Intrapersonal level factors.

The intrapersonal level of the socio-ecological conceptual framework comprises individual knowledge, attitudes, beliefs, perceptions, and skills that influence behaviour [35]. It further includes age, sex, income, mental health, education, substance use, spirituality, comorbidities and past experiences [36].

Interpersonal level factors.

Here the participant’s social network or relationships with other people including family, friends, peers, intimate partner and health care providers [35] are included. This level also highlights trust and communication as factors that builds a relationship between patients and care givers or treatment supporters [25].

Community level factors.

The components of the community-level factors that may influence patient’s adherence to ART incorporate cultural views and social norms towards ART access, HIV related stigma, poverty, and available support services within the community [36].

Policy level factors.

The public policy level is shaped by local, and government laws regarding access and adherence to ART. The socio-ecological conceptual framework reports on awareness of macro level factors which includes health policies, HIV treatment guidelines and best practices, [22,36].

Aim.

To describe the barriers and facilitators to adherence for patients receiving first-line and second-line ART and different adherence strategies utilised.

Overarching hypothesis.

A socio-ecological framework combined with multi-model data collection will identify the barriers and facilitators to ART adherence and retention in care for PLHIV needed to strengthen adherence interventions.

Study design.

The study will employ a quantitative retrospective cohort study using secondary data analysis of data on people with HIV taking ART (18 years and older) recorded in the TIER.Net database.

TIER.Net is the monitoring and evaluation system used by the South African Department of Health for ART patient and data management. It comprises limited demographic information and all treatment and laboratory information from the time of commencement on HIV treatment. This is elaborated on in the data collection section.

The South African Department of Health started providing ART in the public health setting on 01 April 2004. From TIER.Net, we will extract a list of all patients who were initiated on ART from 01 April 2004 to 29 February 2020 in the urban setting (city of Johannesburg region F). The cut-off period of 29 February 2020 was chosen to give ART cohorts a minimum of a 12-month follow-up during which a full clinical assessment could be completed as per guidelines.

Study setting.

This study will be conducted in seven health facilities in the city of Johannesburg region F.

This includes all levels of care

1. Primary Health Care: Jeppe Clinic, Malvern Clinic, Rosettenville Clinic, Yeoville Clinic
2. Community Health Centre: Hillbrow Community Health Center (HCHC)
3. Hospitals: Charlotte Maxeke Johannesburg Academic Hospital (CMJAH) and South Rand Hospital (SRH)

Sampling/Sample size.

For this study, all records of people with HIV who were ever initiated on ART between 01 April 2004 (the inception of the South African national HIV treatment programme in the public health setting) and 29 February 2020 from the seven public health facilities will be included. Based on the TIER.Net database, about 130 000 adult patients were initiated on first-line and second-line ART in the seven facilities selected in the city of Johannesburg region F.

Data collection.

Data will be extracted from the TIER.Net database (as an MS Excel export file). TIER.Net captures demographic information such as patient age, sex, facility name and contact details and also HIV specific information such as HIV diagnosis date, ART start date, regimen at baseline, ART visit dates, CD4 count, viral load done and viral load suppression The system also records all treatment related information, including ART switch. All these variables will be extracted (see S1 Appendix study codebook). The data will be exported to STATA 15.1 for data cleaning and analysis. Records with missing data will be excluded in the final analysis. This will be after conducting all necessary data quality checks, verifications, and triangulations with other data sources.

Outcome. VL count is categorized into suppressed (<1000 copies/ml) or unsuppressed (≥1000 copies/ml) [48,49]. The status on retention in care for patients will be categorized into active in care, LTFU, transferred out, or recorded dead. For this study, LTFU is defined as having missed a scheduled medical appointment by 90 days or more, as defined by the South African Department of Health [50].

Data analysis.

Data will be coded and analysed using STATA version 15.1. Tests of association (Chi-square and t-test) between outcome variables and selected demographics characteristics and clinic indicators will be conducted. Outcome variables will include viral load detectability and retention outcomes (active in care, transferred-out, lost to follow-up and dead),. Regression analysis, univariate and multivariate analyses between variables will be built for outcome variables to identify independent predictors. Variables such as age, sex, health facility, baseline regimen, ART start date, baseline CD4 count, most recent CD4 count, ART visit dates, and months on ART will be considered as predictor variables or independent variables. Analysis will include survival analysis which will consider different entry time points into the ART program. Subsequently, patients will not be grouped all together but will be followed up in a 12-month interval when measuring the outcomes (virological failure and lost to follow up).

