PloS one 2017 09 2212(9) e0184848 doi 10.1371/journal.pone.0184848
Sociodemographic correlates of engagement in human immunodeficiency virus (HIV) care are well studied, however the association with accessing drug resistance testing (DRT) and the development of drug resistance have not been characterized. Between 1996-2014, 11 801 HIV patients accessing therapy in British Columbia were observed longitudinally. A subset of 9456 patients had testable viral load; of these 8398 were linked to census data. Sociodemographic (census tract-level) and clinical (individual-level) correlates of DRT were assessed using multivariable General Estimating Equation logistic regression adjusted odds ratios (aOR). The mean number of tests per patient was 2.1 (Q1-Q3; 0-3). Separately, any drug resistance was determined using IAS-USA (2013) list for 5703 initially treatment naïve patients without baseline resistance; 5175 were census-linked (mean of 1.5 protease-reverse transcriptase sequences/patient, Q1-Q3; 0-2). Correlates of detecting drug resistance in this subset were analyzed using Cox PH regression adjusted hazard ratios (aHR). Our results indicate baseline CD4 <200 cells/μL (aOR: 1.5, 1.3-1.6), nRTI-only baseline regimens (aOR: 1.4, 1.3-1.6), and unknown (therapy initiation before routine pVL in BC) baseline pVL (aOR: 1.8, 1.5-2.1) were among individual-level clinical covariates strongly associated with having accessed DRT; while imperfect adherence (aHR: 2.2, 1.9-2.5), low baseline CD4 count (aHR: 1.9, 1.6-2.3), and high baseline pVL (aHR: 2.0, 1.6-2.6) were associated with a higher likelihood of developing drug resistance. A higher median income (aOR: 0.83, 0.77-0.89) and higher percentage of those with aboriginal ancestry (aOR: 0.85, 0.76-0.95) were census tract-level sociodemographic covariates associated with decreased access to DRT. Similarly, aboriginal ancestry (aHR: 1.2, 1.1-1.5) was associated with development of drug resistance. In conclusion, clinical covariates continue to be the strongest correlates of development of drug resistance and access to DRT for individuals. Regions of high median income and high aboriginal ancestry were weak census-level sociodemographic indicators of reduced DRT uptake, however high aboriginal ancestry was the only sociodemographic indicator for development of drug resistance.