Tweets are not always supportive of patients with mental disorders

Abstract

Introduction: The Internet and the social media offer great potential for empowering patients, who can obtain updated information, peer support, and Internet-based treatment [1-3]. However, initial research and feedback from some patients have suggested that it is unclear to what degree the social media, including Twitter, overall are beneficial to patients [4, 5]. We therefore examined to what degree tweet content could be perceived of as negative or non-supportive of various disorders. Methods: We used the Twitter search engine to sample 150 random tweets containing either #schizophrenia, #bipolar, or #breastcancer. We classified the tweets as supportive, non-supportive or as a misuse of the disease term. As we examined the tweets, we became particularly interested in their affective content, thinking that this was of importance to how patients would react to the tweets. We therefore sampled 1000 random tweets relating to the same three disorders, and in addition, 1000 random tweets including #depression, #HIV, #AIDS, #COPD, #diabetes, and #Alzheimer, respectively (in total 9000 tweets). We subsequently used the computer program LIWC [6] to analyze the tweets. We focused specifically on the affective categories of ‘negative emotion’, ‘positive emotion’, ‘anxiety,’ and ‘sadness’. Results: Nearly all of the breast cancer tweets were supportive, while a majority of the tweets on schizophrenia and bipolar disorder were non-supportive or in the misuse of term category. Interestingly, the tweets including the terms #depression and #bipolar ranked highest in the categories of ‘negative emotion’, ‘anxiety’, and ‘sadness’, with #schizophrenia in third (‘anxiety’) and fourth place (‘sadness’ and ‘negative emotion’). #AIDS came in third for ‘negative emotion’ and diabetes third for ‘sadness’. For the ‘positive emotion’ category, #breast cancer ranked highest, followed by #depression, #HIV, and #bipolar, suggesting that there are also many tweets with positive affective content related to depression and bipolar disorder. Discussion, limitations, and lessons learned: While this is an initial, relatively small study, it suggests that social media such as Twitter may not consistently be a source of support or comfort to all patient groups, and patients and clinicians should be aware of this.