Abstract
Objective:
Evaluate whether the Body Project prevention program adapted for young women with type 1 diabetes (Diabetes Body Project) reduces eating disorder (ED) risk factors and symptoms.
Methods:
Young women (aged 15–30) at high-risk for EDs due to having type 1 diabetes and body image concerns (N = 55) were randomized to virtually delivered Diabetes Body Project groups or an educational control condition, completing measures at pretest, posttest, and 3-month follow-up.
Results:
Diabetes Body Project versus the control participants showed significantly greater reductions in thin-ideal internalization, body dissatisfaction, diabetes distress, diabetes eating pathology, and ED symptoms by posttest, and greater reductions in diabetes eating pathology and ED symptoms, and greater improvements in quality of life by 3-month follow-up, which were medium to large effects (d’s ranged from −0.43 to −0.90). Although control participants showed a worsening of glycemic control (time in range) verses Diabetes Body Project participants, this difference was non-significant (d = 0.26).
Conclusions:
Virtually delivered Diabetes Body Project decreased ED risk factors and symptoms in young women with type 1 diabetes. A well powered randomized controlled trial is warranted to evaluate this intervention over longer follow-up.
Keywords: Type 1 diabetes, Eating disorder, Body dissatisfaction, Prevention, Virtual, Body image
1. Introduction
Young women with versus without type 1 diabetes show a 2- to 4-times higher prevalence of eating disorders (EDs), including threshold and subthreshold bulimia nervosa and binge eating disorder [1,2]. One prospective study of adolescent girls with type 1 diabetes found that the lifetime incidence of EDs was 60% over a 14-year follow-up [1], which is 4.6-times greater than unaffected adolescent girls [3]. Youth with type 1 diabetes are also at elevated risk for diabetes eating pathology, such as insulin omission, intentionally keeping blood sugar levels high or eating a low carbohydrate diet for weight control, and ED symptoms, such as binge eating and fasting, excessive exercise, and vomiting for weight control [4].
Critically, ED pathology is associated with poorer glycemic control [2]. Intentional insulin omission for weight control is reported by 37% of females with type 1 diabetes [5] and is associated with a 3-times greater mortality [6]. Among women with type 1 diabetes, 61% of those with an ED history report intentional insulin omission versus 26% of those without an ED history [7]. Insulin omission increases risk for diabetic ketoacidosis (DKA) and frequent presence of ketones can cause heart complications, kidney failure, cerebral edema, coma, and death [8]. Young women with type 1 diabetes and EDs have 3-times the risk of DKA and 6-times the risk of death compared with their peers without EDs [9]. Comorbid type 1 diabetes and EDs also increases risk for retinopathy and nephropathy [10]. To date, no intervention has been shown to prevent this co-morbidity.
Pressure for thinness, pursuit of the thin appearance ideal, body dissatisfaction, dieting, negative affect, and female sex predict future onset of EDs [11–13]. Theoretically, these risk processes are more pronounced among young women with type 1 diabetes because weight loss occurs at disease onset, weight gain occurs with insulin treatment, and carbohydrate counting is a key element of type 1 diabetes management [14]. Type 1 diabetes also contributes to negative affect and diabetes distress, which increase risk for EDs [1,13].
Dozens of ED prevention programs have been evaluated, but most have not reduced ED symptoms and future ED onset [15–17]. A recent meta-analysis [18] found that the Body Project is the only prevention program that has produced reductions in ED symptoms and ED onset in multiple efficacy trials conducted by independent teams compared to both minimal intervention control conditions and credible alternative interventions [19–24]. The Body Project has also affected objective measures: in controlled trials it has reduced brain reward region response to thin models [25], positive implicit attitudes toward thin models [21], and attentional bias for thin models [24]. Further, it has shown efficacy for various racial and ethnic groups in several countries, for heterosexual and sexual minority females and males (with a version for males), and for a range of ages [18]. In the group-implemented Body Project individuals at high-risk for EDs due to body image concerns collectively critique the thin appearance ideal in verbal, written, and behavioral exercises, which theoretically generates cognitive dissonance that prompts participants to reduce pursuit of the unrealistic ideal because people align their attitudes with their publicly displayed behaviors. The Body Project targets young girls with body image concerns, making it a selective prevention program, though we also have included participants with eating disorders at baseline because we have found that participants with versus without an eating disorder at baseline have shown significantly greater reductions in symptoms, which suggests the Body Project might be better conceptualized as a blended selective and indicated eating disorder prevention program.
Emerging evidence suggests that the Body Project may be more effective when implemented virtually. Virtually implemented Body Project groups delivered by peer educators produced a 77% reduction in future onset of EDs compared to a placebo intervention over 2-year follow-up [19], which is larger than the 54% reduction in future onset of EDs for in-person Body Project groups led by peer educators compared to an educational control condition over 4-year follow-up [23]. Further, the average reduction in outcomes was 49% larger in a trial that evaluated virtual peer educator delivered Body Project groups (mean d = 1.13; [26]) versus in-person peer educator delivered Body Project groups from a previous trial (mean d = 0.53; [23]). Cohen’s d reflects relative change in standard deviation units. An implementation support study likewise found that peer-educator led Body Project groups produced significantly larger reductions in ED risk factors and symptoms when implemented virtually versus in-person [27]. Virtual delivery could markedly expand the reach of this prevention program for youth with type 1 diabetes.
