Migration adversely affects antiretroviral adherence in a population-based cohort of HIV/AIDS patients

Migration among persons with HIV/AIDS is common; however, it is not clear how migration relates to antiretroviral adherence, a key determinant of treatment efficacy. Therefore, our objective was to determine the scale of regional migration and its association with adherence patterns over time among HIV-infected individuals in British Columbia (BC), Canada. Participants initiated HAART in August 1996-November 2004, and were followed until November 2005. Adherence was defined as the number of days worth of antiretrovirals dispensed divided by the number of days of follow-up (expressed as a percentage), and considered a binary time-dependent outcome: 'non-adherence' (less than 95%) versus 'adherence' (95% or more). Migration was calculated as the cumulative number of times a patient's residential address changed during the course of treatment, and treated as a time-dependent variable. Non-linear mixed-effects models were used to estimate the association between migration and adherence over time. All analyses were adjusted for relevant fixed and time-dependent variables. A total of 2421 participants were followed during the study period. Descriptive analysis demonstrated high stability in adherence over time, with more than 55% of patients moving at least once during the course of their treatment. We observed that those individuals migrating at least 3 times were 1.79 times more likely to be in the 'non-adherence' group than individuals who did not migrate. Our results demonstrate that migration in BC is not homogeneous across subpopulations. These results suggest that proactive strategies are needed to ensure that antiretroviral therapy remains available on a continued basis to highly migrant populations.