Received: 22 April 2024 Revised: 4 November 2024 Accepted: 17 November 2024

DOI: 10.1002/lrh2.10475

EXPERIENCE REPORT

Learning sites for health systems research: Reflections on five

programs in Africa, Asia, and Central America

Sophie Witter 1 | Shophika Regmi 2 | Joanna Raven 3 | Jacinta Nzinga 4 |

Maria van der Merwe 5 | Walter Flores 6,7 | Lucia D'Ambruoso 8,9,10,11,12
1
Institute for Global Health and Development, Queen Margaret University, Edinburgh, UK
2
HERD International, Kathmandu, Nepal
3
Liverpool School of Tropical Medicine, Liverpool, UK
4
Health Economics Research Unit, KEMRI Welcome Trust Research Programme, Nairobi, Kenya & Liverpool School of Tropical Medicine, Liverpool, UK
5
Independent Consultant, White River, South Africa
6
School of International Service, American University, Washington, DC, USA
7
Centro de Estudios para la Equidad y Gobernanza de los Sistemas de Salud, Guatemala City, Guatemala
8
Centre for Health Data Science, School of Medicine, Medical Sciences and Nutrition, and Centre for Global Development, School of Education, University of
Aberdeen, Aberdeen, UK
9
MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), School of Public Health, University of the Witwatersrand Johannesburg,
Johannesburg, South Africa
10
Department of Epidemiology and Global Health, Umeå University, Umeå, Sweden
11
Public Health, National Health Service (NHS) Grampian, Aberdeen, UK
12
Department of Global Health, Stellenbosch University, Stellenbosch, South Africa

Correspondence
Sophie Witter, Institute for Global Health and Abstract
Development, Queen Margaret University,
Introduction: Learning sites have supported intervention development and testing in
Musselburgh, Edinburgh, EH21 6UU, UK.
Email: [email protected] health care, but studies reflecting on lessons relating to their deployment for health
policy and system research (HPSR) in low- and middle-income settings are limited.
Funding information
Foreign Commonwealth and Development Methods: This experience report draws from learning over three continents and five
Office, Grant/Award Number: PO 8610;
research and community engagement programs—the oldest starting in 2010—to
European Union Horizon 2020 Research and
Innovation Programme for PERFROM2Scale, reflect on the challenges and benefits of doing embedded HPSR in learning sites, and
Grant/Award Number: 733360; Health
how those have been managed. Its objective is to generate better understanding of
Systems Research Initiative from Department
for International Development (DFID); Medical their potential and constraints. The report draws from team members' experiential
Research Council (MRC); Wellcome Trust;
insights and program publications.
Economic and Social Research Council (ESRC);
VAPAR, Grant/Award Numbers: MR/ Results: Challenges relating to initial engagement in the sites included building and
N005597/1, MR/P014844/1
maintaining trust, managing partner expectations, and negotiating priority topics and
stakeholders. Once the embedded research was underway, sustaining engagement,
and managing power dynamics within the group, supporting all participants in devel-
oping new skills and managing rapidly changing settings were important. Finally, the
complexity of reflecting on action and assessing impact are outlined, along with

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2024 The Author(s). Learning Health Systems published by Wiley Periodicals LLC on behalf of University of Michigan.

Learn Health Sys. 2024;e10475. wileyonlinelibrary.com/journal/lrh2 1 of 13

https://doi.org/10.1002/lrh2.10475
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 of 13 WITTER ET AL.

potential approaches to managing all of these challenges and the variety of gains that
have been noted across the programs.
Conclusions: We highlight the potential of learning sites to develop relationships,
capacities, and local innovations which can strengthen health systems in the long
term and some lessons in relation to how to do that, including the importance of sta-
ble, long-term funding as well as developing and recognizing facilitation skills among
researchers. Supporting spaces for learning is particularly important when health sys-
tems face resource constraints and everyday or acute stressors and shocks.

KEYWORDS
health policy and systems research, learning sites, low- and middle-income countries,
participatory action research

1 | I N T RO DU CT I O N 2 | BACKGROUND ON THE PROGRAMS

Learning sites are research platforms established in a specific geo- Table 1 gives a summary overview of the five programs, which were
graphical area, where researchers and local actors collaborate over an conducted in sites in Nepal, South Africa, Kenya, Ghana, Uganda,
extended period of time to develop contextually tailored interven- Malawi, and Guatemala. The size of sites varied, and most were rural,
tions.1 The type of research conducted and the actors involved will with the exception of ReSYST in South Africa. Two of the programs
depend on the focus of the learning sites, but typically, in HPSR, the are completed (PERFORM2Scale and ReSYST), one is ongoing but
process of enquiry is emergent within a broad framework, with a com- time-limited (ReBUILD), one is currently being scaled-up (VAPAR), and
mitment to joint ownership and co-production of knowledge (Box 1). one is a long-term research center, with a variety of funding sources
However, while potentially a very powerful strategy, such an (CEGSS).
approach is not without ethical, practical and other challenges. For
example, Oliver et al. highlight risks of co-production relating to par-
tiality (e.g., where researchers only report what is judged acceptable
to policy or management partners).2 Learning sites also need to man-
BOX 1 Typical features of learning sites.
age power relationships, working politically as well as technically, and
managing system constraints (including limited resources) in relation
1. They are based on long-term collaboration between
to implementation and research. Sustainability, institutionalization,
researchers and policy-makers and/or communities—
and scale-up are also typically areas of challenge, given the intensity
of effort devoted to relationship building in many learning sites. building trust through sustained engagement.

Although this approach has been adopted by a number of 2. They provide a platform for action research—knowledge
for local use.
research programs, there is relatively little published reflection on
health system learning sites in low- and middle-income settings, how 3. They focus on co-production of knowledge, recognizing

they function, what their strengths are, what challenges they face, the importance of local lived experiences, experiential
knowledge, tacit knowledge about health systems and
and how to overcome them. In this paper, we draw together reflec-
decision making, and local relationships.
tions from five research programs which have employed methods that
fit within a broad learning site approach. We reflect on emerging les- 4. They recognize and attempt to rebalance power dynam-
ics in the program (between international and national,
sons, considering the different models that have been adopted, the
but also community and system actors, and sometimes
diverse objectives and varied settings in which they have been
employed (in programs based in three continents), as well as the com- researchers versus implementers).
5. Researchers are often embedded in or interacting regu-
plex and long-term processes of health system improvement and
larly with the health system; through research activities,
change.
The paper was developed following a panel session on this topic they come to understand the daily routines and chal-

in the Global Symposium on Health Systems Research 2022 and lenges faced by health managers and other actors.

builds on reflections within and across research teams, as well as the

participant questions and feedback which were shared during Source: (Gilson et al.,1); authors.

the panel discussion.

