Patient and health worker experiences of differentiated models of care for stable HIV patients in Malawi: A qualitative study
Patient and health worker experiences of differentiated models of care for stable HIV patients in Malawi: A qualitative study
Margaret L. Prust 1 2
Clement K. Banda 0 2
Katie Callahan 2
Rose Nyirenda 2
Frank Chimbwandira 2
Thokozani Kalua 2
Michael Eliya 2
Peter Ehrenkranz 2
Marta Prescott 1 2
Elizabeth McCarthy 1 2
Elya Tagar 2
Andrews Gunda 0 2
0 Clinton Health Access Initiative, Inc. , Lilongwe, Malawi, 3 HIV , TB and Health Financing Team, Clinton Health Access Initiative, Inc. , New York, NY , United States of America, 4 Department of HIV and AIDS, Ministry of Health , Lilongwe, Malawi, 5 Bill & Melinda Gates Foundation, Seattle, WA , United States of America
1 Applied Analytics Team, Clinton Health Access Initiative, Inc. , Boston, MA , United States of America
2 Editor: Anna Grimsrud, International AIDS Society , SWITZERLAND
Several models of differentiated care for stable HIV patients on antiretroviral therapy (ART) in Malawi have been introduced to ensure that care is efficient and patient-centered. Three models have been prioritized by the government for a deeper and broader understanding: adjusted appointment spacing through multi-month scripting (MMS); fast-track drug refills (FTRs) on alternating visits; and community ART groups (CAGs) where rotating group members collect medications at the facility for all members. This qualitative study aimed to understand the challenges and successes of implementing these models of care and of the process of patient differentiation.
Data Availability Statement: The qualitative code
report files used for this study contain identifying
information of the study participants. Furthermore,
even de-identified code report files could possibly
still be used to determine participant identities.
Therefore, these data are made available on
request, as approved by Malawi National Health
Sciences Research Committee, from the Clinton
Health Access Initiative
() or from the
corresponding author of the study
( or margaret.
A qualitative study was conducted as a part of a broader process evaluation in 30
purposefully selected ART facilities between February and May 2016. Semi-structured, in-depth
interviews with 32 health workers that managed and coordinated ART clinics and 30 focus
groups were held with 216 ART patients. Interviews and focus groups were audio recorded,
transcribed, and coded thematically.
Participants reported that the models of differentiated care have yielded key benefits,
including: reduced patients' travel and visit time, decongestion of facilities, and enhanced social
support. Participants suggested that these benefits could lead to improved HIV treatment
outcomes for patients. At the same time, some challenges were reported, such as
inconsistent stocks of drugs, which can inhibit implementation of MMS. For CAGs, the group-based
nature of the model presented some unique problems, such as conflicts within groups or
Funding: The process evaluation was made
possible through funding support from the Bill &
Melinda Gates Foundation (BMGF) (OPP1134641).
More information about BMGF is available at
https://www.gatesfoundation.org. The BMGF grant
manager provided input into study design, data
collection, and manuscript development. However,
the views expressed in this paper are the opinions
of the study investigators and do not necessarily
reflect the official policies of BMGF.
Competing interests: The authors have declared
that no competing interests exist.
concerns about privacy. Health workers also described some of the reasons why eligible
patients may not receive the models or conversely why ineligible patients sometimes get the
Documenting patient and health worker perspectives on models of differentiated care is
critical to understanding and improving these models. While these models can offer important
benefits, the models may not be appropriate for all sites or patients, and patient status and
needs may change over time. Key challenges should be recognized and addressed for
optimal utilization of the models.
Not all HIV treatment patients have the same types of needs, and the movement toward
differentiated models of HIV care acknowledges that, by tailoring services to different patient types,
it may be possible to improve service quality and efficiency from both the patient and health
system perspective [
]. Previous guidelines for antiretroviral therapy (ART) were developed
at a time when a large portion of the ART patient population had severe clinical manifestations
of disease and many ART drugs had serious side effects. However, several factors have lead to
an increase in the proportion of clients on HIV treatment that are stable and healthy.
Specifically, eligibility to start HIV treatment has expanded significantly in recent years , ART
programs in many countries have been operating for many years and are very established, and
ART drugs have improved, leading to fewer side effects. In this context, models of
differentiated care are an important tool to ensure the provision of patient-centered care [
In Malawi, several models have been developed to provide more streamlined services to
patients that are stable. The MOH and other in-country stakeholders sought to understand
more about three models of differentiated care that were already widespread or had the
potential for national scale-up. These included two facility-based models focused on individual
clients, multi-month scripting (MMS) and fast-track refills (FTR), as well as one group-based,
community-level model, community ART groups (CAGs). In the MMS model, stable patients
are given three-month refills rather than one-month refills. For FTRs, stable patients are also
given three-month refills, but only two of their four visits are required to be clinical visits with
a nurse or doctor. The other two visits in the year are refill-only visits where medications are
dispensed by a lower-level health worker. CAGs, which have also been launched in other
], are peer-led groups that meet monthly at the community level for ARV distribution
and peer-led discussions. Each month a different group member visits the facility to pick up
ARV refills for the entire group, with each person having at least two opportunities for a
twiceannual clinical visit.
