Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Aug 1.
Published in final edited form as: Behav Res Ther. 2015 May 28;71:20–26. doi: 10.1016/j.brat.2015.05.012

Effectiveness Trial of a Selective Dissonance-based Eating Disorder Prevention Program with Female College Students: Effects at 2- and 3-Year Follow-up

Eric Stice a, Paul Rohde a, Meghan L Butryn b, Heather Shaw a, C Nathan Marti a
PMCID: PMC4501883  NIHMSID: NIHMS697749  PMID: 26056749

Abstract

Objective

An efficacy trial found that a dissonance-based prevention program reduced risk factors, eating disorder symptoms, and future eating disorder onset, but smaller effects emerged when high school clinicians recruited students and delivered the program under real-world conditions in an effectiveness trial. The current report describes results at 2- and 3-year follow-up from an effectiveness trial that tested whether a new enhanced dissonance version of this program produced larger effects when college clinicians recruit students and deliver the intervention using improved train and supervision procedures.

Method

Young women from eight universities (N = 408, M age = 21.6, SD = 5.64) were randomized to the prevention program or an educational brochure control condition.

Results

Dissonance participants showed greater decreases in risk factors, eating disorder symptoms, and psychosocial impairment by 3-year follow-up than controls, but not healthcare utilization, BMI, or eating disorder onset.

Conclusions

This novel multisite effectiveness trial found that the enhanced dissonance intervention and improved training and supervision procedures produced an average effect size at 3-year follow-up that was 290% and 160% larger than effects observed in the high school effectiveness trial and efficacy trial respectively. Yet, the lack of eating disorder onset effects may imply that factors beyond pursuit of the thin ideal now contribute to eating disorder onset.

Keywords: prevention, body dissatisfaction, eating disorder, effectiveness trial, dissonance

Eating disorders affect 13-15% of females and are marked by chronicity, relapse, distress, functional impairment, and risk for future obesity, depression, suicide attempts, anxiety disorders, substance abuse, and mortality (Allen, Byrne, Oddy, & Crosby, 2013; Arcelus, Mitchell, Wales, & Nielsen, 2011; Stice, Marti, & Rohde, 2013b; Swanson, Crow, Le Grange, Swendsen, & Merikangas, 2011). Accordingly, a critical public health priority is to develop and broadly implement effective eating disorder prevention programs.

To date, only three prevention programs have significantly reduced DSM-IV eating disorder symptoms (Atkinson & Wade, 2014; Stice, Marti, Spoor, Presnell, & Shaw, 2008; Stice, Rohde, Shaw, & Marti, 2013c) and only three prevention programs have significantly reduced future eating disorder onset (Martinsen et al., 2014; Stice et al., 2008; Stice et al., 2013c). However, only the Body Project has produced effects in multiple trials conducted by independent research teams. In this dissonance-based eating disorder prevention program young women with body dissatisfaction voluntarily critique the thin beauty ideal in verbal, written, and behavioral exercises, which theoretically reduces thin-ideal internalization because humans seek to maintain consistency between their behaviors and attitudes, which putatively decreases body dissatisfaction, unhealthy weight control behaviors, negative affect, eating disorder symptoms, and future eating disorder onset. The Body Project produced greater reductions in eating disorder risk factors, eating disorder symptoms, functional impairment, mental health service utilization, and eating disorder onset over a 3-year follow-up relative to assessment-only control conditions and three alternative interventions in efficacy trials conducted by independent teams (Becker, Smith, & Ciao, 2005; Halliwell & Diedrichs, 2014; Matusek, Wendt, & Wiseman, 2004; Mitchell, Mazzeo, Rausch, & Cooke, 2007; Stice et al., 2008; Stice, Rohde, Durant, & Shaw, 2012; Stice, Shaw, Burton, & Wade, 2006). In support of the intervention theory, reductions in thin-ideal internalization mediate the effects of the Body Project on symptom reductions (Seidel et al., 2009; Stice et al., 2007) and completing high- versus low-dissonance versions of this program resulted in greater symptom reductions (Green et al., 2005; McMillan et al., 2011). Moreover, completion of the Body Project eliminated the adverse effects of exposure to thin models on young women (Halliwell & Diedrichs, 2014) and reduced responsivity of brain reward region to thin models (Stice & Yokum, 2015).