 TABLE 1 Summary of five research programs using learning site approaches.

Program Objectives Funder and dates Setting and coverage Focal topic and approach
ReBUILD for To provide contextualized FCDO Learning sites have been developed in a municipality of Kapilvastu Since federalization in 2017, planning and delivery of
WITTER ET AL.

Resilience (www. evidence on how to develop 2020–2026 district, Nepal; also in the Bekaa valley, Lebanon; a similar approach is health care has been devolved to the municipality
rebuild. resilience capacities for health being deployed in Kailahun and Moyamba districts in Sierra Leone. level in Nepal; however, capacities at this level
consortium.com) systems in fragile and shock- This paper focuses on the Nepal experience, as this was the first to be remain to be built and systems (and connectors) in
prone settings initiated. this new arrangement need to be established.
Kapilvastu is a semi-urban district in Lumbini province, in the lowland An embedded approach of implementation research
(terai) area of Nepal, bordering India. The municipality is fragile in involving a researcher stationed within the
many ways due to its poor health service indicators compared to municipality and working with local health systems
national average, weaker health infrastructure and access to routine stakeholders has been adopted. The action research
health services and cross-border migration, which increases steps are ongoing (12–18 months) and include:
vulnerability to disease and infection. The main source of income is • Partnership development
agriculture, and lack of opportunities for work have increased labor • Participatory mapping of resilience capacities
migration. The population of the learning site is around 77 000. • Co-creation of interventions to strengthen
resilience capacities
• Evaluation of implementation and the PAR cycle
Although the resilience mapping is broad (including
most health system blocks), key areas emerging to
date have been governance and use of local data for
planning.
PERFORM2Scale To strengthen district level EC (Horizon 2020) PERFORM2Scale program worked in learning sites in Ghana, Malawi, The management strengthening intervention (MSI)
www. management to address 2017–2022 and Uganda. used a PAR approach to enable the district health
perform2scale. workforce performance and In each country, we established three learning sites which were management teams to analyze their own problems
org service delivery challenges groups of three district health management teams that were in related to workforce performance and service
neighboring districts. Willingness to participate, established delivery and develop appropriate work plans (plan);
relationships, and avoiding districts where similar interventions are implement the work plans (act) and learn about
being implemented were important criteria for selection. The districts management from the experience (observe and
were mainly rural but some with a city or large town. Populations of reflect). The MSI was facilitated by national research
the districts varied, with Ghana's districts being smaller at around teams and government officials in each country
100 000, whereas in Uganda populations they were between through short workshops, joint meetings of DHMTs,
300,00–500 000, and in Malawi, populations ranged from 200 000 to and follow-up support by visits, e-mail, and
1 million. telephone/Whatsapp. The cycle took about
10 months and then moved into another cycle,
either adapting the strategies to address the same
problem, or addressing another problem.
ReSYST https:// To examine health system DFID (now FCDO) Two sites were developed: Kilifi County in Kenya, and in South Africa, ReSYST focused on leadership practices and
resyst.lshtm.ac. routine stresses and everyday 2011–2018 two health districts located in different provinces (Sedibeng, Gauteng; organizational relationships, with the aim of building
uk/ resilience, including how health and the Mitchell's Plain area of Cape Town, Western Cape). These and supporting soft skills (such as communication,
managers' leadership practices, sites were initially chosen because of prior experience and motivation and responsibility) among leaders and at
organizational relationships and engagement with managers. Research teams live locally and have multiple levels of the county health system (e.g.,
their underpinning values maintained trusting relationships with health system colleagues over through coaching initiatives and participatory
influence health system time. training in communication with colleagues and
resilience Kilifi is a rural setting compared to the urban contexts of the South emotional management). Core approaches were
African sites, and Mitchell's Plain is a sub-district rather than a district, collective enquiry and reflective practice.
3 of 13

(Continues)

23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 TABLE 1 (Continued)
4 of 13

Program Objectives Funder and dates Setting and coverage Focal topic and approach
with a population (around 500 000 people) about half that of the More specifically, in selected facilities, it tracked and
other sites. In Kilifi, the population (1.5 m) is relatively poor. Research investigated specific chronic stressors (e.g.,
teams live locally and have maintained trusting relationships with absenteeism, patient complaints) and potential
health system colleagues over time. alleviators of chronic stress (e.g., new public finance
policies or supervision approaches and guidelines)
identified with mid-level and facility managers as
possibly having important implications for system
resilience.
Beyond tracking the stressors, it continued tracking
implementation and impact of past initiatives (South
African sites), and development and tracking of new
initiatives (all sites) aimed at strengthening health
system governance. Some of these initiatives
included coaching interventions at senior & facility
management levels (implemented in previous phase
of learning site work) and adaptation of a soft skills
communication and emotional management training
course for the managers.
CEGSS (Centro To demonstrate that Internationally CEGSS is present in 35 municipalities, working together with the Demonstration sites are municipalities in rural
de Estudios para participatory governance of funded through Network of Community Health Defenders (REDC-SALUD), which are indigenous areas in which CEGSS, together with
la Equidad y municipal healthcare facilities is private over 150 organized community leaders. However, not all grassroots organizations, reflect, design, implement,
Gobernanza de feasible and brings better results. foundations, municipalities are demonstration sites. Demonstration sites are evaluate and learn about strategies for citizen-led
los Sistemas de In the sites, the program also universities and chosen where there is combination of authorities' openness, health accountability. These processes are based on
Salud) https:// develops and field tests methods research institutes grassroots capacities, and channels of engagement between users of PAR cycles. In all sites, there is collaboration with
cegss.org.gt/en/ and tools for citizen monitoring 2010-present services and authorities. There are currently 8 in areas with diverse providers and authorities. However, in all places the
and authority responsiveness. geography and ethnic population groups. In other sites, CEGSS works process is led by organized users of services. The
to generate or strengthen these conditions. approach is based on longstanding mentoring of
Municipalities vary in population size (between 5 K and 15 K community leaders, whose families and neighbors
inhabitants). However, the focus of the demonstration sites is on the use public services.
network of public healthcare facilities in the territory. On average, The purpose of the sites is to generate learning on
there are one or two health centers per site and several health posts, effective strategies and capacities to improve
which are the most peripheral healthcare facilities. democratic governance of the health system.
The learning sites also host visits from grassroots
organizations, providers, and health authorities.
These visits motivate other actors to replicate the
approach and adapt strategies and methods to their
own context. In addition, the reflection and learning
among all sites – including those that are not
demonstrative – is shared and disseminated through
the REDC-SALUD. The goal for CEGSS and REDC-
SALUD is that through action research, community
organizing, and capacity building, all the 35
municipalities will eventually achieve the favorable
conditions of the learning sites.
WITTER ET AL.

