Strengthening understanding of effective adherence strategies for first-line and second-line antiretroviral therapy (ART) in selected rural and urban communities in South Africa

Abstract

South Africa accounts for approximately 20% (4.8 million) of the worldwide population of individuals who are HIV-positive and receiving antiretroviral medication (ART). In 2019, approximately 15%-20% of individuals receiving first-line antiretroviral therapy (ART) and up to 30% of individuals receiving second-line ART in the HIV treatment program in South Africa encountered virological failure. In addition, over 40% of individuals receiving first-line antiretroviral therapy (ART) and up to 20% of individuals receiving second-line ART experienced loss to follow-up (LTFU). While there is a significant amount of research on adherence to antiretroviral therapy (ART), there is still a lack of studies examining the various elements at different levels that influence adherence to treatment and the processes that shape adherence behaviour, specifically in South Africa. Furthermore, there exists a dearth of documented information regarding the efficacy and consequences of the measures presently utilized to enhance adherence among individuals living with HIV (PLHIV) who are undergoing antiretroviral therapy (ART). This thesis employed a multilevel socio-ecological framework to elucidate the risk factors that influence treatment adherence across various levels. Additionally, it conducted a comprehensive evaluation of existing research that examined the impact or impacts of intervention techniques on enhancing treatment adherence. The studies presented in this thesis identified the barriers to and facilitators of adherence for people living with HIV on ART and assessed the impact of different adherence intervention strategies that aimed to promote treatment adherence. This was achieved by examining the five research questions: 1. What is the uptake rate of ART, and the individual-level predictors of virological failure and being lost to follow-up (LTFU) in PLHIV taking ART in urban communities? 2. What are the individual-level factors that predict virological failure, low CD4 count, and retention in care for patients on second-line ART in urban communities? 3. What are the individual, relationship or interpersonal, and community-level factors associated with self-reported adherence, pill count, and virological failure to ART in rural communities? 4. What are the different treatment-taking behaviours and perspectives on adherence to ART between virally suppressed and unsuppressed patients on second-line ART in urban communities? 5. What treatment adherence strategies and interventions have been implemented and evaluated in sub-Saharan Africa for ART, hypertension, and Diabetes Mellitus? vii Chapter 2, a protocol paper, detailed the rationale, study aims, research designs, and methods employed in the studies reported on in this thesis. By adapting a multi-level socio-ecological framework to identify factors existing at various levels (including individual, relationship/interpersonal, and community level factors) and describing their interplay chapter 2 demonstrated how an existing socio-ecological conceptual framework can be used as a tool to provide guidance regarding facilitators and barriers to ART adherence. In the study reported in chapter 3, we described the ART uptake and the individual level predictors of virological failure and being LTFU in PLHIV taking ART in Johannesburg. In this retrospective cohort study, we presented analyses based on the TIER.Net database for a large cohort of HIV- infected adult patients who are taking first-line and second-line ART in Johannesburg, South Africa. TIER.Net is the ART monitoring and evaluation system used by the South African National Department of Health for recording ART patient-level information. Records were reviewed for patients on ART from seven high-volume public health facilities in Johannesburg. Study data included medical records of people with HIV who started ART between 01 April 2004 (the inception of the South African national HIV treatment program in the public health system setting) and 29 February 2020. This cut-off period was chosen to give the cohort patients a minimum of one year to receive their annual standard-of-care viral load test. In this study, factors such as age at ART start, current age, sex, duration on ART, baseline CD4 cell count, and retention in care were analyzed as covariates of outcomes (viral load and LTFU). Of the total study cohort, 95% (n=117 260/123 002) were on a first-line regimen and 5% (n=5 742/123 002) were on a second-line regimen. Most patients (59%, n=72 430/123 002) were initiated on an efavirenz-based, tenofovir disoproxil fumarate-based and emtricitabine-based regimen (fixed-dose combination). 