oa Predicting Weight Loss in Online Social Media

Obesity is a major public health problem which adversely impacts mortality and quality of life, also it is associated with significantly increased risk of more than 20 chronic diseases and health conditions [Thiese et al., 2015]. According to the World Health Organization, the prevalence of obesity has nearly doubled within the last 30 years, which has led to an impressive estimate of 402 million obese people worldwide. The etiology of obesity is complex and encompasses a wide range of genetic, physiological, behavioral, cultural, social and environmental factors. Before the appearance of online social media, factors associated with obesity could only be measured in the real-world. However, with social environments moving online, the escalating number of interactions in online communities has created a great opportunity to study huge amount of user-generated content that comprises various topics related to obesity. These topics include people's experiences, recommendations and feedback about certain medications, medical procedures, diets or exercises, and emotional support in the form of encouragement, sympathy, and success stories. Analyzing and exploring these topics can give health practitioners various insights into community dynamics, such as the effects of online social support on community members and the profile of influential members, as well as provide important information to design effective online health intervention strategies [Bennett and Glasgow, 2009]. Online communities can be used to understand and promote health behavior as well as disseminate health innovations. But still little is known about how these communities can help enhance health conditions, such as weight loss. Advantages of online communities include access to many peers with the same health concerns, convenient communication spanning geographic distances, and anonymity (if desired) for discussion of sensitive issues [Hwang et al., 2010]. In this study, we are interested in answering the research question of whether it is possible to use online user generated content to predict success or failure of weight loss and weight maintenance. Concretely, this work investigates if there is a relation between online users behavior and the likelihood of them losing weight in an online Reddit weight loss community, namely “loseit”. Data collected include posts, comments and other metadata (i.e., timestamp, user name, number of upvotes) from August 2010 to November 2014.

In total, we obtained 70, 949 posts and 922, 245 comments. These data were generated by 107,886 unique users. The community encourages users to post their weight and progress along time, sharing experiences. Our aim is to show that social media can be exploited to help health practitioners to understand obesity dynamics, delivering more personalized treatments and improving patient-centered care. In this direction, our findings aid health practitioners to design early warning systems or effective online health interventions strategies that can be incorporated into social media platforms and lead to more effective treatment. These systems may provide great benefit to patients, for example, by integrating recommendation systems that can help users make important decisions, such as choosing the right type of diet or exercise for their obesity condition. We believe that exploring novel approaches to understand and address obesity is crucial to realize Qatar's National Vision 2030 of a healthy population.