23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 TABLE 1 (Continued)

Program Objectives Funder and dates Setting and coverage Focal topic and approach
VAPAR https:// To develop inclusive knowledge Medical Research VAPAR was based at the MRC Wits/Agincourt Health and Socio- VAPAR supported a series of reflection/action cycles
WITTER ET AL.

www.vapar.org/ partnerships to strengthen Council/HSRI Demographic Surveillance System (HDSS) in Bushbuckridge sub- to generate information on collectively identified
health systems; evaluate (with additional district, Mpumalanga, South Africa. In the HDSS area (population health concerns in terms of associated disease
changes in health, health equity, funds from the 120 000), rural homesteads experience multigenerational deprivation. burdens, their social and health systems
and empowerment; and build Newton Fund, There is little formal sanitation, unaffordable electricity, high determinants (using VA), and on norms, practices and
sustainability and transferability GCRF and Scottish unemployment, and a limited economic base resulting in labor priorities for action from the perspectives of directly
Funding Council) migration and reliance on social grants. Orphaned youth characterize affected and under-represented groups (using PAR).
Development the population: school drop-out is 40%, 16% of the provincial Local data were developed together with community
grant: 2015–2016; population is illiterate, and district unemployment is 37%. The burden capabilities for these data. Data were then
main grant: 2017– of disease is complex and dynamic: while HIV/AIDS and maternal collectively analyzed in a multi-level process of
2023 mortality and child deaths are decreasing, non-communicable health systems engagement including with DHMT,
conditions, including deaths owing to injuries, accidents and violence, through which recommendations, local action plans,
and health disease, are increasing, together with chronic, complex implementation, evaluation, reflection and adaption
comorbidities in an aging population. There are two community were co-developed, and which fed learning into the
health centers and seven PHC clinics. next cycle. Co-produced evidence was fed into
action in the local health system. Themes were
initially nominated by communities as alcohol and
drugs, and water security (Cycles 1–2). Later, cycles
(Cycles 3–5) prioritized dialogue with local health
systems actors and focused on CHW capabilities,
roles, and functions; specifically, HIV/TB treatment.
CHWs reported improved relationships with
communities, peer support, and recognition by the
health system. The process is currently being
adapted for scale-up from the rural sub-district
(population approx. 500 000) to the province
(population 4.8 million).
5 of 13

23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 of 13 WITTER ET AL.

In terms of focus, three were focused on the health system (sup- participants assumed control of the process: identifying priority health
ply-side), while two (VAPAR and CEGSS) had a stronger community- concerns, directing expansion of the participant base, and controlling
led approach, albeit also engaging with health system strengthening. practical aspects such as dates, times, and venues of workshops.”4
Two took resilience as their core theme—one in more stable contexts The importance of positive relationships was also highlighted in
(ReSYST) and one in fragile settings (ReBUILD), while PERFORM2S- ReSYST—ensuring that participant ideas are incorporated, that feed-
cale focused on strengthening management to address workforce per- back is regular, and that activities are flexible and agile to respond to
formance and service delivery challenges and CEGSS on participatory heavily politicized contexts, demanding job environments and fre-
governance and social accountability. VAPAR has focused on more quent turnover of post-holders. Facing and ultimately overcoming
specific health topics, nominated by community stakeholders in col- challenges also served to build mutual understanding and trust in all
laboration with health systems actors at district, sub-district, clinic and settings.5
community levels, which framed community-nominated priorities in Capacity building through training and mentoring also developed
terms of their social determinants, thereby connecting the process trust across programs. In VAPAR, collective capabilities, mutual under-
with sectors adjacent to health. standing, and trust developed as participants' familiarity were built
All used participatory action research (PAR) methods, seeking with public speaking, analysis (including mapping causes and conse-
leverage points for change and focused on positive system and com- quences of local health priorities, and in selecting, appraising, and cap-
munity assets and capabilities, and using the cycle of collective analy- tioning visual data), consensus building, peer and non-peer
sis, co-creation of responsive and evidence-based plans, their deliberations, and in co-facilitation and recording of meetings.6
implementation, and reflection and feedback on action as an active In Nepal, the use of local data and their intersectional analysis
3
intervention in its own right. All programs had additional elements, proved to be a key instrument in building trust and establishing rela-
such as coaching, mentoring and training (all five programs). In addi- tionships with local counterparts. It supported showing the strengths
tion, VAPAR drew on verbal autopsy (VA) data generated by the and capabilities of the research team, which the local government,
Health and Demographic Surveillance Site (HDSS) in which it is based transitioning to federalization, acknowledged and in response sought
to support participants in planning and priority setting with robust technical support to strengthen their health system.7
local data on burden of disease. VAPAR also extended the VA method
to collect and interpret data accounting for social and health systems
determinants of mortality. 3.1.2 | Managing expectations
Research methods used were/are primarily qualitative and partici-
patory, including document review, individual and group interviews, Particularly at initiation but also on an ongoing basis, managing partici-
critical reflection, use of diaries and case studies, observation, reflec- pant expectations is challenging, particularly in contexts, such as
tive meetings, participatory workshops (including group model build- Nepal's, where financial dependence on external development part-
ing), and photovoice. ners is high and there is a perception that resource limitations are the
main obstacles to improving health services.
This can be managed through continuous and frequent engage-
3 | C H A R A C T E R I S T I C I S S U E S F A C E D BY ment, being transparent about the approach and facilitating role of
L E A RN I N G S I T E S A N D H O W T H E S E W E R E the research team, helping to connect partners to wider resources but
MANAGED also helping them to identify what can be done more effectively
within existing means.8 Over time, partners have come to see the high
A number of challenges were experienced across the learning sites, value of the role played by the learning platform and have adopted
across the different stages of the learning process. We highlight these entrepreneurial approaches to resourcing planned activities.9
here with some examples of how they were managed by the programs
and some lessons learned (see summary in Table 2).
3.1.3 | Priority-setting dilemmas

3.1 | Initial engagement The challenge of prioritizing focal problems to address arose for some
of the programs. For example, in VAPAR there were some initial chal-
3.1.1 | Building and maintaining trust lenges as community-nominated topics in the first two cycles such as
access to water were seen as less tractable by health system actors.
Trust was highlighted as crucial to developing the learning site rela- However, the district health system actors appreciated the rare
tionships. Approaches to earning trust included regular revisiting of opportunity to connect cross-sectorally.10,11 Latterly, in response to
PAR principles to support shared vision and purpose and allowing adaptation during COVD-19, cycles focused on the roles and func-
time for relationships to deepen. For example, in VAPAR “engagement tions of CHWs. This worked well as this group connects communities
gradually improved as core principles were transmitted, discussed, to the health system and so engaged the key points of the community,
revisited, owned, and taken up. Ownership was supported as system, and researcher triangle with a shared priority (addressing loss
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
WITTER ET AL. 7 of 13

TABLE 2 Challenges, how they were managed and lessons learned.