Currently in Malawi, patients are eligible for MMS if they: are 18 years or older, have been
on ART at least six months, are on first-line ART, have no adverse drug reactions or
opportunistic infections, have a viral load less than 1000 copies per mL and have good adherence.
Patients are eligible for FTRs and CAGs if they meet those same criteria and are also not
pregnant or lactating.
Systematic assessments of the implementation of these models in Malawi were lacking, and
more evidence was needed about the models to inform decision-making around model
scaleup and improvement. Through a process evaluation, we sought to assist policy makers in
2 / 15
defining the existing models, understanding the extent to which patients in Malawi were
differentiated according to clinical stability criteria, and analyzing the characteristics and costs of
the models of differentiated care they received. The purpose of this paper is to describe the
qualitative component of the process evaluation that explored patients and provider
perspectives on the key benefits and challenges associated with models of differentiated care for stable
In other settings, researchers have explored the acceptability of differentiated models of
care from the patient and health worker perspective. CAGs in Mozambique, similar to those
implemented in Malawi, were suggested by patients in 2012 to offer cost and time savings,
improved ART access, and enhanced peer support [
], and in 2015, CAG participants in
Malawi reported similar benefits, along with implementation challenges related to low
awareness about CAGs, patient disclosure concerns, and conflicts among CAG members [
Another model for integrated HIV and non-communicable disease adherence clubs has been
found to be generally acceptable to patients in Kenya, but challenges with recruitment, patient
understanding of the model, and practical implementation of the program emerged through
interviews and focus groups [
]. Understanding and documenting these challenges, along
with patient and provide input on potential benefits and general acceptability, is important for
expanding and improving these models of care. There has been limited qualitative research
focused on the FTR model and the somewhat more simple, but widespread, MMS model, and
patient and provider perspectives on this model are critical to understand.
This qualitative study was part of a broader mixed methods process evaluation conducted in
30 purposefully-selected facilities [
] to assess the government-led implementation of these
models of care. These models were implemented by the government and partners, and the
purpose of this observational was to describe and understand the current models better. For
the qualitative aspect of the study, data was collected between February and May 2016 by two
teams of five data collectors trained in qualitative methods.
This study received ethical approval from the Malawi National Health Sciences Research
Committee (approval number 1513) and from the Chesapeake IRB in the United States (approval
We used heterogeneity sampling to achieve diversity of facilities in the sample based on: model
of differentiated care offered, region, patient volume, facility owner (i.e. MOH or Christian
Health Association of Malawi [CHAM]), facility type, and implementing partner support.
Within each facility, one interview was conducted with the health worker in charge of the ART
clinic. Health workers were therefore sampled based on those who held the specific title ªART
in-chargeº. In two sites, an implementing partner-supported clinical mentor was also
interviewed separately. One patient focus group was conducted in each of the 30 study sites.
Among the 30 sites, all were implementing MMS, 8 were implementing CAGs along with
MMS, and 4 were implementing FTRs with MMS.
The facility staff members were asked to support recruitment of four to eight patients from
among those that were visiting for a regularly-scheduled ART appointment. Health workers
were asked to refer all willing and eligible patients to the study team at the end of their
appointment, until we reached the point when we had a sufficient number of participants. Patients
3 / 15
were eligible for inclusion in the focus group if they received one of the relevant ART models
of differentiated care offered at their health facility (though since some facilities were offering
multiple models, participants were included that were accessing one model but not all models
offered by that site). In conversations with health workers, the study team emphasized that the
questions were about the models of care and did not constitute a staff or facility performance
review, so as to reduce potential for health workers to bias the sample by choosing only
patients that would provide positive input.
Interviews and focus groups asked questions about the model process, health worker and
patient satisfaction with the models, and any challenges faced. Interviews were conducted in
either English or local languages. Focus groups were conducted in local languages (Chichewa,
Tumbuka and Yao). Written informed consent was obtained for all participants, and
participants were not compensated.
All interviews were documented through field notes and audio recordings. Achievement of
theoretical saturation [
] was assessed through daily debriefings of individual data collection
teams and several collective debriefings with the whole study team to discuss findings and
areas lacking in clarity. Theoretical saturation for each of the three models of care was
specifically assessed in terms of whether any new concepts were continuing to appear or required
further development, and responses unique to each model of care were assessed individually.
Additionally, transcription of audio files and initial analyses were carried out in parallel with
the data collection, allowing study investigators to provide feedback to data collectors on areas
requiring further exploration to achieve theoretical saturation. The study team made plans to
potentially expand data collection if theoretical saturation was not reached by end of planned
data collection in the 30 sites, but this was determined not to be necessary based on study
procedures described here to assess theoretical saturation.
