We will conduct a quasi-experimental study (25,26) consisting of two arms—intervention and control for 24 months. Approximately 2,000 stable people living with HIV in the intervention arm will receive services from the CAD model. Another 2,000 people living with HIV in the control arm will receive MMD—a default care model for stable people living with HIV to refill ART every three to six months with a routine six-monthly clinical review. We will adopt an effectiveness-implementation hybrid design type II (27), where a mixed-methods approach will run in parallel to evaluate the effectiveness (quantitative) and implementation strategy (qualitative).
This study will include 10 ART clinics in Phnom Penh capital city and four provinces (Kampong Thom, Kampot, Koh Kong, and Takeo) for the intervention arm (Table 1). We will include 10 other ART clinics in Phnom Penh and five provinces (Kampong Cham, Pailin, Preah Sihanouk, Siem Reap, and Prey Veng) for the MMD control arm (Table 1). These sites have been purposively selected based on (i) the availability of implementing partners (Cambodian People Living with HIV Network, ARV User Association, and Partner in Compassion), (ii) the number of eligible people living with HIV for enrollment in the intervention and control arms, and (iv) advice from the national HIV program on the accessibility of the clinics.
We define stable people living with HIV as individuals who (i) are ≥15 years, (ii) have received first-line ART for at least one year, (iii) did not report ART-related adverse reactions or drug interactions requiring regular monitoring; (iv) do not have tuberculosis (presumptive/confirmed), other opportunistic infections, and are not taking any prophylactic treatment; (v) are not pregnant or breastfeeding (for women); (vi) have a good understanding of lifelong treatment and adherence to medication; and (vii) have achieved treatment success—two consecutive undetectable viral load and/or CD4 counts above 200 cells/mm3 (11,19,28). We will conduct in-depth interviews and focus group discussions with people living with HIV, representatives of the communities, government officials at different levels (community, health centers, operational districts, provincial health departments, and national programs), local and international NGOs, donors and other developmental partners in Cambodia to explore the acceptability, barriers, and facilitators to implementing the CAD model.
Sample size calculation
A target sample size of approximately 2,000 stable people living with HIV is required to provide at least 80% power to detect a 10% relative difference retention in care or maintenance of viral suppression. The sample size calculation was based on 10% attrition and conservative estimates of retention in care or viral-suppression maintenance: 50% and 55% maintenance of viral suppression in the MMD (control arm) and CAD (intervention arm), respectively. Findings from a recent randomized controlled trial in high- and medium-HIV-prevalence settings of people living with HIV with detectable viral load showed that CAD increased viral suppression compared to the control group (74% vs. 63%, RR 1.18, 95% CI 1.07–1.29). Under similar scenarios, our proposed study's power with 2,000 participants in each arm will be at least 90%.
We will develop a list of participating ART clinics in consultation with the Database Management Unit of the National Center for HIV/AIDS, Dermatology, and STD (NCHADS) and the other implementing partners. Potential people living with HIV will be jointly identified by the study team and clinicians on-site based on the eligibility criteria. We will seek consent from eligible people living with HIV and recruit them into the study. To measure the impact of CAD on the work burden of healthcare providers, we will recruit all staff working for the HIV program at the intervention and control sites—nurses, counselors, laboratory technicians, pharmacists, and clinicians. For the qualitative study, we will purposively recruit interviewees at the national level and intervention sites for diverse views.
Intervention: CAD model
People living with HIV in the intervention arm will receive antiretrovirals (ARVs) from the community action workers (CAWs) through the CAD model. We will recruit people living with HIV in the community to be CAWs. Training will be provided by the ART clinics and implementing partners on ART dispensing, drug storage, vital signs assessment and documentation, HIV education and counseling, ART adherence and measurement, referral systems, and aspects such as mental health, stigma and discrimination, and sexual and reproductive health. CAWs will obtain pre-packaged ARVs once a month from the attached ART clinic and dispense them to people living with HIV in the community. CAWs will conduct group follow-up sessions monthly to record vital signs, monitor adherence, and provide health education and counseling. Those who are unwell will be referred to the ART clinic. CAWs will also submit the relevant records to the ART clinic at their next visit. Trained clinicians will conduct ARV regimen and HIV clinical management in the ART clinics. People living with HIV in the intervention arm will visit the ART clinics for routine clinical review every six months and on an ad-hoc basis if medical attention is required. Clinical management of people living with HIV will be done by trained personnel at the ART clinic following the national guideline (28). The study flowchart is presented in Figure 1.
Control: MMD model
People living with HIV in the control arm will visit ART clinics and collect their ARVs from the facility-based service providers through the MMD model, the standard care for stable people living with HIV. People living with HIV will refill their prescriptions at the ART clinic every three to six months. The exact frequency might differ between individuals, decided at the discretion of individual ART clinics. People living with HIV will visit the ART clinic on an ad-hoc basis for consultation if needed. They will otherwise undergo routine clinical review every six months. Clinical management of people living with HIV will be done by trained personnel at the ART clinic following the national guideline (28). The study flowchart is presented in Figure 1.