Challenges encountered How these were managed (examples of strategies) Lessons learned
Initial engagement
Building and maintaining • Developing a shared vision of purpose and process, and adapting • Trust is gained and retained by continuous
trust this collectively to changing needs and circumstances enactment of respectful and responsive
• Allowing time for relationships to develop contributions of the members of the
• Progressively shifting control to participants platform, while acknowledging positionality
• Capacity building of participants (communities, health system and limitations
actors, and researchers) • Important to help forge networks for
• Using local data to show utility of work and engagement with local participants, both horizontally (linking with
context other areas to share and support) and
• Working flexibly around participant constraints vertically (given that many issues can only be
• Demonstrating value through effectiveness in engagements addressed effectively at higher levels).
• Bringing in other actors when appropriate and requested • Systems tend to work in silos; learning
Managing expectations • Being transparent about roles platforms aim to break these silos
(of additional resources) • Linking partners to wider resources and building their skills to be
able to access them
Priority-setting dilemmas • Enabling/empowering participants to select and frame local
priority health concerns
• Building on these to identify and progress shared agendas with
authorities
• In some cases, prescribing focal areas which linked to sectoral
interests
Choosing level of • Working at multiple levels of the health and wider system to
engagement ensure that systemic factors can be addressed
• Building confidence of local stakeholders to address priority issues,
despite limited decision space
• Building networks and sharing lessons across them
Managing the process
Maintaining commitment • Demonstrating agility and adaptability to respond to local needs, • In constrained environments, pragmatic use
from participants and including by changing focus during crises of existing structures and fora is important to
funders • Showing wider utility by providing support to local priority encourage engagement and sustainability
activities not included in program plans • We also noted the importance of working
• Building collective capabilities and mindsets multisectorally, especially in decentralized
• Including local leaders as participants and co-authors environments where health is not necessarily
• Using formal tools such as MoUs to institutionalize engagement, a priority for local leaders.
even as local leaders change • Empowerment by learning and doing is key
• Using local platforms and aligning to planning and budget cycles
• Effectively including community members, supporting a rights-
orientation
• Extending engagement to a wide network (planning for expected
attrition over time)
• Demonstrating collective effectiveness
• Critically engaging in academic and funder debates on value
of PAR
• Creating supportive environment for researchers (e.g., capacity
development and equitable publications)
Managing power • Sensitive but assertive facilitation • Facilitation skills involve a complex mix of
dynamics within group • Jointly establishing values of democratic participation, voice, and technical and inter-personal skills; these can
(to ensure participation mutual respect and regularly reinforcing these be developed over time but need recognition
of marginalized voices) • Separating groups where appropriate and nurturing
• Acknowledging power differentials, also for and within researcher • Creating a collective and problem-solving
group mindset can help to overcome differences of
• Acknowledging systemic challenges, even when these are sensitive positionality and power
Developing new skills for • Training and practice in PAR, data analysis and writing for • Spaces for dialogue that link different groups
researchers and participants (democratizing research methods and adapting them and resources are often missing; by creating
participants for use by those dealing with the burden of disease) and in these, opportunities for learning are created
planning and supporting implementation for researchers
• Embedded approaches—learning by working alongside one
another, gaining better mutual understanding of opportunities and
constraints

(Continues)
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 of 13 WITTER ET AL.

TABLE 2 (Continued)

Challenges encountered How these were managed (examples of strategies) Lessons learned
Reflection on action
Complexity of • Avoiding additional workload by building reflection into • Continuous critical reflection is needed to
monitoring, reflection existing fora ensure that all voices are heard, collectively
and learning • Holding regular reflection and sensemaking meetings, with validated and acted on
systematic and transparent recording • Engaging wider networks to reflect on
• Establishing virtual platforms and shared repositories as useful lessons learned in one area can enhance
tools relevance of learning for wider geographies
Assessing impact • Theory-based evaluative approaches can help assemble evidence • While valuing health and health equity gains,
on the complex changes sought and (variably) achieved, using it is important not to overlook the
mixed methods importance of “intermediate” stages in the
• PAR generates data on a continuous basis, which can be used for theory of change—such as strengthened
cycle evaluations, as well as regular reflections relationships and demand for evidence—
• Capacity assessment is complex but can be addressed using especially since these shape the
multiple qualitative methods, for example, observations, sustainability of the learning platform and its
interviews, and researcher logs ability to support future and wider benefits
• Learning sites produce intrinsic as well as
instrumental benefits, but conventional
evaluation metrics tend to neglect or
underplay the intrinsic

to follow up for TB and HIV treatment).4,6 Other programs resolved health system and other sectors for financial and other support, work-
the challenge by defining clear areas of engagement—for example, ing in their own teams but also sharing lessons across sites (which can
PERFORM2Scale focused on workforce performance and service also encourage co-learning and exchange).
delivery, but within that each district chose the problem that they
wanted to address. This fostered more engagement in finding solu-
tions and implementing them. District managers were used to devel- 3.2 | Managing the process
opment partners and others identifying the problems that should be
addressed. 3.2.1 | Sustaining engagement

Maintaining continuity of commitment, including from partners and

3.1.4 | Choosing level of engagement funders, is a challenge across all learning sites. For partners, it is
important to demonstrate the utility of the approach, while keeping
Assessing whether to stay focused or work at multiple levels of the expectations grounded. Many programs have responded adaptively to
system is one of the choices facing learning sites. All programs worked acute and chronic crises, providing additional support during COVID-19,
at multiple levels of the health system, recognizing that change for example. One approach adopted in Nepal was to demonstrate com-
involves mobilizing stakeholders with different roles in the health sys- mitment to local government by supporting them not only in activities
tem and at community level. However, getting traction at higher levels outlined in the action plan but also in their day-to-day programmatic and
can be challenging. In Guatemala, for example, the demonstration site administrative tasks. This significantly contributed to their capacity
work can be seen as threatening by national authorities. The sites enhancement and strategic planning and execution.
encourage an active role by citizens which can threaten the power of More profoundly, it is fundamental that collective mindsets and
officials. However, some authorities and elected officials are support- capabilities for joint action and learning are built, which can sustain
ive of the work all the same. the engagement over the longer term. In VAPAR over five reiterative
Engaging those with some degree of power to act was critical. action/learning cycles, the local action plans became more specific,
Some programs addressed this by working at multiple levels of the strategic, focused, and feasible, reflecting capabilities built through
health system (e.g., in ReSYST), while others engaged a wide range of sustained engagement in the process. Local leaders at multiple levels
actors from beyond the health sector (such as local government in were included as partners from the start of the program and were
ReBUILD and PERFORM2Scale and multiple sectors in VAPAR). How- encouraged to participate in all stages of the cycles, including as co-
ever, structural and political issues constraining decision space of local authors of outputs and publications, but also in training and program
managers are typically hard to challenge.12 The programs noted the development.
importance of building confidence of managers to be able to analyze, In dynamic contexts, such as the decentralized health system in
strategize, and implement plans for improvements and to be able to Nepal, changing office-holders and uncertainty of roles creates addi-
articulate these improvements and advocate to higher levels of the tional operating challenges for learning sites, given their long-term
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
WITTER ET AL. 9 of 13