Audio recordings from the interviews and focus groups were transcribed. Sessions that took
place in a language other than English were simultaneously translated into English. Three
members of the research team systematically coded the transcripts, using the constant
comparative method [
]. Initially, codes were established based on themes explored in the
interview guides (such as satisfaction or challenges with the models or drug storage) and then
during the coding process additional codes or sub-codes were added based on the key concepts
emerging from participant responses. The same coding structure was used for both health
worker interviews and patient focus groups but some themes were only explored in interviews
(such as patient data systems and implementing partner support). The coders independently
coded batches of transcripts and then met to agree upon standard themes and codes and
ensure inter-coder reliability. In this way, half of the transcripts were double- or triple-coded
in batches and the coding key was adjusted iteratively. When coding was consistent and no
emergent themes were identified in new transcripts, the code structure was finalized. The
remaining transcripts were coded by one coder and the final code structure was applied to all
transcripts coded early in the process.
Dedoose software was used to facilitate coding, organization, and retrieval of information.
The research team interpreted the findings according to thematic area and wrote code
summary reports to document key findings. In an effort to synthesize a large volume of qualitative
data and support policy makers in prioritizing the most critical information, we developed a
system for weighting the challenges and benefits of the models. In many cases, the presentation
4 / 15
of results from qualitative data in a quantitative format is not appropriate given that sampling
in qualitative studies often is not designed to be broadly representative; however in large
datasets or for qualitative data sampled from a broader study sample, it can be appropriate to
consider code frequencies as a valuable measure [
]. In this study, the qualitative data has been
organized and prioritized based on the number of participants raising the issue and the degree
to which the item had the potential to influence the overall success of the model and the goals
of the ART program. Specifically, each attribute was assigned a ªfrequency weightº, which was
a measure of how often a particular issue was raised or how many participants shared the view
in interviews with health care workers or focus groups with patients. A frequency weight of
three reflected that the issue was raised in more than half of interviews or focus groups in sites
offering the model. A frequency weight of two or one reflected that the issue was raised in 25±
50% or less than 25% of sessions in sites offering the model, respectively. Additionally, each
attribute was assigned an ªimpact weightº, which was a measure applied by the study team in
consultation with the MOH (as opposed to study participants) of the degree to which the
MOH believed that this item had potential to influence the overall success of the models and
its goals. An overall weight was calculated as the sum of the patient and health worker
frequency weights multiplied by the impact weights. These weights were used to help
policymakers interpret the information from this study and prioritize areas for action.
The characteristics of participants in the final sample for each data collection method are
shown in Table 1. Overall, we held 32 ART in-charges interviews and 30 focus group
discussions with 216 patients. 43.1% of our focus group sample was male, compared to 36.0%
of all ART patients ever enrolled in care in Malawi as of March 2016 according to government
]. The characteristics of the facilities in the sample are reported elsewhere [
Successes and challenges of observed models of differentiated care
The feedback received on the MMS, FTR and CAG models was generally positive and
participants could confidently testify how beneficial models had been, but also had suggestions about
aspects of implementation that needed further improvements.
Multi-month scripting. The most commonly discussed advantages of the MMS model
were related to reduced burden on patients and health workers. From the health worker
perspective, MMS helped to lessen workload and decongest the facilities, while still providing an
appropriate level of care to patients. For patients, MMS reduced travel costs and time spent at
appointments and alleviated issues with absenteeism from work for clinic appointments:
ªThey started giving me three bottles, it makes me happy because I think about the distance
I was travelling. From where I stay to here we pay 1000 kwacha [approximately US$1.60]
on public transport. . . Now they are giving me three bottles so I find time to rest.º±Patient,
Many participants also suggested that by reducing the burden on patients, MMS can help
improve adherence and retention. Patients described experiences of knowing that they had an
appointment but not being able to go to the clinic because of challenges with missing work,
impassable roads in rainy season, not having money for transport or other logistical problems.
When refills were longer, patients explained that they could more feasibly maintain the
designated appointment schedule. Some health workers thought that MMS encouraged good
adherence because patients were aware that if they come to appointments regularly and took
medication as advised, then they would be able to get a three-month refill:
ª[MMS] promotes good adherence because they know that if they have a poor adherence, it
means they will be taking like one month supply. It means they will be coming to clinic
frequently so most of the time the multi-month refills help a lot on good adherence.º±Health
worker, Central Region
At the same time, some health workers suggested that MMS could cause patients to be
more likely to miss appointments because of the long length of time between appointments.