The effectiveness of the Body Project led to a call to test its efficacy for young women with type 1 diabetes [8]. Wisting et al. (2021, [28]) tested an adapted version of the Body Project (Diabetes Body Project) that added 2 sessions focused on diabetes specific content in an uncontrolled pilot trial. Diabetes specific content was generated in collaboration with two peer educators with type 1 diabetes and included how to live well with diabetes and maintain self-care, and the advantages of insulin. Young women who completed virtual Diabetes Body Project showed significant reductions in pursuit of the thin ideal (d = 0.81), body dissatisfaction (d = 0.67), dieting (d = 0.63), diabetes eating pathology (d = 0.83), diabetes distress (d = 0.48), and diabetes illness perceptions (d = 0.41). The reduction in Hemoglobin A1c (HbA1c; d = 0.23), was not significant but encouraging, as a meta-analysis of other interventions targeting HbA1c reported a very small average effect size (d = 0.11; [29]). Because poor glycemic control is linked to morbidity and mortality, it is important to evaluate the effect of the Diabetes Body Project on glycemic control. Trojanowski and colleagues (2022, [30]) subsequently adapted the Body Project for adolescent girls with type 1 diabetes and also found medium-to-large within-condition reductions in outcomes from pretest to 3-month follow-up in an uncontrolled trial. However, neither prevention program has been evaluated in a controlled trial. We used qualitative data from our pilot trial to refine the intervention to improve acceptability of the intervention we developed for young women and launched a randomized trial to evaluate the effects of the Diabetes Body Project compared to a diabetes/ED education control condition for young women with type 1 diabetes.
2. Method
2.1. Participants and procedure
Participants were 55 adolescent girls and young women (M age = 22.8; SD = 4.3; range 15–30) recruited throughout the US. Participants completed assessments at pretest, posttest, and 3-month follow-up. We focused on this age group because EDs often emerge during late adolescence and young adulthood [3]. An a priori power analyses indicated that with an N of 60 and a 2-tailed alpha of 0.05, that we would have a power of 0.80 to detect a moderately large effect size (d = 0.60), which we selected because the average between-subjects effect for virtual Body Project groups compared to an educational control condition was d =0.79 [26]. The sample was 68% White, 9% Hispanic, 11% Asian, 2% African American, and 6% mixed heritage (4% did not to disclose).
Participants were recruited between 4/2022 and 9/2022 using email messages, ResearchMatch.com, social media (Reddit), and MyChart, a Stanford Medicine service. Exclusion criteria included: not in age-range, no Type 1 Diabetes diagnosis or a diagnosis less than 6 months ago, no body image concerns, not currently taking insulin, not having a diabetes care provider, and ED-related DKA or hospitalization. We included individuals with current EDs because participants with versus without an ED at baseline have shown significantly larger reductions in ED symptoms in response to the Body Project but excluded more severe ED cases because of potential medical instability that required monitoring. Participants or their parents (for minors) provided written informed consent. The Stanford University institutional review board approved this project.
Eligible participants completed the baseline assessment (diagnostic interview, surveys, and two-week wear of a continuous glucose monitor). After the baseline assessment, participants were randomly assigned via a random number sequence to virtual Diabetes Body Project groups co-led by a clinician and a peer educator with type 1 diabetes (n = 30) or to a control condition in which they watched educational videos about eating disorders and type 1 diabetes (n = 25; see Table 1). Participants were paid $15 for the pretest assessment, $20 for the posttest assessment, and $25 for the 3-month follow-up assessment and $25 for each complete wear of glucose monitoring at pretest and 3-month follow-up.
Table 1.
Educational Control Videos.
| Week | Topic | Length |
|---|---|---|
|
| ||
| One | Dying to be Thin1 | 53 min |
| Two | Video: Internal Battle of an Eating Disorder, Part 12 | 9 min |
| Video: Internal Battle of an Eating Disorder, Part 22 | 5 min | |
| Understanding Compulsive Exercise2 | 3 min | |
| Binge Eating Disorder2 | 9 min | |
| Emotional Eating2 | 5 min | |
| Emotions and Binge Eating: Part l2 | 15 min | |
| Emotions and Binge Eating: Part 22 | About 8 min | |
| Three | What is Type 1 Diabetes?3 | 6 videos, totals about 65 min |
| Four | Managing Hypoglycemia4 | 2 videos—total about 10 min |
| What is Insulin4 | 3 videos—total about 13 min | |
| The Plate Method for Diabetes4 | 1 video about 4 min | |
| Exercise and T1D3 | 2 videos, totals 10 min | |
| Five | Psychology and T1D3 | 4 videos, totals about 32 min |
| Pregnancy and Type 1 Diabetes3 | 8 videos, totals just under 30 min | |
| Six | Living with Type 1 Diabetes3 | Playlist of 10 videos, totals about 45 min |
Note:
Video is from Novo, PBS
Videos taken from Duke Psychiatry and Behavioral Sciences website, which are publicly available: https://psychiatry.duke.edu/duke-center-eating-disorders/educational-resources
Videos produced by JDRF and made available on YouTube: https://www.jdrf.org/t1d-resources/living-with-tld/exercise/
Videos were produced by Boston Medical Center and are publicly available: https://www.bmc.org/diabetes/managing-hypoglycemia#hypoglycemia. Contact authors for direct links if necessary.