nature. A memorandum of understanding to agree on roles, regardless VAPAR. Including affected people and those whose voices were
of changes in leadership, as well as participating in and contributing to excluded required very careful management of group dynamics to
routine municipality meetings and workshops is how the research ensure inclusion, participation, and respectful engagement. In VAPAR,
partners (HERD International) worked to maintain relationships in this it was beneficial to start by spending significant time building commu-
context. Maintaining multiple relationships and being adaptive to nity capabilities for community voice, together with local, actionable
changes is however resource-intensive, so budget sufficiency and data, as the basis of engagement with the authorities.
flexibility is needed. For example, the research teams need to align During workshops with representatives of the authorities, some
the research pace with the interests and priorities of local counter- stakeholders were dominant and disruptive, leading to others feeling
parts, particularly in the initial phase of the project. They cannot intimidated to raise their opinions. In some instances, VAPAR
impose planned research activities over their priorities, which leads to observed local politicians using the platform to promote current prior-
delays. It took 1 year for the Nepal team to co-create the action plan ities or debates. It dealt with this with sensitive, but assertive, facilita-
(following resilience mapping and participatory workshops). Country- tion, reinforcing principles of democratic participation, voice,
wide elections, local planning, and several other factors affected the representation, and respect. Over time, the regular negotiating of
process.7 Equally, COVID required a pivot in focus and activities for these principles supported mutual understanding, supported agency
programs such as VAPAR and PERFORM2Scale, where managers and more equal participation. Facilitation skills of the engagement
applied problem analysis and solving to addressing COVID-19 in their process were also very key in all programs, blending the technical
districts.13 knowledge and expertise of the research team and the art and skills of
For community members, feeling recognized, respected, and facilitation. Separate workshops with different sets of stakeholders
valued in their contributions, as well as seeing responses in terms of were another strategy to manage power dynamics and enable stake-
action and follow-up, are key to continued engagement. Changing holders to express themselves (e.g., in Nepal, health workers and
mindsets is key here too—citizens need to come to view health care elected official met separately due to the nature of their relationship
as a right and believe that they can claim better services, as with one another).
highlighted in the CEGSS model. In addition, recognizing that com- Power imbalances also relate to researchers, who potentially have
munity activists are volunteers, who may reduce their participation different access to opportunities and resources within their network,
and engagement at different times, additional training is done by compared to communities and system actors. This has the potential to
CEGSS with communities to allow for replacements. The number of undermine the relationship, if not acknowledged and addressed, in
community defenders (as the activists are called) has increased by our experience. Highlighting systemic challenges can also be uncom-
40% in the past few years, which contributes to sustaining fortable for local stakeholders, and researchers need to be able to
the work. address such topics in a sensitive but open way.
According to our experience in learning sites, researcher engage-
ment also needs to be sustained, as academic environments do not
typically support, enable, or reward embedded and PAR approaches, 3.2.3 | Developing new skills for participants and
which are seen as delivering very contextual and “low grade” evi- researchers
dence, with considerable energy devoted to local relevance and
uptake (more than high-ranking publications). Funding does not typi- In learning sites, researchers actively facilitate and support partners.
cally make provisions for stakeholder/policy engagement activities This is a marked departure from the conventional positivist role of
and particularly for the substantial periods that are needed to demon- researchers, which is to objectively study phenomena free of sub-
strate impact. In part, this is addressed by selection of researchers jective influence or “contamination.” This was challenging and
with a particular orientation to applied research, but also through pro- involved capacity building for researchers and partners, who are
viding a supportive environment in terms of training opportunities also being supported to develop a more problem-solving and analyt-
and support for equitable authorship and networking. It is also impor- ical mentality. The research team performed the dual roles of
tant to engage in critical debate in academic spaces to highlight the researchers and implementation support practitioners, which
value and rigor of enquiry paradigms concerning knowledge for required knowledge and skillsets, as well as a deep understanding of
action, plurality of knowledge, cooperative learning, and expertise the context, stakeholders, and power relations. Without this type of
from the margins. It is also important to engage funders on the value implementation software work, progress with implementation is
of learning site approaches and to sustain their support as long-term challenging.
engagement is a key requirement for learning site effectiveness. These new skills are supported by learning through doing, but also
through training, mentorship, and the development of supportive
tools and toolkits (e.g., PERFORM2Scale developed a toolkit for facili-
3.2.2 | Managing power dynamics within group tation [14], while VAPAR developed a community mobilization toolkit
in collaboration with the Department of Health for all participants)15
All programs aimed to include marginalized voices, but especially so and a post-graduate health policy and systems research module to
for those which focused at community level, such as CEGSS and support wider application of methods and tools.
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 of 13 WITTER ET AL.

Strengthening capacity and institutionalizing learning are the core 3.3.2 | Demonstrating impact
principles of engagement in the learning sites. Research partners pro-
viding technical support, such as HERD International in ReBUILD, It is complex to ascertain and demonstrate impact from learning sites,
have been careful to transfer skills and knowledge (e.g., in developing given the complexity of the environment, the intrinsic as well instru-
evidence-based local plans), focusing on institutionalization of prac- mental role that they can play, and the dynamic engagement over
tices to avoid dependency. Equally, where capacity gaps become evi- time. However, this is important, including to funders. All programs
dent, research partners have been able to respond flexibly to support used end-of-cycle evaluations using participatory methods. One
filling these, given their embedded position (e.g., in Mpumalanga, response was to develop a theory-based evaluation of the platform
VAPAR supported the capacity building and development of research (e.g., for VAPAR), which allowed the different domains to be tracked
governance resources and training with the provincial research ethics to assess the contribution of the program.18 While health effects are
committee, which had not been planned originally). 16
In ReSYST, dis- important, the significance of building “intermediate” domains, such
cussion of the learning site concept and sharing of literature, narra- as trust, capacity, confidence, and systemic relationships is highlighted
tives, and ideas were used to increase participant confidence and in these evaluations, tracked largely qualitatively through interviews
support progress with analysis and writing up of learning. and process documentation. These have benefits beyond the immedi-
ate period and (potentially) constitute real system strengthening.19