ªThere is an increase in the number of defaulters. . . because the patients are given a lot of
drugs and they end up forgetting their appointment date.º±Health worker, Southern
Patients reported that in addition to promoting good adherence, MMS can also encourage
patients not in care to seek services. As explained by the following patient, her husband is
unlikely to get tested and start ART if he thinks that HIV treatment will involve monthly
ªIt should be 3 months because my husband really wants to come for HIV test, but he fail[s]
because he is afraid of his boss. So [if] the time to refill drugs gets extended, he can come for
a test and start on ART. It can really help.º±Patient, Central Region
6 / 15
Most patients in the focus groups explained that they felt comfortable and knew when to
access more care if they needed it. Specifically, patients reported that if they felt sick between
ART appointments, they would come to the clinic even if they still had ARVs remaining to get
care for their symptoms. For example:
ªWhen you get your three months medication and you have a problem even before your
appointment date you have to come to the hospital that is what we are told, that we can be
given many months but if you get sick you have to come to the hospital so that you get
helped.º±Patient, Central Region
Although many patients reported that they would return to the clinic if they got sick
between ART appointments, some health workers expressed a concern that patients were not
coming back to the clinic promptly to report any problems, but rather were waiting until their
next appointment date:
ªWe give health talks on the importance of attending OPD [out-patient department] but
most of them they prefer to stay at home while sick waiting for the next appointment. . .
They say they don't want to visit the hospital regularly so they choose to remain at home
until they finish the medication.º±Health worker, Central Region
Other challenges with implementing the MMS model were also described by patients and
health workers. The most common challenge described was related to the perceived availability
of ARVs and cotrimoxazole. While it is not clear whether the MMS model may have
influenced stock security, participants frequently pointed out that limited drug stocks prevented
full implementation of the MMS model. A patient in the Southern Region reported,
ªsometimes they give us for 2 months. . . they tell us there is low stock of drugs.º Based on feedback
from health workers, it seems that the most commonly prescribed regimen is nearly always in
stock; but many facilities experienced low stocks of other first-line regimens and second-line
regimens. Additionally, patients are supposed to receive cotrimoxazole prophylaxis along with
their ARVs, but stockouts and low stock levels of cotrimoxazole frequently meant that patients
either did not receive cotrimoxazole or received only a small amount.
ªLike the [cotrimoxazole], that affects the amount of drugs we dispense to the client
sometimes if we don't have in our stock, they go with nothing in terms of the cotrimoxazole
sometimes when we have few, now we have to reduce the months, the number of months
they are going to get, say a person has come for a three-months visit, they get ARVs for 3
months and cotrimoxazole for 1 month.º±Health worker, Northern Region
While some patients report returning to the facility for more cotrimoxazole even if they
have ARVs, other patients simply go without the cotrimoxazole. A few patients described
using traditional medicines when cotrimoxazole was not provided.
Another challenge is that the policies about MMS may be implemented differently across
facilities, creating frustrations for patients that transfer between facilities. There were several
accounts of patients who switched from one ART clinic to another and were confused or
frustrated by differences in refill length across facilities. Several patients described receiving
threemonth refills at one clinic, but then transferring to a new clinic where the longest refills offered
were for two months.
Some stakeholders in Malawi had concerns about whether giving larger quantities of
medications at the same time would increase the possibility of medications being misused by
7 / 15
patients, so participants were prompted to talk about how they stored their medications and
whether they had ever heard of cases of ART drugs being sold or used for any purpose other
than as prescribed. Patients described taking precautions to safeguard their medications by
storing them in a safe place. Patients were particularly careful to make sure that children did
not have access to tamper with the medications. Most participants had never heard of any
form of misuse of ART drugs in their community and even responded that they were unable
to imagine how or why this would happen.
ªThis medicine being sold? I have never heard. Maybe the other medicine [other than
ARVs] and the one who sells is not sick. . . like me here I should take my medicine and sell?
That means I want death.º±Patient, Northern Region
While there were no reported cases of patients selling ARVs, two participants reported that
health workers may be involved in selling drugs outside of health facilities. Patients suggested
that illegal sales of drugs must be supported by health workers because the drugs were available
in full cartons that individual patients would not have access to. A few participants reported
having heard rumors that ARVs were being used in brewing traditional beer to make it strong
and given to domestic animals to make them grow quickly and be strong. These rumors could
not be verified and were only reported by a couple of the participants.
Fast-track refills. Because FTRs are implemented alongside or in addition to MMS, the
benefits and challenges of MMS apply also to the FTR model. In addition to the topics
described in relation to MMS, there were several other points that participants brought up that
are unique to the FTR model. For example, some participants from FTR sites felt that the FTR
model further reduced the workload for clinical staff such as clinicians and nurses. The model
allowed those clinical staff to spend more time with patients that have a complaint or problem
and need more care:
ªNow the system is working very well and has reduced congestion to the clinician because
previously we were meeting almost 60 or 70 or even 80 something [established] patients. . .
