DIFFERENTIATED SERVICE DELIVERY

FOR HIV TREATMENT:

SUMMARY OF PUBLISHED EVIDENCE
Nov 2020
 Recommended citation:
International AIDS Society, Nov 2020, “Differentiated service delivery for HIV treatment:
Summary of published evidence”, www.differentiatedservicedelivery.org

Photo credit: Paul Odongo/MSF, Peter Casaer/MSF, Isabel Cortheir/MSF

ii Differentiated service delivery for HIV treatment

OVERVIEW

The evidence for differentiated service delivery (DSD) for HIV treatment is summarized and
presented by model type:

Healthcare worker-managed group.....................................................................................page 2

Client-managed group.........................................................................................................page 6

Facility-based individual model.......................................................................................... page 10

Out-of-facility individual model........................................................................................ page 14

This summary of evidence includes published data up until October 2020, including from peer-
reviewed publications and conference abstracts. To read more, download the original research
through the hyperlinks in the References (pages 18-21) and visit www.differentiatedservicedelivery.org

ACRONYMS
AC - adherence club LTFU - loss to followup
aHR - adjusted hazard ratio MACs - medication adherence clubs
AIDS - Acquired immunodeficiency syndrome MMD - multi-month ART refills
aOR - adjusted odds ratio MMPs - multi-month prescriptions
aRD - adjusted risk difference PODI - community ART distribution points
ART - antiretroviral therapy PRP - pharmacy-only refill programme
C-BART - community-based ART R6M - Rendez-vous de Six Mois
CAGs - community ART groups RD - risk difference
CARGs - community ART refill groups RR – risk ratio
CCLAD - client-led ART delivery model SEARCH - Sustainable East Africa Research in
CDDP - community drug distribution point Community Health study

CI – confidence interval SMA - six-monthly appointments

DSD - differentiated service delivery SOC - standard of care

DRC - Democratic Republic of the Congo TASO - The AIDS Support Organization

HBD – home-based delivery TB - tuberculosis

HIV - human immunodeficiency virus UAGs - urban adherence groups
HR - hazard ratio VL - viral loads
IQR - interquartile range WHO - World Health Organization

Summary of published evidence (Nov 2020) 1

 HEALTHCARE WORKER-
MANAGED GROUPS
Photo credit: Miguel Cuenca/MSF

Evidence on healthcare worker-managed groups highlights antiretroviral therapy (ART) adherence club (AC) model in
improved client outcomes, both at individual sites and at Cape Town, South Africa. Details of the model expansion
scale, and suggests benefits to specific client populations, show that from January 2011 to March 2015, 32,425 clients
including children, their caregivers and adolescents and were in an AC (25.2% of the total ART cohort). Fifty-five
increasingly to breastfeeding women and their infants, key facilities were offering a total of 1,308 ACs (1). In a cluster
populations and clients who have struggled with adherence random sample of 10% of the Cape Town ACs (3,216 adults)
previously.
Details of the model expansion show that
from January 2011 to March 2015, 32,425
Quantitative
clients were in an AC (25.2% of the total
The majority of evidence for healthcare worker-managed ART cohort).
group models comes from implementation of the

2 Differentiated service delivery for HIV treatment

from non-research-supported ART sites, retention was A costing study found the Khayelitsha
95.2% (CI 94-96.4) at 12 months and 89.3% (CI 87.1-91.4) at piloted AC model cost effective, with a
24 months. In the 13 months prior to database closure, 88.1%
cost per client year of $300 versus $374
of clients had viral loads (VL) taken, with VL ≤400 copies/ml
for standard of care.
in 97.2% (CI 96.5-97.8) (2). An evaluation of disengagement
from care in Khayelitsha, Western Cape, reported AC
participation as highly protective against disengagement
(hazard ratio 0.27 (Cl 0.24-0.30) (3). A 2019 evaluation of Four additional studies described various aspects of the

long-term virologic outcomes of 8,058 clients ever enrolled AC model in the Western Cape. The first describes the AC

in AC care in Khayelitsha showed high annual VL completion model and the strategy used by local health authorities

over 40 months (82-85%), with 6% experiencing an elevated to scale the model (10). The second outlines the quality

VL at a median of 363 days from AC enrolment (IQR 170- improvement approach embedded in the scale-up strategy

728) (4). (11). The third describes the adjusted AC model implemented
in Gugulethu, where ACs were run at a community venue
With ACs recognized since 2015 as one of three endorsed instead of the clinic (12). The Western Cape AC model
differentiated ART delivery models in South Africa, national changed from providing two-monthly ART refills (six
retention and viral load suppression outcomes were times a year) to providing a four-monthly ART refill at year
evaluated in a 2019 study comparing 24 randomly allocated end (five total visits per year) to accommodate year-end
intervention and control facilities. AC clients had higher holiday migration. A comparison study found no difference
12-month retention (89.5% versus 81.6%, aRD 8.3%; CI 1.1% in short-term retention or viral rebound risk comparing
to 15.6%) and comparable sustained viral suppression (<400 clients who receive two months versus four months of
copies/mL any time ≤ 18 months) (80.0% versus 79.6%, ART over the holiday period (13). A non-inferiority cluster
aRD 3.8%, CI −6.9% to 14.4%). Retention associations were randomized trial comparing retention, VL completion and VL
stronger for men than women (men RD 13.1%, CI 0.3% to suppression outcomes of experienced AC clients receiving
23.5%; women RD 6.0%, CI−0.9% to 12.9%) (5). six-monthly ART refill in their ACs compared with those in
the aforementioned Western Cape AC model found similar
Four comparison cohort studies report client outcomes
24-month retention (intervention 93.1% (CI 91.2-94.7); SOC
from ACs (6, 7). In the pilot study in Khayelitsha, Cape Town,
94.0% (CI 92.4-95.2)), higher VL completion (94.5% vs.
retention at study end was 97% for those stable clients
89.3%) and similar VL suppression (96.3% vs. 97.5%) (14).
who enrolled in an AC versus 85% for those who did not.
Loss to followup (LTFU) was reduced by 57% (hazard ratio A non-inferiority cluster randomized trial
HR 0.43, 95% CI 0.21-0.91) and viral rebound by 67% (HR comparing retention, VL completion and VL
0.33, 95% CI 0.16-0.67) (6). In Gugulethu, Cape Town, suppression outcomes of experienced AC clients
94% were retained at 12 months post AC enrolment, with
receiving 6-monthly ART refill in their ACs
3% experiencing viral rebound by study end (7). After
compared with those in the aforementioned
adjustment, AC participation was associated with a 67%
Western Cape AC model found similar
reduction in the risk of LTFU (aHR 0.33, 95% CI 0.27-0.40)
24-month retention (intervention 93.1% (CI
compared with clients in the standard of care. In the rural
91.2-94.7); SOC 94.0% (CI 92.4-95.2)), higher
Cape Winelands, a 2019 retrospective cohort study of all
VL completion (94.5% vs. 89.3%) and similar VL
adult clients starting ART in 2014-2015 found lower loss
to followup in those attending an AC (aHR 0.25, CI 0.11 to
suppression (96.3% vs. 97.5%).
0.56). This finding was confirmed on analysis restricted to
those eligible for AC referral (aHR 0.28, CI 0.12 to 0.65) (8). In Zambia, urban adherence groups (UAGs) are also part of
A costing study found the Khayelitsha piloted AC model cost national DSD policy. A study using a matched-pair cluster
effective, with a cost per client year of $300 versus $374 randomized study design, only enrolling DSD eligible clients
for standard of care (9). willing to join a UAG at intervention and control clinics,