3.3 | Reflection on action

4 | GAINS ACHIEVED
3.3.1 | Complexity of reflection, monitoring, and
learning Gains documented by the programs included important domains such
as giving voice to communities, increasing skills and confidence of
All programs built reflection, monitoring, learning, and evaluation into managers, and creating and sustain spaces and processes to support
their PAR cycles. However, challenges were noted in application of cadres that link these two groups, such as CHWs. Increased
some of the tools for reflection with partners, such as use of diaries researcher skills and increased demand for local evidence have also
and e-diaries, which were an added duty for them. Reflection in rou- been identified, along with improved relationships between different
tine meetings with partners, across partners (e.g., across sites) and in groups, with learning sites often providing much-needed spaces for
standard fora was more feasible, though still demanding in terms of collaboration. These in turn can trigger innovations, improvements in
time for participants. In most programs, researchers took responsibil- service funding and delivery, and changes to policy and practice.
ity for eliciting, synthesizing, and reporting the reflexive elements and
working to enact the adaptations deliberated over and agreed through
the consensus building processes. 4.1 | Giving voice to communities and connecting
For monitoring and learning, as interactions are frequent and con- them to service providers
tinuous, there is a challenge involved in maintaining good records and
managing routine documentation. The routine documentation of CEGSS has done several participatory impact assessments in which
events, interactions, meetings, research activities, and outputs make organized users define their own impacts. In these exercises, impacts
up a huge resource of information and analyzing it, using the right included increased knowledge and capacity to navigate the public sys-
information in the right place, needs careful planning from the start. A tem and learning about methods and tools to monitor public policies
number of the programs developed regular (e.g., monthly) reflection and services.20,21 Equally, VAPAR has documented community stake-
and planning meetings with researchers and local partners to record holders realizing and developing shared capacities to use their voice,
and reflect on joint activities. Being systematic and transparent in while service providers found that the learning platform was a safe
analysis and sense-checking across teams and with partners is crucial. way to engage with users in spaces which had hitherto been lacking,
Use of virtual platforms and spaces and shared repositories for pro- with discontent frequently expressed in violent protest given the
gram data have also been supportive, especially during times of crisis absence of constructive engagement channels. The process shifted
such as COVID. otherwise disconnected actors toward more constructive dialogue
Another challenge lies in developing messages which have reso- and collective mindsets.4
nance beyond the specific sites, which requires that wide-ranging par-
ticipation is engaged to ensure that lessons are developed which can
inform a broader geography and are shared in a range of formats 4.2 | Increasing skills and confidence among
beyond the traditional incentivized “products” from research (papers managers
and conference presentations). VAPAR, for example, developed a
radio series on CHW roles and functions in TsSonga, research briefs Programs focused on managers (e.g., PERFORM2Scale) noted
in English and local languages, podcasts, YouTube content, and improved confidence and independence in problem-solving and strat-
blogs.17 egy development in this group. Participants also reported more
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
WITTER ET AL. 11 of 13

innovative and creative thinking (in part because of the lack of addi- and community) are rare in most settings, and where they exist can be
tional resources, which necessitated this), as well as more regular poorly functional,23 and this was a key feature that the learning sites
meetings of the management groups. In ReSYST, an evaluation of a aimed to address. Once established and experienced, this was typi-
multifaceted leadership development program for mid-level managers cally highly appreciated by participants, who were able to develop
and facility managers embedded within the learning site in Kenya was new relationships and gain increased appreciation of structures,
undertaken.22 It found that managers reported greater recognition of opportunities, and responsibilities for collective action.24
the importance of health system software (values, belief systems, and A number of the programs engaged with a wide variety of sectors
relationships) and that the training also created spaces for managers and local actors (beyond health) and documented strengthened relation-
to share experiences, reflect upon, and nurture social competences. ships and teamwork between them. For example, in VAPAR, work on
the issues of access to water, drugs, and alcohol mobilized partnerships
across sectors, a feature that was noted as beneficial in working effi-
4.3 | Empowering linking cadres such as CHWs ciently and avoiding duplication by those involved.8 The focus on the
municipality in ReBUILD improved coordination across wards and local
As a cadre connecting health systems with communities, CHWs have leaders, with potential benefits beyond the health sector. This included
been engaged and played central roles, in learning sites. For example, sensitizing the non-health stakeholders in the municipality – mainly the
in VAPAR and at the onset of the COVID-19 pandemic, the process ward level officials who are the key decision makers – on health and
was co-re-designed to focus to CHWs' roles, functions, and relation- health systems, which are a new area of responsibility for them in the
ships with both communities and the health system. In the context of decentralized system. Different sectoral sections within the municipality
suboptimal integration of CHWs and poor working conditions, an also started holding regular coordination meetings – the only platform
evaluation of the VAPAR engagement of CHWs found improvements where all sectors come together to discuss progress, challenges, and
to key CHW capabilities in community mobilization as well as new explore opportunities for collaboration and integration.
skills and confidence in complex analysis, public speaking, and report- In Guatemala, as the network of Community Defenders acquired
ing, alongside greater role clarity. In this evaluation, CHWs reported a more experience and knowledge, they became more effective in moni-
“triple benefit”: strengthened relationships with communities, better toring services, mediating and solving conflicts (when there are com-
peer-to-peer relationships and support, and improved recognition by plaints between providers and users of services), and helping health
the health system.6 providers and authorities in communicating key messages and priorities
to communities. As a result, their recognition by providers and authori-
ties grew, leading to demand for them to support other programs such
4.4 | Benefits for researchers and increasing as nutrition, environmental health and school health. The increase in
demand for local data recognition and demand also contributes to sustaining the work.

