Experiences of Social Media Platforms' Policies and Restrictions Related to Self-Harm Content: Mixed Methods Study.

Abstract

Many features of social media platforms can influence safety in relation to self-harm/suicide-related content. In 2019, platforms including Meta updated their policies regarding self-harm/suicide content. This included community guidelines, content restrictions, and signposting. The impact of these safety measures has yet to be fully evaluated. This study aimed to better understand the perspectives of social media users on safety measures related to self-harm/suicide content, including perspectives of those with a history of self-harm. This was a cross-sectional mixed methods study. Participants were recruited through convenience sampling. Quantitative data were collected using a fixed-response cross-sectional survey (5294 respondents aged 16-84 years, largely women/girls or nonbinary). Questions related to the experience of self-harm and suicide content, the change policy in 2019, and specific features of platforms. Qualitative data were collected via semistructured interviews (17 participants aged 16-56 years) with interview guides broadly following the topic areas of the survey. Interviews were analyzed using thematic analysis. Half of the survey respondents (49.8%, 95% CI 47.8%-51.8%; 2476/4970) stated the current restrictions made social media feel safer, 34.9% (95% CI 32.6%-37.2%; 1736/4981) reported that this changed what they saw on social media, and just 9.9% (95% CI 7.5%-13.0%; 492/4479) stated that this changed what they posted. Furthermore, 64.3% (95% CI 62.6%-65.9%; 3212/4997) stated that they would be "very likely" to click on a post marked with a generic content warning, with 15.1% (95% CI 12.6%-17.9%; 753/4998) being "very likely" to click on a post with a self-harm specific content warning. Of the respondents who had content removed due to mental health or self-harm content, ~90% indicated that this was harmful or very harmful to them at the time, particularly where content was removed because of visible scars (extremely harmful 58.3%, 95% CI 54.3%-62.3%; 599/1027; harmful 31.3%, 95% CI 26.3%-36.7%; 321/1027). Thematic analysis of in-depth interviews revealed 13 subthemes organized into 7 overarching thematic areas. Participants described the experience of social media platforms before and after the policy changes, with opinions divided on whether existing safety features were effective. Participants described the distress caused by inappropriate censoring of images, for example, where healed self-harm scars were visible. Suggestions were made for how platforms can continue to improve, including age verification procedures for young people, tailored signposting, and increased control over content. To our knowledge, this is the first study to explore the 2019 change in platform policies with respect to self-harm/suicide-related content from the perspectives of those with a history of self-harm. This research adds to the evidence of the risks and unintended consequences of imposing untested blanket restrictions in online settings. Full evaluation of these restrictions is essential to allow platforms to continue to improve, encouraging a safe space for supportive communities while mitigating potential harm.