91% (n=76 737/84 252) achieved viral suppression at least once since initiating ART and 59% (n=57 981/98 071) remained in care as at the end of February 2020. Findings from the univariate, multivariable logistic regression analysis and fixed effect model showed that younger patients, male patients, patients with low CD4 cell counts, and patients who were initiated on ART between 2004 and 2010 all had poorer clinical, treatment and retention outcomes, particularly those on second-line ART. While national ART guidelines and efforts to initiate PLHIV on treatment have contributed to a higher uptake of ART over time, much still needs to be done to improve retention in care. Although slight efforts have been made to address similar findings in sub-Saharan Africa, these demographic and clinical characteristics must be considered when designing/implementing treatment support strategies and models to improve treatment outcomes, retention in care, and subsequently treatment failures which lead to switching to more complex ART regimens. viii In the study reported in chapter 4, we aimed to identify individual-level factors that predict virological failure, low CD4 count, and retention in care for patients on second-line ART in Johannesburg. In this retrospective cohort study, we conducted analyses of secondary data that was exported from the TIER.Net database. Variables extracted included ART start dates, ART switch dates, treatment retention, viral load, and CD4 cell count results. This retrospective study of 825 PLHIV on second-line ART reported viral load suppression of 83% (n=570/686) among patients on second-line ART, demonstrating lower suppression rates compared to historic first- line treatment (92% suppression rate) in Johannesburg. Just under three-quarters (72%, n=597/825) of the patients remained in care over the reported period, slightly lower than the reported retention rate of 78% in a first-line treatment cohort from Johannesburg. Results from the multivariable logistic regression analysis reported that being <25 years of age, male sex, and geographical transfer (started initial treatment in a different region) independently predicted low CD4+ cell counts and virological failure on second-line treatment. Being younger than 25 years of age, male sex, and transferred-in patients, are easily identifiable factors that may trigger the need for added adherence and support interventions, which include targeted adherence and retention support programs, using mobile health solutions for patient communication, education, and appointment reminders. The study presented in chapter 5 investigated individual, relationship or interpersonal, and community-level factors associated with self-reported adherence, pill count, and virological failure to ART of patients accessing care at the Ndlovu Medical Centre, Limpopo Province. This study was performed as a sub-study of the Intensified Treatment Monitoring Strategy to Prevent Accumulation of Drug Resistance (ITREMA) randomized clinical trial, a well-characterized cohort of 501 participants on antiretroviral treatment, that received prospective long-term follow-up for 96 weeks. In this study, markers of adherence and virological suppression status were periodically assessed. A comprehensive assessment of multilevel risk factors at the baseline of this trial enabled us to characterize their association with study outcomes (viral load, self-reported adherence, and pill count). The multilevel factors included demographic information, employment status, income composition, household composition, partnership status, food security, adherence, actual support from household members, actual family support, coping abilities, clinician trust, health literacy, mental health, and stigmatization. We found that over half (53%, n=243/458)) of the participants reported difficulties with adherence, and over one-third (35%, n=162/458) had suboptimal adherence measured through pill count (pill count<95%) at any point during follow-up. Virological failure appeared infrequently and occurred in 16% (n=68/436) of participants. Using tests of association and multivariable logistic regression analysis (stratified by sex), we found that being male was an independent risk factor for self-reported difficulties with adherence, suboptimal adherence measured through pill ix count, and virological failure. PLHIV who experienced moderate or severe depressive symptoms or had low household income were at increased risk of poor adherence and/or virological failure and may benefit from additional ART adherence support. In the stratified analysis, we found that the risk of virological failure was higher among male participants with food insecurity. We also found that while the prevalence of depressive symptoms was similar between males and females, the association was significant among female participants only. Task-oriented coping was associated with suboptimal adherence as indicated by pill count<95%. Our findings reported in chapter 5 contribute to the available knowledge on risk factors for adverse outcomes of ART in rural populations. The study findings may also contribute to the ongoing development of ‘rural proof’ healthcare policies currently being introduced in South Africa, such as the National Health Insurance and the new 2030 Human Resources for Health Strategy. These strategies seek to promote comprehensive access to healthcare services and also highlight the need for the government to take decisive steps to improve access to care for all individuals seeking healthcare services. Chapter 6 reports a cross-sectional study that sought to describe the different treatment-taking behaviours and perspectives of adherence to ART between virally suppressed and unsuppressed patients using second-line ART in Johannesburg. This study was conducted between July 2018 and August 2018, in five public health facilities (two hospitals, one community health center, and two primary healthcare clinics). We randomly sampled 10% of the population of 1 500 eligible patients and they were invited to participate in this study in one of two ways; telephonically or in facility recruitment where researchers met them at the facility during their scheduled clinic visit. The study sample comprised 149 participants; of which 48% (n=71/149) were virally unsuppressed. The majority of participants (63%, n=94/149) had disclosed their HIV status to their relatives and/or partners within one week of diagnosis. However, 28% (n=42/149) took longer