Given the support for the Body Project from efficacy trials, the next step is to conduct effectiveness trials of this intervention. In contrast to efficacy trials, which test whether preventive interventions produce effects under carefully controlled experimental conditions, wherein research clinicians are thoroughly trained and supervised, the intervention is delivered in adequately staffed settings, and participants are homogenous, effectiveness trials test whether interventions produce effects when delivered by endogenous clinicians (e.g., school counselors) who receive less supervision under real world conditions in natural service provision settings with heterogeneous populations (Glasgow, Lichtenstein, & Marcus, 2003). The first effectiveness trial confirmed that the Body Project reduced eating disorder risk factors and symptoms when high school clinicians recruited students with body image concerns and delivered the intervention at schools, with significant eating disorder symptom reductions persisting through 3-year follow-up (Stice, Rohde, Gau, & Shaw, 2009; Stice, Rohde, Shaw, & Gau, 2011b). Yet the average effect size was 32% smaller than observed in our large efficacy trial (Stice et al., 2006, 2008) and unlike the efficacy trial, there were no significant reductions in healthcare utilization and eating disorder onset over 3-year follow-up, suggesting the need to improve procedures for training and supervising clinicians. In addition, it is crucial to conduct effectiveness trials of eating disorder prevention programs in colleges because eating disorders often emerge during college (Stice et al., 2013b), there are over 10 million female college students in the U.S. (Department of Education, 2008), and colleges typically have an existing infrastructure for delivering prevention programs.

Thus, we initiated an effectiveness trial to evaluate the Body Project when college clinicians recruit female students at risk for eating pathology and deliver the intervention under ecologically valid conditions at universities. To maximize effects, we worked with clinicians who had experience delivering group interventions, improved the clinician training and supervision, and used a new enhanced dissonance version of the Body Project (McMillan et al., 2011). Another novel feature of the present study is that it is the first multisite trial of the Body Project. The report focusing on the acute effects found that Body Project participants showed significantly greater decreases in risk factors, eating disorder symptoms, and functional impairment than controls at posttest and 1-year follow-up, resulting in medium average effect size (d = .60), though there were no effects for healthcare utilization (Stice, Butryn, Rohde, Shaw, & Marti, 2013a). It was encouraging that the average effects were 83% larger than effects observed in the high school effectiveness trial. Herein we report the effects at 2- and 3-year follow-up from this effectiveness trial.

Methods

Participants and Procedure

Participants were 408 young women (M age = 21.6, SD = 5.6) recruited from 8 universities in Pennsylvania, Texas, and Oregon. Female undergraduate and graduate students, as well as university staff, could enroll if they reported body image concerns, who were targeted because they are at increased risk for future eating disorder onset (Beato-Fernandez et al., 2004; Jacobi et al., 2011; Stice, Marti, & Durant, 2011a). The sample was 58% European American, 17% Asian, 13% Hispanic, 7% African American, 4% Native Americans, and 1% Native Pacific Islander, which was representative of the universities from which we sampled (Stice et al., 2013a). Average parental education was 14% high school graduate or less, 25% some college, 32% college graduate, and 28% advanced graduate or professional degree. Participants were randomly assigned to the Body Project (n = 203) or an educational brochure control condition (n = 205) via a random number table. The Body Project consisted of 4 weekly 1-hour group sessions with 5-9 participants. Facilitators delivered the intervention using a scripted manual. The acute effects report provides details regarding recruitment, screening, concent procedures, facilitator characteristics, and training and supervision of facilitators (Stice et al., 2013a).

Participants completed assessments at pretest, posttest, and at 1-, 2-, and 3-year follow-ups (see Figure 1 for a participant flowchart). They were paid $30 to $40 for completing each assessment. Female assessors, who had a B.A. or M.A. in psychology, were blinded to condition. Assessors attended 24 hours of training, wherein they received instruction in interview skills, reviewed eating disorder diagnostic criteria, observed simulated interviews, and role-played interviews. They also attended annual refresher trainings. Assessors had to demonstrate inter-rater agreement (kappa [k] > .80) with supervisors using 12 audio-recorded interviews before collecting data. Weekly consensus meetings resolved diagnostic ambiguities. The institutional review board at each campus approved this project.

Figure 1.

Figure 1

Open in a new tab

Participant Flow Throughout Study

Interventions

Body Project

In session 1 participants collectively define the thin ideal, discuss costs of pursuing this ideal, and are assigned home exercises (e.g., write an essay about the costs associated with pursuing the thin ideal). In session 2 participants discuss each home exercise, dissuade facilitators from pursuing the thin ideal in role-plays, and are assigned more exercises (e.g., generate a top-10 list of things young women can do to challenge the thin ideal). In session 3 participants discuss home exercises, conduct role-plays challenging thin-ideal statements, discuss personal body image concerns, and are assigned home exercises (e.g., engage in a behavior that challenges their body image concerns). In session 4 participants discuss home exercises, plan for future pressures to be thin, discuss perceived benefits of the group, and are assigned exit home exercises (e.g., write a letter to a younger adolescent girl about avoiding body image concerns). To increase dissonance-induction, we underscored the voluntary nature of participation, increased accountability (e.g., by videotaping sessions), and increased the amount of effort required on the part of participants relative to the original Body Project.