Fig. 1 provides a participant flowchart. Female assessors with at least a bachelor’s degree completed a 4-hour training in which the diagnostic criteria for ED symptoms were reviewed and they observed supervisors conducting 4 diagnostic interviews; they were then asked to rate 12-audio-recorded diagnostic interviews and were required to show at least a 0.80 κ agreement with supervisors before starting data collection. Assessors were masked to condition. Diagnostic interviews were conducted over the phone, following standardized procedures. Surveys were administered via Redcap.
Fig. 1.
Difference in Adjusted Means by Outcome by Session by Condition. Note. Error bar corresponds to 95% confidence interval. Bars overlapping with 0 indicate no significant difference by condition at that session. Eating disorder symptoms was square root transformed.
2.2. Intervention
The Diabetes Body Project script was adapted from the standard Body Project script in collaboration with young women with type 1 diabetes. After the uncontrolled pilot study [28], qualitative feedback from participants [31] guided refinements to the script. The Diabetes Body Project consisted of 6 weekly 1-hour virtually delivered group sessions with 5–6 participants facilitated by a clinician and peer educator. In sessions, participants voluntarily engage in verbal, written, and behavioral exercises in which they critique pursuit of the thin ideal, as well as discuss advantages of insulin and costs of not giving the body sufficient insulin. Also, participants engaged in discussions on how to maintain a positive body image while taking care of their diabetes and on costs of pursuing the thin ideal when one has type 1 diabetes. Home exercises were completed between sessions. Participants who missed a session were asked to discuss the key activities of the session and their homework exercises over the phone with a facilitator before the next session, following standard procedures for the Body Project. Sessions were recorded to increase accountability, which increases dissonance induction, and so that sessions could be rated for intervention fidelity and therapeutic competence. See https://sticebodyprojectsupport.weebly.com/ for the manual.
The first author provided a 6-hour training during which the intervention theory and evidence-base for this intervention were presented and group facilitators practiced delivering the sessions virtually and received supervision, following established procedures [26]. Facilitators were clinical psychology graduate students and peer educators with type 1 diabetes. The last author viewed video-recordings of the group and rated fidelity and therapeutic competence for the first group provided by a facilitator team and 50% of subsequent sessions. Key elements of each session were rated for degree of accurate presentation (100-point scale from 10 = No adherence; the section was skipped to 100 = Perfect; all material in the section was presented as written). Facilitator competence was rated with 12 items (e.g., leaders express ideas clearly and at an appropriate pace) using a 100-point scale with behavioral anchors for each item (e.g., 20 = Poor; leaders are difficult to follow and session proceeds at an uncomfortable pace, 100 = Superior; leaders are unusually articulate and express ideas in way that all group members understand; perfect pace). The last author’s fidelity and competence ratings have shown inter-rater agreement with independent raters; intraclass correlation coefficient (ICC) was 0.78 for fidelity and 0.84 for competence [32]. The last author provided supervision to facilitators based on the fidelity and competence ratings.
2.3. Measures
Thin-ideal internalization (TII).
The 8-item Thin-Ideal Internalization scale assessed endorsement of the thin ideal espoused for women [13] using a response scale ranging from 1 = strongly disagree to 5 = strongly agree. Items were averaged for this scale and those described below. This scale has shown internal consistency (α = 0.91), 2-week test–retest reliability (r = 0.80), predictive validity for future ED onset, and sensitivity to detecting intervention effects [13].
Body dissatisfaction (BS).
Ten items from the Satisfaction and Dissatisfaction with Body Parts Scale [33] assessed satisfaction with body parts with a response scale ranging from 1 = extremely dissatisfied to 6 = extremely satisfied (responses were reverse scored). This scale has shown internal consistency (α = 0.94), 3-week test–retest reliability (r = 0.90), predictive validity for future ED onset, and sensitivity to detecting intervention effects [13].
Eating disorder symptoms (EDS).
The semi-structured Eating Disorder Diagnostic Interview (EDDI; [13]) assessed eating disorder symptoms, which include the frequency of binge eating, vomiting, laxative/diuretic use, fasting, and excessive exercise, as well as degree of overvaluation of weight/shape, feeling fat, and fear of weight gain. Participants were also asked about distress regarding binge eating, rapid eating, eating until uncomfortably full, eating large quantities of food when not hungry, eating alone because of embarrassment, and feeling disgusted, depressed, or guilty after binge eating. Items assessing symptoms in the past month were summed to form a composite (scores can range from 0 to over 100 for the most severe cases). This composite has shown internal consistency (α = 0.92), inter-rater agreement (ICC r = 0.93), 1-week test–retest reliability (ICC r = 0.95), predictive validity, and sensitivity to detecting intervention effects [13].
Diabetes Eating Problem Survey-Revised (DEPS-R).
The DEPS-R [34] is a self-report scale that assesses insulin omission, meal skipping, self-induced vomiting, and body/weight concerns. Scores of 20 or higher on the DEPS-R indicate a level of disturbed eating behavior. This scale has shown internal consistency (α = 0.86) and convergent validity with HbA1c [34] and with the EDDI (r = 0.78) in the present sample.
Diabetes distress (DD).