Summary of published evidence (Nov 2020) 3

 found the rate of late drug pick-up was lower in UAG with 83% in CAGs, 95% in UAGs, 79% in home delivery and
participants compared to clinic-based care participants (aHR 69% in mobile outreach. Provider costs per person retained
0.26, 95% CI 0.15-0.45) (15). Median medication possession was higher in DSD models than in clinic-based care (23). In
ratio was 100% in intervention participants compared with KwaZulu-Natal, South Africa, clients were offered a choice
96% in control participants. Although 18% of the UAG group between community ART groups (CAGs), community ACs,
meeting visits were missed, on-time drug pick-up (within in-facility individual fast-lane pick-up or out-of-facility
seven days) still occurred in 51% of these missed visits individual pick-up. Overall DSD model retention was high
through alternate means (use of buddy pick-up or early at 12, 24 and 36 months when compared with those who
return to the facility). qualified for a DSD model but remained in clinic-based care,
but viral suppression was significantly lower for those who
had participated in the group models by 36 months (24).
Qualitative
Six qualitative studies looked at healthcare worker and/or Specific populations
client perceptions around ACs. The first explored the
acceptability of community-based ACs from a health worker Four studies from Cape Town and one from Uganda,
perspective, including enablers and barriers to roll out (16). reported outcomes for specific populations: adolescents,
The second explored the perceptions of AC members and children and their caregivers, postnatal women and men
non-members at two sites in Khayelitsha and Gugulethu, who have sex with men receiving their care from youth-
Cape Town (17). The third explored the perceptions of clients specific ACs, family ACs or adults ACs, respectively. For
who were enrolled in either community or facility-based youth ACs, ART client outcomes were good. Retention at 12
ACs in Witkoppen, Johannesburg (18). The fourth took place months for youth stable on ART was 94.3% (CI 85.4-96.8);
in Lusaka, Zambia, where client and provider perceptions for youth newly initiated on ART, it was 86.4% (CI 78.7-91.4);
of UAGs were explored (19). In Cape Town, two recent and for youth ineligible for ART, it was 52.9% (CI 40.0-64.2)
qualitative studies were undertaken. One nested within (25). For family ACs, child and caregiver retention was 93.7%
a non-inferiority cluster randomized trial explored client, (CI 88.7-96.6) and 93.9% (CI 85.9-97.4) at 12 months and
healthcare worker and key informant experiences and 86.1 (CI 79.5-90.8) and 89.7 (CI 80.4-94.8) at 36 months
perceptions of receiving six-month ART refills in ACs (26). For women initiated on ART during pregnancy and
(20). The second explored how participation in postnatal who chose to join an existing community-based adult
ACs affected knowledge transmission, peer support, AC immediately post-delivery (84/129; 65%) compared
health-seeking behaviour and satisfaction with the care with ART at their clinic, viral loads above 1,000 copies/ml
provided (21). were lower at 12 (AC 16% and clinic 23%) and 24 months
(AC 29% and clinic 37% (27). In a further study, postnatal
Two studies in Zambia and one in South Africa compared
women and their infants were offered enrolment into
outcomes after providing clients with a choice of
postnatal clubs. These included both stable and high-risk
differentiated ART delivery model. Within the PopART
mother-infant pairs until the infant reached 18 months.
study in Lusaka, Zambia, clients in two study arms were
offered a choice for collecting a three-month ART refill in:
i) clinic-based care or home delivery (HBD); or ii) clinic-
For women initiated on ART during pregnancy
based care or community-based AC (22). Twelve-month
and who chose to join an existing community-
viral suppression was non-inferior in the community DSD
based adult AC immediately post-delivery
models (above 98% in all three arms). More clients were lost
(84/129; 65%) compared with ART at their
to care in the clinic-base care arm (52/781; HBD 18/825; AC
20/808) with more deaths in the HBD arm (17; clinic-based
clinic, viral loads above 1,000 copies/ml were
2; AC 7). In a retrospective outcomes analysis of clients lower at 12 (AC 16% and clinic 23%) and 24
who enrolled in DSD models in Zambia from 2015-2017, months (AC 29% and clinic 37%).
12-month retention was 81% in clinic-based care compared

4 Differentiated service delivery for HIV treatment

Eighteen-month retention was 79.2% (28), with 76% of In Mozambique, ACs were implemented for clients
mothers with a VL taken between 12 and 18 months and with a history of HIV treatment failure. Retention at
viral suppression of 94%. Eighty-one percent of infants 12 and 24 months was 98.9% (95% CI 98.2-99.7) and
completed nine-month HIV testing and 64% 18-month HIV 96.4% (95% CI 94.6-98.2), respectively. Concurrently,
testing compared with 51% and 32% of historical controls 85.8% (95% CI 83.1-88.2) and 80.9% (95% CI 77.8-
(29). In Uganda, ACs were implemented for men who have 84.1) of clients maintained VL suppression.
sex with men with 100% viral suppression maintained after 11
months of follow up (30).
In 2020, a study in Cape Town compared retention and
viral suppression outcomes of 503 AC clients who had
Among previously clinically unstable experienced viraemia and had either been referred back to
clinic-based care or erroneously remained in ACs. Those who
Two studies reported on client outcomes for clients who remained in ACs had the same 12-month retention (93%),
had previously struggled with adherence. In Cape Town, slightly lower VL completion (77% versus 84%) and higher VL
clients who had re-suppressed after a nurse-led intervention re-suppression (62% versus 53%) (33).
and were immediately referred into an adult AC, 12-month
retention and viral suppression after AC enrolment was
94.8% (CI 89.8-97.4) and 85.2% (CI 78.0-90.1), respectively Among those with hypertension or
(31). In Mozambique, ACs were implemented for clients diabetes
with a history of HIV treatment failure. Retention at 12
and 24 months was (95% CI 98.2-99.7) and 96.4% (95% CI Two early studies have evaluated medication adherence clubs
94.6-98.2), respectively. Concurrently, 85.8% (95% CI 83.1- (MACs), which adjusted the AC model to incorporate stable
88.2) and 80.9% (95% CI 77.8-84.1) of clients maintained hypertension or diabetic clients. A retrospective descriptive
VL suppression. Among 90 clients attending AC and study of 1,432 clients in MACs reported 3.5% LTFU in the
simultaneously having VL rebound, 64 (71.1%) achieved VL first year after enrolment (34). A qualitative study found the
re-suppression, 10 (11.1%) did not re-suppress and 14 (15.6%) model acceptable to both clients and healthcare workers,
had no subsequent VL result (32). saving clients time and reducing queues at the clinic (35).

Summary of published evidence (Nov 2020) 5

 CLIENT-MANAGED
GROUPS
Photo credit: Peter Casaer/MSF

The most common example of a client-managed group is a were 97.7%, 96.0%, 93.4% and 91.8%, respectively, with a
self-forming group of people living with HIV who meet at mortality rate of 2.1 and LTFU rate of 0.1/100 per client year.
an agreed community location and nominate a member to Data from three qualitative studies found cost and time
collect ART for the group from the facility on a rotational savings for clients and improved certainty of ART access and
basis. That member then distributes ART to the group at mutual peer support, including health educational benefits,
the agreed community location. Data from client-managed
group models have shown improved client outcomes, with The earliest evidence for client-managed groups
qualitative evidence supporting reduced costs and increased came from a large cohort of clients enrolled
time savings. in community ART groups (CAGs) in Tete,
The earliest evidence for client-managed groups came from
Mozambique. In a 2014 descriptive cohort study,
a large cohort of clients enrolled in community ART groups retention outcomes at 12, 24, 36 and 48 months
(CAGs) in Tete, Mozambique. In a 2014 descriptive cohort were 97.7%, 96.0%, 93.4% and 91.8%, respectively.
study (36), retention outcomes at 12, 24, 36 and 48 months

6 Differentiated service delivery for HIV treatment

which facilitated better adherence (37, 38, 39). A descriptive In Lesotho, a small mixed-method comparison cohort study
editorial explains the step-wise scale-up approach that found 12-month retention of 98.7% (95% CI 94.9-99.7)
was taken from the pilot site, to the district, and eventually among stable clients who joined a CAG (n=199, median
nationally in Mozambique (40). 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) (45). More recently, a cluster randomized
Quantitative trial in 30 facilities in Lesotho compared three-monthly
clinical consultations and ART refill collection from a health
Scaled-up CAG model outcomes were published in late
facility (n=1,898), three-monthly ART refills from CAGs
2016 in an evaluation of trends observed after a decade
as per the national CAG model (n=1,558) and six-monthly
of ART scale up in Mozambique (41). From 2004 to 2013,
individual ART refill collection from community pick-up
455,600 people over 15 years of age had initiated ART, with
points (n=1,880). Both latter community DSD models
6,766 enrolling in a CAG at 69 facilities from 2011 to 2013.
required annual clinical consultations at the health facility.
CAG participation was associated with a 35% lower LTFU
Twelve-month retention was similar across arms (94.9%
but similar mortality. Incidence of LTFU and mortality after
vs. 95.4 vs. 93.3%) and achieved the pre-specified non-
ART initiation for CAG and non-CAG participants was 2.9%
inferiority limit (-3.25%) with viral suppression above 98%
and 0.3% at two years and 10.1% and 1.4% at four years. In a
across arms (46). The associated costing study found that
further study (42) reporting outcomes for the same cohort
the two community DSD models reduced provider costs per
of CAG clients matched with eligible non-CAG clients (37%
client by approximately 7% and, importantly, client costs by
of cohort) at facilities offering the CAG model, eligible non-
approximately 60% (47).
CAG clients had a significantly higher LTFU rate (hazard
ratio, HR 2.36; 95% CI 1.54-3.17) but also similar mortality.
The associated costing study found that the two
Interestingly, the study also compared outcomes of clients
community DSD models reduced provider costs
in CAGs who were eligible for CAGs with those in CAGs
per client by approximately 7% and, importantly,
who were ineligible (19% of cohort). One-year retention
client costs by approximately 60%.
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 In Zimbabwe, community ART refill groups (CARGs) have

Mozambique assessed all ART clients over 15 years of age been endorsed in the National Operational and Service

who were eligible to join a CAG (n=1,306) from 2010 to Delivery Manual. Clients collect ART refills every three

2015 for associations between baseline characteristics and months; members attend once a year as a group for clinical

total days late for appointments in the first six months on review and viral load. Qualitative work has explored patient