Educational brochure control condition

Participants in the control condition received a brochure from the National Eating Disorders Association describing negative and positive body image, noting that negative body image increases risk for eating disorder onset, and offering 10 steps for achieving a positive body image. They also received a brochure from the American Psychological Association describing eating disorders and factors that predict onset of these disorders. Participants in both conditions also were given a referral list of local mental health treatment providers with expertise in eating disorders at each assessment.

Measures

The measures used to assess thin-ideal internalization, body dissatisfaction, dieting, negative affect, eating disorder symptoms, psychosocial impairment, and health/mental health care utilization were described in the initial report from this project, along with relevant psychometric information (Stice et al., 2013a). The semi-structured Eating Disorder Diagnostic Interview (EDDI) assessed DSM-IV eating disorder diagnoses have shown test-retest reliability (κ = .96), inter-rater agreement (κ = .86), and sensitivity to intervention effects (Stice et al., 2008, 2009).

Statistical Methods

Missing data

Following best-practice recommendations from Graham (2009), multiple imputation was used to replace missing data. Values were imputed using the R Amelia package (Honaker, King, & Blackwell, 2011), which uses all available data to impute longitudinal missing data via bootstrapping. Distributions of the observed and imputed data were inspected to ensure they showed comparable distributions. 20 data sets were created that contained the non-missing data and imputed values for missing data. Each data set was analyzed separately and model parameters and standard errors were derived following Rubin (1987). We also assessed whether attrition was associated with pretest measures of the primary outcomes by fitting generalized linear mixed models in which missingness at each follow-up wave was coded 1 if missing and 0 otherwise. Missingness was regressed on the baseline measure of each of the primary outcomes.

Preliminary analyses

Dependent variable distributions were examined with histograms. In the event of skewed or kurtotic distributions, we transformed data so that distributions better approximated normal distributions to minimize possible undesirable model characteristics, including non-linearity, heterogeneous variance, and outliers. We compared outcomes and demographic variables across conditions at pretest on to evaluate whether randomization created equivalent groups.

Model building

Linear mixed effects models were fit with the lme function in the nlme package from the R project (Pinheiro, Bates, DebRoy, Sarkar, & R Core Team, 2013). These models accommodate multilevel data structures and unevenly spaced longitudinal data. Prior to model-building, we examined intervention groups and study sites as a source of non-independence by treating them as level-3 random factors in a multilevel structure, in which level-1 units were time points, and level-2 units were participants. Intervention group random effects were evaluated using a partially clustered design (Baldwin, Bauer, Stice, & Rohde, 2011) that accounts for group variability in study designs in which participants in one condition are in clusters (i.e., Body Project groups) and participants in the other condition are not (i.e., educational brochure controls). The intervention effect is treated as random on prevention groups with a fixed level-3 intercept, to reflect the fact that only the intervention participants exhibited this random effect. The 3-level model examining site variability nested participants within sites in a standard multilevel structure. A deviance test that compared models with and without a level-3 random coefficient was used to assess where there was significant variability and the term was retained if significant.

Following Singer and Willett (2003) the longitudinal portion of the model was constructed using the following steps: (a) empirical growth plots were examined; (b) unconditional means models were fit; (c) unconditional linear growth models were fit; (d) unconditional non-linear models were fit; (e) models in the previous two steps were compared using the Akaike Information Criterion (AIC); and (f) level-2 predictors and cross-level interactions were added to the models. For the fourth step, we evaluated several nonlinear models. A longitudinal elevation change model assessed elevation change (i.e., mean difference) between pretest and follow-up time points (posttest, 1-, 2-, and 3-year follow-ups). A longitudinal elevation and slope change model added a slope parameter to the previous model that represented change in the outcome following the intervention. Natural-log and quadratic time models were also assessed. The longitudinal elevation change model had the lowest AIC value with the exception of the BMI model and was thus used to model time with the exception of the BMI model; henceforth, time effects are simply referred to as time. Longitudinal change in elevation change model is dummy coded (pretest = 0, posttest time points = 1), serving to contrast pretest values with the average across follow-up time points. In the final model-building step, the pretest outcome value, condition (Body Project condition = 1), and time x condition interaction were added.

Eating disorder onset

Onset of eating disorders was assessed in a proportional hazard model in which time to onset of a sub- or full-threshold DSM-IV eating disorder was regressed on intervention condition among participants who did not have an eating disorder at baseline (n = 380). At baseline 17 (8.4%) of participants in the Body Project condition met criteria for a subthreshold DSM-IV eating disorder, versus 11 (5.4%) of participants in the control condition, which was not a significant difference.

