In this paper, we investigate whether social support exchanged in an online healthcare community benefits patients’ mental health. We propose a nonhomogeneous Partially Observed Markov Decision Process (POMDP) model to examine the latent health outcomes for online health community members. The transition between different health states is modeled as a probability function that incorporates different forms of social support that patients exchange via discussion board posts. We find that patients benefit from learning from others and that their participation in the online community helps them to improve their health and to better engage in their disease self-management process. Our results also reveal differences in the influence of various forms of social support exchanged on the evolution of patients’ health conditions. We find evidence that informational support is the most prevalent type in the online healthcare community. Nevertheless, emotional support plays the most significant role in helping patients move to a healthier state. Overall, the influence of social support is found to vary depending on patients’ health conditions. Finally, we demonstrate that our proposed POMDP model can provide accurate predictions for patients’ health states and can be used to recover missing or unavailable information on patients’ health conditions.

T he United States has the highest rate of obesity in the world. To help address this problem, social support is gaining credibility as a powerful tool to facilitate weight loss because it can affect people's behavior. Although social support has long been recognized for its effectiveness in promoting health, we argue, in this study, that social support may not always lead to good outcomes. Specifically, we differentiate between support providers and support seekers, and examine whether providing and receiving support affect individuals' weight-loss outcomes differently. By analyzing a group of individuals participating in an online weight-loss community, we show that providing and receiving support does affect weight-loss outcomes in different ways. First, the influences are dynamic. Second, while providing support is positively associated with weight-loss progress, receiving support could hinder weight-loss outcome for a person with high self-efficacy in weight-loss progress. Third, by categorizing social support into different types, we find evidence suggesting that the match between needed and received social support type also influences individuals' performance in the weight-loss process. Furthermore, mismatches of social support could negatively affect weight-loss outcomes. These findings have implications for maximizing the usefulness of social support for participants in the online environment as well as for clinicians who refer individuals to online weight-loss communities and for those who design them.

Disaster relief organizations increasingly engage in social conversations to inform social media users about activities such as evacuation routes and aid distribution. Concurrently, users share information such as the demand for aid, willingness to donate and availability to volunteer through social conversations with relief organizations. We investigate the effect of this information exchange on social engagement during disaster preparedness, response, and recovery. We propose that the effect of information on social engagement increases from preparedness to response and decreases from response to recovery. Some of the information exchanged in social conversations is actionable as well. We propose, however, that the effect of actionable information reaches its lowest point during disaster response. To test our theory, we use Facebook data from five benchmark organizations that responded to Hurricane Sandy in 2012. We analyze all of the organizations’ posts and users’ comments during a three‐week period before, during and after Hurricane Sandy. Our findings support our theory. Furthermore, we identify an opportunity for relief organizations to improve their use of social media for disaster management. While relief organizations focus on informing disaster victims about aid distribution, most users are asking about how they as individuals can donate or volunteer. Thus, besides posting information directed to victims, organizations should post more information targeting potential donors and volunteers.