Understanding global health research collaborations: an Indigenous realist evaluation of the African Research Initiative for Scientific Excellence programme

Abstract

Introduction

Global North–South research collaborations are complex and dynamic despite their centrality to global health research. We conducted an Indigenous-inspired realist evaluation to examine how and why research collaborations under the ‘African Research Initiative for Scientific Excellence’ programme work to strengthen African researchers’ health research capacity. With funding from a Global North partner, the programme supports multiple research projects, each led by an African principal investigator (PI).

Methods

We adopted a multicase study design with three cases covering laboratory-based, policy-based and innovation-based health research projects. Realist-informed qualitative methods were used to test an initial programme theory (IPT). We conducted 32 realist interviews with the PIs, collaborators, research support staff, masters/PhD trainees and 4 programme-level staff. Retroductive theorising guided the testing of the IPT through context–mechanism–outcome configurations (CMOs). We adopted a realist approach to thematic analysis, identifying experiential, inferential and dispositional themes to formulate CMOs. Through retrodiction—theoretical abstraction—we refined the IPT using CMOs from the within-case theories.

Results

Multiple mechanisms (ie, trust, empowerment, motivation, inspiration, sense of agency and responsibility, and psychological readiness for policy leadership) were triggered to generate varied outcomes (ie, equitable research collaboration established, improved knowledge exchange and learning and appointment of the PI to policy leadership role) across the three cases. Where the African PI has autonomy over research resources, but research expertise and equipment are locally unavailable, the PI’s access to research funding and liberty to choose collaborators from anywhere enacts a sense of agency to establish equitable, collaborative relationships with Global North partners to leverage resources. A longstanding relationship between the PI and the collaborator engenders trust, while a sense of responsibility is enacted if the collaborator is committed and renowned in their research field. Also, continued government engagement motivates and engenders ‘psychological readiness for policy leadership’, leading to meaningful collaborations.

Conclusion

Increasing direct funding support to African PIs and giving them the autonomy to define research priorities and implement their research initiatives can enhance their ability to conduct impactful research addressing local health concerns.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Global health research collaborations/partnerships are complex, and the experiences of the researchers differ from one context to another.

• Power imbalances stemming from the colonial structures and legacies and exacerbated by the funding flows from the Global North to the South add to the complexity of global health research collaborations.

WHAT THIS STUDY ADDS

• This study identifies important context conditions for establishing transformative and equitable collaborations in global health.

• It also highlights the critical role that sense of agency, inspiration, trust, sense of responsibility, motivation and empowerment play in fostering effective and equitable research partnerships.

• Employing an Indigenous realist evaluation approach can help us understand how global health research collaborations work, why, for whom and under what conditions.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Understanding what mechanisms are triggered under what conditions is important for health research funding and implementing actors (in Africa and generally Low- and Middle-Income Countries) in their efforts to address power imbalances in research collaborations/partnerships and pursuit of equity.

• Understanding what works, for whom, and under what conditions can help research partnership advisory boards to oversee the operationalisation of health research partnerships, ensuring mutual respect, value and benefit, and are fit for purpose considering the uniqueness of each partner ecosystem.

Introduction

Research collaborations are central to (Global) health research capacity strengthening efforts.¹ The value of collaborations in global health research is widely acknowledged,2,4 and translational science demands establishing effective research collaborations given science’s complex and interdisciplinary nature.⁵ Researchers have different motivations for developing collaborations, including inter alia sharing research resources and expertise, improving funding potential by working on joint grant applications, learning tacit knowledge from established expertise and enhancing productivity through co-authorship.^{6 7}

Research collaborations are complex and messy as they are undertaken in complex research environments.^{8 9} One complexity that researchers, specifically in the Global South, have to contend with is asymmetry informed by the legacy of colonialism or coloniality.¹⁰ Coloniality-informed power asymmetry is compounded by global health research funding, mainly from the North to the South.^{11 12} Dodsworth¹³ argues that coloniality not only shapes a researcher’s choice of a collaborator but also determines who makes that choice from a position of power. Inequity in research collaborations is a consequence of ideological and structural factors grounded in coloniality that produce and reproduce power imbalances and which researchers—seeking to make collaborative projects in Africa more equitable—must proactively and continuously contend with throughout the life cycle of their research projects.¹³

In their article, Molosi-France and Makoni¹⁴ wondered what would happen if research funders directly funded researchers in the Global South and encouraged them to determine their collaboration relations. To establish an even playing field in global health research, we must promote equitable collaborations between researchers from the Global South and those from the Global North, which requires addressing existing power imbalances.^{15 16} One way to address power imbalances within collaborations and reduce exploitative forms of global health research is to promote the ethical principle of shared decision-making.¹⁷ When collaborations are equitable, both the funding and implementing partners look at each other’s contributions as a critical part of a global and collective struggle for health and well-being for all,¹⁸ regardless of their economic or geographical divide.