Results

Preliminary analyses

A natural log transformation was used to normalize negative affect, eating disorder symptoms, health care utilization, and mental health care utilization. Participants in the two conditions did not differ significantly on demographics or outcomes at pretest. Table 1 provides means and SDs for outcomes at each time point across conditions. Data were complete at pretest, 1% were missing at posttest, 8% were missing at the 1-year follow-up, 8% were missing at the 2-year follow-up, and 15% were missing at the 3-year follow-up. None of the pretest outcomes were significantly associated with missingness. Assessment of level-3 random coefficients indicated no significant variability attributable to groups in the partially clustered models; however, significant variability attributable to site was present for the dieting and eating disorder symptoms. Thus, all models were fit as two-level models in which time points were nested within individuals with the exception of dieting and eating disorder symptoms, which contained a level-3 random effect for site.

Table 1.

Means and standard deviations for outcomes by condition at pretest, posttest, 1-year follow-up, 2-year follow-up, and 3-year follow-up

Variable Pretest Posttest 1-year follow-up 2-year follow-up 3-year follow-up
Thin-ideal internalization
    Brochure controls 3.84 (0.55) 3.80 (0.55) 3.74 (0.65) 3.67 (0.61) 3.69 (0.60)
    Body Project 3.90 (0.58) 3.43 (0.68) 3.53 (0.61) 3.55 (0.73) 3.44 (0.66)
Body dissatisfaction
    Brochure controls 3.31 (0.7) 3.16 (0.72) 3.09 (0.75) 2.99 (0.73) 2.96 (0.79)
    Body Project 3.40 (0.73) 2.81 (0.76) 2.80 (0.79) 2.92 (0.82) 2.84 (0.78)
Dieting
    Brochure controls 2.76 (0.87) 2.58 (0.87) 2.56 (0.87) 2.41 (0.85) 2.31 (0.87)
    Body Project 2.87 (0.88) 2.25 (0.90) 2.26 (0.85) 2.38 (0.89) 2.29 (0.85)
Negative affect
    Brochure controls 11.58 (8.59) 10.40 (8.58) 9.90 (9.20) 9.77 (9.43) 8.83 (8.76)
    Body Project 12.54 (9.10) 7.72 (8.19) 8.06 (8.12) 7.73 (7.12) 8.04 (8.47)
Eating disorder symptoms
    Brochure controls 10.99 (9.03) 9.53 (10.13) 10.13 (9.96) 9.82 (11.36) 8.67 (9.14)
    Body Project 12.7 (10.53) 7.47 (8.52) 10.47 (21.96) 8.88 (13.44) 8.07 (11.49)
BMI
    Brochure controls 24.15 (5.36) 24.07 (5.28) 24.28 (5.61) 24.32 (5.82) 24.45 (5.45)
    Body Project 24.72 (4.69) 24.98 (5.22) 25.05 (5.24) 25.33 (5.55) 25.34 (5.51)
Psychosocial impairment
    Brochure controls 2.22 (0.48) 2.16 (0.49) 2.17 (0.50) 2.18 (0.53) 2.12 (0.48)
    Body Project 2.27 (0.47) 2.05 (0.45) 2.07 (0.49) 2.11 (0.47) 2.14 (0.48)
Health service utilization
    Brochure controls 1.61 (5.63) 0.84 (2.52) 1.53 (3.75) 0.89 (2.30) 1.28 (3.91)
    Body Project 1.15 (2.17) 0.88 (2.81) 1.07 (2.38) 1.38 (3.39) 0.9 (1.99)
Mental health service utilization
    Brochure controls 2.02 (5.90) 1.37 (4.39) 1.76 (6.18) 1.31 (5.48) 1.13 (3.29)
    Body Project 1.71 (5.97) 1.41 (3.87) 3.02 (23.73) 2.74 (13.09) 1.25 (3.44)
Open in a new tab

Primary outcome analyses

Significant condition × time effects through 3-year follow-up were present for all five primary outcomes: thin-ideal internalization (t [395] = −6.12, p < .001), body dissatisfaction (t [389] = −5.29, p < .001), dieting (t [393] = −4.78, p < .001), negative affect (t [387] = −4.25, p < .001), and eating disorder symptoms (t [398] = −5.22, p < .001). In each case, the interaction indicated that the decrease between pretest and the two follow-up measures was greater for Body Project participants than control participants.

