New research suggests how social media draws vulnerable users back to eating disorder content

Paula Saukko for The Conversation

People recovering from eating disorders often use social media for support, seeking out recovery content, body-positive creators and others with similar experiences.

But recent research my colleagues and I have conducted suggests these platforms can also steer users back towards the very content they are trying to avoid.

We carried out in-depth interviews with people who had experienced eating disorders. Participants described how diet, fitness and body-focused posts repeatedly appeared in their social media feeds, even when they were actively trying to follow recovery content.

Supportive and potentially harmful material often surfaced side by side during the same scrolling session.

Participants said they used social media to manage their mental health, following recovery accounts and blocking triggering material. At the same time, many felt recommendation systems continued to introduce weight-loss content, fitness imagery and appearance-focused posts.

Some felt this exposure had contributed to setbacks in their recovery or reinforced unhealthy thought patterns, although these are self-reported experiences rather than causal findings.

This qualitative research captures how people experience social media during recovery. It does not show that social media causes eating disorders, or that exposure to specific content leads directly to relapse.

It does, however, highlight how users navigate platforms where recovery and diet content coexist, and how recommendation systems shape their feeds.

A growing body of research suggests this wider environment matters.

Studies have linked social media use with body dissatisfaction and disordered eating symptoms, particularly among young people and women, though these relationships are complex and cannot establish causation.

Exposure to idealised body imagery, “fitspiration” and diet content has been associated with increased concern about weight and appearance in observational research.

Platform dynamics are also part of the picture. One study found that TikTok’s recommendation system delivered substantially more diet content to users who had indicated eating disorder experiences than to those who had not.

Recommendation systems determine what appears in a person’s feed based on patterns of viewing and interaction, which can reinforce existing interests and vulnerabilities.

Other work by researchers in this area has shown how users can become caught in cycles of appearance-focused content online.

Research on Instagram and other visual platforms suggests that repeated exposure to diet, beauty and fitness posts can narrow what users see, keeping them within loops of body-related material rather than directing them towards a broader range of interests.

Our interviews add depth to this evidence by showing how these patterns are experienced in everyday life. Participants described moving between recovery content and more harmful material, sometimes within minutes.

Several said this constant switching made it harder to disengage from weight-focused thinking, even when they were trying to avoid it.

At the same time, many emphasised the positive role social media played in their recovery. Online content offered reassurance and provided access to experiences and perspectives that were difficult to find offline.

The same platforms that exposed users to triggering material also enabled connection and support. This push-pull dynamic is central to understanding social media’s role in eating disorders.

Rather than being purely harmful or beneficial, platforms create environments where supportive and risky content coexist, shaped by both user choices and recommendation systems.

NHS survey data indicate that around one in five girls aged 17–19 in England has an eating disorder. This underlines how common these conditions are among young people, who also tend to be heavy users of digital platforms.

These findings raise questions about how social media environments are structured, particularly as governments consider age-based restrictions, such as those introduced in Australia and proposals currently debated in the UK.

We argue our research suggests attempts to make online spaces safer need to go beyond a focus on who can access social media.

These efforts need to look at how content is curated and amplified, as platform design and recommendation systems appear to play an important role in shaping exposure.

There are already social media literacy initiatives in schools and elsewhere that aim to help young people critically evaluate idealised body images.

Strengthening these programmes to include greater understanding of how recommendation systems work could help users better navigate environments where supportive and harmful content can appear side by side.

As Meta, YouTube and TikTok face lawsuits alleging that aspects of their platform design encourage compulsive use, debates about regulation are likely to intensify.

The experiences described in our research suggest that, for people vulnerable to eating disorders, what matters is not simply time spent online, but how feeds are structured and how difficult it can be to avoid appearance-focused material once it enters the algorithmic loop.

The author, Paula Saukko, is a reader in Social Science and Medicine, Loughborough University.