ART (prior to CAG eligibility) and CAG participation. It found and provider perspectives of CARGs (48), male engagement

no associations other than female sex. Only 13.8% joined a in CARGs (49), and client and provider preferences for TB
CAG, with CAG participation reducing mortality by 55.1% preventative treatment integration into the CARG model
(adjusted hazard ratio, aHR 0.449; 95% CI 0.264-0.762) (50). Combined with a discrete choice experiment, it was
and reducing the risk of LTFU by 84.3% (aHR 0.157; 95% CI found that clients in urban areas preferred facility-based
0.086-0.288) (43). A 2019 conference abstract compared individual models to community-based group models
viral suppression rates after the introduction of routine VL (51). A three-arm, cluster randomized non-inferiority
monitoring among persons receiving ART for more than six trial across 30 health facilities compared three-monthly
months at 83 health facilities, including those in CAGs (12% clinical consultations and ART refill collections from the
of the sample, n=1,823). The overall viral suppression rate facility, three-monthly ART refills from CARGS as per the
was 76%, with significantly higher suppression rates among national CARG model and six-monthly ART refills from
people in CAGs than those not in CAGs (OR 1.16 95% CI CARGs. Twelve-month retention was similar across arms
1.03-1.30) (44). (clinic-based 91%; three-monthly CARG 93.3%; six-monthly

Summary of published evidence (Nov 2020) 7

 CARG 93.6%) and met the pre-specified noninferiority limit In Zambia, qualitative work demonstrated that
(-3.25%, risk difference (RD)). VL completion at 12 months both healthcare workers and clients favoured
was poor across all arms (below 50%) but particularly in CAGs due to their ability to decongest the clinics
the six-monthly CARG arm (8%). Intention-to-treat VL and reduce workload. Several health system issues
suppression was above 99% for clinic-based and three- were, however, cited as problematic. Challenges
monthly CARGs and marginally reduced in six-monthly
included inadequate supplies of ARVs and the
CARGs, mostly driven by poor VL completion (92.9%) (52).
inability to have monitoring tests performed
In Uganda, The AIDS Support Organization (TASO), a non- according to the CAG schedule due to stock-outs
governmental organization supporting more than 100,000 of specimen bottles.
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)
(53). In Eswatini, health facilities offered three different
Specific populations and contexts
models (CAGs, mobile outreach and facility ACs). Among
Three studies reported outcomes for specific populations
those enrolled, 12-month retention was high at 93.7%, but
and contexts. A small study, which offered clients with
retention by model varied substantially (CAG 70.4%; mobile
elevated VLs enrolment in a CAG alongside attendance at
outreach 86.3%; facility AC 90.4% (p<0.001)) (54). In a small
a dedicated VL clinic, reported high CAG uptake (89.6%),
study in Haiti, cross-sectional retention for a cohort of 80
but limited re-suppression (27.8%) among those with a
CAG clients was 88.4% (55).
documented follow-up VL (58). In the Central African
Republic and the Democratic Republic of the Congo, CAGs,

Qualitative combined with extended refills, enabled continuity of care

throughout several outbreaks of violence (59). In Uganda,
Qualitative studies in Malawi and Zambia also explored CCLADs were introduced for clinically stable female sex
client and provider perceptions of CAGs. In Malawi, positive workers. Two CCLADs of seven members each were formed.
experiences regarding peer support were reported, but CAG Retention rates of 100% were achieved in each group and all
uptake was hindered by limited awareness of the existence female sex workers remained virologically suppressed. ART
of CAGs or how they functioned (56). In Zambia, qualitative adherence improved from 75% to 95% (60).
work demonstrated that both healthcare workers and
In the Central African Republic and the
clients favoured CAGs due to their ability to decongest the
clinics and reduce workload. Several health system issues
Democratic Republic of the Congo, CAGs,
were, however, cited as problematic. Challenges included combined with extended refills, enabled
inadequate supplies of ARVs and the inability to have continuity of care throughout several
monitoring tests performed according to the CAG schedule outbreaks of violence.
due to stock-outs of specimen bottles (57).

8 Differentiated service delivery for HIV treatment

 Photo credit: Paul Odongo/MSF

Summary of published evidence (Nov 2020) 9

 FACILITY-BASED
INDIVIDUAL MODELS
Photo credit: Albert Masias/MSF

The principle of differentiating between the need for Median waiting time was reduced from 102 to
a clinical visit versus an ART refill visit, combined with 20 minutes, with increased client and provider
extended ART refills, has been used in a number of facility- satisfaction in the intervention group compared
based individual models of ART delivery. These models go with the standard of care.
beyond only extending ART refills to reducing time spent at
the facility setting up fast-track or quick pick-up services.

(six-monthly clinical reviews and two-monthly ART refills

Quantitative, including cost data from the pharmacy) (61). The PRP was less costly (US$520/
year versus $655/year) and more cost effective than the
Evidence of the effectiveness of facility-based individual standard of care (62). The second study assessed clinic
models has been reported from four studies in Uganda efficiencies after implementation of a fast-track system
(61-64). The first was a cost-effectiveness study conducted (six-monthly clinical visits with two-monthly ART refills after
after implementing a pharmacy-only refill programme (PRP) seeing a triage nurse). Median waiting time was reduced

10 Differentiated service delivery for HIV treatment

from 102 to 20 minutes, with increased client and provider The Sustainable East Africa Research in Community Health
satisfaction in the intervention group compared with the (SEARCH) study, a test-and-treat trial in Kenya and Uganda,
standard of care. The third was a descriptive study after streamlined HIV care for adults (≥15 years; CD4 ≥350 cells/
implementation of a refill pick-up system (six-monthly μl) and children (2-14 years; CD4 ≥500 cells/μl), including
clinical review and ART refill of 30-90 days at clinician nurse-driven triage and referral for visits with physician for
discretion) (63). There were significant reductions in missed complex cases; three-month combined clinical and ART refill
appointments from 24.4% to 20.3% (adjusted odds ratio, visits for stable clients; consolidation of multiple chronic
aOR 0.67; 95% CI 0.59-0.77) and medication gaps of three disease services at encounter; client appointment flexibility;
days or more from 20.2% to 18.4% (aOR 0.69; 95% CI and missed appointment tracing from ART start at first visit.
0.60-0.79) in the intervention group compared with the This resulted in 48-week retention and viral suppression
standard of care. The fourth evaluated PRP revised to six- among adults of 92% (897/972) and 93% (778/838) and
monthly clinical reviews and three-monthly refills directly retention and viral suppression among children of 89%
from the pharmacy (64). Overall retention was 99.3% and (74/83) and 92% (65/71) in Uganda and Kenya, respectively
among those who completed 12 months in the PRP, viral (67). There were also significant reductions in time spent
suppression was maintained at 98.8%. at the health facility and away from work or other usual
activities. Out-of-pocket expenses for clients from baseline
Data reported from Malawi describes a growing cohort
to one year later were reduced in Uganda, but not in Kenya
of clinically stable clients receiving multi-month ART
(68). Costing of streamlined HIV care was similar or lower
refills (MMD) or enrolled in a fast-track clinic system
to standard of care cost estimates after accounting for VL
(six-monthly clinical review and three-monthly ART refills
testing and VL result counselling session costs (69). In the
from lay healthcare workers) known as the six-monthly
Western Cape, South Africa, a “quick pick-up” model for
appointment (SMA) strategy. In a 2017 mixed-methods
clinically stable clients documented that 12 months after
process evaluation, 100% of 730 Malawian ART sites
joining the model, 96% of clients were still in care, with 85%
offered multi-month ART refills with 72.9% of eligible clients
of them remaining in the model (70).
accessing MMD (65). Only 11 (1.5%) facilities offered SMA,
with 77.7% of eligible clients at these facilities enrolled in
SMA. A 2018 retrospective study assessed all clinically Those in the fast-track model were more
stable clients eligible for the SMA model between 2008 likely to be retained at 12 months (RR 1.52)
and 2015 (n=22,633) at these 11 facilities (66). It found that and maintain viral suppression (RR 1.07).
81% enrolled in SMA with median time from eligibility to
enrolment of 12 months (interquartile range 3-27 months)
and median cumulative time on SMA was 14.5 months. The Two studies in Zambia evaluated fast-track facility DSD
cumulative probability of retention in care one year after models. In 2020, a study compared all clients in routine
first SMA eligibility was 86.8% (CI 85.6-87.8%) among those facility-based care (n=83,764) with those in the fast-track
who never enrolled, 97.3% (CI 96.8-97.6%) among early model (n=3,671) where clients went directly to a dedicated
SMA enrolees and 99.8% (CI 99.7-99.9%) among late SMA room where they received an expedited clinical visit and
enrolees. The corresponding figures at five years were 47.4% ART refill every three months. Those in the fast-track model
(CI 45.0-49.7%), 85.5% (CI 84.0-86.9%) and 93.4% (CI were more likely to be retained at 12 months (RR 1.52) and
92.8-94.0%). Among eligible clients enrolling for SMA, the maintain viral suppression (RR 1.07) (71). An analysis of
adjusted hazard of attrition was 2.4 (95% CI 2.0-2.8) times 62,084 clinically stable clients (on treatment for >6 months
higher during periods of SMA discontinuation than during with CD4 >200 cells/μl and not on TB treatment or unwell)
periods on SMA. Male gender, younger age, more recent showed that the longer the appointment interval and ART
SMA eligibility and WHO Stage 3/4 conditions in the past refill (up to six months), the less likely the client was to have
year were also independently associated with attrition from missed appointments, have a gap in medication or become
SMA. Approximately 26,000 consultations were “saved” lost to follow up (72). Associated qualitative work to explore
during 2014 alone. healthcare workers and client experiences of a fast-track

