Socioeconomic and clinical factors explaining the risk of unstructured antiretroviral therapy interruptions among Kenyan adult patients.

A cross-sectional study was conducted to assess the extent of unstructured HIV treatment interruptions (TIs) and investigate the effects of socioeconomic, socio-demographic, HIV treatment-related and clinical factors on the magnitude and rate of the same among adult patients at a Kenyan regional referral center. Four hundred and twenty-one adult patients actively receiving antiretroviral therapy at Nyeri County Referral Hospital since 2003 were randomly selected to complete a health survey questionnaire. Electronic records were used to obtain their HIV treatment utilization history. The marginal effects of selected determinants on prevalence and rate of TI were assessed by fitting multiple Poisson log-linear regression models. In total, 392 patients participated in the study. HIV TI was prevalent with 64.5% having had at least one TI of 3 months or more during treatment. The risk of TI was significantly higher in those longer on treatment (prevalence ratio = 1.2, 95% confidence interval [CI] 1.12-1.28). Greater risk of TI was also associated with lower income (prevalence rate ratio [PRR] = 0.9, 95% CI 0.83-1.00), low medication adherence (PRR = 0.3, 95% CI 0.13-0.72), inconsistent treatment engagement (PRR = 0.4, 95% CI 0.19-0.75) and, contrarily, fewer adverse drug reactions (PRR = 0.9, 95% CI 0.90-0.97). Unstructured HIV TIs appear to be fairly common at the study site. The results suggest that efforts to minimize HIV TI could benefit from treatment-continuity monitoring strategies that target the high-risk sub-samples identified.