The most common example of a client-managed group is a self-forming group of people living with HIV who meet at an agreed community location and nominate a member to collect ART for the group from the facility on a rotational basis. That member then distributes ART to the group at the agreed community location. Data from client-managed group models have shown improved client outcomes with qualitative evidence supporting reduced costs and increased time savings.
The earliest evidence for client-managed groups came from a large cohort of clients enrolled in community ART groups (CAGs) in Tete, Mozambique. In a 2014 descriptive cohort study , retention outcomes at 12, 24, 36 and 48 months were 97.7%, 96.0%, 93.4% and 91.8%, respectively, with a mortality rate of 2.1 and loss-to-follow-up (LTFU) rate of 0.1/100 per client year. Data from three qualitative studies found cost and time savings for clients and improved certainty of ART access and mutual peer support, including health educational benefits, which facilitated better adherence [2,3]. The most recent of these qualitative studies across 10 health facilities also showed that CAG participation preserved social capital for members by providing a supportive extended family, access to mutual savings and financial assistance from other group members . A descriptive editorial explains the step-wise scale-up approach that was taken from the pilot site, to the district, and eventually nationally in Mozambique .
Scaled-up CAG model outcomes were published in late 2016 in an evaluation of trends observed after a decade of ART scale up in Mozambique . From 2004 to 2013, 455,600 people over 15 years of age had initiated ART, with 6,766 enrolling in a CAG at 69 facilities from 2011 to 2013. CAG participation was associated with a 35% lower LTFU but similar mortality. Incidence of LTFU and mortality after ART initiation for CAG and non-CAG participants was 2.9% and 0.3% at two years and 10.1% and 1.4% at four years. In a further study  reporting outcomes for the same cohort of CAG clients matched with eligible non-CAG clients (37% of cohort) at facilities offering the CAG model, eligible non-CAG clients had a significantly higher LTFU rate (hazard ratio, HR: 2.36; 95% CI: 1.54-3.17) but also similar mortality. Interestingly, the study also compared outcomes of clients in CAGs who were eligible for CAGs with those in CAGs who were ineligible (19% of cohort). One-year retention was 92.5% and 86.4% respectively (LTFU: 6.7% and 9.6%; mortality: 0.8% and 4%).
A recent retrospective study undertaken in northern Mozambique assessed all ART clients over 15 years of age who were eligible to join a CAG (n=1,306) from 2010 to 2015 for associations between baseline characteristics and total days late for appointments in the first six months on ART (prior to CAG eligibility) and CAG participation. It found no associations other than female sex. Only 13.8% joined a CAG, with CAG participation reducing mortality by 55.1% (adjusted hazard ratio, aHR: 0.449; 95% CI: 0.264-0.762) and reducing the risk of LTFU by 84.3% (aHR: 0.157; 95% CI: 0.086-0.288) . A 2019 conference abstract compared viral suppression rates after the introduction of routine VL monitoring among persons receiving ART for more than 6 months at 83 health facilities, including those in CAGs (12% of the sample, n=1,823). The overall viral suppression rate was 76% with significantly higher suppression rates among people in CAGs compared to those not in CAGs (OR 1.16; 95% CI: 1.03-1.30) .
In Lesotho, a small mixed-method comparison cohort study found 12-month retention of 98.7% (95% CI: 94.9-99.7) among stable clients who joined a CAG (n=199, median time on ART 54 months) versus 90.2% (95% CI: 86.6-92.9) for those who did not join the CAG (n=397, median time on ART 21 months) . More recently, a cluster randomized trial in thirty facilities in Lesotho compared three-monthly clinical consultations and ART refill collection from a health facility (n=1,898), three-monthly ART refills from CAGs as per the national CAG model (n=1,558) and six-monthly individual ART refill collection from community pick-up points (n=1,880). Both latter community DSD models required annual clinical consultations at the health facility. 12-month retention was similar across arms (94.9% vs. 95.4 vs. 93.3%) and achieved the pre-specified non-inferiority limit (-3.25%) with viral suppression above 98% across arms . The associated costing study found that the two community DSD models reduced provider costs per patient by approximately 7% and importantly client costs by approximately 60% .
In Zimbabwe, community ART refill groups (CARGs) have been endorsed in Zimbabwe’s National Operational and Service Delivery Manual. Clients collect ART refills every three months; members attend once a year as a group for clinical review and viral load. Qualitative work has explored patient and provider perspectives of CARGs , male engagement in CARGs , client and provider preferences for TB preventative treatment integration into the CARG model  and combined with a discrete choice experiment found clients in urban areas preferred facility-based individual models to community-based group models . A three-arm, cluster randomized non-inferiority trial across thirty health facilities compared three-monthly clinical consultations and ART refill collections from the facility, three-monthly ART refills from CARGS as per the national CARG model and six-monthly ART refills from CARGs. 12-month retention was similar across arms (clinic-based: 91%; three-monthly CARG: 93.3%; six-monthly CARG: 93.6%) and met the pre-specified noninferiority limit (-3.25%, risk difference [RD]). VL completion at 12 months was poor across all arms (below 50%) but particularly in the six-monthly CARG arm (8%). Intention- to-treat VL suppression was above 99% for clinic-based and 3-monthly CARGs and marginally reduced in six-monthly CARGs mostly driven by poor VL completion (92.9%) .
In Uganda The AIDS Support Organization (TASO), a non-governmental organization supporting more than 100,000 people living with HIV, has reported encouraging adherence outcomes (89%) among a sample of clients (n=2,799 ) in its community client-led ART delivery model (CCLAD) . In Eswatini, health facilities offered three different models (CAGs, mobile outreach and facility ACs). Among those enrolled, 12-month retention was high at 93.7% but retention by model varied substantially (CAG: 70.4%; mobile outreach 86.3%; facility AC 90.4% (p < 0.001)) . In a small study in Haiti, cross-sectional retention for a cohort of 80 CAG clients was 88.4% .
Qualitative studies in Malawi and Zambia also explore client and provider perceptions of CAGs. In Malawi, positive experiences regarding peer support were reported, but CAG uptake was hindered by limited awareness of the existence of CAGs or how they functioned . In Zambia, qualitative work demonstrated that both healthcare workers and clients favoured CAGs due to their ability to decongest the clinics and reduce workload. Several health system issues were, however, cited as problematic. Challenges included inadequate supplies of ARVs and the inability to have monitoring tests performed according to the CAG schedule due to stock-outs of specimen bottles .
Three studies reported outcomes for specific populations and contexts. A small study, which offered clients with elevated viral loads (VLs) enrolment in a CAG alongside attendance at a dedicated VL clinic, reported high CAG uptake (89.6%), but limited re-suppression (27.8%) among those with a documented follow-up VL . In the Central African Republic and the Democratic Republic of the Congo, CAGs, combined with extended refills, enabled continuity of care throughout several outbreaks of violence . In Uganda, CCLADs were introduced for clinically stable female sex workers. Two CCLADs of seven members each were formed. Retention rates of 100% were achieved in each group and all female sex workers remained virologically suppressed. ART adherence improved from 75% to 95% .