Summary of published evidence (Nov 2020) 11

 model demonstrated that healthcare workers and clients Qualitative
viewed the model as being able to decongest the clinic and
reduce waiting times (73). Overall, the model was highly Three qualitative studies explored six-monthly ART refills. In
applicable and acceptable. There were requests to carry Malawi, a study assessed provider and client perceptions of
out additional activities, such as taking weight and blood six versus three monthly refills. Both clients and providers
pressure, that were continued in the dedicated fast-track reported larger supply had more benefits. Providers’
service room. concerns regarding medication storage challenges and the
risk of sharing ART were not supported by clients (76). A
Associated qualitative work to explore second study, in Zambia, determined provider perceptions
healthcare workers and client experiences only and established that providers perceived multi-month
of a fast-track model demonstrated that dispending to hold significant benefits with advantages
healthcare workers and clients viewed the of six- over three-monthly dispensing (77). In Ethiopia,
model as being able to decongest the clinic focus groups were held with clients eligible for six-monthly
and reduce waiting times. refills, some of whom had enrolled and other who had not.
It showed high model satisfaction for those who enrolled,
but importantly, that this model did not suit everyone.
One study in KwaZulu-Natal, South Africa, compared
Decreased facility visits, lack of private space for medication
outcomes after providing clients with a choice of
storage and mistrust of the healthcare systems were reasons
differentiated ART delivery model (24). Clients were offered
for not enrolling (78).
a choice between in-facility individual fast-lane pick-up (also
known as spaced fast-lane appointment), out-of-facility
individual pick-up, community adherence groups (CAGs) and Specific contexts
community adherence clubs (ACs). Retention was high at 12,
24 and 36 months across all DSD models when compared Appointment spacing has also been shown to have benefits
with those who qualified for a DSD model but remained in in low-prevalence settings. In Guinea in West Africa, the
clinic-based care, but was highest at 90.6% at 36 months SMA model was piloted in 2013 and expanded in 2014
for facility individual pick-up. Viral suppression was high in following the outbreak of the Ebola virus disease (79). The
both individual DSD models at 36 months (facility individual six-monthly spacing approach, Rendez-vous de Six Mois
pick-up at 92.6%, community individual pick-up at 93.8% (R6M), was scaled up to 60% of the cohort (n=1,166). Clients
compared with routine clinic care at 88.6%) and significantly outside of the capital city of Conakry received six-monthly
lower for those who had participated in the group models. clinical visits and ART refills, and those in Conakry received
three-monthly ART refills and six-monthly appointments.
The R6M group had a 60% reduction in the risk of attrition
Implementation compared with the standard of care after adjusting for

Importantly, an increasing number of studies are evaluating duration on ART and TB co-infection.

DSD implementation. In Zimbabwe, a mixed-method DSD

Outside of sub-Saharan Africa, a facility-based individual
model implementation evaluation in a rural district with 26
differentiated ART delivery model implemented in Yangon,
health facilities found that only 31% has implemented a fast-
Myanmar, has reported good early outcomes (80). Clients
track individual facility refill model, but that clients spent
were differentiated between unstable, short-term stable
40% less time at the facility than those in routine care (74).
(29.2% of cohort) and long-term stable (51.2% of cohort).
To date, the majority of differentiated ART delivery models Short-term stable clients received three-monthly combined
have provided two or three months of ART. In Ethiopia, clinical review and ART refills visits, alternating between a
six-monthly refills were introduced at health facilities with physician and nurse. Long-term stable clients received six-
biannual clinical visits. In total, 51% of clients were assessed monthly clinical reviews from a nurse and three-monthly
to be eligible for this model, of whom 49% enrolled (75). fast-tracked ART refills from a pharmacist or dispenser.

12 Differentiated service delivery for HIV treatment

In politically unstable settings, such as the the implementation of multi-month prescriptions (MMPs)

Central African Republic, South Sudan and the for children across six sub-Saharan African countries, clients
aged 0-19 years were transitioned to MMPs when they
Democratic Republic of the Congo, the ability to
were defined as clinically stable (82). The study analysed
provide extended refills of three to six months
outcomes from more than 22,000 children, 66% of whom
has also enabled continuity of ART delivery
were transitioned to MMPs. Of those transitioned, 2.6%
during periods of acute conflict.
were lost to follow up and 2% died. Virological suppression
remained high over the first five years in MMPs, ranging
The number of clients that a team, made up of a physician,
by year from 79% to 85%. These results provide reassuring
nurse and counsellor, could manage increased from 745
evidence that children and adolescents who are clinically
in 2011 to 1,627 in 2014, averting 41,116 physician visits.
stable can have good outcomes with reduced visit
Aggregated 12-month retention for both clinically stable
frequencies and extended ART refills.
groups was 98.7%, with clinical treatment failure of 0.8% and
immunological treatment failure of 5.8%. A children-focused DSD model implemented in Tanzania
utilized MMPs and also introduced a fast-track component
In politically unstable settings, such as the Central African
where children could go directly to the pharmacy to collect
Republic, South Sudan and the Democratic Republic of the
their ART refills after an initial triage (83). Clients in this
Congo, the ability to provide extended refills of three to six
model received ART refills every two months and had a
months has also enabled continuity of ART delivery during
clinical visit every four months. A total of 51.3% of the
periods of acute conflict (81).
paediatric, adolescent and young adult ART clients were
able to be enrolled in this model, with 98.8% remaining in

Specific populations care. Reduced clinical visits and extended ART refills for
clinically stable adults, children and adolescents should be
Extended ART refills and fast-track service delivery models a priority model of differentiated service delivery that can
have also shown benefits for children. In a study assessing yield benefits in both high- and low-prevalence settings.

Summary of published evidence (Nov 2020) 13

 OUT-OF-FACILITY
INDIVIDUAL MODELS
Photo credit: Peter Casaer/MSF

Out-of-facility individual models vary according to where Democratic Republic of the Congo (DRC), external pick-up
in the community the services are provided, as well as what points in South Africa, community pick-up points in Zambia
services are delivered and by whom. They can be divided into and community private pharmacy collection in Nigeria and
three categories: fixed community points; (including private Uganda.
pharmacies) mobile outreach ART delivery; and home
In Uganda, cross-sectional outcomes for clients initiating
delivery.
ART from 2004 to 2009 (median time on ART 5.7 years;
interquartile range, IQR 4.1-7.2 years) in the CDDP model
Fixed community points were that 69% were retained in care, 17% had died, 6%
were transferred out and 9% were lost to follow up (LTFU).
Promising results have been found in models using fixed Among CDDP clients, viral load suppression (<1,000
community points. Evidence has come from the community copies/mL) was 93% (median time on ART 7.0 years; IQR
drug distribution point (CDDP) model in Uganda, 5.0-8.0) (84) . In a subsequent conference abstract, LTFU
the community ART distribution points (PODI) in the was reported as 16.5% in the facility arm and 4.28% in the

14 Differentiated service delivery for HIV treatment

CDDP arm (85). A costing comparison study in Uganda Interestingly, the study reported increased viral suppression
put the model from TASO, including TASO-run clinics for among men (RD 11.1%; 95% CI −3.4% to 25.5%) (5). A
new initiations and the CDDP model for stable clients, at qualitative study of external pick-up points reported quicker
US$74/visit and $332/client/year compared with a mobile and more convenient ART collection in the community,
ART delivery model utilizing expert clients to dispense ART seen as a reward for taking ART well and reduced disruption
(US$45/visit; $404/client/year) and a facility-based nurse- in client life activities. At private pharmacies, some clients
led model (US$38/visit; $257/client/year) (86). reported receiving inferior care compared with paying
customers, and some worried about inadvertently revealing
In the DRC, a retrospective cohort analysis found LTFU
their HIV status. Clients and healthcare workers had to
and death among PODI clients to be at 2.2% and 0.1% at six
negotiate problems with Central Chronic Medication
months,4.8% and 0.2% at 12 months and 9% and 0.3% at 24
Dispensing and Distribution implementation, such as delayed
months, respectively, with overall crude attrition of 5.66/100
SMS reminders, ART not being available at the external pick-
person years with little variation over time (87). Two 2018
up point and a few private pharmacies placing restrictions
conference abstracts also reported on PODI outcomes. The
on ART pick-up times (90).
first on 576 clients enrolled in a PODI from October 2016 to
December 2017 reported 12-month retention of 98% (88). Nigeria is expanding an out-of-facility model
The second on 1,484 ART clients enrolled at the four PODI where community private pharmacies are
houses resulted in decanting of linked facilities by 44-47%. linked with public health facilities and clients
The four PODI houses show high retention rates of 92-100%
can elect to collect two-month ART refills from
at three, six and nine months and VL suppression above 90%
these pharmacies.
(89).