We present the conceptual framework reflecting the outcome measures and their effects on the continuum of care in Figure 2. Our primary outcomes will include (i) viral load suppression, (ii) retention in care, and (iii) adherence to ART. Viral load suppression is defined as at least 90% of the study participants in the intervention arm will have a viral load of <1000 RNA copies/mL at the end of the intervention period. We define care retention as the proportion of people living with HIV who will remain in HIV care and treatment 12 and 24 months after study commencement. Treatment adherence will be self-reported, and we will present the proportion of people living with HIV with good adherence to ART based on standard cut-offs.
We will also assess secondary outcomes, including (i) the impact of CAD on the work burden of healthcare providers at the ART sites, (ii) cost-effectiveness of this model and its impact on (iii) the quality of life, (iv) mental health, (v) social support, and (vi) stigma and discrimination faced by people living with HIV in the community. We will report the time and volume of stable people living with HIV seen at the ART sites as a measure of the workload of healthcare providers in HIV care delivery. We will present quality of life, mental health, social support, stigma and discrimination, and workload (burnout) as scores measured using validated tools described in the section below. We will perform an economic evaluation to determine if the intervention is cost-effective. We will also determine the acceptability, feasibility, barriers, and facilitators to implementing the CAD model in Cambodia using the qualitative method.
All people living with HIV will be interviewed face-to-face by trained data collection teams using structured questionnaires. We will interview all participants in the intervention and control arms at baseline (0 months), midline (at 12 months of the intervention), and endline (at 24 months of the intervention). We will collect qualitative data in private locations using semi-structured interview guides. The data collection timelines for quantitative surveys and qualitative in-depth interviews and focus group discussions are illustrated in Figure 3.
Variables and measurements
We will collect information on sociodemographic (age, sex at birth and gender, membership of key population groups, income, residence, education, family, and dependency) and HIV treatment history and health status (duration of taking ART, transportation mode, relationship, and satisfaction with health care workers). We will adapt questionnaires previously implemented by NCHADS to measure medication adherence and retention in care and treatment (28,29). Viral load will be obtained from medical records at the attached ART clinics.
We will measure depressive symptoms using the Center for Epidemiologic Studies Depression Scale (CES-D) (18), quality of life using the WHO Quality of Life BREF (WHOQOL-BREF) for people living with HIV and EQ-5D-5L (30–32) and stigma and discrimination using the People Living with HIV Stigma Index (18). We will also measure self-efficacy and levels of social support using the Perceived Self-Efficacy for Receiving Antiretroviral Therapy Scale (PSEARTS) (33–35) and Berlin Social Support Scale (BSSS) (34,36,37), respectively. For the economic evaluation, we will collect direct and indirect costs for follow-up care and ARV refills. We will assess healthcare providers' workload at ART clinics using two different measures. First, we will evaluate burnout (extent of physical and psychological fatigue and exhaustion) using the Maslach Burnout Inventory (MBI) – Human Services Survey for Medical Personnel (38). Second, we will measure the number of people living with HIV (both stable and unstable) served per day and week and the time spent per people living with HIV at the different service points in the clinic. For the qualitative assessment, we will explore the perception of people living with HIV, HIV service providers, and other key stakeholders on the advantages and disadvantages, acceptability, feasibility, and sustainability of CAD in the HIV continuum of care.
We will collect and manage data using REDCap (39,40). Data coding, quality control, and data entry will be done. We will enter all data into the database within one week of collection. Subsequently, we will export the database into Microsoft Excel (Microsoft Corp., Redmond, Washington, USA) to check for consistency periodically. All questionnaires will be checked for errors by the team leaders, and necessary corrections will be made before data entry and analyses.
We will compare the characteristics of the participants in the intervention and control arms using descriptive statistics. In the presence of significant differences, we will adjust for them in the subsequent between-group (i.e., intervention vs. control) analyses. Key outcome indicators will be compared using inferential statistics such as Chi-square tests for categorical variables and Student's t-test for continuous variables. To evaluate the effectiveness of the CAD model intervention, we will compare the outcome indicators at baseline, midline, and endline within and between the intervention and control arm using the difference-in-differences method (41). We will apply multivariable regression analyses using mixed models, logistic and Cox regression, considering repeated measurements design and the nature of the response variables. Statistical analyses will be performed in STATA (Stata Corp LP, Texas, United States of America) and R (R Foundation for Statistical Computing, Vienna, Austria). For qualitative data, we will transcribe audio recordings verbatim, and the transcripts will be coded using NVIVO (QSR International). Content analyses will be performed to identify emerging categories, themes, and common and divergent patterns pertinent to the study’s objectives.
We will inform all participants clearly about the study’s objectives and the participation risks and benefits before enrolling them in the study. If a participant cannot read or write, the CAWs or data collectors will read the information sheet to the participant during the consent-taking process. Participants will be able to withdraw from or discontinue the study at any time. All participant records—written, recorded, and transcribed data—will be stored securely. We will assign coded identifiers to participant names (with a master list stored separately). Participants will receive a token equivalent to US$3 for their time and transportation compensation. It is anticipated that the study results will benefit people living with HIV and their communities, and the harm to the participants will be minimal.