The 28-item Type 1 Diabetes Distress Scale (DDS; [34]) assesses seven areas tied to living with this condition that might cause distress: powerlessness, negative social perceptions, physician distress, friend/family distress, Hypoglycemia distress, management distress, and eating distress. Items were scored on a scale of 1 (not a problem) to 6 (a very serious problem). This scale has shown internal consistency (α = 0.91), 9-month test–retest reliability (ICC = 0.74), and convergent validity with alternative measures [34].
Quality of Life (QOF).
Participants completed the 5-item World Health Organization Well-Being Index (WHO-5; [35]) that assessed health related quality of life. Items are scored on a scale of 0 = at no time to 5 = all the time. The total raw score can range from 0 to 25 and is then multiplied by 4 for a final possible score of 0 (worse possible well-being) to 100 (the best possible well-being). This scale has shown internal consistency (α = 0.86), convergent validity with alternative measures, and discriminant validity [36]. We also assessed negative affect with the Positive Affect and Negative Affect Scale-Revised (PANAS-X; [37]) but omitted this outcome because it was highly correlated with quality of life (r = −0.67); the effects were identical for these two outcomes.
Glycemic control.
Objectively assessed time-in-range (TIR), reflecting the percentage of time that glucose values are within the target range (70–180 mg/dL), were collected by continuous glucose monitoring sensors. CGM data were accessed through the participant’s cloud sharing system (e.g., Dexcom Clarity) and TIR was calculated based on 14 days of data. Participants used their own CGMs, all of which are factory calibrated and FDA approved. TIR has garnered significant attention in the past decade and now is considered a primary metric for understanding glycemic health of people with type 1 diabetes [39].
2.4. Statistical Methods
For all outcomes except TIR we fit a linear mixed effect model where we regressed the outcome onto the pretest score for the outcome, treatment group (coded as treatment group 1, 2, 3, 4, or 5 or 0 for control), assessment (coded 0 for posttest and 1 for follow-up), and the interaction of condition (coded 0 for control or 1 for treatment) and assessment. Using treatment group instead of condition as a main effect enabled us to control for the partially nested design, where observations were nested within participants for the control condition and where observations were nested within participants nested within groups for the treatment condition. We included treatment group as a fixed effect rather than a random effect because the number of groups (n = 5) and the number of participants per group were small (n = 6). To control for repeated measurements for each participant we included a random intercept. Models tested for condition differences at posttest and 3-month follow-up. Because TIR was measured only at pretest and 3-month follow-up, we fit a multiple regression model where we regressed TIR at 3-month follow-up onto the pretest TIR score and treatment group. For all models, we examined modeling assumptions and transformed data as necessary. We square root transformed eating disorder symptoms to meet normality and homogeneity of variance assumptions.
Missingness at pretest ranged from 0% to 11% for TIR; at posttest ranged from 11% for EDS to 15% for TII, BD, DD, and QOL; and at 3-month follow-up ranged from 7% for EDS to 18% for TIR. Missingness was assumed to be missing at random and full information maximum likelihood was used to handle missing data, consistent with an intent-to-treat approach. Models were fit in R using the lavaan package [38].
3. Results
Table 2 presents demographic and outcome data at pretest for participants randomized to the two conditions, which did not differ on pretest versions of the outcomes or demographics, suggesting that randomization resulted in equivalency across conditions. The average duration of diabetes for the intervention condition was 11.6 years and for the educational control condition was 13.2 years. This difference was not significantly different (t = 0.96, df = 52.83, p = 0.341. Table 3 presents the unadjusted means and standard deviations at pretest, posttest, and 3-month follow-up by condition and the correlations between the outcomes at pretest.
Table 2.
Participant Characteristics at Baseline.
| Control | Treatment | p-value | |
|---|---|---|---|
|
| |||
| Sample Size | 25 | 30 | |
| Age1 | 22.06 (4.66) | 23.50 (3.83) | 0.223 |
| Thin-Ideal Internalization | 3.22 (0.77) | 3.48 (0.52) | 0.151 |
| Body Dissatisfaction Index | 3.46 (0.83) | 3.66 (0.61) | 0.330 |
| Negative Affect | 2.38 (0.89) | 2.15 (0.73) | 0.318 |
| Eating Disorder Symptoms | 16.00 (18.12) | 12.83 (8.65) | 0.429 |
| Diabetes eating pathology | 1.43 (0.81) | 1.34 (0.65) | 0.650 |
| Diabetes Distress | 2.44 (1.01) | 2.44 (0.82) | 0.990 |
| Quality of Life | 2.07 (0.99) | 2.33 (0.99) | 0.347 |
| Time-In-Range | 0.62 (0.19) | 0.63 (0.21) | 0.942 |
| Ethnicity | 0.537 | ||
| Asian | 3 (12%) | 3 (10%) | |
| - | |||
| Black | 0 | 1 (3.3%) | |
| White | 18 (72%) | 25 (83.3%) | |
| 2 or More Ethnicities | 1 (4%) | 1 (3.3%) | |
| Other | 1 (4%) | 0 | |
| Did Not Disclose | 2 (8%) | 0 | |
| Hispanic | 3 (2%) | 2 (6.7%) | 0.650 |
| Education | 0.257 | ||
| Grade School | 1 (4%) | 0 | |
| Some High School | 1 (4%) | 3 (10%) | |
| High School Degree/Equivalent | 4 (16%) | 1 (3.3%) | |
| Some College, no Degree | 7 (28%) | 4 (13.3%) | |
| Associate’s Degree | 1 (4% | 2 (6.7%) | |
| Bachelor’s Degree | 7 (28%) | 14 (46.7%) | |
| Master’ s Degree | 3 (12%) | 4 (13.3%) | |
| Ph.D. or Higher | 0 | 2 (6.7%) | |
| Did Not Disclose | 1 (4%) | 0 | |
Note:
For continuous measures, two-sample t-tests were conducted and for categorical measures, Fisher exact tests were conducted to test for condition differences at baseline. One control participant did not provide data on ethnicity/race or caretaker education. Some percentages do not sum to 100% due to rounding.
