Natural Language Processing for Identifying Patients With Inflammatory Bowel Disease on Twitter and Learning From Their Personal Experience

Abstract

This study provides a framework for identifying patients with Inflammatory Bowel Disease (IBD) on Twitter and learning from their personal experiences. First, we built a user classifier that distinguishes IBD patients from other Twitter users. We constructed classification features from the user's behavior on Twitter, the content of their tweets, and their social network. We compared the performances of five algorithms within two classification approaches. One classified each tweet and deduced the user's class from their tweet-level classification. The other aggregated tweet-level features to user-level features and then classified the users themselves. Both approaches showed promising classification results. Then, we used the classifier to analyze patients' tweets related to health and nutrition. We identified frequently mentioned lifestyles and the patients' sentiments toward them. The findings correlated with what is known about suitable nutrition for IBD. The methods can be adapted to other diseases and enhance medical research regarding chronic conditions.