Secondary outcome analyses

There was a significant condition × time effect for psychosocial impairment (t [395] = −3.28, p < .001), indicating greater improvements in Body Project versus control participants over follow-up. The condition × time effect was not significant for BMI (t [353] = −0.71, p = .479), health care utilization (t [385] = −0.05, p = .956), or mental health care utilization (t [390] = 1.18, p = .239). Parameter estimates, coefficients, 95% confidence intervals, and p-values of the terms from the models that tested for the effects on primary and secondary outcomes are provided in the Supplementary Material section. Table 2 provides the effect sizes for the significant intervention effects.

Table 2.

Effect Sizes (d) for change in outcome in the Body Project intervention group relative to the brochure control condition

Outcome Pre-to-post Pre-to-1-year follow-up Pre-to-2-year follow-up Pre-to-3-year follow-up
Thin-ideal internalization -0.77 -0.48 -0.33 -0.54
Body dissatisfaction -0.64 -0.54 -0.24 -0.30
Dieting -0.50 -0.47 -0.15 -0.15
Negative affect -0.53 -0.38 -0.29 -0.22
Eating disorder symptoms -0.54 -0.40 -0.33 -0.39
Psychosocial impairment -0.34 -0.32 -0.26 -0.07
Open in a new tab

Eating disorder onset

There was no difference in time to onset of DSM-IV eating disorders across conditions (z = 1.39, p =.165). Twenty-four Body Project participants (12.9%) met a DSM-IV diagnosis during the 3-year follow-up (bulimia nervosa [BN] = 5, binge eating disorder [BED] = 2, atypical anorexia nervosa [AN] = 2, subthreshold BN = 18, subthreshold BED = 7, purging disorder = 3; 9 participants exhibited more than one eating disorder diagnosis during follow-up). Twenty-five control participants (12.9%) met a DSM-IV diagnosis over follow-up (BN = 2, BED = 2, subthreshold BN = 17, subthreshold BED = 8, purging disroder = 6; 8 participants exhibited more than one eating disorder diagnosis during follow-up). The pattern of eating disorder diagnosis was similar across conditions (e.g., most common diagnosis was subthreshold BN followed by having multiple eating disorder diagnoses).

Discussion

This is the first effectiveness trial to evaluate the Body Project when college clinicians recruit high-risk female college students for this selective eating disorder prevention program and deliver it under ecologically valid conditions in typical service provision settings (Stice et al., 2013a). It is also the first trial of the new enhanced dissonance version of the Body Project and the first multisite trial of this prevention program. The Body Project produced significantly greater reductions in eating disorder risk factors and symptoms than observed in educational brochure controls. The average effect size at 3-year follow-up was d = .32, which indicates that the Body Project produced approximately a 1/3rd standard deviation reduction in these continuous outcomes, which is clinically meaningful. The average effect size was 290% larger than the parallel average effect size from the high school effectiveness trial (d = .11; Stice et al., 2011) and 160% larger than the parallel average effect size from the efficacy trial (d = .20; Stice et al., 2008). No other eating disorder prevention program has produced significant reductions in eating disorder symptoms at 3-year follow-up, although the Healthy Weight prevention program did produce significant reductions in this outcome relative to control participants at 2-year follow-up (d = .26; Stice et al., 2013c). Thus, results suggest that the brief 4-hour Body Project prevention program produces effects over 3-year follow-up when college clinicians deliver this intervention under typical service delivery conditions. The larger effects in the present trial apparently emerged because we used the new enhanced dissonance version of the Body Project, improved facilitator training and supervision procedures, and worked with college clinicians who had greater expertise delivering group-based interventions than their high school counterparts. Results imply that it would be prudent to use the new enhanced dissonance script, employ the new training and supervision procedures, and work with clinicians experienced in delivering group interventions.

Results also suggested that the Body Project produced significant improvements in psychosocial impairment relative to the control condition. This is important because it implies that this brief prevention program improved functioning in the family, peer, school, and work domains, producing effects that persisted over long-term follow-up. The reductions in body dissatisfaction, negative affect, and eating disorder symptoms from participating in the Body Project may have increased the participant's comfort in social engagement and ability to focus on school and work, which improved relations with family, peers, teachers, and co-workers that account for the improved functioning in these domains. This finding is novel in that only one past trial has found reductions in psychosocial impairment from an eating disorder prevention program: The Body Project reduced psychosocial impairment through 3-year follow-up relative to assessment-only controls and two alternative active interventions in our large efficacy trial (Stice et al., 2008). The fact that the previous high school effectiveness trial did not observe significant effects for psychosocial impairment (Stice et al., 2011b) suggests that the steps we took to improve intervention effects from the Body Project in the present trial may have contributed to improved effects regarding functional impairment.

