body Areas Satisfaction Scale; BAT = Body Attitudes Test; BES = Body Esteem Scale; BIAQ = Body Image Avoidance Questionnaire; BIATQ = Body Image Automatic Thoughts Questionnaire; BITE = Bulimic Investigatory Test Edinburgh; BSQ = Body Shape Questionnaire; RS = Restrained Scale; SIBID = Situational Inventory of Body Image Dysphoria; BDI = Beck Depression Inventory; PANAS = Positive and Negative Affect Schedule; EDNOS = eating disorder not otherwise specified; CT = cognitive therapy; DC = dietician control; DAS = Dysfunctional Attitudes Scale; LCB = Locus of Control of Behaviour.

^b Patient self-report.

^c Patient rated.

^d Clinician rated.

^e Russell et al. (1987) consisted of three groups; the adolescent AN group is reported in this table (see Table 3 for the adult AN and adult BN group results).

Note: Results from Bachar et al. (1999) are not reported because analyses included adult women with bulimia nervosa; results from Perpina et al. (1999) are not included because statistics were not reported. Global EDE scores were not reported in Lock et al. (2005). Positive within-treatment effect sizes reflect increases in weight, decreases in body image disturbance, decreases in bulimic symptoms, and decreases in composite scores (i.e., better outcome). Positive between-treatment effect sizes indicate better outcome associated with the first treatment listed compared to the second treatment listed in the comparison. Superscripts indicate which measure was used to calculate Cohen's d.

^a Body mass index (BMI).

^b Bulimic symptoms.

^c Eating Disorders Inventory (EDI) total score.

^d Eating Attitudes Test (EAT) total score.

^e Percentage average body weight

^f EDI drive for Thinness subscale.

^g EDI Body Dissatisfaction subscale.

^h EDI Bulimia subscale.

ⁱ Eating Disorders Examination (EDE) Weight Concerns subscale.

^j EDE Shape Concerns subscale.

^k EDE objective binge frequency.

^l EDE compensatory behavior frequency.

^m EDE global score.

ⁿ Between-treatment effect sizes were not calculated because of nonrandom assignment to conditions.

^o Inappropriate weight and shape concerns.

^p Effect size reported by Schmidt et al. (2007) for difference in abstinence from binge eating assessed by EATATE interview.

^q Between-treatment BMI effect size for Serfaty et al. (1999) reflects BMI at last observation because all participants in the dietician control condition dropped out. This study did not report pretreatment BMI in the dietician control condition. Thus, no effect size for change in weight is reported for this condition.

∗Bold face indicates effect sizes for comparisons between treatments.

¹This study also included patients with BN who were mostly adults (M age = 24 years). Among all patients, participants randomized to self psychology therapy achieved significantly greater reductions in EAT scores compared to participants randomized to cognitive orientation therapy.

Note: tx = treatment; ind = individual; fam = family; CBT = cognitive behavioral therapy; ERP = exposure plus response prevention; IPT = interpersonal psychotherapy; PE = psychoeducational treatment; BT = behavioral therapy; wk = week; DBT = dialectical behavioral therapy; SH = self help; WT = weight loss treatment; BED = binge eating disorder.

^a Random assignment to condition utilized blocked, stratified, restricted randomization or assignment to group with greatest mean difference from participant to avoid significant pretreatment differences among groups.

^b Russell et al. (1987) consisted of three groups: bulimia nervosa (BN), adult anorexia nervosa, and adolescent anorexia nervosa (see Table 1).

^c Palmer et al. (2002) included BN (n = 71), BED (n = 28), and eating disorder not otherwise specified-partial BN (n = 22).

Note: AN = Anorexia nervosa.

^a There are no well-established psychosocial treatments for adolescents with bulimia nervosa (BN). However, cognitive behavioral therapy (CBT) represents a well-established treatment for older adolescents/young adults with BN (Cooper & Steere, 1985; Fairburn et al., 1991; Fairburn, Kirk, O'Connor, & Cooper, 1986; Garner et al., 1993).