All programs note the benefits for researchers of engaging in learning

sites, including gaining a richer understanding of resources and chal- 4.6 | Supporting tailored innovations to strengthen
lenges in the local health system, but also deepening relationships systems
with policy-makers (beneficial for research and also its uptake) and
identifying opportunities and priorities for future research All of the programs noted gains in relation to their core ambition to cre-
(as highlighted, e.g., in5). ate conditions for dialogue and innovation to strengthen local systems.
In addition, in many settings, health data are fed upward but not For example, ReSYST has documented the co-production of local
used locally for priority setting and resource allocation, which in interventions—small wins that increase the confidence and agency of
decentralized contexts is particularly problematic. Learning sites managers to make a difference in their sphere of influence and can
encourage more active use of co-developed, relevant, and owned have ripple effects to other areas of engagement. In Kenya, the plat-
local data (in its many forms, including community and managerial per- form brought together staff across levels of the now-fragmented
ceptions, as well as surveys and routine sources). In ReBUILD, for decentralized health system to share concerns and perspectives, result-
example, appreciation of support in analyzing local data was noted ing in practical improvements such as new TORs for health facility in-
from the municipality, which has become a champion in this area and charges and new induction plans. In PERFORM2Scale, health workforce
has started to allocate funds for research in annual plans.7 and service delivery problems were addressed, including implementing
changes to areas such as induction, attendance, supervision, perfor-
mance appraisal, and rewards and sanctions. In ReBUILD, the co-
4.5 | Building local relationships and created action plans were owned by the local government and the
intersectoral work actions (immediate ones) were incorporated into the municipality
annual workplan and budget, while agreeing to gradually incorporate
Spaces for collaborative reflection, dialogue, and planning for action other actions in longer-term plans and practice. These measures
across organizational silos and district teams (and between system address strengthening governance mechanisms at municipality and
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12 of 13 WITTER ET AL.

health facilities and improving planning using local evidence. These well recognized, in the form of partiality and challenges of generaliz-
gains were achieved through the strategies highlighted above, including ability and scale-up, as the engagement needs to be embedded, inten-
continuous reflective engagement, respectful partnership, creating safe sive and sustained. However, if health policy and systems research
spaces, and capacity development. aims at impact, then this approach is worth investing in, as the process
of research is itself the means through which capacity can be built and
change achieved, breaking down the barrier commonly found
4.7 | Improvements in service funding and delivery between research and research uptake and empowering local actors
(addressing problematic and exploitative norms of extractive and
The advocacy of organized users of services has resulted in improved decontextualize working that some research has adopted in the past).
funding for local healthcare services in Guatemala. Still, many bottle- There has been a focus on the importance of fostering “learning
necks are caused by actions at central levels, and therefore, CEGSS is health systems” in low- and middle-income countries in recent years,
supporting organized users to also engage at national level. with a review finding that learning comes from the connection
In other sites, there is evidence of improved service delivery linked between information, deliberation, and action, which can be fostered
to learning site activities. For example, improvements in access to ser- by creating spaces and resources for communities, staff, and managers
vices were shown in a district in Ghana which selected the problem of to share experiential knowledge.25 This is exactly the role which learn-
low outpatient department (OPD) attendance (PERFORM2Scale). By ing sites have been trying to take on, and it remains crucial, particu-
strengthening Community Health Committee meetings and the regular larly when systems face resource constraints and everyday or acute
engagement of health staff with the community, outpatient attendance stressors and shocks.
increased, as well as community participation in health campaigns, such The programs, which are only a small subset of existing learning
as mass drug administration for neglected tropical diseases. These sites but which do include a wide range of geographies and focal areas,
improvements were recognized in the annual district review and resulted highlight some important contextual factors which support effective
in that district being awarded “best performing district in the region.”9 learning sites, including pre-existing relationships and infrastructure
Improvements in patient adherence to HIV/AIDS treatment and better (such as the HDSS platform in Mpumalanga and prior programs for
patient tracking were among the benefits documented for VAPAR.24 ReBUILD, PERFORM2Scale, and ReSYST), along with supportive orga-
nizational environments, local champions, and stable, longer-term fund-
ing.26 Within the sites, regularity of engagement over time (e.g., over
4.8 | Contributing to broader policy and practice repeated PAR cycles), building respectful relations, fostering capabilities
and mutual connections, adaptability, and linking participants to rele-
In addition to local benefits, all learning sites aim to develop lessons that vant external resources (including peer-to-peer learning) are highlighted.
can be shared horizontally (e.g., to other municipalities or districts) but Attention to power relations within the group is also key, although
also vertically, in informing national policy and practice. This is often there can be tensions to be carefully managed here, including between
done through engagement of participants in technical working groups giving voice to marginalized groups but also ensuring that those with
at different levels. For example, in Kenya, the ReSYST learning site con- power to act on problem areas are involved. Engaging wider actors at
tributed to the Kilifi County Health Facilities Improvement Fund Bill in local and higher system level is also required, as constraints are gener-
2016 and the Ministry of Health Guidelines for County Level Health ally driven by multi-level factors that local actors alone cannot address.
Sector Annual Work Planning and Performance Review Processes in It is clear that the role of learning site facilitation is highly skilled,
2018. There has also been considerable effort put into scaling up and and includes the ability to build relationships with and across partners,
sharing lessons more widely—for example, VAPAR has seeded a new create constructive and respectful engagement, maintain group cohe-
learning platform in a neighboring province, and is now responding to sion and enthusiasm, find opportunities to input to local plans without
demand from the provincial health authority to implement the CHW over-committing, be responsive to local requests, communicate effec-
training across Mpumalanga. In PERFORM2Scale, scale-up strategies tively to multiple actors, be reflective about progress and lessons, and
were developed in each country with integration of the management be able to document systematically and share findings widely. These
strengthening intervention into policies and routine practice in Uganda roles can be shared across the team and partners and can be taught,
and Malawi, all in the absence of additional financial support. nurtured, and supported, where not inherent.
Sustainability is partly achieved through the enhanced capabilities
and connections highlighted above, but in addition, the learning sites
5 | DISCUSSION undertake deliberate efforts to encourage incorporation over time in
routine local processes, which can support learning activities beyond
Reflecting across the five programs, which incorporate a number of the funding timetable of research programs. This is challenging, but
sites, it is clear that the learning site approach can be potentially very crucial as learning sites require long-term engagement. Evidence of
powerful in developing relationships, capacities, and local innovations. demand for continuation of the platforms at local level in these pro-
The aim is to break down the traditional relationship between grams, such as the scale-up of the VAPAR work to provincial level and
researcher and researched, with researchers working with local actors the growth and long timespan of the CEGSS program, is encouraging
to co-create and study participatory solutions. The risks of this are in that respect.
 23796146, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lrh2.10475 by Cochrane Costa Rica, Wiley Online Library on [10/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
WITTER ET AL. 13 of 13