In South Africa, external pick-up points were endorsed

In Zambia, clinically stable clients were able to select
as one of the country’s three differentiated ART delivery
community pick-up points for ART refill collection. ART
models. This model allows clients to choose a non-facility
refills were pre-packed by a central dispensing unit. Six
based venue for ART refill collection. These include fixed
participating clinics enrolled 6,303 clients to collect from 19
community points, private pharmacies and, more recently,
community pick-up points (rural markets, shopping malls and
lockers. National retention and viral load suppression
some at clinics) with a 96% retention rate, 94% ART refill
outcomes were evaluated in a 2019 study comparing 24
pick-up compliance rate and 93% viral suppression rate (<50
intervention and control facilities. External pick-up point
copies/ml) (91).
clients had lower 12-month retention (81.5% versus 87.2%,
aRD −5.9%; CI −12.5% to 0.8%) and comparable sustained Nigeria is expanding an out-of-facility model where
viral suppression (<400 copies/mL any time ≤18 months) community private pharmacies are linked with public health
(77.2% versus 74.3%, aRD −1.0%; CI −12.2% to 10.1%). facilities and clients can elect to collect two-month ART
refills from these pharmacies. An early evaluation reported
a 100% prescription refill rate and 99.7% retention in care
In Zambia, clinically stable clients were able
(92). Uganda has also started utilizing private pharmacies,
to select community pick-up points for ART
with six serving as community ART refill points for stable
refill collection. ART refills were pre-packed
clients from four high-volume health facilities (>8,000
by a central dispensing unit. Six participating
ART clients) (93). A nurse-dispenser distributes ART refills
clinics enrolled 6,303 clients to collect from
with six-monthly clinical reviews at the health facility. Over
19 community pick-up points (rural markets, a 30-month period (January 2017 to June 19), 9,921 (29%
shopping malls and some at clinics), with a men) clients enrolled, representing 30% of clients at the
96% retention rate, 94% ART refill pick-up four facilities. Of these, 96% had received ART refills as
compliance rate and 93% viral suppression rate scheduled, and the average waiting time at the pharmacy
(<50 copies/ml). was <10 minutes. The 12-month retention rate was 98%, and
>99% of enrolled clients remained virally suppressed.

Summary of published evidence (Nov 2020) 15

 In northern Namibia, two high HIV-burden districts with between the virological failure rates for home versus for
far distances to clinics implemented C-BART sites (94). facility care (rate ratio 1.04, 95% CI 0.78-1.40; equivalence
Community members constructed basic structures close to shown) (97). Mortality rates were also similar between the
their homes where healthcare workers visited quarterly to groups (rate ratio 0.85, 95% CI 0.71-1.28). Health services
provide HIV clinical assessment, ART refills and VLs. Clients and client cost per year were less for home delivery
did not need to attend health facilities. In a retrospective compared with facility refill (US$793 vs. $838 for health
cohort analysis of 909 adults (≥15 years) and 122 children services and $18 vs. $54 for client). In Kenya, no significant
enrolled from 2007-2017, 12-, 24-, 36-, 48- and 60-month intervention-control differences were found with regard
retention remained at 97% for adults and 81.5% for children. to detectable viral load, mean CD4 count, change in ART
regimen, new opportunistic infections or pregnancy rates.
Intervention clients made half as many clinic visits as did
Mobile outreach ART delivery controls (98).

There is limited published evidence of utilizing mobile In Tanzania, a non-inferiority cluster trial randomized
outreach services to distribute ART refills outside of the 24 health facilities to: i) clinic-based ART delivery for all
health facility. In Eswatini, health facilities were offered a clients or: ii) the offer of home-based delivery (HBD) by lay
choice of three ART delivery models for implementation healthcare workers for stable adult clients with an annual
(mobile outreach, CAGs and facility-based ACs). One health clinical review visit at the facility (99). In the intervention
centre and one clinic implemented mobile outreach to arm, 516 (44.4%) of the clients took up an offer to receive
support remote communities. Among those enrolled in DSD ART refills in or close to their homes (87.4% of stable
models, 12-month retention was high at 93.7% but retention clients). At the end of the study period (mean follow-up time
within models varied substantially (mobile outreach 86.3%; was 326 days), loss to follow up was 18.9% in the intervention
CAGs 70.4%; ACs 90.4%) (95). clinics and 13.6% in the control clinics, with 9.7% (91/943) of
intervention clients and 10.9% (95/872) of control clients
In South Africa, comprehensive ART services (including
failing virologically.
ART refills) were provided by mobile outreach on South
African-Zimbabwean border farms to vulnerable, highly In Lusaka, Zambia, within the PopART study, clients in two
mobile Zimbabwean migrant farm workers and their families study arms were offered a choice for collecting a three-
(96). The intervention piloted a travel package, including month ART refill within: i) clinic-based care or home delivery,
a 3-month ART refill. Viral suppression was 91.2%, and of or ii) clinic-based care or community-based AC (22). Twelve-
those clients who indicated planned travel to Zimbabwe, month viral suppression was non-inferior in the community
only 2% did not return within three months of their planned DSD models (above 98% in all three arms). More clients
return date. In a retrospective outcomes analysis of clients were lost to care in the clinic-base care arm (52/781; HBD
who enrolled in DSD models in Zambia from 2015 to 2017, 18/825; AC 20/808) with more deaths in the HBD arm (17;
12-month retention was 81% in clinic-based care compared clinic-based 2; AC 7).
with 69% in two-monthly mobile outreach to rural health
centres with both cohorts including non-stable clients. Two South African studies report on home ART refill

Provider costs per person retained was unsurprisingly delivery. A retrospective cohort study of clients utilizing

much higher in the mobile outreach model (US$291 versus private healthcare through a private health management

US$124) (23). scheme compared outcomes between clients receiving their

ART refills through a courier service at home (n=14,620) and
those who collected their ART refills at a private pharmacy
Home delivery (n=19,202) (100). The likelihood of viral suppression was
higher for the home refill group (81% vs. 71%).
There are four cluster randomized controlled trials from
Kenya and Uganda reporting outcomes from home ART In a marginal structure model addressing time-varying
delivery models. In Uganda, there was no difference aspects and causality, home refill was associated with

16 Differentiated service delivery for HIV treatment

 Photo credit: Sydelle Willow Smith/IAS

an even higher benefit (aHR 0.66, 95% CI 0.55-0.78). the COVID-19 pandemic at a single clinic in the Tshwane
A descriptive study reports on eligibility and uptake of health district (101). Thirty-two percent of 1,727 clients
home delivery of ART and other chronic disease refills evaluated were eligible for home delivery, of whom 432
(including hypertension, diabetes, mental health conditions, (79%) accepted.
dyslipidaemia, osteoarthritis, asthma and epilepsy) during