Table 3.
Correlations, Means and Standard Deviations for Outcomes by Condition at Pretest, Posttest, and 3-Month Follow-Up.
| Correlations at Baseline |
Summary Statistics by Assessment Mean (Standard Deviation) |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Outcome | TII | BD | EDS | DEPS-R | DD | QOL | TIR | Condition | Pretest | Posttest | 3-Month |
|
| |||||||||||
| TII | 1 | 0.32 | 0.31 | 0.32 | 0.24 | − 0.15 | 0.24 | Control | 3.22 (0.77) | 3.15 (0.66) | 3.10 (0.65) |
| Treatment | 3.48 (0.52) | 2.92 (0.75) | 3.12 (0.71) | ||||||||
| BD | 0.32 | 1 | 0.48 | 0.51 | 0.30 | − 0.37 | 0.02 | Control | 3.46 (0.83) | 3.46 (0.50) | 3.29 (0.70) |
| Treatment | 3.66 (0.61) | 2.98 (0.56) | 3.09 (0.71) | ||||||||
| EDS | 0.31 | 0.48 | 1 | 0.78 | 0.49 | − 0.26 | − 0.27 | Control | 16.00 (18.12) | 11.87 (12.42) | 12.30 (14.00) |
| Treatment | 12.83 (8.65) | 6.50 (8.10) | 5.82 (7.14) | ||||||||
| DEPS-R | 0.32 | 0.51 | 0.78 | 1 | 0.66 | − 0.49 | − 0.39 | Control | 1.43 (0.81) | 1.02 (0.44) | 1.06 (0.69) |
| Treatment | 1.34 (0.65) | 0.64 (0.39) | 0.70 (0.48) | ||||||||
| DD | 0.24 | 0.30 | 0.49 | 0.66 | 1 | − 0.49 | − 0.22 | Control | 2.44 (1.01) | 2.08 (0.82) | 2.18 (1.22) |
| Treatment | 2.44 (0.82) | 1.81 (0.88) | 1.92 (0.96) | ||||||||
| QOL | − 0.15 | − 0.37 | − 0.26 | − 0.49 | − 0.49 | 1 | 0.10 | Control | 2.07 (0.99) | 2.55 (0.88) | 2.17 (0.93) |
| Treatment | 2.33 (0.99) | 2.94 (0.91) | 2.75 (0.99) | ||||||||
| TIR | 0.24 | 0.02 | − 0.27 | − 0.39 | − 0.22 | 0.10 | 1 | Control | 0.62 (0.19) | − | 0.61 (0.20) |
| Treatment | 0.63 (0.21) | − | 0.63 (0.21) | ||||||||
Note. TII = thin-ideal internalization, BD = body dissatisfaction, NA = negative affect, EDS = eating disorder symptoms, DD = diabetes distress, QOL = quality of life, TIR = time-in-range, DEPS-R = diabetes eating pathology.
Among the 30 participants who were assigned to the intervention condition, 90% (n = 27) attended all 6 sessions or missed a session but made it up; 3.3% (n = 1) attended 3 sessions; and 6.6% (n = 2) attended 1 session. Fidelity (M [SD] = 74.6 [20.8]) and competence (M [SD] = 82.8 [9.5]) ratings of session delivery were high. Among the participants assigned to the educational control condition, 82% reported watching all videos.
Table 4 presents the tests of condition differences by assessment for each outcome. Fig. 1 graphs the differences in model-implied means (with standard errors). There were significant effects for thin ideal internalization, body dissatisfaction, eating disorder symptoms, diabetes distress, and quality of life. For thin ideal internalization, there was a significant difference at posttest (p =.001), where Diabetes Body Project participants had an average score 0.44 lower than controls (d = −0.62). For body dissatisfaction, there was a significant difference at posttest (p =.001), where Diabetes Body Project participants had an average score 0.53 lower than controls (d = −0.90). For eating disorder symptoms, there was a significant difference at posttest (p =.007), where Diabetes Body Project participants had an average score 0.79 lower than controls (d = −0.52) and at 3-month follow-up (p =.013), where Diabetes Body Project participants had an average score 0.72 lower than controls (d = −0.46). For the diabetes eating pathology, there was a significant difference at posttest (p =.002), where Diabetes Body Project participants had an average score 0.33 lower than controls (d = −0.74) and at 3-month follow-up (p =.002), where Diabetes Body Project participants had an average score 0.35 lower than controls (d = −0.58). For diabetes distress, there was a significant difference at posttest (p =.049), where Diabetes Body Project participants had an average score 0.37 lower than controls (d = −0.43). For quality of life, there was a significant difference at 3-month follow-up (p =.038), where Diabetes Body Project participants had an average score 0.46 higher than controls (d = 0.46).
Table 4.
