Primary prevention aims to limit the incidence of disease, in this case of eating disorders, by managing causes and exposure to risk factors or the patient’s increased resistance to these. Clearly, the first step to establish primary prevention measures is to identify relevant exposures and assess their impact on the patient’s and the population’s risk of developing the disease.

Specific primary prevention interventions for the most frequent eating disorders can be classified in the following 5 types: psychoeducational, media literacy, dissonance-induction techniques, and interventions focused on eliminating eating disorder risk factors and strengthening the host.

In order to implement primary prevention measures there are two strategies, which often complement each other: the population strategy and the high-risk strategy. The population strategy is general and is aimed at the entire population with the objective of reducing its average risk. The high-risk strategy is individual and is aimed at those people who most need it, meaning the most vulnerable individuals or those presenting additional risk as a consequence of being exposed to certain risk factors. In the case of eating disorders, groups are considered of risk due to their age (adolescence), sex (female) or engagement in risk activities, be it as a hobby or at a professional level (gymnasts, athletes, models, ballerinas, skaters, elite athletes, etc.)¹³³.

5.1. What is the efficacy of primary prevention interventions in avoiding eating disorders? Are there any negative effects?

In order to respond to these questions, 4 high quality (1++) quantitative SRSE or metaanalysis (MA) on eating disorder prevention programmes have been identified.

Of the four MAs, one focuses the review on children and adolescents and only includes RCT¹³⁴; another one focuses on reviewing RCT and quasi-experimental studies (Q-RCT) on Internet-based prevention programmes¹³⁵; the others perform a general review of the topic, including RCT and quasi-experimental studies¹³⁶. The potential iatrogenic effect (negative impact) of primary prevention programmes was another objective of two of these 4 MAs^{134, 136}.

The MA conducted by Stice E, et al. (2007)¹³⁷ includes all studies considered in the MA carried out by Newton MS and Ciliska D published in 2006¹³⁵; 24/46 studies included in the MA conducted by Cororve FM, et al. (2006) ¹³⁶; 10/12 studies included in the fourth MA conducted by Pratt BM and Woolfendern SR (Cochrane Review), which was published earlier (2002) ¹³⁴.

The search for RCTs has identified eight studies that were either published after or not included in previous MAs^138-145. The eight RCTs assess high-risk strategies given that primary prevention programmes are aimed at adolescents (men and women) ¹³⁸; at adolescent women with body dissatisfaction^{141, 143, 145}, at female university students with subclinical levels of eating pathology or who wish to improve their body image^{139, 140, 144} and female athletes¹⁴². The most administered intervention type was psychoeducational^{138-140, 145} (in three of the studies the Internet was used), followed by dissonance-induction techniques^{143, 144}, critical assessment of media content (media literacy) ¹⁴¹ and a specific programme called ATHENA (The Athletes Targeting Healthy Exercise and Nutrition Alternatives)¹⁴².

The NICE CPG (2004)³⁰ did not address this aspect and only some protocols, recommendations and other documents on eating disorders elaborated by other organizations in our setting (See Annex 6.1.) tackled this aspect^{10, 11, 13, 19, 146-151}, even though only the PAPPS¹² group of experts formulates general recommendations.

In the following section, the scientific evidence on primary prevention programmes for eating disorders is described.

Scientific Evidence

In a RCT (Stice, 2008; United States) conducted on a sample of 481 adolescent women (mean age: 17 years) with body dissatisfaction, the following interventions were compared: dissonance-based thin-ideal internalisation reduction programme (group 1) vs. the healthy weight control programme (group 2) vs. expressive writing control condition (group 3) vs. assessment-only control condition (group 4). Group 1 showed significant reduction of thin-ideal internalisation, body dissatisfaction, negative affects and psychosocial impairment and lower risk for eating pathology onset at 2- and 3- year follow-up vs. control subjects. In group 1 there was a greater decrease in thin-ideal internalisation, body dissatisfaction and psychosocial incapacity when compared to group 3. Group 2 showed greater reduction in all results, less increases in weight and lower risk of eating pathology onset at 2- or 3- year follow-up vs. control subjects. Group 2 showed greater decrease in weight and in thin-ideal internalisation than group 3.

In a RCT (Becker, 2008)¹⁴⁴ a population comprised of 188 female university students (age range: 18-21 years) who had participated in a specific programme on body image but who did not have an eating disorder was studied. Cognitive dissonance moderated by trained peer leaders (N=88) was compared to mobilisation and social and political activism (media advocacy) (N=85) with 8-month follow-up. Results indicated that both interventions reduce thin-ideal internalisation, body dissatisfaction, dietary restraint and bulimic pathology at 8 months, although higher and lower risk groups responded differently. Both interventions were effective in the high-risk group. However, only the dissonance-based technique was effective in lower risk participants.