Study design.

This cohort study using secondary data analysis will be conducted as a sub-study of the Intensified Treatment Monitoring Accumulation (ITREMA) study (Clinicaltrials.gov Identifier NCT03357588) [51]. The ITREMA database will be used to extract ART information. ITREMA is an open-label randomised clinical trial evaluating different treatment monitoring strategies for first-line ART, which ran from June 2015 to January 2019 [51]. ITREMA was conducted at the Ndlovu medical centre in Limpopo province, South Africa. The ITREMA study enrolled adult PLHIV and assessed an intensified HIV-treatment monitoring strategy in a randomised comparison with a control group receiving standard-of-care HIV treatment in accordance with the South Africa National Department of Health guidelines

Study setting.

Patient enrolment for the ITREMA trial was done between June 2015 and August 2017 at the Ndlovu Medical Centre in Elandsdoorn, Limpopo Province, South Africa.

Sampling/Sample size.

All the records from ITREMA database will be used. There are 501 ART patients in the ITREMA database.

Data collection.

Data will be extracted from the ITREMA databases. The ITREMA database contains fields for socio-demographics and psychosocial characteristics. The control variables include sex, age, level of education, employment status, sources of income, household members, food security, mental health, HIV self-efficacy and HIV related stigma (see S1 Appendix study codebook and S6 Appendix ITREMA questionnaire). The data will be exported to STATA 15.1 for data cleaning and analysis.

Outcome. Self-reported non-adherence will be measured using three items from the ACTG questionnaire [52]: “How often do you have difficulty in taking your medication on time?, with responses given on 4-point scale (All the time, Most of the time, Rarely, Never), “On average how many days per week would you say that you missed at least one dose of your medication?”, with responses given on a 6-point scale (every day, 4–6 days per week, 2–3 days per week, Once a week, Less than once a week, Never), and “When was the last time you missed taking any of your medications?”, with responses also given on a 6-point scale (Past week, 1–2 weeks ago, 2–4 weeks ago, 1–3 months ago, More than 3 months ago, Never) (see S6 Appendix ITREMA questionnaire). Responding ‘never’ to all three questions will be taken to indicate good self-reported adherence. Suboptimal adherence measured using pill count is defined as a pill count <95%. This threshold is aligned with the WHO cut-off, which considers a pill count ≥ 95% as good adherence for patients taking ART [53]. Virological failure will be defined as viremia ≥1000 copies/ml within 96 weeks of follow-up [48,49].

Data analysis.

Data will be coded and analysed using STATA version 15.1. Tests of association (Chi-square and t-test) between outcome variables and selected socio-demographics and health related characteristics will be conducted. Outcome variables will include (but not limited to) viral load detectability, retention outcomes (active in care, transferred-out, lost to follow-up and dead), side effects and treatment interruptions (stop and restarting treatment). Regression analysis, univariate and multivariate analyses between variables will be built for outcome variables to identify independent predictors. Variables such as age, sex, health facility, beliefs, education, economic status, employment status, religious status, disclosure, and months on ART will be considered as predictor variables or independent variables.

Study design.

The study will employ a qualitative study design approach. Active patients from phase I and II (in both urban and rural settings) will be invited to participate in the in-depth interviews (IDIs) to explore factors including (but not limited to) treatment history, current use of ART, treatment regimen, financial/economic factors, risk behaviours (substance use), psychosocial characteristics cultural beliefs, spirituality and), relationship related factors (treatment support), community level factors (societal norms, stigma, discrimination, disclosure), and policy level factors (understanding/awareness of HIV/ART policies and treatment guidelines).

Study setting.

This study will be conducted in seven health facilities in the city of Johannesburg region F in Gauteng Province and one in Limpopo Province (Ndlovu Medical Centre).

The seven health facilities in the city of Johannesburg include:

1. Primary Health Care: Jeppe Clinic, Malvern Clinic, Rosettenville Clinic, Yeoville Clinic
2. Community Health Centre: Hillbrow Community Health Center (HCHC)
3. Hospitals: Charlotte Maxeke Johannesburg Academic Hospital (CMJAH) and South Rand Hospital (SRH)

Sampling/Sample size.

