Introduction

Studies on the use and content of social networking sites (SNS) in the setting of eating disorders (ED) is a rapidly growing area of research [1]. To date, research has examined the role of pro-eating disorder social networking sites in perpetuating disordered eating behaviour and body image concerns among youth with EDs or those at risk for EDs [2]. Content analyses of these social networking sites have shown support for continued disordered eating behaviours, excessive exercise, and other compensatory behaviours which may maintain or worsen eating disorders. [3,4,5,6,7,8]

Viewers of pro-eating disorder SNS content have higher Eating Disorder Examination Questionnaire (EDE-Q) scores, more extreme weight loss behaviour, worse ED-related quality of life, and increased harmful behaviour post website usage [10]. In a recent sample of 405 participants who engaged with pro-eating disorder social media, 84% met criteria for a clinical/subclinical eating disorder [10].

Although most research has focussed on the impact of pro-eating disorder content on SNSs, research has also found a significant, albeit smaller, amount of pro-recovery and anti-eating disorder content [11,12,13]. Research has shown that these two online communities (pro-eating disorder and pro-recovery) are largely distinct, with minimal interaction between the two communities and different online content [11,12,, 14]. Posts of those who are recovering used more language suggesting insight (know, realize), past tense verbs (was, went) and displayed less psychological concerns with eating, school and death [15]. The recovery content focussed on inspiring others to recover, showed increased social connection and provision of reassurance, as well as highlighted the difficulties in recovery, including some posters’ ambivalence towards recovery [13, 14].

Further research is needed to understand the role of SNS in ED recovery or as a vehicle to encourage the receipt of professional help [1, 11,12,, 12]. Pro-recovery posts on SNSs may aid in recovery through functioning as a source of peer support through validation and reassurance, informational support through advice sharing and a forum for individuals to document their illness and recovery. While posting ED recovery content on SNSs appear to have higher rates of eating disorder recovery and more psychological evidence of recovery in their posts than those posting pro-eating disorder content, little is known regarding the prevalence of potential ongoing eating disorder behaviour in this SNS community which may be more oriented towards recovery.

Therefore, further research is needed regarding the presence of ongoing eating disorder symptoms or orientation towards professional help-seeking among these posters.

Research has suggested that SNS users with eating disorders may be receptive to technology-based interventions for treatment [10]. As such, standardized frameworks are necessary to identify warning signs of eating disorders in social media content. Previous SNS research has used the DSM-V guidelines [4] and EDE-Q as frameworks for identifying ED behaviours and cognitions. For use in the broader non-medical community, the Eating Disorder First Aid (EDFA) Guidelines were developed as a set of evidence-based guidelines to assist in early identification and treatment of eating disorders [16] as part of Mental Health First Aid (MHFA). No research to date has examined whether the EDFA guidelines can be applied to social network sites to identify ED warning signs. EDFA has the potential to be a useful identification tool as it was designed to be used by a broader community beyond health professionals treating eating disorders and it highlights behavioural, physical, and psychological warning signs and provides resources to encourage professional help-seeking.

While various SNS have been examined in past research, recent studies have focussed on Instagram content analysis of ED posts [6, 17,18,19,20]. Instagram is a popular SNS among adolescents and young adults, the demographic most affected by eating disorders [21,22,23,24] and has also been identified as the main SNS for advice and support [18, 25]. Previous studies that examined the role of SNS in ED recovery utilized SNS platforms that featured images less prominently such as Tumblr or Twitter [11,12,13]. Very few studies have included analysis of images and their content in the context of ED recovery [14].

This paper, therefore, seeks to gain deeper insight into ED recovery content on Instagram by analysing the quotations and images found in Instagram posts for evidence of ongoing EDs and analysis of ED recovery themes. This will provide a better understanding of eating disorders and the recovery process as portrayed on ED recovery hashtags. Furthermore, we will explore the utility of using the EDFA guidelines to identify ED-associated behaviour on Instagram. A better understanding of ongoing ED-associated behaviour, the ability of guidelines developed for a lay audience to identify the behaviour and the portrayal of recovery in SNS focussed on ED recovery can inform interventions that utilize SNS.

