Exploring Self-reported Adherence Measures to Screen for Elevated HIV Viral Load in Adolescents: A South African Cohort Study

Abstract The timely identification of ART non-adherence among adolescents living with HIV presents a significant challenge, particularly in resource-limited settings where virologic monitoring is suboptimal. Using South African adolescent cohort data (N = 933, mean age 13.6 ± 2.89 years, 55.1% female, follow-up = 2014–2018), we examined the association between elevated viral load (VL ≥ 1000 copies/mL) and seven self-reported adherence measures on missed doses, and clinic appointments –with varying recall timeframes. The best performing measures, which were significantly associated with elevated viral load in covariate-adjusted models are: any missed dose –past 3 days (sensitivity = 91.6% [95%CI: 90.3–92.8], positive predictive value (PPV) = 78.8% [95%CI: 77.2–80.4]), –past week (sensitivity = 87% [95%CI: 85.4–88.6], PPV = 78.2% [95%CI: 76.5–79.9]), –past month (sensitivity = 79.5% [95%CI: 77.5–81.4], PPV = 78.2% [95%CI: 76.4–79.9]), any past-month days missed (sensitivity = 86.7% [95%CI: 85.1–88.3], PPV = 77.9% [95%CI:76.2–79.6]), and any missed clinic appointment (sensitivity = 88.3% [95%CI: 86.8–89.8], PPV = 78.4% [95%CI: 76.8–79.9]). Combining the three best performing measures missed dose –past 3 days, –past week , and any past-year missed clinic appointment increased sensitivity to 96.4% while maintaining a PPV of about 78%. The discriminatory power of simple and easy-to-administer self-reported adherence measures in detecting elevated viral load warrants consideration in resource-limited settings and may contribute to the aims of the new Global Alliance to End AIDS in children and adolescents by 2030.