In this study, purposive sampling of patients currently taking ART will be employed to ensure maximum variation among the study sample. We will ensure a diverse sample by considering viral load status (suppressed and unsuppressed time on ART, age (18 years and older), sex (both males and females) and education. Sample size will depend on when saturation is reached; we anticipate that a maximum of 60 IDIs will be conducted across both study settings and viral load status groups (suppressed and unsuppressed). An anticipated number of study participants to recruit is 15 per viral load status group in each study setting (making a total of 30 participants in each study setting).

Data collection.

Patients for this phase will be contacted using the contact information that they provided for their facility records or databases used in phase I and II. IDIs will be conducted following a semi-structured interview guide. The guide will comprise open-ended questions covering treatment history, current use of ART and multilevel factors derived from the socio-ecological framework. This will include individual level factors including treatment-related factors (ART regimen, use of non-ART medication), financial and economic factors risk behaviours (substance use), psychosocial factors (cultural beliefs, spirituality), interpersonal-level factors (relationship between patients and treatment supporters or caregivers, such as intimate partners, family members, friends and health care workers), community level factors (social norms regarding HIV and ART, HIV-related stigma and discrimination, HIV-status disclosure), health-system factors (access to HIV and ART services including adherence counselling), and policy level factors (HIV testing and treatment guidelines and policies). Additional probes will be included for each question to promote sharing of detailed information regarding their perspectives and experiences, and to ensure clarification if required (see S1 Appendix study codebook and S2 Appendix interview guide). All IDIs will be audio recorded and transcribed verbatim. Transcripts will be translated into English.

Deductive themes. Deductive themes may include treatment history, current use of ART, treatment regimen, financial/economic factors, risk behaviors (substance use), psychosocial characteristics cultural beliefs, spirituality and), relationship related factors (treatment support), community level factors (societal norms, stigma, discrimination, disclosure), and policy level factors (awareness of HIV/ART policies and treatment guidelines).

Data analysis.

Transcripts will be imported and analysed using NVIVO. Data coding will be undertaken using deductive (top down) and inductive approaches (bottom up) [54]. A deductive approach is driven by researchers’ analytic interest in the study, reflecting the broad issues addressed in the interview guide. An inductive approach is used to identify the detailed themes related to the overarching issues that can be identified in the data [54]. Thematic analysis will be used, which is a method for identifying, analysing, and reporting patterns (themes) within data [54,55]. Transcripts will be read while noting similar topics that will be grouped into major topics or themes. Data will be analysed as transcripts become available shortly after interviews are conducted. This will ensure the early identification of emerging themes and assist in the identification of data saturation. Analysis of the IDIs will follow the phases of thematic analysis which are familiarization, generating initial codes, searching for themes, reviewing themes and interpretation [54]. Familiarization (getting grounded into the data collected), will be achieved by reading the transcripts and field notes repeatedly. During the process of generating codes, key emerging ideas and words from the familiarization phase will be recorded from which we will search, identify, and review themes, concepts, categories, and sub-categories. This will be done in keeping with socio-ecological framework, views and experiences that persist from the data. Finally, factors that influence adherence to ART will be identified and grouped into main categories. Analysis will be guided by the codebook which will be developed by the study team post familiarization with the data. There will be multiple independent coders to ensure the reliability of the coding.

Study design.

This systematic review will be designed and reported according to the PRISMA [56] (see S5 Appendix PRISMA checklist), following the registered protocol (CRD42019127564) on the international prospective register of systematic reviews, Prospero [57] (see S4 Appendix Systematic review protocol). The study will use PICO criteria as the search strategy tool.

A systematic review on the impact of treatment adherence interventions in chronic conditions (HIV, hypertension, DM) in sub-Saharan Africa will be conducted to provide context to adherence in sub-Saharan Africa. In this region, HIV remains the leading cause of death more especially in the young and middle-aged adults. However, the burden of non-communicable diseases (NCDs), particularly DM and hypertension, has increased rapidly in recent years [58–60].

Study setting.

All information from sub-Saharan Africa only will be included for the systematic review.

Sampling/Sample size.

All interventions described as chronic conditions adherence interventions (HIV/ART, hypertension, DM). Inclusion in the systematic review will be dependent on the criteria set out in the systematic review protocol [57] and reported using the PRISMA reporting guidelines.

Data collection.

A pre-defined data sheet will be developed for data extraction. The tool will include (but not be limited to): reference (author, title), year of publication, setting or location, sample size, intervention description, participants receiving adherence (in case of comparison) (see S1 Appendix study codebook). The form/tool will be tested before conducting the final searches. One reviewer will conduct all the data extraction while a second reviewer will be responsible for data quality assurance on the extraction and also conduct full text review of the included material.

