Experiences of health-related stigma can discourage people from seeking medical treatment and can adversely affect their mental and physical well-being. Research suggests that up to half of people living with chronic illness in the UK report experiences of stigma connected to their health condition.

Online Health Stigma

These days, a significant portion of discourse around health, and the associated stigma that comes with it, takes place on social media platforms. Platforms like Twitter (X) have received strong criticism for facilitating the dissemination of contentious or divisive content.

It is important that researchers study health stigma in order to identify the narratives and communication features that proliferate online. Examining tweets that include health stigma can shed light on the prevalence of potentially stigmatizing narratives online. Such insights can assist health communicators in delivering precise and impactful health messages to counteract detrimental and potentially stigmatizing content.

Our Research

Our research at the Psychology and Communication Technology (PaCT) Lab at Northumbria University investigated over 1.8 million Tweets about potentially stigmatising health conditions and disorders. We captured the prevalence of different types of stigma, and explored how different health conditions are discussed online.

One third of a smaller sample of health-related tweets were coded as potentially stigmatising and there were notable differences in the prevalence of stigma between health conditions. For example, tweets about substance-use disorders typically contained messages suggesting societal danger. Such tweets often included terms like 'crime', 'criminal', 'homelessness', and mentions of financial and familial problems. Previous research has indicated that substance-use disorders are predominantly framed as issues of morality and criminality, rather than as health concerns. Though perceptions towards physical diseases have become more positive in recent years, the negative connotations and societal dangers associated with substance-use disorders suggest that the stigma surrounding addiction remains largely unchanged.

Stigma relating to eating disorders was far more likely to include physical descriptions, or ‘marks’ of health. Words like 'fat' and 'skinny', and targeted comments about someone’s appearance, were the most common in relation to these disorders. This is likely to reinforce prevailing stereotypes about disordered eating. Such reinforcement is concerning, as perpetuated stereotypes might result in treatment biases, especially when an individual doesn't display the typical 'marks' associated with eating disorders, such as being thin, white, and female.

Despite the large presence of health-related stigma on Twitter, approximately one-fifth of all health-related tweets contained what we describe as ‘Anti-Stigma or Advice’. This suggests that public health communicators are actively using social media platforms to counter potentializing stigmatizing narratives online, although there is a lot of work still to do.

Automatically Detecting Stigma Online

Artificial intelligence holds the promise of real-time detection of specific language features, making it especially valuable for identifying harmful speech online. We compared human coding of tweets with natural-language processing approaches to assess the suitability of automatic detection for picking up on health-related stigma. Our findings suggest that automated tools, especially machine learning approaches, could be instrumental in identifying health-related stigma on a broader scale. Such methods could be valuable in gauging the effectiveness of efforts to diminish online stigma on popular platforms such as Twitter (X) and Facebook.

Machine-learning techniques have not yet been utilized for pinpointing health-related stigma on the internet. The consistency between our human coding and automated analysis highlights the potential for digital innovation to help combat the high prevalence of health-related stigma online.