Methods

Relevant Instagram posts were identified from ED recovery hashtags #EDrecovery, #EatingdisorderRecovery, #anorexia recovery, #bulimia recovery and #bingeeatingdisorder recovery from 17–18 July 2019. The two most popular hashtags, #EDrecovery (4.0 million at the time of writing) and #EatingDisorderRecovery (2.07 million posts at the time of writing), were selected as the above two hashtags were commonly tagged with the other ED recovery hashtags such as #anorexia recovery, #bulimia recovery and #bingeeatingdisorder recovery. The most recent 600 posts were chosen to simulate real-world Instagram viewing behaviour and to ensure thematic saturation [6]. The username, post caption, hashtags used, location, and number of likes were collected for each post. These data fields were extracted automatically using the Octoparse software [26]. Manual data entry was required for age, gender, height and weight (if present). Images were manually screened for and coded based on the coding scheme explained below. Given that Instagram allows users to post a series of multiple images to a single post, all images within a given post were analysed and coded.

Our inclusion criteria, therefore, were the most recent 600 posts from #EDRecovery and #EatingDisorderRecovery found on the public domain. The exclusion criteria were as follows: if posts were removed, an increase in privacy settings that limited access to post, if the post was unrelated to eating disorders and finally non-English language posts.

Ethical approval and waiver of consent for this study was granted by the Nanyang Technological University Institutional Review Board. The information extracted from Instagram was anonymized and collected in such a manner that subjects could not be identified directly or through identifiers linked to the subjects, and subjects were neither contacted nor re-identified. Quotes that were extracted from Instagram posts and used in this manuscript were paraphrased to protect user privacy. Datasets were secured using strong password protection and encryption, with access restricted to study authors only.

The data extracted from each post were subsequently coded into various categories. The coding scheme was adapted from Tiggemann and Zaccardo [6] and Deighton-Smith and Bell [27] in their analysis of fitspiration images on Instagram and there were similarities in the type of posts found in both fitspiration images and #EDRecovery and #EatingDisorderRecovery with the predominance of food, people and quotes images. The posts were categorised into food, quotes, people and Others. Images featuring food were sub-categorised based on whether calories were displayed, if all food groups were present or if vegetarian or vegan options were portrayed. ‘Unhealthy food’ was defined as any food that is low in fibre and vitamins, high in salt and fats or high in sugar [28]. Images featuring quotes were sub-categorised as inspirational, factual or potentially negative. Images featuring people were sub-categorised by gender, ethnicity, body shape (in terms of muscularity and adiposity), action type and objectification (which includes focussing on a specific body part or posing). As a single post may feature elements from multiple categories, posts were sorted into multiple categories as applicable; hence percentages exceed 100%. We categorized the post’s account type into ‘Personal’ and ‘Practitioners/Advocates/Organizations/Influencers’.

In addition, all captions and images were mapped onto the EDFA Guidelines and analysed for behavioural, psychological and physical signs of eating disorders. Behavioural warning signs included dieting, binge eating or excessive exercise. Psychological warning signs included body dissatisfaction and distorted body image. Physical warning signs included negative changes in weight or menstruation.

Posts were also analysed for their orientation towards seeking professional help and were coded as follows: encourages, neutral or discourages help-seeking behaviours.

Data analysis

Quantitative data looking at the frequency of the various categories and their subcategories were presented in proportions. The frequency of behavioural, psychological and physical signs of ED and help-seeking behaviours was also presented in proportions. Demographics data of users were reported in mean with standard deviations or proportions as appropriate.

The posts were coded by author (AG) and 10 percent of the posts (n = 40) were coded independently by author (CCSE) to ensure coding reliability. There was high level of agreement between the authors across all domains of the posts (k = 0.79, p < 0.005).

To better understand the thematic content present in the posts, the qualitative data (text) were analysed using the process for a descriptive, inductive analysis as outlined by Braun and Clarke [29]. Themes were generated inductively after review of the data and through discussion of all authors. Themes and subthemes were refined through consensus among all the authors. Thematic connections between the various categories of the posts and signs of ED using EDFA guidelines were identified with the use of a thematic map. Data achieved saturation as no new themes arose with new data generated from the posts included in final analysis.

Results

Six hundred posts were initially collected with 405 unique posts included in final analysis (Fig. 1).

