A Study on the Source and Truthfulness of User-Generated Content of Health Information on Social Media: A Case for Colorectal Cancer on Twitter

Abstract

In this tech-savvy era, people consult their well-being to Dr. Google and hashtags (#). They can almost self-diagnose themselves using a search engine, and micro-posts on Twitter can predict influenza outbreaks. If any misinformation is found in this vein, its impact can be fatal. This necessitates an in-depth analysis of consumer health information online, with a particular focus on its quality.

This study analyzes the source and truthfulness of user-generated content of health information on Twitter, one of the most popular social networking services since its inception in 2006. Particularly, it explores who says what on Twitter about colorectal cancer, the third leading cause of cancer deaths worldwide. Using Twitters Public Streaming API, all tweets containing colorectal cancer-related keywords were collected from August 1st, 2014 to October 31st, 2014. The tweets were classified by their content and user, and the quality of informative content was assessed. Notable findings of this study include: 1) the tweets included approximately twice more informative content (65.20%) than communicative content

2) a significant number of tweets (90.62%) were written by individual users

3) though smaller in tweet count, organizations showed a tendency to tweet information more

4) most tweets (73.40%) included URLs to external websites such as news articles and medical information resources

and finally 5) sampled informative tweets were mostly medically correct information (84.52%), reviewed by board-certified physicians.

This study makes the following contributions. First, it has quantitatively analyzed the content and source of colorectal cancer information delivered on Twitter, with a focus on assessing its quality. Second, it has shown that Twitter can potentially aid social health information seeking, narrowing the gap between doctors and patients. Third, this study has suggested that more popular retweets may filter medically correct information, hinting a possibility for collective intelligence in information sharing on Twitter. Yet one needs much caution subscribing to health information online, as its quality is hard to assess without a trained eye. Lastly, this study has confirmed the need for future system features to be equipped with a sanity check on the quality of any health information.

In sum, this study concludes that using Twitter for seeking and sharing health information is both an opportunity and a risk. While a well-structured piece of health information in tweets may foster better informed public, it is often vague whether tweets, limited to 140 characters, are telling accurate, up-to-date health information. Future research opportunities are bountiful, such as detecting fraud in health information and automatically identifying information sources.