A Latent Transition Analysis of Longitudinal Patterns of Substance Use Among Vulnerable Persons Living with HIV/AIDS

Introduction: Substance use problems are common among people living with HIV/AIDS (PLHIV). And substance use often interferes with antiretroviral (ART) adherence. Substance use can also strain supportive social relationships important to health and well-being.

Method: Drug use behaviors and other socio-environmental characteristics were assessed in a sample of 383 PLHIV who resided in Baltimore, MD, had used hard drugs such as heroin or cocaine in their lifetimes and were taking ART. Longitudinal patterns of substance use were examined using Latent Transition Analysis (LTA). Four survey items assessed substance use (heroin or cocaine in the previous 6 months, or daily or binge drinking in the previous 30 days). First, we ran latent class analyses (LCA) at both timepoints and determined a 3-class model had optimal fit at each time-point. Partial measurement invariance was obtained by releasing the constraint on the cocaine variable. The LTA was run and initial status and transition probabilities were regressed on HIV-related stigma.

Results: The sample was primarily African American/Black (84.5%), earned less than $1,000 per month (81.4%), more than half (57.4%) was male, and currently used substances (58.9%). Class one, labeled the polysubstance use group, had particularly high levels of binge drinking (.95). The second class had high cocaine (.87) and heroin (.64) use, while the third class used little if any substances. While the majority of respondents remained in their respective classes over time, 19% of the polysubstance use class and 28% of the heroin and cocaine class moved to the low use class at follow-up. Participants who perceived fewer network members available to provide them health advice or social participation (enjoyable social activity) had more than 20 times the odds (OR = 20.17, p < .05 and OR = 29.20, p< .05, respectively) of switching from the polysubstance class to the heroin and cocaine class than to the reference (low use) class.

Discussion: The results indicate high levels of hard drug use and hazardous drinking in this sample of PLHIV. The results suggest that having less health informational or enjoyable social participation support were at greater risk of transitioning to a more serious substance use class marked by a higher proportion of heroin users. These results emphasize the importance of specific forms of network support that can provide protective functions for reducing substance use longitudinally. Interventions to promote PLHIVs’ engagement with supportive persons knowledgeable about HIV and ART and in non-drug-related enjoyable activities are warranted for reducing their (severity of) substance use.