We will search using several electronic databases. These will include PubMed/Medline, Web of Science, Google Scholar, Scopus, and CINAHL. If necessary, we will contact study authors and request more information on individual studies. Citations and bibliographies of records will be reviewed to identify additional relevant material.

The basic search terms included will be:

“Chronic conditions” OR “hypertension” OR “high blood pressure” OR “blood pressure” OR “arterial hypertension” OR “mellitus diabetes type I” OR “mellitus diabetes type II” OR “Diabetes” OR “Sugar” OR “HIV” OR “Antiretroviral Therapy” OR “Antiretroviral Treatment” OR “ART” OR “ART Programs” OR “ART Programmes” AND “adherence” OR “compliance” AND “interventions” OR “strategies” OR “odds ratio” OR “risk ratio” OR “evaluation” OR “impact” OR “effectiveness” OR “outcome” AND “sub-Saharan Africa” OR “sub Saharan Africa” OR “sub-Saharan African” OR “sub Saharan African” OR “Africa” (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Methodological aspect of the systematic review.

https://doi.org/10.1371/journal.pone.0261107.t001

The search terms will be adjusted to suit the database being searched. An inventory with the database searched, the corresponding search criteria used, the date when the searches were conducted, and the results will be maintained. A second reviewer will run the searches separately for comparison. The strength of the body of evidence (quality of evidence), the risk of bias and magnitude of effect will be rated and assessed using Grading of Recommendations Assessment, Development and Evaluation (GRADE) [61,62].

Outcome. The primary outcome will be adherence to antiretroviral therapy, defined as the proportion of patients meeting the defined adherence criteria. The secondary outcome will be proportion of patients achieving viral suppression, as defined by the study. Outcome (and impact) measures will be reported in terms of changes in the prevalence or reduction in the relative risk. Should there be adequate statistical reporting, a meta-analysis will be considered [63].

Data analysis.

All adherence interventions or strategies will be described, based on the type of intervention implemented and the setting. The different evaluations methods will then be described in detail by comparing the type of assessments and outcome measures (adherence to ART and viral load). If appropriate, outcome measures will be reported in terms of changes in the prevalence or reduction in the relative risk. Whenever necessary, we will calculate unadjusted risk ratios (RRs) and 95% confidence intervals (CIs) from data provided and present the outcome indicator results in forest plots. Furthermore, we will perform a sensitivity analysis to measure the robustness of our results to the choice of summary statistic and calculated unadjusted risk differences. We will apply a random- effects model to calculate summary RRs and 95% Cl. To test the robustness of the findings, we will re-run the analysis using a fixed effects model. Data will be coded and analysed used STATA version 15.1.

Details for study criteria are presented in Table 1.

Strengths and limitations.

To our knowledge this will be the first study using a social-ecological perspective and a convergent parallel mixed method design to report on combination ART in South Africa. Combining quantitative and qualitative data from public health settings, a controlled environment, the patient perspective and evidence from the region will enable us to make recommendations for a comprehensive, acceptable, and appropriate adherence strategy for the country.

The study presents a fewlimitations. The study will be 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). Therefore, findings may not be generalizable to other municipalities and districts in South Africa, or to other country settings. Furthermore, although there are efforts to ensure good quality of data by the South African Department of Health, supporting partners and research staff, secondary data are subject to quality issues, due to data inconsistencies and missing data. Other potential predictors of adherence, such as disclosure, stigma, and self-reported adherence, will be assessed through a standard questionnaire. This may lead to reporting bias (memory and social desirability biases). Policy-level factors might not probe for relevant factors due to how this study is designed (with it focus on the individual’s awareness of health policies and guidelines which is insufficient to fully assess policy-level factors).

In conclusion, our study will demonstrate how an existing socio-ecological conceptual framework can be used as a tool to provide guidance regarding facilitators and barriers to ART adherence. By populating this framework through secondary data analysis, participant interviews of PLHIV who are taking ART and a systematic review comparing adherence intervention strategies for the chronic conditions, this mixed method study will provide evidence on factors affecting treatment adherence at different socio-ecological levels and guide context-specific intervention strategies to improve ART adherence. We believe that the use of our study results to strengthen adherence intervention will subsequently improve health outcomes and decrease the number of patients switching to complex treatment such as second-line and third-line regimens.