From online engagement to offline health outcomes—analyzing user behaviors in online health communities

The prevalence of internet provides a great lens to study “Information-Technology-People” interactions. The global healthcare system is undergoing a radical shift as more and more patients use internet to take control of their personal well-being. Online health communities (OHCs) have become a major resource for people with similar health concerns to get together. Taking advantage of text mining, regression models, and machine learning, I investigate how to foster continued online engagement for users in OHCs, and how to better support OHC users’ pursuits of better health outcomes in their offline lives. I first examine how the user’s first experience seeking support affects their early-stage engagement in OHCs. Results suggest that both the quantity and quality of the received social support are significant predictors of new users’ continued engagement. I also propose a general and flexible framework to predict user posting activities. The prediction of user online engagement finds that discussions about companionship and emotional support at the top of OHC managers’ recommendation lists can retain user posting activities in OHCs. Last but not the least, I study the connection of user online activities and offline health outcomes. When promoting user online participation, a useful practice is to recruit a core group of active users (i.e., seed users) to create new content and encourage user participation. Findings suggest that seed users who have successful experience in quitting smoking can better support health outcomes. The outcomes of our findings help better understand the dynamics of OHCs, promote successful online communities, and improve individual health via interventions.