Uptake of WHO Recommendations for First- Line Antiretroviral Therapy in Kenya, Uganda, and Zambia

INTRODUCTION: Antiretroviral therapy (ART) guidelines were significantly changed by the World Health Organization in 2010. It is largely unknown to what extent these guidelines were adopted into clinical practice. METHODS: This was a retrospective observational analysis of first-line ART regimens in a sample of health facilities providing ART in Kenya Uganda and Zambia between 2007-2008 and 2011-2012. Data were analyzed for changes in regimen over time and assessed for key patient- and facility-level determinants of tenofovir (TDF) utilization in Kenya and Uganda using a mixed effects model. RESULTS: Data were obtained from 29507 patients from 146 facilities. The overall percentage of patients initiated on TDF-based therapy increased between 2007-2008 and 2011-2012 from 3% to 37% in Kenya 2% to 34% in Uganda and 64% to 87% in Zambia. A simultaneous decrease in stavudine (d4T) utilization was also noted but its use was not eliminated and there remained significant variation in facility prescribing patterns. For patients initiating ART in 2011-2012 we found increased odds of TDF use with more advanced disease at initiation in both Kenya (odds ratio [OR]: 2.78; 95% confidence interval [CI]: 1.73-4.48) and Uganda (OR: 2.15; 95% CI: 1.46-3.17). Having a CD4 test performed at initiation was also a significant predictor in Uganda (OR: 1.43; 95% CI: 1.16-1.76). No facility-level determinants of TDF utilization were seen in Kenya but private facilities (OR: 2.86; 95% CI: 1.45-5.66) and those employing a doctor (OR: 2.86; 95% CI: 1.48-5.51) were more likely to initiate patients on TDF in Uganda. DISCUSSION: d4T-based ART has largely been phased out over the study period. However significant in-country and cross-country variation exists. Among the most recently initiated patients those with more advanced disease at initiation were most likely to start TDF-based treatment. No facility-level determinants were consistent across countries to explain the observed facility-level variation.