Fig. 1
figure 1

Flow diagram on analysis of posts

Full size image

Demographics of Instagram users

Of the 360 unique users whose posts were included in the analysis, the majority were personal accounts(68%) with the largest proportion of users originating from the USA (47.1%). The majority of users were young adolescent females (97.2%) with mean age of 19.0 years (SD = 3.2) (Table 1).

Table 1 Demographics of Instagram users
Full size table

Post content

The majority of posts contained elements pertaining to food (49.6%), followed by quotes (24.2%), people (22.7%) and others (12.3%). Food posts were notable for the high percentage of healthy foods (51.7%) and vegan/vegetarian meals (38.1%) depicted. In images of people, 31.8% revealed thin adiposity, while the most common action included posing (73.8%) (Fig. 1).

EDFA Guidelines: warning signs of eating disorders

A total of 172 posts suggested psychological and behavioural signs of ED with 91(53%) posts illustrating psychological signs of ED being the most common and was commonly found in food posts. This was followed by behavioural signs (n = 73; 42%) and physical signs (n = 8; 5%) (Table 2). In total, 127 users (31.3%) demonstrated at least 1 symptom of ED.

Table 2 Warning Signs of ED and help-seeking behaviours with examples of quotations (N = 172)
Full size table

Qualitative analysis

Six key themes were observed from the Instagram posts analysed, namely (1) Recovery as a difficulty process with relapses, (2) Self-validation and acceptance, (3) Motivation for recovery, (4) Role of treatment, (5) Increased awareness and stigma around ED, and (6) Developing a supportive community within Instagram. These six themes described the main content noted on these Instagram hashtags. A thematic map (Fig. 2) shows the connections between the post contents with eating disorder warning signs identified using EDFA and elucidation of the recovery themes. The six themes identified and the frequency of these six themes with an example of the quotations are found in Table 3.

Fig. 2
figure 2

Thematic map of posts with associated signs of ED and recovery themes

Full size image
Table 3 Recovery themes with examples of quotations
Full size table

Theme 1: Recovery as a difficult process with relapses

This is the most prevalent theme as Instagram users monitor their progress and ED recovery journey and their relapses. Sharing of critical factors in the recovery journey also featured prominently. Users commonly shared both the challenging and rewarding moments in their recovery and emphasised that recovery was an ongoing process that required persistence and vigilance. Posts detailed significant milestones, including days since their last relapse to remind themselves of the obstacles and struggles that they continue to face in their recovery process. These posts are commonly found in the ‘Food’ posts where users shared about anxiety of meals or fear foods, accompanied by photos of their meals. These posts also commonly share the rewards of recovery with improved physical and mental health and social functioning. These posts are commonly seen in the ‘Quotes’ category, accompanied by photos of motivational quotes to highlight the rewards of recovery.

Relapse was an important recovery theme including the recognition of triggers for the relapse and strategies on relapse management. Identified triggers included stressful life events such as moving to a new city or a new job.

“Since getting back to London I’ve really been spiraling… so many behaviours are sneaking back in, incessant calorie counting, weighing & body checking, exercise….”

Theme 2: Self-validation and acceptance

Self-validation and acceptance was another key theme identified with users posting the importance of forming their own identity and maintaining a sense of personal worth and value in their recovery. Several users posted that they are less self-critical, challenge negative self-talk and are more accepting of their body shape. They also celebrated personal success and accomplishments through their posts. These posts also talked about the need to ask for help. These users commonly post under pictures of ‘People’ posts with ‘before and after’ photos to highlight the importance of health, weight gain and body shape acceptance. For example:

“throughout the whole of my illness and recovery and every moment in between i have searched for validation from those that i know and those that know of my illness”

The role of an improved relationship with themselves was also key to the recovery process. Users share on how reconnecting with themselves had helped with their body image and eating habits and common strategies included yoga, art and development of awareness of body hunger cues.

“Yoga helped me learn how to heal this relationship with how I saw my body in the mirror”

Conversely, posts on self-criticism, fear over failures, sense of worthlessness and hopelessness were in the minority. These posts commonly talked about seeking outside validation with self-evaluation heavily based on the opinions and views of others. These posts also commonly commented on ongoing body image concerns and were accompanied by ‘People’ photos of themselves and body parts of concern such as arms, waists and face.

