An Interactive Internet-based Program for Prevention of Eating Disorders on a broad basis
Meta-analyses have shown that interactive (versus didactic) preventive programs that target selected groups (versus universal) are more effective. Selection, especially among youth, might lead to stigmatization and a tendency to avoid participation in such programs. Current approaches to implement such programs reach very few.
We will be adapting the DBI to be delivered digitally, while it remains interactive, which is more innovative compared to other Internet-based interventions that are highly psychoeducational. Such a major adaptation decreases stigmatization as the selection can be done in private, booster sessions can be implemented easily, which maintains the long-term effects, and the intervention can be implemented with high cost-effectiveness on a broad basis.
Within a randomized controlled study, we recruit 400 girls (15-17 years old) to participate in DBI (4 session), expressive writing or a wait-list condition. In addition to mediators and moderators, the outcome will be assessed post-intervention as well as 6, 12 and 18 months later.
The study was completed successfully in line with the plans in the application. The main aim of the study was to investigate the efficacy of a scalable intervention for prevention of eating disorders. In addition, we aimed to investigate different strategies for recruitment to create a model for largescale implementation of the intervention in the future in case we would obtain promising outcome. Finally, we would also conduct an analysis of its cost-benefits.
We obtained ethical approval for the study, registered it on Clinical Trials (NCT02567890), and translated the Dissonance-based intervention (DBI) and adapted the DBI for delivery through virtual groups. Then, we investigated the reliability and stability of different platforms for virtual groups. Most programs such as Skype showed to be not secure or stable enough for groups up to 5-6 participants. Google Hangouts which is a free application showed to be a suitable option.
To test different recruitment strategies, we sent a letter about the project to all the high school principals in Sweden. Although several principals expressed interest in the study, the procedures to reach the pupils at schools, demanding approval of school psychologist, school nurses and student health teams and several rounds of internal information, showed to be very slow, and very few participants could be recruited that way. We sent an amendment to the ethical application to recruit participants at 15-20 years of age instead of 15-18 to widen the pool from which we could recruit. Next, we educated psychology students to act as ambassadors for the project, visiting the schools and informing about the study while we also improved the website of the study to recruit from Facebook. Recruitment through ambassadors showed to be too slow, and although we recruited some more participants via Facebook, the numbers were small. After consulting with experts on digital technic and social media, a new website for the study was created (www.sbodyproject.se) along with banners for ads in Facebook and Instagram. After systematic tests and optimizing the banners, we started the sharp recruitment. Within 24 hours, we managed to reach 70.000 in our target group, and after just a few days we had reached 240.000. A total of 1678 potential participants responded to our ads, and we closed the application form directly as it was impossible to interview too many subjects within a reasonable time frame. Our aim was to recruit 400 participants. We closed the application form after recruiting 43 participants. We randomized the 431 eligible participants to DBI, expressive writing (EW) or a waitlist control group. The wait-list group was offered the DBI after 6 months of waiting, and all the participants were followed-up 24 months post-intervention (pre-assessment, post-assessment, and 6, 12, 18 and 24 months follow-up).
At 24 months follow-up, 50% in the DBI and 53% in the EW completed the assessments. All the analyses were though intention-to-treat. All randomized participants were included in all the analyses. The main outcome variable was the incidence of eating disorders across the follow-up.
Across the 24 months follow-up, 3 participants in the DBI and 13 in EW developed an eating disorder based on the DSM-system. This means 73% lower incidence in the DBI compared to a credible placebo (EW), which is a statistically significant difference (Cox's proportional hazard model: B=-1.32, Wald=4.26, p=.039, Hazard Ratio (Exp. B)= 0.27, 95% CI [0.076, 0.94]), and clinically meaningful effect.
DBI resulted in significantly larger short-term (post-treatment and 6 months follow-up) reduction of self-reported risk factors for eating disorders compared to EW and the wait-list condition. In terms of long-term follow-ups, the DBI was superior to EW in terms of reduction in eating disorders symptoms, internalization of the thin ideal and body dissatisfaction.
The outcomes are promising and mark a change in prevention of eating disorders as the study shows that the conditions to implement the DBI on a large scale with good outcome are present.
The cost-benefit analyses are not performed yet, but they will be done given the drastic differences between the conditions in terms of the incidence of ED in DBI versus EW.
The main results are submitted for open source publication. Another manuscript, in addition to the one on cost-benefit analysis, is in preparation to explain strategies for large-scale recruitment for prevention efforts.
The project resulted in closer collaboration with Professor Eric Stice (Stanford University) and his team. We are currently looking at possibilities for implementing the intervention on a large scale and seeking funding. The results from the project have been presented on several international conferences and discussed with other researchers.