than four weeks to disclose to their relatives and/or partners. Using multivariable logistic regression analysis, we found that single and unmarried people living with their partners were more likely to experience virological failure compared to those who were married. The more toxic second-line multi-pill, which is taken multiple times a day, was seen as significantly harder to take than a single tablet daily, well-tolerated first-line regimen. Participants experiencing medication- related difficulties in taking second-line ART and experiencing side effects were also subjective predictors of virological failure. We also found that participants with virological failure were more likely to have treatment-related side effects. Those participants with side effects were more likely to be unemployed. In general, employed individuals are linked to improved access to healthcare and better health outcomes as compared to their unemployed counterparts. However, while the correlation between improved health outcomes and employment exists, the causal relationship is x complicated as the relationship can be bi-directional. Our study results suggest the importance of improving patients’ knowledge about treatment and adherence, and motivation to continue ART use despite the persistence of side effects. Participants interviewed in the study reported in chapter 6 had firm recommendations around improving adherence to second-line ART, largely focused on reduced dosing and pill burden. These included a second-line fixed-dose combination, a dosage taken once a day, and a reduction in the pill size. Furthermore, the participants suggested that education on the benefits of taking ART could improve adherence, whilst a few participants also suggested the implementation of injectable second-line ART. In chapter 7, we report a systematic review that assessed the impact of interventions that aimed to promote adherence to treatment for chronic conditions (ART, hypertension, diabetes mellitus). We systematically searched the PubMed, Web of Science, Scopus, Google Scholar, and Cumulative Index of Nursing and Allied Health Literature (CINAHL) databases to identify relevant publications. Data were extracted from eligible studies for study characteristics and description of interventions for the study populations of interest. We found a relatively large body of evidence on interventions to improve adherence among adults living with chronic conditions in sub-Saharan Africa. Of the 25 473 total studies/records screened, a total of 77 studies were subsequently included, describing a total of 49,364 patients. Of the total included studies, 70% (n=54/77) were related to ART for HIV, 8% (n=6) were anti-hypertensive medication related, 16% (n=12/77) were anti- diabetic medication related and 6% (n=5/77) focused on medication for more than one condition. Of the total 77 studies, 60% (n=46/77) reported improved adherence based on the described study outcomes while 21% (n=16/77) reported no significant difference between studied groups. There is expanded evidence that community-and home-based, digital or mobile health (mhealth) and adherence counselling interventions can improve adherence to treatment for chronic conditions. Our findings underscore the need to develop a gold standard (or uniform measures) for measuring adherence. In the general discussion in chapter 8, the main findings were summarized, collated, and discussed. Chapter 8 provided context to the findings about the research questions and discussed its implications for future research along with recommendations. Thereafter, the strengths, limitations of this thesis, and directions for future research were also discussed. Combining multi-level models, the evidence from the studies presented in this thesis enabled us to identify the barriers to and facilitators of adherence for PLHIV on first-line and second-line ART and make recommendations for comprehensive, acceptable, and appropriate intervention strategies to improve treatment adherence. Our research found that many factors influence the xi ability to successfully engage individuals in HIV care. These factors include being male, being younger, experiencing ART-related side effects, having a low household income, presence of food insecurity, and experiencing moderate or severe depressive symptoms. With a large total sample size of 173 842 people included across all studies, our research ensured that a strong body of evidence was created regarding barriers to and facilitators of adherence to ART and adherence intervention strategies implemented to improve treatment adherence. However, all the research studies included in this thesis were conducted in a total of eight health facilities (seven of over 120 health facilities in one South African metropolitan municipality (urban setting) and one facility in a rural setting). While this ensured that study participants had comparable demographic profiles throughout the different studies in this thesis, these findings may not be generalizable to other regions or municipalities in South Africa, or other country settings. However, the research included in this thesis have sufficient sample sizes to enhance the impact of the findings. Furthermore, the magnitude and direction of the impact remained consistent throughout all chapters, indicating that the study results may be strong despite constraints associated with the study's conditions. Utilizing our study findings to enhance adherence intervention tactics is expected to enhance health outcomes and reduce the rate of patients transitioning to more intricate treatment alternatives, such as second-line and third-line ART regimens.