but now with [FTR] the clinician does meet maybe 40 patients or 35 plus new patients.º±
Health worker, Southern Region
At the same time, the FTR can reduce waiting time for patients. Whereas patients
previously had to wait for long period to see the clinical health workers, they now received services
at some appointments from other cadres of health workers with reduced waiting time. But
even with reduced waiting times, patients like this one report waiting up to 90 minutes for a
fast-track refill and longer for regular appointments:
ªWhen we come and [are] not meeting the doctor, it is faster than when we are meeting the
doctor. Because you can get here at half past six, that means by the time they are opening at
eight o'clock you will be the first to collect drugs. Within ten minutes you are done. But the
day we are meeting nurses we take very long.º±Patient, Southern Region
While some patients described the wait times as being shorter when they did not have to
see the clinician or nurse, other patients talked about how the wait time to see a Health
Surveillance Assistant (HSA) was still long. In particular there were complaints about the HSA
arriving later than the specified opening time, though it is possible that the same could occur if
services were provided by another health worker.
8 / 15
ªThe challenge is the fact that you come as early as 5 o'clock. . . but they don't open on good
time to start dispensing drugs. Sometimes you find that. . . the one who dispenses drugs
hasn't yet arrived so you come early and you still leave late.º±Patient, Southern Region
Community ART groups. CAGs transfer some aspects of care from the facility to the
community level, so that patients do not have to visit the ART clinic as frequently. For patients
in rural areas that may live far away from the clinic, this was a major benefit. In addition to
serving as an ARV delivery mechanism, the CAGs serve an important role in terms of
increasing social support among people living with HIV. One patient from the Southern Region
described the relationships within the CAG as follows: ªWe are like one family, we advise each
other and encourage each other.º In addition to providing social support, the CAGs can form
the foundation for other types of material support among members. This is not a formal part
of the CAG program and is not promoted by facility staff, but can occur as the initiative of the
ªWe meet twice a month, we contribute money for emergencies if one of us is sick we go
visit. And help with food or transport to the hospital . . . It is going well and we rest now
and work normally without coming here often.º±Patient, Southern Region
While many patients reported positive perceptions of and experiences with CAGs, the
model was not a good fit for every patient. For example, participants in the focus groups and
interviews talked about some challenges with CAGs, such as concerns about joining a CAG
due to privacy and confidentiality issues.
ªI am not in a group. I feel there is no privacy. When you go to the groups everyone knows
you are on medication which makes it very difficult for us with the discrimination.º±
Patient, Southern Region
The fact that some patients had concerns about joining a CAG made it difficult for others
to form a group in some cases. Some individuals were eager to be a part of a CAG but they did
not know others in their community with HIV or were not willing to approach others about
forming a group:
ªI would [join a CAG] but we don't have a lot of people who are willing to start in our
village. People who are interested are not from our village, three from another village and two
from the other. . . I think most people hide themselves.º±Patient, Southern Region
Even for patients that were willing to join a CAG, they sometimes faced problems once the
group was established. As with any group-based process, there were sometimes inter-personal
conflicts or issues among members. One patient described how some members of the group
did not follow the guidelines and created problems for other members:
ªSometimes we have problems like other people when it is their turn to collect drugs for the
rest of the members in the group; they send their children to collect for them which is not
allowed.º±Patient, Southern Region
Some patients also made statements that indicated that they may not have fully understood
the guidance provided to them about how the CAG should work. According to the CAG
protocols, if a member of the group is feeling sick, then that member should go to the clinic for
9 / 15
the appointment regardless of whether it is their turn. If needed, that patient may be
accompanied by the person whose turn it is for support. However, several patients described one of the
benefits of CAGs being that they did not have to go to the clinic when they felt too sick to
travel. This indicates that patients have not fully understood the guidance provided about
ªWe think the groups are good. One of us could get sick and still receive the drugs without
going to the hospital. . . Today the one who was supposed to come is not feeling well, so I
volunteered to come and collect the drugs.º±Patient, Southern Region
Other patients recognized that if a group member was sick on or near the appointment
date, that member was supposed to go for the appointment, and if that patient required help to
get to the facility, a group member could accompany them. Still, some health workers
expressed concerns about the reduced frequency of monitoring or seeing patients at the ART
clinic. Despite the role that the rest of the group plays in supporting each other to recognize
symptoms of treatment failure and get care, there continue to be some concerns that patients
may develop problems or have reduced adherence and that these problems would fail to be
recognized by health workers until they had progressed to a severe stage.
Prioritization of benefits and challenges. We prioritized the key advantages and
challenges associated with each model as shown in Table 2. This information should not be used
to compare across models, but rather to prioritize issues for consideration regarding each
Process of patient differentiation
Regardless of the model being offered at a particular facility, these models share the
characteristic of being provided differentially to patients determined to be stable. Across all models, it
was acknowledged that there are some challenges in differentiating patient groups and
ensuring that only those that are eligible receive the model. We explored reasons why ineligible
patients were included in the models. The most common reason suggested by health workers
was that patients often make special requests, especially for MMS, if they live a long distance
from the facility or are travelling. This is allowed by the guidelines; the MOH ART guidelines
state that ªin exceptional cases (e.g. international travel) up to 6 or even 12 months of ARVs
can be dispensedº [
], though the guidance is not specific as to the extent of situations that
may qualify as exceptional. In many cases, were asked in interviews to describe the eligibility
criteria for these models of care, they provided incorrect or incomplete responses, indicating
that health worker knowledge of these eligibility criteria may be incomplete. In other cases,
health workers described the desire to reduce congestion or workload in facilities as a key
factor in the decision of enrolling patient in these differentiated models of care. For example, one
in-charge described how their facility uses a quota system for each ART day and may give
patients longer refills if they meet that quota.