Test of Condition by Session Effects by Outcome.
| Outcome | Session | Treatment | Control | Difference | SE | z | p | Cohen’s d |
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| Thin-ideal internalization | Posttest | 2.87 | 3.32 | − 0.44 | 0.13 | − 3.34 | 0.001 | − 0.62 |
| Follow-Up | 3.08 | 3.29 | − 0.21 | 0.13 | −1.57 | 0.116 | − 0.31 | |
| Body dissatisfaction | Posttest | 3.00 | 3.52 | − 0.53 | 0.16 | − 3.38 | 0.001 | − 0.90 |
| Follow-Up | 3.09 | 3.33 | − 0.24 | 0.16 | −1.55 | 0.121 | − 0.35 | |
| Eating disorder symptoms | Posttest | 2.37 | 3.16 | − 0.79 | 0.29 | − 2.71 | 0.007 | − 0.52 |
| Follow-Up | 2.20 | 2.91 | − 0.72 | 0.29 | − 2.49 | 0.013 | − 0.46 | |
| Diabetes eating pathology | Posttest | 0.720 | 1.052 | − 0.33 | 0.11 | − 3.08 | 0.022 | − 0.74 |
| Follow-Up | 0.759 | 1.105 | − 0.35 | 0.11 | − 3.17 | 0.002 | − 0.58 | |
| Diabetes distress | Posttest | 1.82 | 2.19 | − 0.37 | 0.19 | −1.97 | 0.049 | − 0.43 |
| Follow-Up | 1.96 | 2.14 | − 0.18 | 0.19 | − 0.96 | 0.336 | − 0.17 | |
| Quality of life | Posttest | 2.79 | 2.61 | 0.18 | 0.22 | 0.84 | 0.401 | 0.20 |
| Follow-Up | 2.68 | 2.22 | 0.46 | 0.22 | 2.08 | 0.038 | 0.46 | |
| Time-in-range | Follow-Up | 0.64 | 0.59 | 0.05 | 0.03 | 1.53 | 0.126 | 0.26 |
Note. All models controlled for pretest scores on the outcome, treatment group, and including random intercepts for participants to control for repeated measurements (except time-in-range). Values reported in the treatment and control columns are adjusted means reported at the mean of the pretest.
It should be noted that 2 participants had bulimia nervosa and 2 participants had binge eating disorder at baseline (2 were in the intervention condition and 2 were in the control condition), and that none of the significant intervention effects became non-significant when these 4 participants were excluded.
4. Discussion
Diabetes Body Project participants showed significantly greater reductions in pursuit of the thin ideal, body dissatisfaction, and eating disorder symptoms than participants randomized to the educational video control condition, though the effects for the first two outcomes were no longer significant by 3-month follow-up. The between-condition effect sizes for thin-ideal internalization, body dissatisfaction, and eating disorder symptoms at posttest in the present trial (d = 0.62, 0.90, and 0.52), which were medium-to-large effect sizes, were similar to or larger that the effects for peer-led Body Project groups from a randomized trial (d = 0.68, 0.71, and 0.38; [26]). The observed effect sizes were also similar to the within-condition reductions in thin-ideal internalization and body dissatisfaction (d = 0.81 and 0.67) from a preliminary uncontrolled study [28]. It is also important to note that the average scores on eating disorder risk factors and symptoms in the present sample were similar to the average scores from past Body Project trials in which onset of EDs were higher than observed in representative samples [22,23,26], confirming that the present sample was likewise at high-risk for EDs.
Regarding diabetes-related outcomes, compared to educational video controls, Diabetes Body Project participants showed significantly greater reductions in diabetes eating pathology at posttest and at 3-month follow-up, greater reductions in diabetes distress at posttest (but not at 3-month follow-up) and greater increases in quality of life at 3-month follow-up (but not at posttest). These effects sizes ranged from medium (d = 0.43) to large (d = 0.74). The fact that three of the seven outcomes showed effects at posttest but not 3-month follow-up suggests that the effects faded over follow-up and may have been due to the limited power of this pilot trial. The improvement in quality of life is consistent with the finding that the Body Project has produced improvements in interpersonal functioning [22]. The between-condition effect size for diabetes eating pathology (d = 0.74) and diabetes distress (d = 0.43) were similar to the within-condition effect from the uncontrolled pilot (d = 0.83 and d = 0.48, respectively; [28]).
Although TIR decreased for controls through 3-month follow-up but not for Diabetes Body Project participants (d = 0.26), this difference did not reach significance in this pilot. Nonetheless, the fact that this effect was similar to the within-condition d = 0.23 for change in HbA1c in the uncontrolled pilot [28] suggests that the Diabetes Body Project may produce a small improvement in glycemic control.
In total, 90% of participants completed all sessions of the Diabetes Body Project, suggesting high acceptability. The fact that attendance was higher than was observed in the uncontrolled study, where 74% attended all sessions [28], implies that the refinements that were made to this intervention based on the qualitative input from the uncontrolled trial may have improved acceptability. The main feedback from participants was to include an early recognition that everyone had type 1 diabetes and the impact of this chronic medical condition on body image and eating behaviors. Fidelity (M [SD] = 74.6 [20.8]) and competence (M [SD] = 82.8 [9.5]) ratings of virtual Diabetes Body Project groups were slightly higher than ratings from Body Project groups implemented in-person by peer educators (M [SD] = 70.2 [5.2] and M [SD] = 65.1 [8.6]; respectively, [13]).
