Author's School

Brown School

Author's Department

Social Work

Language

English (en)

Date of Award

7-18-2024

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Chair and Committee

Fred Ssewamala

Abstract

Globally, 1.8 million adolescents are living with HIV (ALHIV), of whom 89% (1.5 million) are in Sub-Saharan Africa (SSA) (UNAIDS, 2021b). Unfortunately, despite advances in prevention—such as the prevention of mother to child transmission (PMTCT) and pre-exposure prophylaxis—new infections continue to occur. For example, in 2021 alone, the number of new HIV infections among adolescents aged 10 to 19 was 160,000 cases. Moreover, the rate of HIV-related deaths among adolescents has only declined by 10% in the last decade, which is not fast enough to meet the 2030 global targets. Despite antiretroviral therapy (ART) playing a crucial role in controlling HIV, ALHIV have poor ART adherence and subsequently face low levels of viral suppression, which calls for more efforts towards reversing the trends among ALHIV. However, the ALHIV (10 – 19 year) have received less attention, evidenced by the relatively limited research on interventions to improve outcomes in this group. Guided by a combination of behavioral theories, including asset theory, health belief model, and the socioecological model, this dissertation addressed three research aims. First, the dissertation aimed to compare the performance of three adherence measures —self-reports, pill counts, and electronic adherence measured using Wisepill devices—in monitoring ART adherence and predicting viral suppression (Aim 1). Secondly, it aimed to develop and validate a model to predict the risk of virologic failures (Aim 2). Finally, this dissertation aimed to examine the pathways through which addressing household economic insecurity using an economic empowerment intervention might influence ART adherence among ALHIV in Uganda by assessing the mediating effects of adolescent social transition, barriers to medical care, and adherence self-efficacy. The dissertation used data from 702 ALHIV enrolled in a National Institutes of Health-funded longitudinal two-group (1:1) cluster-randomized clinical trial, the Suubi+Adherence study (R01HD074949) that tested the impact of an economic empowerment (EE) intervention—comprising incentivized youth savings accounts, financial literacy training, and micro-enterprise workshops—among ALHIV in Uganda. To be eligible, participants had to meet the following criteria 1) aged 10 to 16 years; 2) received medical confirmation of their HIV-positive status; 3) awareness of their HIV status; 4) living within a family setting, as opposed to institutionalized care where distinctive characteristics and requirements may vary; 5) being registered and receiving antiretroviral therapy (ART) from one of the selected collaborating clinics involved in the study. To address Aim 1, baseline data from the Suubi+adherence study was used. To begin, the performance of the three adherence measures (self-reports, pill counts, and Wisepill) was compared using the Kappa and agreement coefficient statistics to determine their concordance. This was followed by the Bland-Altman analysis to determine the bias between the measures. In addition, the sensitivity, specificity, and area under the receiver-operator characteristic (ROC) curve were determined for each adherence measure to ascertain how accurately the adherence measures predicted viral suppression (<200 copies/mL). Finally, separate covariate-adjusted multilevel logistic regression models were fitted to determine the association between each adherence measure and viral suppression. Aim 2 relied on baseline data, as well. Specifically, for this aim (2), guided by theory, I used selected sociodemographic, behavioral, psychological, economic, and treatment-related factors to develop and validate a model to predict virologic failure (defined as having a viral load of  200 copies/mL. Least absolute shrinkage and selection operator (Lasso) regression using 10-fold cross-validation with bootstrapping was used to select the predictors for the final model. Model performance was assessed by determining the discrimination using the c-statistic and calibration by drawing a calibration plot. To address Aim 3, longitudinal data collected at 3 time points—baseline, and years six and seven post intervention initiation—was used to determine the long-term direct and indirect effects of the EE intervention on ART adherence. Sequential mediation via structural equation modeling in Mplus software was used to address Aim 3. Results from self-reported adherence under Aim 1 indicate that at baseline 73% of ALHIV reported good ART Adherence—defined as adherence 90% (Byrd et al., 2019; Ministry of Health, 2022). At the same measurement point (baseline), 67.1% reported virological suppression. Taken as a whole, these results point to poor adherence by ALHIV who were enrolled in Suubi+Adherence at baseline. There was considerable disagreement between the adherence measures employed, which were: self-reports, pill counts, and electronic adherence using Wisepill devices, with kappa values below 0.10 and covariate-adjusted AUC slightly above 0.60. When compared, the three adherence measures exhibited more disagreement when the mean adherence was low, but the agreement improved with increasing mean adherence. Only self-reported adherence had a significant relationship with viral suppression, OR = 2.16 (95% CI: 1.25 – 3.81), p = 0.006. Regarding the risk prediction model (Aim 2), a model with 24 predictors was developed using a lambda value of 0.0071304. Variables retained in the model included participants' age, sex, work status, stigma, depressive symptoms, adherence self-efficacy, HIV knowledge, duration with HIV, time spent on ART, communication with the caregiver, family cohesion, social support, orphanhood status, number of people in the household, HIV disclosure, years spent at the current residence, and household asset ownership. The model predicted virologic failure with an AUC of 73.8 (95% CI: 68.3 – 78.0) and an almost perfect calibration of 0.985. Finally, in Aim 3, the results showed a significant direct effect of the intervention on ART adherence was observed,  = 0.066 (0.007, 0.125), p = 0.028. In addition, the intervention had a significant total indirect effect on ART adherence,  = -0.028 (-0.054, -0.002), p = 0.033. A closer examination of the pathways revealed that the indirect effect was mediated through barriers to medical care,  = -0.030 (-0.057, -0.004), p = 0.026. Specifically, the intervention was efficacious in reducing the barriers to medical care,  = -0.178 (-0.259, -0.096), p < 0.001. However, paradoxically, participants experiencing more barriers to medical care also reported higher levels of ART adherence,  = 0.170 (0.036, 0.304), p = 0.013. Overall, this dissertation adds to our understanding of the challenges and the strategies to enhance ART adherence and viral suppression among ALHIV in Uganda and similar low-resourced high-HIV burden settings. In summary, self-reported adherence was found to significantly predict viral suppression. In addition, a model that can accurately predict the risk of virologic failure among ALHIV was developed using the participant's socioeconomic, behavioral, psychological, and economic information. More importantly, the dissertation showed that providing the ALHIV and their families with financial resources through a family economic empowerment intervention improved their ART adherence and highlighted alternative pathways influencing this relationship. These findings point to a pressing need for more efforts to improve ART adherence among ALHIV and the need to incorporate interventions that address poverty in the programs aimed at controlling HIV.

Download

Included in

Public Health Commons

Share

COinS