ªThe other issue is about how many patients you have booked for a particular day . . . If you
think you need to see 100 patients in a day, as you reach of that number, you may think of
changing this particular patient. Instead of giving 2 months, I will give him 3 months so
that we relieve those that will be working on that particular day.º±Health worker, MMS
only site in Central Region
Finally, ineligible patients may receive these models if a transition from being eligible to
ineligible is not recognized. If patients have previously been included in a model, it was
10 / 15
Frequency weighting2 by
Frequency weighting2 by Impact weighting3 according to
patients policy makers
11 / 15
suggested that some health workers may continue to include them without assessing eligibility
at each visit.
Models of differentiated care are being successfully implemented in Malawi and patients and
health workers reported that these models have yielding key benefits. All three models
investigated in this study, MMS, FTRs and CAGs, can reduce the burden on health workers and
patients. For patients, travel time and visit time can be reduced by fewer or shorter visits to the
ART clinic. For health workers, seeing individual patients less frequently can reduce
congestion in facilities. At the same time, some challenges have been faced in attempting to
implement these models. For MMS and FTRs, inconsistent stocks of ART and other supplemental
drugs can restrict the quantity of medications provided at each visit thereby inhibiting
implementation of the models. For CAGs, the group-based nature of the model presented some
unique problems, such as conflicts within groups or patient concerns about privacy. Other
published literature highlights similar challenges and benefits to the CAG model in Malawi [
and Mozambique [
] though previous studies have not pointed specifically to challenges
related to transitioning patients out of the model if they become unstable or ineligible or to the
issue of limited resources for supervision and implementation. This latter issue on resources
for supervision has likely arisen in the more recent past since implementing partners have
attempted to transition supervision of the model to the government. For all models, some
health workers mentioned concerns about limited ability to identify treatment failure or side
effects when patients are seen less frequently and about whether patients would report to the
facility if they experienced any adverse events before their next appointment.
These results should be viewed in light of several strengths and weaknesses. This study
captured a wide range of in-depth perspectives on three differentiated models of care from
12 / 15
patients and health workers in different types of facilities. Whereas other studies cited in this
work have tended to explore only one model of differentiated care at a time, this study was
able to bring perspectives on multiple models together. At the same time, the facilities and
participants included in this study were selected purposefully and cannot necessarily be
considered to be nationally-representative. Additionally, it is important to note that these findings
may not be generalizable outside of the Malawian context. Policy makers should consider the
lessons of this qualitative study together with other types of evidence regarding differentiated
models of care.
These results have implications both for research and practice in relation to differentiated
care. In order to continually improve the implementation of these and similar models,
mechanisms should be established to identify and address challenges unique to each model. In
Malawi, challenges such as low stocks of ARVs and cotrimoxazole and limited resources for
monitoring of CAGs should be considered by program managers since planning and proper
supervision may be effective in reducing these issues [
]. Additionally, all models were
impacted by challenges with accurately differentiating patients and recognizing changes in
status. As suggested by other experts, the expansion of access to viral load testing may improve
differentiation of patients [
]. Improvements in data system structure may also support
health workers to differentiate patients.
A key question for researchers is related to the adequacy of patient outcomes for those
included in these models. Participants suggested that by reducing the burden on patients to
remain in care, the models may help to improve retention in care. CAGs may play a unique
role in encouraging retention as well due to the social support provided by the group nature of
the model. Research suggests that patient outcomes in these models are acceptable [
determining the impact of these models on patient outcomes such as retention or viral
suppression is challenging due to the differences in the types of patients that are enrolled or not
enrolled in the models. A systematic review suggested that less frequent clinic visits may be
associated with improved treatment outcomes [
] but noted that the evidence base was sparse
and of low quality. Researchers have considered the retention and viral suppression of patients
in models similar to Malawi's CAGs in Mozambique [
], South Africa [
], and Uganda [
in an observational way. While rigorous analyses using a counterfactual should be explored to
understand the impact of these models, randomization of clients into models such as CAGs
may not be appropriate since client choice and motivation to join these models may be a
critical factor in their success [
]. While it remains a priority to refine methodologies to
rigorously assess the impact of these models, it is also important to continue to explore patient
perspectives on these models, since the very goal of the models is to deliver patient-centered
care. Future research should also focus on understanding perspectives of patients not enrolled
in the models and exploring discrepancies in the perspectives of patients and health workers
about these models. For example, patients reported returning to the clinic before their
appointments when they felt sick, but health workers believe that patients would delay returning to the
clinic. Further qualitative and quantitative investigations should assess these accounts since
both patient responsibility and health worker confidence are critical to the success of these
The authors would like to acknowledge the support of the Malawi MOH in carrying out the
process evaluation. We would also like to thank all study participants, including the patients
and staff of the participating health facilities, for sharing their invaluable perspectives on
models of differentiated care. We further extend our acknowledgment to MSF-France and
MSF13 / 15
Belgium for providing details and documentation on the FTRs and CAGs, respectively. We
express deep thanks to our study coordinator, Nurse Nyambi, as well as to the data collection
and data entry teams. We also appreciate the logistical support of the CHAI Malawi office
staff, particularly Alfred Kasito. Without the support of these individuals the study would not
have been possible and we warmly thank them for their contributions.
