2026 Theses Doctoral

An examination of the effects of loneliness on substance use and HIV outcomes

Marziali, Megan Elyssa

Drug overdose is a pressing public health concern in the United States and Canada. Loneliness, defined as a perceived lack of social connection, can harm both physical and mental health; loneliness may also be associated with drug use. In addition, loneliness may increase the risk of adverse HIV prevention outcomes through a mechanism involving higher-risk sexual behavior to manage loneliness.

Throughout this dissertation, and informed by a systematic scoping review, I theorized that loneliness is associated with substance use outcomes, including drug overdose, and HIV outcomes. I began by explaining loneliness as a health priority and delineated the theoretical mechanism through which loneliness may impact both substance use outcomes and HIV outcomes. In Chapter 2, I presented evidence from a systematic scoping review and outlined the state of the literature concerning loneliness and substance use (including drug use, alcohol use, and substance use).

These findings supported the development of a conceptual model to guide my empirical aims. In Chapter 3, the first empirical aim, I examined the effect of persistent loneliness (operationalized as the proportion of study visits where participants reported loneliness) on drug use and HIV risk, among a cohort of sexual and gender-diverse groups in British Columbia, Canada. I also explored the role of loneliness in the risk of drug overdose. Chapter 4 involved exploring the effect of loneliness on drug use, including polydrug use, and HIV prevention outcomes, among a cohort of Black sexual and gender-diverse groups in Chicago, Illinois. This dissertation concludes with a discussion of the main findings, strengths, and limitations of each aim, and directions for future research.

In brief, findings from Chapter 2 informed by qualitative and quantitative literature suggested a relationship between loneliness and substance use. Few studies examined specific classifications of drugs (i.e., opioids, stimulants) or overdose; these studies were all limited by cross-sectional data. Evidence Chapter 3 demonstrated an association between loneliness and drug use, specifically opioid, stimulant, and polysubstance use; evidence suggested a relationship between loneliness and drug overdose. There was support for a relationship between loneliness and HIV risk in sensitivity analyses. The second empirical aim, Chapter 4, similarly identified a relationship between loneliness and stimulant use; loneliness was further associated with polydrug use, and data supported a linear relationship between loneliness and the number of drugs used. The data did not support a relationship between loneliness and HIV prevention outcomes.

Overall, this dissertation applied robust epidemiologic analyses to examine the relationship between loneliness, drug use, and HIV outcomes. Future directions should include research to understand interventions that would alleviate feelings of loneliness among sexual and gender-diverse groups.