However, the Body Project did not significantly affect body mass or healthcare utilization. This intervention significantly reduced obesity onset and health care utilization in the high school/college efficacy trial (Stice et al., 2006), but not in the prior high school effectiveness trial (Stice et al., 2011b). The pattern of findings suggests that although the intervention typically reduces eating disorder symptoms, including binge eating, this prevention effect does not translate into reductions in excess weight. The fact that participants in both conditions had an average BMI at pretest that was in the healthy weight range and only increased slightly implies that participants showed only modest weight gain over follow-up. Results also suggest that it is challenging to reduce health and mental health care utilization with brief eating disorder prevention program. To our knowledge, only two eating disorder prevention programs have significantly reduced unhealthy weight gain (c.f., Jones et al., 2008; Stice et al., 2006, 2008, 2012) and no other eating disorder prevention program has significantly reduced healthcare utilization.

It was also disappointing that the eating disorder onset prevention effects observed in the efficacy trial (Stice et al., 2008) did not replicate in either the high school effectiveness trial (Stice et al., 2011b) or the present college effectiveness trial. Although it is tempting to conclude that this pattern of results suggests that the Body Project does not reduce eating disorder onset when implemented by clinicians under real-world condition, the Body Project did produce larger eating disorder symptom reductions through 3-year follow-up when delivered by college clinicians versus both high school clinicians and research clinicians. Specifically, we observed larger reductions in eating disorder symptoms for Body Project participants than for controls by 3-year follow-up in the present trial (d = .39) than observed in the high school effectiveness trial (d = .30; Stice et al., 2011b) or the high school/college efficacy trial (d = .07; Stice et al., 2008). This pattern of findings implies that college clinicians did deliver the Body Project in an effective manner in terms of significantly reducing eating disorder symptom levels but not in preventing eating disorder onset. The fact that the Healthy Weight prevention program, which aims to prevent eating disorder onset through small but lasting changes in dietary intake and exercise levels, has significantly reduced future eating disorder onset in two separate trials involving college women (Stice et al., 2008; Stice et al., 2013c) suggests that this prevention program produces more robust reductions in eating disorder onset.

Limitations

Several limitations in this study should be considered when interpreting the results. Although the control group received educational information, the conditions were not matched on contact with a clinician. The retention rate was high and similar across conditions and attrition was not associated with any of the outcome measures at pretest, but participants who could not be reached for follow-up assessments may have differed in a systematic way from those who completed the assessments on some unmeasured factor, so attrition bias cannot be ruled out. As this was an effectiveness trial, there were few inclusion or exclusion criteria, but the proportion of total students from each campus who participated was still relatively low, leaving the possibility that a selection bias in participation may have occurred. Finally, all of the outcomes that showed reductions in this trial were based on self-report data, raising the possibility that demand characteristics inherent to randomized trials might have contributed to the observed effects. Future research should therefore attempt to incorporate more biological and objective outcomes.

Implications for Prevention and Future Research

This trial confirmed that it is feasible for college clinicians to successfully implement the Body Project on college campuses and produce clinically meaningful intervention effects that persist over time. Endogenous providers on college campuses were able to recruit young women at high-risk for eating disorders and to be trained and supervised in effective intervention delivery. Body Project participants demonstrated clinically meaningful benefits from the intervention, with meaningful decreases in eating disorder risk factors and symptoms through 3-year follow-up. The size and duration of these effects is especially compelling given that the group-based intervention was only 4 hours in duration. These findings suggest that broad scale dissemination of the Body Project on college campuses is warranted.

Given the promising evidence-based for the Body Project, it would be useful to investigate more cost-effective methods of implementing this prevention program on a broad basis, such as by using peer leaders or the Internet (e.g., Becker et al., 2010; Stice et al., 2012). However, it will also be crucial to conduct systematic research on alternative eating disorder prevention programs that appear more effective in reducing eating disorder onset, such as the Healthy Weight prevention program. As the science of prevention becomes more advanced in multiple domains of mental and physical health, it also may be important to consider how the Body Project could be integrated with other prevention efforts, such as those for obesity, depression, or substance abuse.

Supplementary Material

  • - Tested eating disorder prevention when college clinicians recruit and deliver program

  • - 3-year follow-up data presented on 408 women randomized to group or brochure control

  • - Average effect size for eating disorder risk factors and symptoms was d = .32

  • - Program achieved effects for psychosocial impairment but not eating disorder onset

  • - Real-world clinicians can obtain meaningful eating disorder prevention effects

Acknowledgements

This study was supported by Grant MH086582 from the National Institute of Mental Health. We thank Julie Pope, Madeline Kirch, and Colleen Kase for being exemplary project coordinators.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

We have no financial conflicts to disclose for consideration of this article in Behaviour Research and Therapy and it is not being considered for publication elsewhere.