Conceptualization: Margaret L. Prust, Clement K. Banda, Katie Callahan, Frank
Chimbwandira, Thokozani Kalua, Michael Eliya, Peter Ehrenkranz, Elizabeth McCarthy, Elya Tagar,
Data curation: Margaret L. Prust, Clement K. Banda.
Formal analysis: Margaret L. Prust, Clement K. Banda, Katie Callahan, Marta Prescott.
Funding acquisition: Margaret L. Prust, Katie Callahan, Peter Ehrenkranz, Elizabeth
McCarthy, Elya Tagar, Andrews Gunda.
Investigation: Margaret L. Prust, Katie Callahan, Andrews Gunda.
Methodology: Margaret L. Prust, Marta Prescott, Elizabeth McCarthy.
Project administration: Margaret L. Prust, Clement K. Banda, Peter Ehrenkranz, Elizabeth
McCarthy, Andrews Gunda.
Resources: Peter Ehrenkranz, Elizabeth McCarthy, Andrews Gunda.
Software: Margaret L. Prust.
Supervision: Margaret L. Prust, Rose Nyirenda, Frank Chimbwandira, Thokozani Kalua,
Michael Eliya, Peter Ehrenkranz, Marta Prescott.
Validation: Rose Nyirenda, Frank Chimbwandira, Thokozani Kalua, Michael Eliya, Elizabeth
McCarthy, Elya Tagar, Andrews Gunda.
Writing ± original draft: Margaret L. Prust.
Writing ± review & editing: Margaret L. Prust, Clement K. Banda, Katie Callahan, Rose
Nyirenda, Frank Chimbwandira, Thokozani Kalua, Michael Eliya, Peter Ehrenkranz, Marta
Prescott, Elizabeth McCarthy, Elya Tagar, Andrews Gunda.
14 / 15
1. Global Fund to Fight AIDS , TB and Malaria. A Toolkit for Health Facilities: Differentiated Care For HIV and Tuberculosis . Geneva: GFATM; 2015 .
2. International AIDS Society (IAS). Differentiated Care for HIV: A Decision Framework For Antiretroviral . 2016 .
World Health Organization. Guideline on when to start antiretroviral therapy and on pre-exposure prophylaxis for HIV . Geneva: World Health Organization; 2015 .
World Health Organization. Consolidated Guidelines on HIV Prevention, Diagnosis, Treatment and Care for Key Populations . Geneva: World Health Organization; 2016 .
5. Grimsrud A , Bygrave H , Doherty M , Ehrenkranz P , Ellman T , Ferris R , et al. Reimagining HIV service delivery: the role of differentiated care from prevention to suppression . J Int AIDS Soc. 2016 Dec 1 ; 19 ( 1 ): 21484 . https://doi.org/10.7448/IAS.19.1.21484 PMID: 27914186
6. Duncombe C , Rosenblum S , Hellmann N , Holmes C , Wilkinson L , Biot M , et al. Reframing HIV care: putting people at the centre of antiretroviral delivery . Trop Med Int Health . 2015 ; 20 ( 4 ): 430 ± 47 . https:// doi.org/10.1111/tmi.12460 PMID: 25583302
7. Bemelmans M , Baert S , Goemaere E , Wilkinson L , Vandendyck M , van Cutsem G , et al. Communitysupported models of care for people on HIV treatment in sub-Saharan Africa . Trop Med Int Health . 2014 ; 19 ( 8 ): 968 ± 77 . https://doi.org/10.1111/tmi.12332 PMID: 24889337
8. Rasschaert F , Telfer B , Lessitala F , Decroo T , Remartinez D , Biot M , et al. A Qualitative Assessment of a Community Antiretroviral Therapy Group Model in Tete, Mozambique . PLoS One . 2014 ; 9 ( 3 ): e91544. https://doi.org/10.1371/journal.pone. 0091544 PMID: 24651523
9. Pellecchia U , Baert S , Nundwe S , Bwanali A , Zamadenga B , Metcalf CA , et al. "We are part of a family". Benefits and limitations of community ART groups (CAGs) in Thyolo, Malawi: a qualitative study . J Int AIDS Soc. 2017 Mar 28 ; 20 ( 1 ):1± 6 .