It is important to note the limitations of the present study. First, the sample size for this pilot was only moderate, which limited sensitivity. Second, we only conducted a 3-month follow-up. It will be important for larger future trials to follow participants for a longer period and to directly test whether this intervention significantly reduces future onset of EDs as has been observed in trials of the Body Project. Third, although we collected TIR data to assess changes in glycemic control, we were unable to also collect HbA1c samples from participants given the remote nature of the study and fiscal limitations. Fourth, the sample was somewhat homogeneous with regard to race/ethnicity and parental education, which limits generalizability.
5. Conclusions and directions for future Research
In sum, results provide evidence that the Diabetes Body Project produced medium to large reductions in ED risk factors, ED symptoms, diabetes eating pathology, and diabetes distress, and improvements in quality of life for young women with type 1 diabetes when co-implemented by a clinician and a peer educator with type 1 diabetes. It would therefore be useful to utilize the qualitative feedback from this trial to refine the Diabetes Body Project and conduct a fully powered trial that evaluates the effects on these continuous outcomes and tests whether this intervention significantly reduces future onset of EDs over long-term follow-up in young women with type 1 diabetes.
Funding
This work was supported by pilot funding from the Stanford Diabetes Research Center (P30DK116074).
Footnotes
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- [1].Colton PA, Olmsted MP, Daneman D, Farquhar JC, Wong H, Muskat S, et al. Eating disorders in girls and women with type 1 diabetes: a longitudinal study of prevalence, onset, remission, and recurrence. Diabetes Care 2015;38(7):1212–7. [DOI] [PubMed] [Google Scholar]
- [2].Young V, Eiser C, Johnson B, Brierley S, Epton T, Elliott J, et al. Eating problems in adolescents with Type 1 diabetes: a systematic review with meta-analysis. Diabet Med 2013;30(2):189–98. 10.1111/j.1464-5491.2012.03771.x. [DOI] [PubMed] [Google Scholar]
- [3].Stice E, Marti CN, 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. J Abnorm Psychol 2013;122(2):445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Olmsted MP, Colton PA, Daneman D, Rydall AC, Rodin GM. Prediction of the onset of disturbed eating behavior in adolescent girls with type 1 diabetes. Diabetes Care 2008;31(10):1978–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Rodin G, Olmsted MP, Rydall AC, Maharaj SI, Colton PA, Jones JM, et al. Eating disorders in young women with type 1 diabetes mellitus. J Psychosom Res 2002;53(4):943–9. [DOI] [PubMed] [Google Scholar]
- [6].Goebel-Fabbri AE, Fikkan J, Franko DL, Pearson K, Anderson BJ, Weinger K. Insulin restriction and associated morbidity and mortality in women with type 1 diabetes. Diabetes Care 2008;31(3):415–9. [DOI] [PubMed] [Google Scholar]
- [7].Peveler RC, Bryden KS, Neil HAW, Fairburn CG, Mayou RA, Dunger DB, et al. The relationship of disordered eating habits and attitudes to clinical outcomes in young adult females with type 1 diabetes. Diabetes Care 2005;28(1):84–8. [DOI] [PubMed] [Google Scholar]
- [8].Hanlan ME, Griffith J, Patel N, Jaser SS. Eating disorders and disordered eating in type 1 diabetes: prevalence, screening, and treatment options. Curr Diab Rep 2013; 13:909–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Gibbings NK, Kurdyak PA, Colton PA, Shah BR. Diabetic ketoacidosis and mortality in people with type 1 diabetes and eating disorders. Diabetes Care 2021;44(8):1783–7. [DOI] [PubMed] [Google Scholar]
- [10].Takii M, Uchigata Y, Tokunaga S, Amemiya N, Kinukawa N, Nozaki T, et al. The duration of severe insulin omission is the factor most closely associated with the microvascular complications of type 1 diabetic females with clinical eating disorders. Int J Eat Disord 2008;41(3):259–64. [DOI] [PubMed] [Google Scholar]
- [11].Allen KL, Byrne SM, Oddy WH, Schmidt U, Crosby RD. Risk factors for binge eating and purging eating disorders: Differences based on age of onset. Int J Eat Disord 2014;47(7):802–12. [DOI] [PubMed] [Google Scholar]
- [12].Jacobi C, Fittig E, Bryson S, Wilfley D, Kraemer H, Taylor CB. Who is really at risk? Identifying risk factors for subthreshold and full syndrome eating disorders in a high-risk sample. Psychol Med 2011;41(9):1939–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Stice E, Gau JM, Rohde P, Shaw H. Risk factors that predict future onset of each DSM–5 eating disorder: Predictive specificity in high-risk adolescent females. J Abnorm Psychol 2017;126(1):38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Rodin GM, Daneman D. Eating disorders and IDDM: a problematic association. Diabetes Care 1992;15(10):1402–12. [DOI] [PubMed] [Google Scholar]
