Abstract / Summary
Background: There is a need for evidence as to whether substance use (by type) has a different relation with (1) loss of housing and (2) regaining housing among people who use substances. Methods: Seven cohorts, all part of the HIV SUCCESS consortium, contributed data between 2005 and 2024 to estimate time to change in housing status in longitudinal models with six-month intervals. To examine associations between substance use and loss of housing, we first included all participants at their first housed visit and estimated time to housing loss. To examine associations with rehousing, we included participants who were unhoused at any point prior to their final study visit and estimated time to regain housing. Housing status and substance use were self-reported. For all analyses we used adjusted time-updated Cox models with a separate model for each substance. Cohort-specific results were then pooled using a random effects meta-analysis approach with inverse variance weights. Results: Among 4931 housed participants followed for an average of 4.2 years, current use of methamphetamine/amphetamines (hazard ratio [HR]=1.81, 95% confidence interval [CI]:1.47, 2.23), opioids (HR=1.68, 95%CI:1.39, 2.02) and cocaine (HR=1.39, 95%CI:1.14, 1.69), were associated with a significantly faster rate of becoming unhoused, once adjusted for age, sex, race/ethnicity and year of study enrollment. Among 1901 unhoused participants, followed for an average of 1.4 years, methamphetamine/amphetamines (HR=0.80, 95%CI:0.70, 0.91) opioids (HR=0.85, 95%CI:0.76, 0.95) and cocaine (HR=0.83, 95%CI:0.75, 0.92) were associated with regaining housing at a slower rate. Conclusions: Opioid, methamphetamine/amphetamine, and cocaine use are all associated with adverse housing outcomes.