Predictors of lost-to-follow-up amongst adolescents on antiretroviral therapy in an urban setting in Botswana

Abstract

There has been a recent increase in the proportion of adolescents living with HIV being enrolled on anti-retroviral therapy (ART) in Botswana, with more than 90% accessing this life saving therapy. A significant proportion become lost-to-follow-up (LTFU) from the ART care, reversing the initial gains attained. The factors associated with lost-to-follow-up in this vulnerable HIV positive adolescent population in Botswana are generally unknown, as most studies on this subject are on the adult population. This study investigated the rate and factors associated with lost-to-follow up amongst middle and late adolescents (15-19 years) on ART at Botswana-Baylor clinic in Gaborone, Botswana. The center provides comprehensive HIV prevention, treatment and psychological services to children and adolescents living with HIV. The research intended to assist HIV program managers to retain adolescents in care, which is vital in reducing morbidity, mortality and also new HIV infections amongst this population. The research design was an observational comparison between those LTFU on ART and those retained on ART. Simple random sampling was used and the study had 133 of those LTFU and 133 of those not LTFU. Data on the risk factors were retrospectively extracted from patient records stored in a data-base at the clinic. The data were analyzed using Epi-info 7 statistical software to determine if there were any statistically significant factors associated with lost-to-follow-up amongst adolescents on ART. The study involved a vulnerable population who are HIV positive and a proportion who were less than 18 years. However minimal harm was expected as this was a record review and data was anonymized prior to analysis. Ethical clearance was given by the Botswana Baylor Clinic Ethics Committee and the University of Western Cape Biomedical Ethics Research Committee. The LTFU rate in the clinic amongst adolescents (15-19 years) was 4.6%. Using bivariate analysis there were significant associations between LTFU and the following variables: CD4 count (prior to LTFU), detectable viral load, past history of Tuberculosis, PI-based ART regimen and suboptimal adherence. However after multivariate analysis, only detectable viral load and suboptimal adherence were independent predictors of lost-to-follow-up. Middle and older adolescent patients lost-to-follow-up were 5 times more likely to have a detectable viral load and 4 times more likely to have suboptimal adherence than those not lost-to-follow-up. The findings in our study will assist clinicians at the Baylor-Botswana clinic identify adolescent patients needing extra support to be retained in care and improve clinical outcomes. More studies of this type are needed in solely public sector clinics in Botswana and regionally, as there is a sizeable population of people living with HIV in middle and late adolescents in which lost-to-follow up appears to be greatest.