Summary of published evidence (Nov 2020) 17

 REFERENCES
1. Wilkinson L, Harley B, Sharp J, Solomon S, Jacobs S, Cragg 13. Grimsrud A, Patten G, Sharp J, Myer L, Wilkinson L, Bekker L-G.
C, et al. Expansion of the Adherence Club model for stable Extending dispensing intervals for stable patients on ART. J
antiretroviral therapy patients in the Cape Metro, South Africa Acquir Immune Defic Syndr. 2014;66(2):e58-60.
2011-2015. Trop Med Int Health. 2016;21(6):743-9.
14. Cassidy T, Keene C, Lebelo K, Grimsrud A, Sibanda B, Ndlovu
2. Tsondai PR, Wilkinson LS, Grimsrud A, Mdlalo PT, Ullauri A, S, et al. Twenty-four month retention and viral load outcomes
Boulle A. High rates of retention and viral suppression in the from a non-inferiority cluster randomized trial of extending
scale-up of antiretroviral therapy adherence clubs in Cape ART dispensing intervals to 6-monthly in Adherence Clubs.
Town, South Africa. Journal of the International AIDS Society. AIDS 2020: Virtual; July 6-10 2020.
2017;20(Suppl 4).
15. Roy M, Bolton-Moore C, Sikazwe I, Mukumbwa-Mwenechanya
3. Kaplan SR, Oosthuizen C, Stinson K, Little F, Euvrard J, M, Efronson E, Mwamba C, et al. Participation in adherence
Schomaker M, et al. Contemporary disengagement from clubs and on-time drug pickup among HIV-infected adults
antiretroviral therapy in Khayelitsha, South Africa: A cohort in Zambia: A matched-pair cluster randomized trial. PLOS
study. PLOS Medicine. 2017;14(11):e1002407. Medicine. 2020;17(7):e1003116.
4. Kehoe K, Boulle A, Tsondai PR, Euvrard J, Davies MA, Cornell M. 16. Tshuma N, Mosikare O, Yun JA, Alaba OA, Maheedhariah
Long-term virologic responses to antiretroviral therapy among MS, Muloongo K, et al. Acceptability of community-based
HIV-positive patients entering adherence clubs in Khayelitsha, adherence clubs among health facility staff in South Africa:
Cape Town, South Africa: a longitudinal analysis. Journal of the a qualitative study. Patient preference and adherence.
International AIDS Society. 2020;23(5):e25476-e. 2017;11:1523-31.
5. Fox MP, Pascoe S, Huber AN, Murphy J, Phokojoe M, Gorgens 17. Venables E, Towriss C, Rini Z, Nxiba X, Cassidy T, Tutu S, et al.
M, et al. Adherence clubs and decentralized medication Patient experiences of ART adherence clubs in Khayelitsha and
delivery to support patient retention and sustained viral Gugulethu, Cape Town, South Africa: A qualitative study. PLOS
suppression in care: Results from a cluster-randomized ONE. 2019;14(6):e0218340.
evaluation of differentiated ART delivery models in South
18. Mudavanhu M, West NS, Schwartz SR, Mutunga L, Keyser
Africa. PLoS Med. 2019;16(7):e1002874.
V, Bassett J, et al. Perceptions of Community and Clinic-
6. Luque-Fernandez MA, Van Cutsem G, Goemaere E, Based Adherence Clubs for Patients Stable on Antiretroviral
Hilderbrand K, Schomaker M, Mantangana N, et al. Treatment: A Mixed Methods Study. AIDS Behav.
Effectiveness of Patient Adherence Groups as a Model of Care 2020;24(4):1197-206.
for Stable Patients on Antiretroviral Therapy in Khayelitsha,
19. Topp S, Mwamba C, Bolton-Moore C, Sharma A. Acceptability
Cape Town, South Africa. PLOS ONE. 2013;8(2):e56088.
and feasibility: patient and provider experiences of urban
7. Grimsrud A, Lesosky M, Kalombo C, Bekker LG, Myer L. adherence groups (UAGs) in Zambia. AIDS 2020: Virtual; July
Implementation and Operational Research: Community-Based 6-10 2020.
Adherence Clubs for the Management of Stable Antiretroviral
20. Keene CM, Zokufa N, Venables EC, Wilkinson L, Hoffman R,
Therapy Patients in Cape Town, South Africa: A Cohort Study.
Cassidy T, et al. ‘Only twice a year’: a qualitative exploration of
J Acquir Immune Defic Syndr. 2016;71(1):e16-23.
6-month antiretroviral treatment refills in adherence clubs for
8. Bock P, Gunst C, Maschilla L, Holtman R, Grobbelaar N, people living with HIV in Khayelitsha, South Africa. BMJ open
Wademan D, et al. Retention in care and factors critical for [Internet]. 2020 2020/07//; 10(7):[e037545 p.].
effectively implementing antiretroviral adherence clubs in a
rural district in South Africa. Journal of the International AIDS 21. Duvivier H, Decroo T, Nelson A, Cassidy T, Mbakaz Z, Duran LT,
Society. 2019;22(10):e25396-e. et al. Knowledge transmission, peer support, behaviour change
and satisfaction in post Natal clubs in Khayelitsha, South Africa:
9. Bango F, Ashmore J, Wilkinson L, van Cutsem G, Cleary S. a qualitative study. Reproductive Health. 2020;17(1).
Adherence clubs for long-term provision of antiretroviral
therapy: cost-effectiveness and access analysis from 22. Limbada M, Macleod D, Chileshe K, Muhau E, Shibwela O,
Khayelitsha, South Africa. Trop Med Int Health. 2016;21(9):1115- Floyd S, et al. Viral suppression in stable HIV+ patients in two
23. community models of ART delivery: A cluster-randomized trial
nested within the HPTN 071(PopART) trial in Lusaka, Zambia.
10. Wilkinson L. ART adherence clubs: A long-term retention AIDS 2020: Virtual; July 6-10 2020.
strategy for clinically stable patients receiving antiretroviral
therapy. Southern African Journal of HIV Medicine. 23. Nichols BE, Cele R, Jamieson L, Long LC, Siwale Z, Banda P, et
2013;14:48-50. al. Community-based delivery of HIV treatment in Zambia: costs
and outcomes. CROI 2020; March 8-11 2020; Boston USA.
11. Horwood CM, Youngleson MS, Moses E, Stern AF, Barker PM.
Using adapted quality-improvement approaches to strengthen 24. Shigayeva A, Gcwensa N, Ndlovu CD, Ntumase N, Sabela
community-based health systems and improve care in high S, Ohler L, et al. Retention in care among patients in
HIV-burden sub-Saharan African countries. AIDS (London, differentiated models of HIV care in KwaZulu-Natal, South
England). 2015;29 Suppl 2:S155-64. Africa. AIDS 2020: Virtual; July 6-10 2020.

12. Grimsrud A, Sharp J, Kalombo C, Bekker LG, Myer L. 25. Wilkinson L, Moyo F, Henwood R, Runeyi P, Patel S, de Azevedo
Implementation of community-based adherence clubs for V, et al. Youth ART adherence clubs: Outcomes from an
stable antiretroviral therapy patients in Cape Town, South innovative model for HIV positive youth in Khayelitsha, South
Africa. J Int AIDS Soc. 2015;18(1):19984. Africa. AIDS 2016; 2016 July 18-22; Durban, South Africa.

18 Differentiated service delivery for HIV treatment

26. Tsondai P, Wilkinson L, Henwood R, Ullauri A, Cassidy T, Tutu retroviral care delivery model - a qualitative research study in
S, et al. Retention and viral suppression outcomes of patients Tete, Mozambique. Journal of the International AIDS Society.
enrolled in family ART adherence clubs in Cape Town, South 2014;17(1).
Africa. IAS 2017; 2017 July 22-27; Paris, France2017.
40. Decroo T, Lara J, Rasschaert F, Bermudez-Aza E, Couto A,
27. Odayar J, Lesosky M, Malaba TR, Kabanda S, Allerton J, Hu N-C, Candrinho B, et al. Scaling Up Community ART Groups in
et al. Differentiated care for postpartum ART in South African Mozambique. International STD Research& Reviews. 2013;
women living with HIV: an RCT. CROI 2020; 2020 March 8-11; 1(2):49-59.
Boston USA.
41. Auld AF, Shiraishi RW, Couto A, Mbofana F, Colborn K, Alfredo
28. Myer L, Iyun V, Zerbe A, Phillips TK, Brittain K, Mukonda E, C, et al. A Decade of Antiretroviral Therapy Scale-up in
et al. Differentiated models of care for postpartum women Mozambique: Evaluation of Outcome Trends and New Models
on antiretroviral therapy in Cape Town, South Africa: a of Service Delivery Among More Than 300,000 Patients
cohort study. Journal of the International AIDS Society. Enrolled During 2004-2013. J Acquir Immune Defic Syndr.
2017;20(S4):21636. 2016;73(2):e11-22.
29. Lebelo K, Cassidy T, Duran LT, Bhardwaj V, Mantangana N, 42. Jobarteh K, Shiraishi RW, Malimane I, Samo Gudo P, Decroo
Mdani L, et al. Integrated postnatal clubs show improved T, Auld AF, et al. Community ART Support Groups in
maternal viral load completion and infant testing uptake Mozambique: The Potential of Patients as Partners in Care.
compared to historical controls in Khayelitsha, South Africa. PloS one. 2016;11(12):e0166444-e.
AIDS 2020: Virtual; July 6-10 2020.
43. Jefferys LF, Hector J, Hobbins MA, Ehmer J, Anderegg, N.
30. Atuhura S, Weikum D, Ajulong C, Ganafa S, Kimbugwe M, Ndeloa Improved survival and retention in HIV treatment and care:
C, et al. Leaving no one behind: Assessing the impact of MSM the value of community ART groups for HIV patients on ART
community-based adherence clubs on retention and viral load in rural northern Mozambique. AIDS 2016; 2016 Jul 18-22;
suppression in Uganda. AIDS 2020: Virtual; July 6-10 2020. Durban, South Africa.
31. Sharp J, Wilkinson L, Cox V, Cragg C, van Cutsem G, Grimsrud 44. Van Rompaey S, Melo M, Bilhete F, Tancredo I, Wester C,
A. Outcomes of patients enrolled in an antiretroviral adherence De Schacht C. Improved viral suppression rates among
club with recent viral suppression after experiencing elevated HIV-positive adults receiving antiretroviral therapy (ART)
viral loads. Southern African Journal of HIV Medicine. via community adherence group (CAG) support in Zambézia
2019;20(1). province, Mozambique. 13th INTEREST Conference; 2019,
Accra, Ghana.
32. Finci I, Flores A, Gutierrez Zamudio AG, Matsinhe A, de
Abreu E, Issufo S, et al. Outcomes of patients on second- and 45. Vandendyck M, Motsamai M, Mubanga M, Makhakhe S,
third-line ART enrolled in ART adherence clubs in Maputo, Tunggal S, Jonckheree S, et al. Community-Based ART
Mozambique. Tropical Medicine & International Health. Resulted in Excellent Retention and Can Leverage Community
2020;n/a(n/a). Empowerment in Rural Lesotho, A Mixed Method Study.
HIV/AIDS Res Treat Open J. 2015;2(2):44-50.
33. Cassidy T, Euvrard J, Keene C, Roberts E, Gerstenhaber R,
Boulle A. Patients experiencing viraemia in Adherence Clubs: 46. Tukei BB, Fatti G, Tiam A, Ngorima-Mabhena N, Tukei VJ,
is back-to-clinic always best? CROI 2020; March 8-11 2020; Tshabalala I, et al. Twelve-Month Outcomes of Community-
Boston USA. Based Differentiated Models of Multimonth Dispensing of
ART Among Stable HIV-Infected Adults in Lesotho: A Cluster-
34. Khabala KB, Edwards JK, Baruani B, Sirengo M, Musembi P,
Randomized Noninferiority Trial. Journal of Acquired Immune
Kosgei RJ, et al. Medication Adherence Clubs: a potential
Deficiency Syndromes. 2020;85(3):280-91.
solution to managing large numbers of stable patients with
multiple chronic diseases in informal settlements. Trop Med Int 47. Nichols B, Fatti G, Cele R, Leodeba N, Maotoe T, MV. S, et al.
Health. 2015;20(10):1265-70. Economic evaluation of differentiated service delivery models
35. Venables E, Edwards JK, Baert S, Etienne W, Khabala K, for ART service delivery in Lesotho: cost to provider and cost
Bygrave H. “They just come, pick and go.” The Acceptability to patient. AIDS 2020: Virtual; July 6-10 2020.
of Integrated Medication Adherence Clubs for HIV and Non 48. Bochner AF, Meacham E, Mhungu N, Manyanga P, Petracca
Communicable Disease (NCD) Patients in Kibera, Kenya. PLOS F, Muserere C, et al. The rollout of Community ART Refill
ONE. 2016;11(10):e0164634. Groups in Zimbabwe: a qualitative evaluation. Journal of the
36. Decroo T, Koole O, Remartinez D, dos Santos N, Dezembro International AIDS Society. 2019;22(8):e25393-e.
S, Jofrisse M, et al. Four-year retention and risk factors for 49. Mantell JE, Masvawure TB, Mapingure M, Apollo T, Gwanzura C,
attrition among members of community ART groups in Block L, et al. Engaging men in HIV programmes: a qualitative
Tete, Mozambique. Tropical Medicine & International Health. study of male engagement in community-based antiretroviral
2014;19(5):514-21. refill groups in Zimbabwe. Journal of the International AIDS
37. Rasschaert F, Telfer B, Lessitala F, Decroo T, Remartinez Society. 2019;22(10):e25403-e.
D, Biot M, et al. A Qualitative Assessment of a Community 50. Msukwa M, Apollo T, Mapingure M, Boccanera R, Chingombe
Antiretroviral Therapy Group Model in Tete, Mozambique. PLOS
I, Gwanzura C, et al. Leveraging Zimbabwe’s Community
ONE. 2014;9(3):e91544.
Antiretroviral Groups (CARGs) to deliver TB preventive
38. Rasschaert F, Decroo T, Remartinez D, Telfer B, Lessitala F, treatment is feasible and acceptable: Findings from a mixed
Biot M, et al. Adapting a community-based ART delivery model methods study. AIDS 2020: Virtual; July 6-10 2020.
to the patients’ needs: a mixed methods research in Tete,
51. Strauss M, George G, Mantell JE, Mapingure M, Masvawure
Mozambique. BMC Public Health. 2014;14(1):364.
TB, Lamb MR, et al. Optimizing Differentiated HIV Treatment
39. Rasschaert F, Decroo T, Remartinez D, Telfer B, Lessitala Models in Urban Zimbabwe: Assessing Patient Preferences
F, Biot M, et al. Sustainability of a community-based anti- Using a Discrete Choice Experiment. AIDS and Behavior. 2020.