“Hating myself, hating my body, constantly feeling threatened by other girls..”

Theme 3: Motivation for recovery

A second key sub-theme in recovery was internal motivation for change. Most users articulated a need of a powerful desire to change and to let go of their ED and to move beyond their illness and views recovery as a liberation. Users wrote of a desire to be healthy with the decision being self-initiated and self-directed rather than for other people. For example:

“The only way to beat it is to challenge, rebel, stick your own fingers up in the air and grab your life back from it's filthy hands.”

A minority of users expressed ambivalence towards recovery and felt that the eating disorder allowed them to stay in control. For example:

“all I want is my eating disorder. because it is comfortable and it is safe and it is known”

Theme 4: Role of treatment

Users also commented on the role of treatment in their recovery. Users shared on their positive treatment experience and valued the input and validation of healthcare professionals such as, dietitians, therapist and ED service teams. While acknowledging the difficulty of adhering to healthcare professionals’ advice, they could also see the importance of their advice. Users also posted on the importance of adhering to therapy and medications and highlighted the downsides of stopping both. For example:

“Meds are no joke. Stopping them randomly causes you to crash hard. All it does is balance your chemicals in your brain nothing to be ashamed of.”

Theme 5: Increased awareness of and reducing stigma of ED

Several posts clarified stereotypical beliefs that ED must be characterised by extreme low weight to reduce stigma around EDs. Such posts also commonly highlighted the societal role in emphasis of the thin ideal and the need to reduce stigma about having an ED.

“Anorexia is a mental illness, there is no ‘look’”

“Eating disorders do not discriminate. Eating disorders affect people of every age, weight, race, and gender.”

Conversely, there were a minority of posts that felt they were not deserving of help and felt embarrassed by their diagnosis of ED, reinforcing the stereotypical belief that EDs must be characterised by extreme low weight..

“I’m not sick enough to be anorexic… and i don’t have a bmi under 18”

Theme 6: Development of a supportive community on Instagram

Many users highlighted the encouraging and supportive community within the Instagram posts and the importance in their recovery process. Users tag other users, motivating them to continue their recovery and acknowledging their growth since the beginning. This also took the form of grateful messages to other users, thanking them for their support. In this theme, there were no negative posts and no posts that suggested that SNSs may be negatively influencing their recovery. For example,

“Thank you so much @[username] for making me see that I can do this. Couldn't be doing recovery without you. You inspire me, encourage me, understand me and listen to me. We can do this”

Discussion

Sampling on ED recovery SNS, our study has performed a content analysis of quotations and images found in Instagram posts for warning signs of EDs using the EDFA Guidelines as a framework as well as conducted a thematic analysis regarding recovery themes.

Of concern, this study demonstrates that a significant proportion of users posting under recovery hashtags appear to exhibit ongoing behavioural and/or psychological signs of ED. However, only a limited number of users openly discuss or encourage others toward seeking professional help on Instagram. While we are unable to determine on the prevalence of help-seeking amongst users posting under these hashtags, previous research has shown limited professional help-seeking amongst ED patients who use social media [10].”

Food and people posts were the most common and the EDFA warning signs were a useful framework to identify warning signs of eating disorders in these posts. The warning signs of ED in these posts also provided insights into users’ perception and motivation towards recovery and self-validation. Food posts, in particular, were suggestive of numerous behavioural warning signs, including “exhibiting dieting behaviours” and “changes in food preferences” and 38% of the food posts being vegan. The high prevalence of objectification in people posts suggest that users may be concerned with body shape and weight, which is consistent with behavioural and psychological signs of ED. Identifying warning signs of EDs even amongst a recovery community is unsurprising—given that users would be in various stages of the recovery process, the presence of ED content is expected as users share authentically about their own struggles and relapses. This information may be useful for future intervention for targeted identification of eating disorder warning signs among a community of individuals oriented towards recovery.

This study thus provides valuable information on important themes for recovery and may be helpful for the development of online engagement tools to encourage help-seeking.

Consistent with previous research on the use of Instagram in recovery, our thematic analysis demonstrates that users track their personal growth in recovery on Instagram and demonstrated similar themes including recovery tracking and reducing stigma of EDs [30].