In an RCT (Heinicke, 2007) ¹⁴⁵ a population of 83 female adolescents (mean age: 14.4 years) from schools in Australia who self-identified as having eating and body image problems was studied. The programme (My Body, My Life: Body Image Program for Adolescent Girls) (N=40) was compared to the control group (N=43) at 2- and 6- months follow-up. This programme consists of 6 weekly group sessions delivered via the Internet during 6 weeks and a SH manual (psychoeducational content), coordinated by a trained therapist. Out of those who completed the programme (28 and 26 of the experimental and control groups, respectively), clinically significant improvements were reported in the experimental group in terms of body dissatisfaction, disordered eating and depression at the end of the intervention (p≤0.002). At two months follow-up, clinically significant improvements were obtained in the scores of different specific eating disorder outcome measures and other related measures (also at 6 months follow-up).

In an RCT (Jacobi, 2007) ¹⁴⁰ performed on a population comprised of 100 female university students in Germany (age range: 18-29 years) who wished to improve their body image, an 8-week Internet-based psychoeducational prevention programme (German adapted version of Stanford’s original programme) (N=50) and a wait-list (N=50) were compared, with follow-up at 3 months. The experimental group obtained better results than the control group in all measures. Participants in the experimental group presented sustained effects in terms of their knowledge regarding a healthy diet and exercise, a reduction in their wish to be thin and decreased disordered eating at 3 months follow-up.

This MA review (Stice E, 2007) ¹³⁷ included 68 controlled studies that assessed 51 eating disorder prevention programmes, demonstrating the wide array of programmes available. This also occurs when revising the content of the 15 programmes that had an effect, ranging from promoting self-esteem, stress management skills, body acceptance, healthy weight management practices, to the critical analysis of the thin-ideal.

Based on the results of this MA, 51% of eating disorder prevention programmes reduce eating disorder risk and 29% reduce current or future eating pathology. These overall percentages are favourable when compared to results obtained in other public health programmes (21% obesity prevention and 22% HIV prevention).

In an RCT (Brien, 2006) ¹³⁹ a population of 24 university females from Canada of several different origins (58% Caucasian, 19% Hispanic and the rest of other origins), presenting subclinical levels of eating pathology, was studied. An 8-week psychoeducational intervention (N=13) was compared to self-monitoring control (N=11). Participants in the experimental group showed improvement in the scores of different specific questionnaires on eating disorders and other outcome measures that were significantly different from those observed in the control group.

In the only RCT (Elliot, 2006) ¹⁴² identified in a population of 1,179 female athletes studying in schools in the US (mean age: 14.6 years), the ATHENA programme (N=457) was compared to a control group (standard treatments) (N=471). ATHENA is a programme that promotes a healthy diet and physical exercise as alternatives to eliminate harmful practices. It consists of group sessions with educational material. Significant reductions were observed in the experimental group versus the control group in terms of behaviours related with eating disorders and in the use of weight reduction drugs (p<0.05). Athletes obtained positive changes in strength-training self-efficacy (p<0.005) and in healthy dieting practices (p<0.001). The components of the ATHENA programme were significantly modified in the right direction: mood (p<0.005), refusal skills (p<0.05), belief in the media (p<0.005), perception of close friends’ body shaping drug use (p<0.001).

In yet another identified MA, in this case conducted by Cororve FM, et al. (2006)¹³⁶ 46 studies (32 published and 14 unpublished) were finally included. The interventions assessed in the different studies were classified as purely psychoeducational, CBT-based psychoeducational, or purely interactive/non psychoeducational. There were no differences between the types of interventions. There were no significant different between purely psychoeducational and CBT-based psychoeducational interventions regarding dieting practices, internalisation and body dissatisfaction. Prevention programmes had the greatest beneficial effect on the acquisition of knowledge. The effect was also positive in other outcome variables, indicating an improvement in the symptoms of general eating pathology, dieting behaviour and internalisation of the thin-ideal. Body dissatisfaction also improved, but the effect was not sustained at follow-up.

In the MA (Cochrane Review) published in 2002 by Pratt BM and Woolfenden SR¹³⁴ 12 RCTs were included. Based on the programme’s content there were 4 types of intervention: a) knowledge on eating disorders; b) eating attitudes and behaviours and adolescent issues; c) media literacy and mobilisation and social and political activism (advocacy); and d) interventions related with self-esteem. 2 of the programmes based on media literacy and mobilisation and social and political activism (advocacy) indicate a reduction in the internalisation or acceptance of societal ideals relating to appearance at 3- to 6-month follow-up (SMD:-0.28; 95% CI: -0.51 to -0.05). There is sufficient evidence to support the effect of the 5 programmes that include interventions on eating attitudes and behaviours and other adolescent issues. There is also sufficient evidence to support the effect of two of the programmes designed to improve self-esteem.