10. Venables E , Edwards JK , Baert S , Etienne W , Khabala K , Bygrave H . "They just come, pick and go." The Acceptability of Integrated Medication Adherence Clubs for HIV and Non Communicable Disease (NCD) Patients in Kibera, Kenya . PLoS One . 2016 Oct 20 ; 11 ( 10 ):e0164634. https://doi.org/10.1371/ journal.pone.0164634 eCollection 2016 . PMID: 27764128
11. Prust ML , Banda CK , Nyirenda R , Chimbwandira F , Kalua T , Jahn A , et al. Multi-month prescriptions, fast-track refills, and community ART groups: Results from a process evaluation in Malawi on using differentiated models of care to achieve national HIV treatment goals . J Int AIDS Soc. 2017 Jul 21 ; 20 ( Suppl 4 ): 41 ± 50 .
12. Morse J. The significance of saturation . Qualitative Health Research . 1995 ; 5 : 147 ± 149 .
13. Miles M , Huberman AM . Qualitative data analysis: an expanded sourcebook , 2nd Ed. Thousand Oaks, Calif: Sage Publications; 1994 .
14. Glaser B , Strauss A . The discovery of grounded research: Strategies for qualitative research . Chicago: Aldine; 1967 .
15. Namey E , Guest G , Thairu L , Johnson L . Data Reduction Techniques for Large Qualitative Data Sets . In: Guest G, MacQueen KM , editors. Handbook for Team-based Qualitative Research . Plymouth, United Kingdom: Altmira Press; 2007 .
16. Malawi Ministry of Health. Integrated HIV Program Report, January-March 2016 . Lilongwe: Government of Malawi; March 2016 .
17. Malawi Ministry of Health. Clinical Management of HIV in Children and Adults . 2nd ed. Lilongwe: Government of Malawi; 2014 .
18. Nsanzimana S , Remera E , Ribakare M , Burns T , Dludlu S , Mills EJ , et al. Phased implementation of spaced clinic visits for stable HIV-positive patients in Rwanda to support Treat All . J Int AIDS Soc. 2017 Jul 21 ; 20 ( Suppl 4 ): 21 ± 27 .
19. Phillips A , Shroufi A , Vojnov L , Cohn J , Roberts T , Ellman T , et al. Sustainable HIV treatment in Africa through viral-load-informed differentiated care . Nature . 2015 ; 528 ( 7580 ):S68± 76 . https://doi.org/10. 1038/nature16046 PMID: 26633768
20. Kwarisiima D , Kamya MR , Owaraganise A , Mwangwa F , Byonanebye DM , Ayieko J , et al. High rates of viral suppression in adults and children with high CD4+ counts using a streamlined ART delivery model in the SEARCH trial in rural Uganda and Kenya . J Int AIDS Soc. 2017 Jul 21 ; 20 ( Suppl 4 ): 58 ± 67 .
21. Mutasa-Apollo T , Ford N , Wiens M , Socias ME , Negussie E , Wu P , et al. Effect of frequency of clinic visits and medication pick-up on antiretroviral treatment outcomes: a systematic literature review and meta-analysis . J Int AIDS Soc. 2017 Jul 21 ; 20 ( Suppl 4 ): 88 ± 98 .
22. Decroo T , Koole O , Remartinez D , dos Santos N , Dezembro S , Jofrisse M , et al. Four-year retention and risk factors for attrition among members of community ART groups in Tete, Mozambique . Trop Med Int Health . 2014 ; 19 ( 5 ): 514 ± 21 . https://doi.org/10.1111/tmi.12278 PMID: 24898272
23. Tsondai P , Wilkinson L , Grimsrud A , Mdlalo P , Trivino A , Boulle A . High rates of retention and viral suppression in the Scale-up of Antiretroviral Therapy Adherence Clubs in Cape Town, South Africa . Poster presented at the AIDS Conference; Durban, South Africa; 2016 .
24. Okoboi S , Ding E , Persuad S , Wangisi J , Birungi J , Shurgold S , et al. Community-based ART distribution system can effectively facilitate long-term program retention and low-rates of death and virologic failure in rural Uganda . AIDS Res Ther . 2015 ; 12 : 37 . https://doi.org/10.1186/s12981-015 -0077-4 PMID: 26566390
25. Hanrahan C , Schwartz S , Keyser V , Mudavanhu M , West N , Mutunga L , et al. Retention in community versus clinic-based adherence clubs for stable ART patients in South Africa: 24 month final outcomes from a randomized controlled trial. Oral presentation (Abstract MOAD0106LB) at the 9th International AIDS Society Conference; 2017 Jul 23 ± 26 ; Paris, France.