Interestingly, our thematic analysis displayed common themes as previous qualitative interview research examining the themes of relapse and recovery for individuals in ED receiving treatment. Previous research has identified the themes of recovery as a process that requires vigilance and persistence, the need for internal motivation for change, the need for self-validation and the negative impacts of seeking outside validation [31]. In addition to the role of relationships with family and friends, our data revealed support coming from the virtual SNS community. Other qualitative research studies similarly highlighted the role of stigma as a barrier to utilizing mental health related services [32] and life transitions such as moves as potential triggers for ED behaviour. These similarity in themes between our study and previous related qualitative research provides support for the generalizability and validity of these themes [33].

The potential functions of Instagram for users interested in ED recovery can be leveraged by healthcare providers and mental health organizations for intervention. Firstly, our study demonstrates that Instagram serves as a platform for peer support. SNS communities on Instagram can possibly support recovery through the development of positive peer support communities. While the role of online community is less well studied, previous research highlights the role of community, including peer and family support, in ED recovery [34, 35].

Second, Instagram can serve as a tool to reduce stigma and provide informational support regarding eating disorders and treatment. Many patients already receive health information from SNSs due to its ease of accessibility [36, 37]. Efforts on social media towards stigma reduction could possibly reduce barriers to the barriers of mental health services. Moreover, stigma can be reduced through framing of users’ ED in the context of everyday life, via implicit and explicit references [38]. Rather than expressing their illness as something extraordinary, deviant and to be stigmatized, users referenced it in their posts as part of their daily life. By assimilating recovery into their everyday activities and lives, the eating disorder is destigmatized. This shows readers of these posts that it is possible to continue through life with ED and recover from it.

Instagram itself could potentially be leveraged as an adjunctive treatment to aid users in documenting their own journey in recovery through recording their positive and negative milestones. Ramsey-Wade, Williamson and Meyrick suggest that therapeutic writing in individuals with ED can create a relieving effect through the expression of emotional experiences [39]. These Instagram users documenting their recovery journey are possibly engaging in a process similar to therapeutic writing and this could be leveraged in treatment. Women with eating disorders commonly experience shame and solitude [40] and sharing of lived experiences may allow for Instagram users with ED to share through the relation of lived emotional experiences. However, Instagram is potentially a double-edged sword, populated with both pro-recovery and pro-ED content; the latter is potentially a hindrance to recovery. Further research is hence required to establish the role of Instagram in facilitating ED recovery, and whether the use of SNSs like Instagram could negatively affect some individuals’ recovery from EDs.

Finally, Instagram can be a source of learning and listening for health care providers. While healthcare professionals are proponents of treatment and have ongoing contact with those receiving treatment, posts on self-criticism, fear over failures, sense of worthlessness and hopelessness are valuable source of learning for health care professionals [41]. These stories can allow for a respectful and different kind of listening termed as “listening in the darkness” where providers can learn from the stories of these ongoing struggles.

This paper has several limitations. Both our exclusion of non-English language posts as well as the homogeneity of our posters’ demographics may affect the generalizability of our findings. Second, our data were limited to posts that remained visible in the Instagram public domain. We were unable to analyse private posts or posts removed by individual users or Instagram content monitors due to violation of the Instagram community guidelines. This may led to selection bias as individuals may be more willing to post socially desirable content in a public forum.

Further research is needed to better understand the users who post on ED recovery hashtags. Whilst our analysis suggests that Instagram users demonstrate ongoing signs of EDs when posting on ED recovery SNSs, further research is needed to evaluate the prevalence of active clinical and subclinical eating disorders in this SNS community. Moreover, further research should evaluate the effect of accessing pro-recovery ED SNS on ED cognitions as well as how to leverage communication on these sites to increase appropriate help-seeking for those in need.

What is already known on this subject?

Limited research has evaluated the role of Social Networking Sites (SNS) in eating disorder (ED) recovery. While research has demonstrated the deleterious effects of pro-ED SNS content, less is known regarding the SNS content documenting ED recovery.

What this study adds?

Despite identifying with ED recovery, posts had a high prevalence of ongoing ED behaviour and low prevalence of help-seeking. Thematic analysis emphasized recovery as a journey, role of self-validation, motivation and treatment in ED recovery. Stigma around ED and development of a supportive community in SNS are also key themes identified.