Summary of published evidence (Nov 2020) 19

 52. Fatti G, Ngorima-Mabhena N, Mothibi E, Muzenda T, Choto 65. Prust ML, Banda CK, Nyirenda R, Chimbwandira F, Kalua T,
R, Kasu T, et al. Outcomes of Three- Versus Six-Monthly Jahn A, et al. Multi-month prescriptions, fast-track refills, and
Dispensing of Antiretroviral Treatment (ART) for Stable community ART groups: results from a process evaluation
HIV Patients in Community ART Refill Groups: A Cluster- in Malawi on using differentiated models of care to achieve
Randomized Trial in Zimbabwe. J Acquir Immune Defic Syndr. national HIV treatment goals. J Int AIDS Soc. 2017;20(Suppl
2020;84(2):162-72. 4):21650.
53. Okiror A. ART adherence among patients cared in the 66. Wringe A, Cawley C, Szumilin E, Salumu L, Amoros Quiles I,
community client-led ART delivery (CCLAD) - TASO, Uganda Pasquier E, et al. Retention in care among clinically stable
experience. AIDS 2018; 2018 Jun 23-27; Amsterdam, the antiretroviral therapy patients following a six-monthly
Netherlands. clinical consultation schedule: findings from a cohort study
in rural Malawi. Journal of the International AIDS Society.
54. Pasipamire L, Nesbitt RC, Ndlovu S, Sibanda G, Mamba
2018;21(11):e25207-e.
S, Lukhele N, et al. Retention on ART and predictors of
disengagement from care in several alternative community- 67. Kwarisiima D, Kamya MR, Owaraganise A, Mwangwa F,
centred ART refill models in rural Swaziland. Journal of the Byonanebye DM, Ayieko J, et al. High rates of viral suppression
International AIDS Society. 2018;21(9):e25183. in adults and children with high CD4+ counts using a
streamlined ART delivery model in the SEARCH trial in rural
55. Naslund JA, Dionne-Odom J, Junior Destiné C, Jogerst KM,
Uganda and Kenya. J Int AIDS Soc. 2017;20(Suppl 4):21673.
Renold Sénécharles R, Jean Louis M, et al. Adapting and
Implementing a Community Program to Improve Retention in 68. Jakubowski A, Kabami J, Mwai D, Snyman K, Clark T, Ayieko J, et
Care among Patients with HIV in Southern Haiti: “Group of 6”. al. Provision of streamlined HIV care associated with reduced
AIDS research and treatment. 2014;2014. economic burden of care-seeking among HIV-infected adults.
AIDS 2016; 2016 July 18-22; Durban, South Africa2016.
56. Pellecchia U, Baert S, Nundwe S, Bwanali A, Zamadenga B,
Metcalf CA, et al. “We are part of a family”. Benefits and 69. Shade SB, Osmand T, Mwangwa F, Owaraganise A, Clark TD,
limitations of community ART groups (CAGs) in Thyolo, Malawi: Charlebois ED, et al. Cost of streamlined HIV care in rural
a qualitative study. J Int AIDS Soc. 2017;20(1):21374. Kenyan and Ugandan clinics in the SEARCH Study. CROI 2017;
2017 Feb 13-16; Seattle, USA.
57. Nkwemu S, Mwamba C, Bolton-Moore C, Topp SM, Jere L,
Mukumbwa-Mwenechanya M, et al. Challenges in providing 70. Cassidy T, O’Connell R, Euvrard J, Ndlovu S, Duran LT, Steele
patients centred care through community adherence groups SJ, et al. A “quick pick-up” differentiated model of ART delivery
in three provinces of Zambia. AIDS 2018; 2018 June 23-27; shows good retention in care. AIDS 2018; 2018 July 23-17;
Amsterdam, the Netherlands. Amsterdam, the Netherlands.
58. Mwamba D, Herce M, Mwanza MW, Thulani R. Community 71. Bosomprah S, Zulu I, Herce M, Mulenga L, Shah MP, Sikazwe
Adherence Group (CAG) for HIV viremic patients: Early lessons I, et al. Differentiated service delivery for HIV care: The fast
learnt from Lusaka, Zambia. IAS 2019; 2019 Jul 21-24; Mexico track experience from Zambia. CROI 2020; 2020 March 8-11;
City, Mexico. Boston USA.
59. Ssonko C, Gonzalez L, Mesic A, da Fonseca MS, Achar J, 72. Mody A, Roy M, Sikombe K, Savory T, Holmes C, Bolton-
Safar N, et al. Delivering HIV care in challenging operating Moore C, et al. Improved Retention With 6-Month Clinic
environments: the MSF experience towards differentiated Return Intervals for Stable Human Immunodeficiency Virus-
models of care for settings with multiple basic health Infected Patients in Zambia. Clinical Infectious Diseases.
care needs. Journal of the International AIDS Society. 2018;66(2):237-43.
2017;20(S4):21654.
73. Jere L, Cheelo C, Mukumbwa-Mwenechanya M, Mwamba C,
60. Kagimu D, Egessa J, Oucul L, Senyonga E. Overcoming barriers Nkwemu S, Sharma A, et al. Healthcare workers’ and patients’
to access of HIV/AIDS services among female sex workers experiences during implementation of the Fast Track model:
through differentiated service delivery models, TASO Entebbe Piloting differentiated ART service delivery in Zambia. AIDS
experience. AIDS 2018; 2018 Jun 23-27; Amsterdam, the 2018; 2018 July 23-27; Amsterdam, the Netherlands.
Netherlands.
74. Christ B, van Dijk J, Ballif M, Nyandoro T, Reichmuth ML,
61. Babigumira JB, Castelnuovo B, Stergachis A, Kiragga A, Shaefer Mukondwa WR, et al. Differentiated antiretroviral therapy
P, Lamorde M, et al. Cost Effectiveness of a Pharmacy-Only delivery in rural Zimbabwe: availability, needs and challenges.
Refill Program in a Large Urban HIV/AIDS Clinic in Uganda. OSF Preprints2020.
PLOS ONE. 2011;6(3):e18193.
75. Assefa T, Melaku Z, Amdino W, Abebe A, Rabkin M, Hartsough
62. Alamo ST, Wagner GJ, Ouma J, Sunday P, Marie L, Colebunders K, et al. Implementation of appointment spacing model of
R, et al. Strategies for optimizing clinic efficiency in a differentiated care in Ethiopia-successes and challenges. AIDS
community-based antiretroviral treatment programme in 2018; 2018 July 23-27; Amsterdam, the Netherlands.
Uganda. AIDS Behav. 2013;17(1):274-83.
76. Hubbard J, Phiri K, Moucheraud C, McBride K, Bardon A,
63. Obua C, Kayiwa J, Waako P, Tomson G, Balidawa H, Chalker J, et Balakasi K, et al. A Qualitative Assessment of Provider and
al. Improving adherence to antiretroviral treatment in Uganda Client Experiences With 3- and 6-Month Dispensing Intervals
with a low-resource facility-based intervention. Glob Health of Antiretroviral Therapy in Malawi. Glob Health Sci Pract.
Action. 2014;7:24198. 2020;8(1):18-27.
64. Kaimal A, Castelnuovo B, Atwiine M, Musomba R, Nabaggala 77. Phiri K, McBride K, Siwale Z, Hubbard J, Bardon A, Moucheraud
MS, Ratanshi RP, et al. Experiences with retention in care and C, et al. Provider experiences with three- and six-month
viral suppression in a pharmacy refill program. CROI 2017; antiretroviral therapy dispensing for stable clients in Zambia.
2020 February 13-16; Seattle USA. AIDS Care. 2020:1-7.

