Non-adherence to antiretroviral treatment and unplanned treatment interruption among people living with HIV/AIDS in Cameroon: Individual and healthcare supply-related factors.

In low-income countries, health system deficiencies may undermine treatment continuity and adherence to antiretroviral therapy (ART) that are crucial for the success of large-scale public ART programs. In addition to examining the effects of individual characteristics, on non-adherence to ART and treatment interruption behaviors - i.e. treatment interruption for more than 2 consecutive days during the previous 4 weeks, this study aims to extend our knowledge on the role played by healthcare supply-related characteristics in shaping these two treatment outcomes. These effects are examined using multilevel logistic models applied to a sub-sample of 2381 ART-treated patients followed-up in 27 treatment centers in Cameroon (ANRS-EVAL survey, 2006-2007). Multivariate models show that factors common to both non-adherence and treatment interruption include binge drinking (at the individual-level) and large hospital size (at the healthcare supply-level). Among the individual factors, financial difficulties of paying for HIV-care are the major correlates of treatment interruption [Adjusted Odds Ratio (AOR) 95% confidence interval (CI) = 11.73(7.24-19.00)]. By contrast, individual factors associated with an increased risk of non-adherence include: having a main partner but not living in a couple compared to patients living in a couple [AOR(95%CI) = 1.51(1.14-2.01)]; experience of discrimination in the family environment [AOR(95%CI) = 1.74(1.14-2.65)]; a lack of regular meals [AOR(95%CI) = 1.93(1.44-2.57)], and switching antiretroviral drugs (ARV) regimen [AOR(95%CI) = 1.36(1.08-1.70)]. At healthcare facility-level, the main correlate of ART interruption was antiretroviral stock-outs [AOR(95%CI) = 1.76(1.01-3.32)] whereas the lack of psychosocial support from specialized staff and lack of task-shifting to nurses in medical follow-up were both associated with a higher-risk of non-adherence [respective AOR (95%CI) = 2.81(1.13-6.95) and 1.51(1.00-3.40)]. Results reveal different patterns of factors for non-adherence and treatment interruption behaviors. They also suggest that psychosocial support interventions targeted at the individual patient-level will not be sufficient to achieve favorable treatment outcomes if not combined with interventions focused on strengthening health systems, including appropriate drug supplies and human resources policies, as well as sustainable and equitable financing mechanisms.