- [15].Jacobi C, Hütter K, Völker U, Möbius K, Richter R, Trockel M, et al. Efficacy of a parent-based, indicated prevention for anorexia nervosa: randomized controlled trial. J Med Internet Res 2018;20(12):e296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Tanofsky-Kraff M, Shomaker LB, Wilfley DE, Young JF, Sbrocco T, Stephens M, et al. Targeted prevention of excess weight gain and eating disorders in high-risk adolescent girls: a randomized controlled trial. Am J Clin Nutr 2014;100(4):1010–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Taylor CB, Bryson S, Luce KH, Cunning D, Doyle AC, Abascal LB, et al. Prevention of eating disorders in at-risk college-age women. Arch Gen Psychiatry 2006;63(8):881–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Stice E, Onipede ZA, Marti CN. A meta-analytic review of trials that tested whether eating disorder prevention programs prevent eating disorder onset. Clin Psychol Rev 2021;87:102046. [DOI] [PubMed] [Google Scholar]
- [19].Ghaderi A, Stice E, Andersson G, Enö Persson J, Allzén E. A randomized controlled trial of the effectiveness of virtually delivered Body Project (vBP) groups to prevent eating disorders. J Consult Clin Psychol 2020;88(7):643. [DOI] [PubMed] [Google Scholar]
- [20].Halliwell E, Diedrichs PC. Testing a dissonance body image intervention among young girls. Health Psychol 2014;33(2):201. [DOI] [PubMed] [Google Scholar]
- [21].Kant RN, Wong-Chung A, Evans EH, Stanton EC, Boothroyd LG. The impact of a dissonance-based eating disorders intervention on implicit attitudes to thinness in women of diverse sexual orientations. Front Psychol 2019;10:2611. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Stice E, Marti CN, Spoor S, Presnell K, Shaw H. Dissonance and healthy weight eating disorder prevention programs: long-term effects from a randomized efficacy trial. J Consult Clin Psychol 2008;76(2):329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Stice E, Rohde P, Shaw H, Gau JM. Clinician-led, peer-led, and internet-delivered dissonance-based eating disorder prevention programs: Effectiveness of these delivery modalities through 4-year follow-up. J Consult Clin Psychol 2020;88(5):481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Tobin LN, Sears CR, von Ranson KM. Two eating disorder preventive interventions reduce attentional biases in body-dissatisfied university women: A cluster randomized controlled trial. J Consult Clin Psychol 2022. [DOI] [PubMed] [Google Scholar]
- [25].Stice E, Yokum S, Waters A. Dissonance-based eating disorder prevention program reduces reward region response to thin models; how actions shape valuation. PLoS One 2015;10(12):e0144530. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Stice E, Bohon C, Shaw H, Desjardins CD. Efficacy of virtual delivery of a dissonance-based eating disorder prevention program and evaluation of a donation model to support sustained implementation. J Consult Clin Psychol 2023. [DOI] [PubMed] [Google Scholar]
- [27].Stice E, Rohde P, Gau JM, Shaw H. Implementation factors that predict larger effects from a peer educator delivered eating disorder prevention program at universities. J Consult Clin Psychol 2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Wisting L, Haugvik S, Wennersberg AL, Hage TW, Stice E, Olmsted MP, et al. Feasibility of a virtually delivered eating disorder prevention program for young females with type 1 diabetes. Int J Eat Disord 2021;54(9):1696–706. [DOI] [PubMed] [Google Scholar]
- [29].Hood KK, Rohan JM, Peterson CM, Drotar D. Interventions with adherence-promoting components in pediatric type 1 diabetes: meta-analysis of their impact on glycemic control. Diabetes Care 2010;33(7):1658–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Trojanowski PJ, Frietchen RE, Harvie B, Mehlenbeck R, Fischer S. Internet-delivered eating disorders prevention program for adolescent girls with type 1 diabetes: Acceptable and feasible. Pediatr Diabetes 2022;23(7):1122–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Hage TW, Nilsen J-V, Karlsen KM, Lyslid MH, Wennersberg AL, Wisting L. “I am not alone”. A qualitative feasibility study of eating disorders prevention groups for young females with type 1 diabetes. J Eat Disord 2023;11(1):42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Stice E, Rohde P, Shaw H, Gau JM. Clinician-led, peer-led, and internet-delivered dissonance-based eating disorder prevention programs: Acute effectiveness of these delivery modalities. J Consult Clin Psychol 2017;85(9):883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Berscheid E, Walster E, Bohrnstedt G. The happy American body: A survey report. 1973. [Google Scholar]
- [34].Fisher L, Polonsky WH, Hessler DM, Masharani U, Blumer I, Peters AL, et al. Understanding the sources of diabetes distress in adults with type 1 diabetes. J Diabetes Complications 2015;29(4):572–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].De Wit M, Pouwer F, Gemke RJ, Delemarre-Van De Waal HA, Snoek FJ. Validation of the WHO-5 Well-Being Index in adolescents with type 1 diabetes. Diabetes Care 2007;30(8):2003–6. [DOI] [PubMed] [Google Scholar]
- [36].Hajos TR, Pouwer F, Skovlund S, Den Oudsten BL, Geelhoed-Duijvestijn P, Tack C, et al. Psychometric and screening properties of the WHO-5 well-being index in adult outpatients with Type 1 or Type 2 diabetes mellitus. Diabet Med 2013;30(2):e63–9. [DOI] [PubMed] [Google Scholar]
- [37].Watson D, Clark LA. Affects separable and inseparable: On the hierarchical arrangement of the negative affects. J Pers Soc Psychol 1992;62(3):489. [Google Scholar]
- [38].Team RC. R: A language and environment for statistical computing. 2014. R Foundation for Statistical Computing: Vienna, Austria; 2018. [Google Scholar]