What question did this study set out to answer?

The study aims to identify patterns of long-term adherence to antiretroviral therapy and their impact on viral non-suppression among individuals with HIV.

June 13, 2026Open Access

Longitudinal adherence trajectories measured by electronic dose monitoring and viral non-suppression in a South African HIV cohort study

Key Points

The study aims to identify patterns of long-term adherence to antiretroviral therapy and their impact on viral non-suppression among individuals with HIV.
Analyzed data from a prospective cohort of people with HIV on tenofovir-based ART in Cape Town, South Africa.
Measured daily adherence using electronic dose monitoring over 12 months, summarizing as monthly counts.
Applied group-based trajectory modeling to determine distinct adherence patterns and assessed their association with time to viral non-suppression.
Identified 5 adherence trajectories: persistently high (49.6%), gradually declining (22.7%), delayed declining (9.6%), rapidly declining (8.0%), and consistently suboptimal adherence (10.1%).
Rapidly declining and delayed declining trajectories had a higher incidence of viral non-suppression compared to persistently high and gradually declining trajectories (log-rank p < 0.001).
Younger individuals were more likely to fall into lower-adherence trajectories, while baseline psychosocial characteristics showed no significant differences across groups.

Abstract

Abstract Background Sustained adherence to antiretroviral therapy (ART) is essential for maintaining viral suppression among people with HIV (PWH), yet adherence fluctuates over time and is imperfectly captured by routine viral load monitoring in many settings where measurements are often infrequent. We aimed to identify longitudinal ART adherence trajectories using electronic dose monitoring (EDM) and assess their association with viral non-suppression in a South African cohort. Methods We analyzed data from a prospective cohort of PWH in Cape Town, South Africa on tenofovir-based ART. Daily adherence was measured using EDM over 12 months and summarized as monthly counts of dose events. Group-based trajectory modelling was applied to monthly adherence proportions to identify distinct adherence patterns. Associations between trajectory membership and time to viral non-suppression (HIV-1 RNA > 50 copies/mL) were assessed using Kaplan–Meier methods and log-rank tests. Results Among 238 PWH, 5 adherence trajectories were identified: persistently high (49.6%), gradually declining (22.7%), delayed declining (9.6%), rapidly declining (8.0%), and consistently suboptimal adherence (10.1%). Rapidly declining and delayed declining trajectories were associated with earlier and higher incidence of viral non-suppression compared with persistently high and gradually declining trajectories (log-rank p < 0.001). Younger PWH were more likely to be in lower-adherence trajectories; baseline psychosocial characteristics did not differ across groups. Conclusions ART adherence is dynamic and heterogeneous; early declines in electronically measured adherence are associated with increased viral non-suppression risk. Longitudinal adherence trajectories may offer clinically actionable signals before routine viral load testing, informing differentiated HIV care programmes.

Bookmark

View Full Paper