References

  1. Arcelus J, Mitchell A, Wales J, Nielsen S. Mortality rates in patients with anorexia nervosa and other eating disorders: A meta-analysis of 36 studies. Archives of General Psychiatry. 2011;68:724–731. doi: 10.1001/archgenpsychiatry.2011.74. doi:10.1001/archgenpsychiatry.2011.74. [DOI] [PubMed] [Google Scholar]
  2. Atkinson M, Wade T. Does mindfulness have potential in eating disorder prevention? A preliminary controlled trial with young adult women. Early Intervention in Psychiatry. 2014 doi: 10.1111/eip.12160. Doi:10.1111/eip.12160. [DOI] [PubMed] [Google Scholar]
  3. Baldwin SA, Bauer DJ, Stice E, Rohde P. Evaluating models for partially clustered designs. Psychological Methods. 2011;16:149. doi: 10.1037/a0023464. doi: 10.1037/a0023464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Beato-Fernandez L, Rodriguez-Cano T, Belmonte-Llario A, Martinez-Delgado C. Risk factors for eating disorders in adolescents: A Spanish community-based longitudinal study. European Child and Adolescent Psychiatry. 2004;13:287–294. doi: 10.1007/s00787-004-0407-x. [DOI] [PubMed] [Google Scholar]
  5. Becker C, Smith L, Ciao A. Reducing eating disorder risk factors in sorority members: A randomized trial. Behavior Therapy. 2005;36:245–254. doi:10.1016/S0005-7894(05)80073-5. [Google Scholar]
  6. Becker C, Wilson C, Williams A, Kelly M, McDaniel L, Elmquist J. Peer-facilitated cognitive dissonance versus healthy weight eating disorders prevention: A randomized comparison. Body Image. 2010;7:280–288. doi: 10.1016/j.bodyim.2010.06.004. doi:10.1016/j.bodyim.2010.06.004. [DOI] [PubMed] [Google Scholar]
  7. Glasgow R, Lichtenstein E, Marcus A. Why don't we see more translation of health promotion research to practice? Rethinking the efficacy-to-effectiveness transition. American Journal of Public Health. 2003;93:1261–1267. doi: 10.2105/ajph.93.8.1261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Graham J. Missing data analysis: Making it work in the real world. Annual Review of Psychology. 2009;60:549–576. doi: 10.1146/annurev.psych.58.110405.085530. doi:10.1146/annurev.psych.58.110405.085530. [DOI] [PubMed] [Google Scholar]
  9. Green M, Scott N, Diyankova I, Gasser C, Pederson E. Eating disorder prevention: An experimental comparison of high level dissonance, low level dissonance, and no-treatment control. Eating Disorders. 2005;13:157–169. doi: 10.1080/10640260590918955. doi:10.1037/a0024351. [DOI] [PubMed] [Google Scholar]
  10. Halliwell E, Diedrichs P. Brief report: Testing a dissonance body image intervention among young girls. Healthy Psychology. 2014;33:201–204. doi: 10.1037/a0032585. [DOI] [PubMed] [Google Scholar]
  11. Honaker J, King G, Blackwell M. Amelia II: A program for missing data. Journal of Statistical Software. 2011;45:1–47. [Google Scholar]
  12. Jacobi C, Fittig E, Bryson S, Wilfley D, Kraember H, Taylor C. Who is really at risk? Identifying risk factors for subthreshold and full syndrome eating disorders in a high-risk sample. Psychological Medicine. 2011;41:1939–1949. doi: 10.1017/S0033291710002631. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jones M, Luce K, Osborne M, Taylor K, Cunning D, Doyle A, Taylor C. Randomized, controlled trial of an Internet-facilitated intervention for reducing binge eating and overweight in adolescents. Pediatrics. 2008;121:453–462. doi: 10.1542/peds.2007-1173. doi:10.1542/peds.2007-1173. [DOI] [PubMed] [Google Scholar]
  14. Martinsen M, Bahr R, Borresen R, Holme I, Pensgaard A, Sundgot-Borden J. Preventing eating disorders among young elite athletes: A randomized controlled trial. Medicine & Science in Sports & Exercise. 2014;46:435–447. doi: 10.1249/MSS.0b013e3182a702fc. Doi:10.1249/MSS.0b013e3182a702fc. [DOI] [PubMed] [Google Scholar]
  15. Matusek J, Wendt S, Wiseman C. Dissonance thin-ideal and didactic healthy behavior eating disorder prevention programs: Results from a controlled trial. International Journal of Eating Disorders. 2004;36:376–388. doi: 10.1002/eat.20059. doi:10.1002/eat.20059. [DOI] [PubMed] [Google Scholar]
  16. McMillan W, Stice E, Rohde P. High- and low-level dissonance-based eating disorder prevention programs with young women with body image concerns: An experimental trial. Journal of Consulting and Clinical Psychology. 2011;79:129–134. doi: 10.1037/a0022143. doi:10.1037/a0022143. [DOI] [PubMed] [Google Scholar]
  17. Mitchell K, Mazzeo S, Rausch S, Cooke K. Innovative interventions for disordered eating: Evaluating dissonance-based and yoga interventions. International Journal of Eating Disorders. 2007;40:120–128. doi: 10.1002/eat.20282. doi:10.1002/eat.20282. [DOI] [PubMed] [Google Scholar]
