Abstract/Summary

Digital Health Social Networks (DHSNs), otherwise known as online support groups or peerto- peer patient groups, have been in existence since the mid-1980s. However, they have only recently been recognized as important tools in healthcare. This dissertation focuses on superusers, a subset of DHSN participants who create the majority of content, and who are essential to the health and vibrancy of a network. The three essays in this dissertation assess the feasibility of quantitatively identifying superusers though theoretical models rooted in econometrics and graph theory. The data sources are four, long-standing DHSNs. Two of the four DHSNs focus on mental health (depressive disorder and panic disorder), and the remaining two on addictions (problem drinking and smoking cessation). The first essay examines associations between demographic characteristics, indication severity, and posting behaviour. The second investigates whether the distribution frequency in the four DHSNs follows properties of power laws. The third explores the feasibility of applying the Gini coefficient to measure DHSN inequality. This dissertation has two main contributions to theory and practice. The first is that superusers cannot be predicted through demographic or indication-specific characteristics. The second is that graph theory can be used to detect and track superusers in real time. Collectively, the three essays contain unique insights into DHSN utility and function. These insights, and related metrics, can be leveraged by researchers, moderators, managers, and funders to quantify the growth, stagnation, or decline of their networks.