ACKNOWLEDGMENTS 12. Witter S, van der Merwe M, Twine R, et al. Evaluating a community
We would like to thank all of our colleagues in the learning sites and and health system learning platform: findings from the VAPAR experi-
ence in Mpumalanga, South Africa. Submitted to Wellcome Open
programs, whose experience is reflected in this report. In addition, we
Research. 2024a.
acknowledge the funders of these research and other programs, 13. Raven J, Mansour W, Aikins M, et al. From PERFORM to
including FCDO/UK Aid for ReSYST and for ReBUILD for Resilience PERFORM2Scale: lessons from scaling-up a health management
(PO 8610), the European Union Horizon 2020 Research and strengthening intervention to support universal health coverage in
three African countries. Health Policy Plan. 2024;39:841-853.
Innovation Programme for PERFROM2Scale (Grant Number 733360),
14. PERFORM Consortium. PERFORM2Scale, toolkit for the management
and the Health Systems Research Initiative from Department for strengthening intervention. Open-source version 1.0. PERFORM
International Development (DFID)/Medical Research Council (MRC)/ Consortium. 2020.
Wellcome Trust/Economic and Social Research Council (ESRC) 15. Mpumalanga Health Policy and Systems Learning and Platform. Commu-
nity health workers community mobilisation training manual. 2022.
(MR/N005597/1, MR/P014844/1) for VAPAR.
https://www.vapar.org/_files/ugd/532615_
d50539e3dc82416980f5c4732e11eb10.pdf
CONF LICT OF IN TE RE ST ST AT E MENT 16. VAPAR Mpumalanga Health Policy and Systems Research Learning
The authors have no conflict of interest to declare. Platform. VAPAR Mpumalanga health policy and systems research
learning platform. 2022. https://www.vapar.org/training
17. VAPAR Mpumalanga Health Policy and Systems Research Learning.
ORCID
Media. 2024. https://www.vapar.org/media-1
Sophie Witter https://orcid.org/0000-0002-7656-6188 18. Witter S, van der Merwe M, Twine R, et al. Verbal autopsy with par-
ticipatory action research (VAPAR) programme in Mpumalanga,
RE FE R ENC E S South Africa: protocol for evaluation. BMJ Open. 2020;10(2):
e036597. doi:10.1136/bmjopen-2019-036597
1. Gilson, L, Nxumalo, N., Molyneux, S. Health system learning sites:
19. Kok M, Bulthuis S, Dieleman M, et al. Using a theory of change in
understanding health systems through research collaboration and
monitoring, evaluating and steering scale-up of a district-level health
engagement. 2016. https://resyst.lshtm.ac.uk/resources/health-system-
management strengthening intervention in Ghana, Malawi, and
learning-sites-understanding-health-systems-through-research-collaboration
Uganda – lessons from the PERFORM2Scale consortium. BMC Health
2. Oliver K, Kothari A, Mays N. The dark side of coproduction: do the
Serv Res. 2022;22(1):1001. doi:10.1186/s12913-022-08354-y
costs outweigh the benefits for health research? Health Res Policy
20. Hernandez A, Hurtig AK, San Sebastian M, Jerez F, Flores W. “History
Syst. 2019;17:33. doi:10.1186/s12961-019-0432-3
obligates us to do it”: political capabilities of indigenous grassroots
3. Loewenson R, Laurell AC, Hogstedt C, D'Ambruoso L, Shroff Z. Key
leaders of health accountability initiatives in rural Guatemala. BMJ
features of participatory action research. Participatory Action Research
Glob Health. 2022;7(5):e008530. doi:10.1136/bmjgh-2022-008530
in Health Systems: A Methods Reader. The University of Aberdeen;
21. Hernández A, Ruano AL, Hurtig AK, Goicolea I, San Sebastián M,
2014:11-36. https://www.researchgate.net/publication/283715907_
Flores W. Pathways to accountability in rural Guatemala: a qualitative
Participatory_Action_Research_in_Health_Systems_A_Methods_
comparative analysis of citizen-led initiatives for the right to health of
Reader (Accessed: 3 April 2024).
indigenous populations. World Dev. 2019;113:392-401. doi:10.1016/
4. Mabetha D, Ojewola T, van der Merwe M, et al. Realising radical
j.worlddev.2018.09.020
potential: building community power in primary health care through
22. Nzinga J, Boga M, Kagwanja N, et al. An innovative leadership devel-
participatory action research. Int J Equity Health. 2023;22(1). doi:10.
opment initiative to support building everyday resilience in health
1186/s12939-023-01894-7
systems. Health Policy Plan. 2021;36:1023-1035. doi:10.1093/
5. Barasa E, Boga M, Kagwanja N, et al. Learning sites for health system
heapol/czab056
governance in Kenya and South Africa: reflecting on our experience.
23. Hove J, D'Ambruoso L, Kahn K, et al. Lessons from community
Health Res Policy Syst. 2020;18:44. doi:10.1186/s12961-020-00552-6
participation in primary health care and water resource governance
6. D'Ambruoso L, Abruquah NA, Mabetha D, et al. Expanding commu-
in South Africa: a narrative review. Glob Health Action. 2022;15:
nity health worker decision space: learning from a participatory action
2004730.
research training intervention in a rural south African district. Hum
24. Witter S, van der Merwe M, Twine R, et al. Opening decision spaces:
Resour Health. 2023;21(1):66. doi:10.1186/s12960-023-00853-1
analysis of opportunities and constraints in the public health sector of
7. Regmi S, Bertone MP, Shrestha P, et al. Co-creation as a promising
Mpumalanga Province, South Africa. PLoS One. 2024b;19:e0304775.
approach for building a resilient local health system in Nepal.
25. Witter S, Sheikh K, Schleiff M. Learning health systems in low-income
Manuscript in draft. 2024.
and middle-income countries: exploring evidence and expert insights.
8. D'Ambruoso L et al. Rethinking collaboration: developing a learning
BMJ Glob Health. 2022;7:e008115.
platform to address under-five mortality in Mpumalanga province,
26. Bulthuis S, Kok M, Onvlee O, et al. How to scale-up: a comparative
South Africa. Health Policy Plan. 2019;34(6):418-429. doi:10.1093/
case study of scaling up a district health management strengthening
heapol/czz047
intervention in Ghana, Malawi and Uganda. BMC Health Serv Res.
9. Raven J, Aikins M, Banda H, et al. Perform 2 scale synthesis report,
2023;23(1):35. doi:10.1186/s12913-023-09034-1
Improving Health Workforce Performance. pp. 1–49. 2022. https://
www.kit.nl/wp-content/uploads/2022/05/Perform_To_Scale_Synthesis-
report.pdf
10. Hove J, D'Ambruoso L, Mabetha D, et al. Water is life’: developing
community participation for clean water in rural South Africa. BMJ How to cite this article: Witter S, Regmi S, Raven J, et al.
Glob Health. 2019;4:e001377. Learning sites for health systems research: Reflections on five
11. Oladeinde O, Mabetha D, Twine R, et al. Building cooperative learn- programs in Africa, Asia, and Central America. Learn Health
ing to address alcohol and other drug abuse in Mpumalanga,
Sys. 2024;e10475. doi:10.1002/lrh2.10475
South Africa: a participatory action research process. Glob Health
Action. 2020;13(1):1726722.