20 Differentiated service delivery for HIV treatment

78. Mantell J, Assefa T, Molla M, Zech J, Block L, Deguma D, et decongestion through community-based individual drug
al. “You are not benefiting us by keeping us away”: Why do distribution (PODIs) in DRC. AIDS 2018; 2018 July 23-17;
some eligible people on antiretroviral therapy (ART) decline Amsterdam, the Netherlands.
to participate in Ethiopia’s appointment spacing model with
90. Dorward J, Msimango L, Gibbs A, Shozi H, Tonkin-Crine S,
6-month ART dispensing? AIDS 2020: Virtual; July 6-10 2020.
Hayward G, et al. Understanding how community antiretroviral
79. Bekolo CE, Diallo A, Philips M, Yuma J-D, Di Stefano L, Drèze delivery influences engagement in HIV care: a qualitative
S, et al. Six-monthly appointment spacing for clinical visits as a assessment of the Centralised Chronic Medication Dispensing
model for retention in HIV Care in Conakry-Guinea: a cohort and Distribution programme in South Africa. BMJ Open.
study. BMC Infectious Diseases. 2017;17(1):766. 2020;10(5):e035412.
80. Mesic A, Fontaine J, Aye T, Greig J, Thwe TT, Moretó-Planas 91. Hendriksz F, Dube T, Strydom B, Banoo S. Centralized
L, et al. Implications of differentiated care for successful ART dispensing and alternative pick up points increases access
scale-up in a concentrated HIV epidemic in Yangon, Myanmar. to ARVs in Zambia. IAS 2019; 2019 July 21-24; Mexico City,
Journal of the International AIDS Society. 2017;20(Suppl 4). Mexico.
81. Ssonko C, Gonzalez L, Mesic A, da Fonseca MS, Achar J, 92. Kambai Avong Y, Ayodele Kayode G, Jatau B, Bosede Avong E,
Safar N, et al. Delivering HIV care in challenging operating Adekanmbi V, Abimiku A, et al. Providing antiretroviral therapy
environments: the MSF experience towards differentiated outside the hospital in a low-resource setting: a pilot study. The
models of care for settings with multiple basic health care Lancet Global Health. 2018;6:S3.
needs. J Int AIDS Soc. 2017;20(Suppl 4):21654.
93. Ssuuna M, Nakade S, Kabanda J, Zalwango S, Okello DA,
82. Kim MH, Wanless RS, Caviness AC, Golin R, Amzel A, Ahmed S, Mugara C, et al. Community private pharmacy ART Refill in
et al. Multimonth Prescription of Antiretroviral Therapy Among Kampala, Uganda. CROI 2020; 2020 March 8-11; Boston USA.
Children and Adolescents: Experiences From the Baylor
94. Shoopala N, Baughman AL, Mengistu AT, Mitruka K, Woelk G,
International Pediatric AIDS Initiative in 6 African Countries.
Mutandi G, et al. Outcomes of community-based antiretroviral
Journal of Acquired Immune Deficiency Syndromes. (1999).
treatment program in Namibia. CROI 2020; 2020 March 8-11;
2018;78 Suppl 2(Suppl 2):S71-S80.
Boston USA.
83. Bacha JM, Aririguzo LC, Mng’ong’o V, Malingoti B, Wanless RS,
Ngo K, et al. The Standardized Pediatric Expedited Encounters 95. Pasipamire L, Nesbitt RC, Ndlovu S, Sibanda G, Mamba
for ART Drugs Initiative (SPEEDI): description and evaluation S, Lukhele N, et al. Retention on ART and predictors of
of an innovative pediatric, adolescent, and young adult disengagement from care in several alternative community-
antiretroviral service delivery model in Tanzania. BMC Infect centred ART refill models in rural Swaziland. J Int AIDS Soc.
Dis. 2018;18(1):448. 2018;21(9):e25183.

84. Okoboi S, Ding E, Persuad S, Wangisi J, Birungi J, Shurgold S, et 96. Matambo T, Mwongera C, Wilkinson L, Van Cutsem G,
al. Community-based ART distribution system can effectively Bauernfeind A, Metcalf C, et al. Targeted adherence strategies
facilitate long-term program retention and low-rates of death for provision of cross border antiretroviral therapy (ART) to
and virologic failure in rural Uganda. AIDS Res Ther. 2015;12:37. migrant farm workers in Musina, South Africa. AIDS 2012; 2012
July 22-27; Washington, USA.
85. Mpima D, Birungi J, Makabayi R, Kanters S, Luzze C. Community
Antiretroviral Therapy (ART) delivery models for high patient 97. Jaffar S, Amuron B, Foster S, Birungi J, Levin J, Namara G, et
retention and sustaining good adherence: The AIDS Support al. Rates of virological failure in patients treated in a home-
Organisation (TASO) operational reseach findings, CDC/ based versus a facility-based HIV-care model in Jinja, southeast
PEPFAR funded project in Uganda. IAS 2013; 2013 30 June-3 Uganda: a cluster-randomised equivalence trial. Lancet.
July; Kuala Lumpur, Malaysia. 2009;374(9707):2080-9.

86. Vu L, Waliggo S, Zieman B, Jani N, Buzaalirwa L, Okoboi S, et 98. Selke HM, Kimaiyo S, Sidle JE, Vedanthan R, Tierney WM, Shen
al. Annual cost of antiretroviral therapy among three service C, et al. Task-shifting of antiretroviral delivery from health care
delivery models in Uganda. Journal of the International AIDS workers to persons living with HIV/AIDS: clinical outcomes of
Society. 2016;19(5 Suppl 4):20840-. a community-based program in Kenya. J Acquir Immune Defic
Syndr. 2010;55(4):483-90.
87. Vogt F, Kalenga L, Lukela J, Salumu F, Diallo I, Nico E, et al. Brief
Report: Decentralizing ART Supply for Stable HIV Patients to 99. Geldsetzer P, Francis JM, Sando D, Asmus G, Lema IA, Mboggo
Community-Based Distribution Centers: Program Outcomes E, et al. Community delivery of antiretroviral drugs: A non-
From an Urban Context in Kinshasa, DRC. JAIDS Journal of inferiority cluster-randomized pragmatic trial in Dar es Salaam,
Acquired Immune Deficiency Syndromes. 2017;74(3). Tanzania. PLOS Medicine. 2018;15(9):e1002659.

88. Kamerhe D, Kibwa Milenge P, Monga F, Thior I, Mwamba R, 100. Leisegang R, Calkins K, Cotton M, Cleary S, Karamchand S,
Canagasabey D, et al. Improved HIV treatment retention Hammann F, et al. HIV+ patients receiving ARVs through home
among patients enrolled in a differentiated care model at delivery: Does it matter? A Causal analysis. CROI 2020; 2020
Kenya General Reference Hospital in Haut Katanga province March 8-11; Boston, USA.
of the Democratic Republic of the Congo. IAS 2019; 2019 July
101. Louw JM, Rantloane B, Ngcobo S, Brey Z, Hugo J, Basu D, et al.
21-24; Mexico City, Mexico.
Home delivery of medication as part of reducing congestion
89. Mothibi E, Mangoma J, Kabamba D, Monteith L, N’siesi FX, in primary healthcare in Tshwane District Health Services.
Firth J, et al. Improving retention, viral suppression, and facility Southern African Journal of Public Health. 2020; 4(2).

Summary of published evidence (Nov 2020) 21

 DIFFERENTIATED
SERVICE DELIVERY
www.differentiatedservicedelivery.org

22 Differentiated service delivery for HIV treatment