  18. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team [February 16, 2015];nlme (Version 3.1-119) [Computer program and manual] 2013 from http://cran.r-project.org/web/packages/nlme/nlme.pdf.
  19. Rubin DB. Multiple Imputation for Nonresponse in Surveys. John Wiley & Sons, Inc; New York: 1987. [Google Scholar]
  20. Seidel A, Presnell K, Rosenfield D. Mediators in the dissonance eating disorder prevention program. Behaviour Research and Therapy. 2009;47:645–653. doi: 10.1016/j.brat.2009.04.007. doi:10.1016/j.brat.2009.04.007. [DOI] [PubMed] [Google Scholar]
  21. Singer J, Willett J. Applied longitudinal data analysis: Modeling change and event occurrence. Oxford University Press; New York: 2003. [Google Scholar]
  22. Stice E, Butryn M, Rohde P, Shaw H, Marti N. An effectiveness trial of a new enhanced dissonance eating disorder prevention program among female college students. Behaviour Research and Therapy. 2013a;51:862–871. doi: 10.1016/j.brat.2013.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Stice E, Marti N, Durant S. Risk factors for onset of eating disorders: Evidence of multiple risk pathways from an 8-year prospective study. Behaviour Research and Therapy. 2011a;49:622–627. doi: 10.1016/j.brat.2011.06.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Stice E, Marti C, Rohde P. Prevalence, incidence, impairment, and course of the proposed DSM-5 eating disorder diagnoses in an 8-year prospective community study of young women. Journal of Abnormal Psychology. 2013b doi: 10.1037/a0030679. doi:10.1037/a0030679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Stice E, Marti C, Spoor S, Presnell K, Shaw H. Dissonance and healthy weight eating disorder prevention programs: Long-term effects from a randomized efficacy trial. Journal of Consulting and Clinical Psychology. 2008;76:329–340. doi: 10.1037/0022-006X.76.2.329. doi:10.1037/0022-006X.76.2.329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Stice E, Presnell K, Gau J, Shaw H. Testing mediators of intervention effects in randomized controlled trials: An evaluation of two eating disorder prevention programs. Journal of Consulting and Clinical Psychology. 2007;75:20–32. doi: 10.1037/0022-006X.75.1.20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Stice E, Rohde P, Durant S, Shaw H. A preliminary trial of a prototype Internet dissonance-based eating disorder prevention program for young women with body image concerns. Journal of Consulting And Clinical Psychology. 2012;80:907–916. doi: 10.1037/a0028016. doi:10.1037/a0028016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Stice E, Rohde P, Shaw H, Gau J. An effectiveness trial of a selected dissonance-based eating disorder prevention program for female high school students: Long-term effects. Journal of Consulting and Clinical Psychology. 2011b;79:500–508. doi: 10.1037/a0024351. doi:10.1037/a0024351. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Stice E, Rohde P, Shaw H, Marti N. Efficacy trial of a selective prevention program targeting both eating disorders and obesity among female college students: 1- and 2-year follow-up effects. Journal of Consulting and Clinical Psychology. 2013c;81:183–189. doi: 10.1037/a0031235. Doi: 10.1037/a0031235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Stice E, Shaw H, Burton E, Wade E. Dissonance and healthy weight eating disorder prevention programs: A randomized efficacy trial. Journal of Consulting and Clinical Psychology. 2006;74:263–275. doi: 10.1037/0022-006X.74.2.263. doi:10.1037/0022-006X.74.2.263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Stice E, Yokum S. Dissonance-based eating disorder prevention program reduces reward region response to thin models: How actions shape valuation. 2015 doi: 10.1371/journal.pone.0144530. Manuscript submitted for publication. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Swanson S, Crow S, Le Grange D, Swendsen J, Merikangas K. Prevalence and correlates of eating disorders in adolescents: Results from the national comorbidity survey replication adolescent supplement. Archives of General Psychiatry. 2011;68:714–723. doi: 10.1001/archgenpsychiatry.2011.22. doi:10.1001/archgenpsychiatry.2011.22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. U.S. Department of Education . Digest of Educational Statistics, 2007. National Center for Education Statistics; 2008. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS697749-supplement.docx (33.7KB, docx)

RESOURCES