While the global health research community, including research funders, encourages interdisciplinary and collaborative research, the researchers’ experiences with collaborative processes have been understudied.¹⁹ This study uses the ‘African Research Initiative for Scientific Excellence’ (ARISE) programme as a real-world case. The ARISE programme is a 5-year initiative (2022–2026) jointly implemented by the African Academy of Sciences (AAS) and the African Union with financial support from the European Commission (EC). The programme seeks to (1) strengthen the capacities of the emerging African research leaders committed to a research and teaching career in Africa, (2) strengthen institutional research management and support systems for research to thrive and (3) support the generation of cutting-edge research that will contribute towards transforming lives in Africa. The ARISE programme supports African early-career researchers, also called principal investigators (PIs), who have demonstrable potential to become research leaders. The PIs are offered up to €500 000 to implement research projects exploring various research questions related to health, climate change, among other thematic areas. They are expected and encouraged to establish and leverage research collaborators’ resources and expertise, engage research assistants (ie, PhD and master’s trainees) and research support staff to implement their research projects effectively. In this study, we aimed to explore how and why these PIs within the ARISE programme form Global North collaborations and the power dynamics at play in these collaborations using indigenous realist evaluation.

Indigenous realist evaluation

Realist evaluation investigates what works, how, why, for whom and under what circumstances²⁰ through the formulation of context, mechanism and outcomes (CMO) configurations.²¹ It entails working with different stakeholders to gather information to generate theories about how and why programmes work. The Indigenous realist evaluation integrates the Indigenous research principles described by Chilisa and Mertens²² and Pawson and Tilley’s²⁰ realist evaluation approach to obtain an indigenous sensitive theory-driven research approach among Indigenous and/or formerly colonised settings.²³

Nine principles guide indigenous research (1) relationality which emphasises the role of the programme beneficiaries in prioritising their needs and judging the relevance of programmes, (2) responsibility which emphasises that the evaluator has a critical role in addressing inequities and social injustice by unravelling them and speaking on behalf of the voiceless, (3) reverence which recognises that spirituality and values contribute to ways of knowing and the evaluator must examine how both intersect, (4) reciprocity which emphasises that the evaluation must examine the value added to the beneficiaries’ lives by the programme and how the beneficiaries respond to the programme resources, (5) respectful representation which emphasises that the evaluation cycle should be beneficiary centred and give beneficiaries ownership and access to the evaluation data, (6) reflexivity which inspires the evaluator to continuously reflects on their positionality and how it may affect their attitude, judgement or decisions throughout the evaluation, (7) responsivity which inspires the evaluator to learn from the process and adapt their approach, methodologies and tools to become context and culturally sensitive, (8) rights and regulations principle which emphasises that the ethical protocols should grant stakeholders and beneficiaries the right to prioritise their needs, participate in the evaluation and have ownership of the data and report and lastly, (9) decolonisation principle which emphasises that the evaluation should examine power structures and dynamics and challenge the Western values and standards in evaluations conducted in Africa.²²

While an amalgam informs the realist evaluation approach of critical and scientific realism principles,²⁴ the Indigenous realist evaluation would take on a multiparadigmatic perspective that includes realism and postcolonial Indigenous paradigms.²⁵ Realist evaluation is focused on explaining social phenomena and understanding how and why programmes work or not through formulating theories and models.²⁴ Conversely, the Postcolonial Indigenous paradigm examines (post)colonial ideologies and power dynamics when researching Indigenous or formerly colonised societies.²⁶ Indigenous research principles promote ethical research conduct, emphasising the need for evaluators to carry out transformative evaluations that address inequities and socioeconomic injustices and centre the voices of the Indigenous programme participants.²² An Indigenous realist evaluation approach would, therefore, go beyond addressing the question of what works, how, why, for whom and under what circumstances to ensure that the evaluation centres the participants’ voices and contributes to addressing social injustice and inequities. We recently proposed how Indigenous-inspired realist evaluation could be operationalised in the context of (global) health research capacity strengthening²⁵ and presented an Indigenous realist evaluation framework.²³

Overview of ARISE research collaborations/partnerships

The ARISE programme supports the establishment of effective collaborations with different actors (eg, academia, industry, policymakers, innovators, etc) to implement the scientific research component and strengthen the research capacity of the stakeholders. A review of the Programme’s Description of Action—a document that describes the ARISE programme architecture—includes a clause stipulating that the PIs are encouraged, but not required, to forge collaborations with Global North counterparts.

We first elicited an initial programme theory (IPT) in line with the realist methodological approach, as summarised in online supplemental figure 1. We reviewed six ARISE programme documents, including one funding proposal, the programme description of action document, the Monitoring & Evaluation (M&E) framework and three grantees’ progress reports. Following the review, a rough IPT was drafted. We then revised the IPT by conducting three focus group discussions (FGDs) with the project-level participants (one FGD per case) and three interviews with the programme partners. Each FGD comprised different participants as shown in online supplemental table 1.

The realist interviewing technique was used, and the draft IPTs were discussed during the focus groups and interviews. Pawson et al²¹ argue that programme designers and/or practitioners should be consulted during the IPT development process, while beneficiaries/ programme participants should be consulted when testing the theories. However, since the Indigenous research principles emphasise the need to involve all participants throughout the evaluation process,²² it is on this basis that the ARISE beneficiaries were consulted at the IPT development stage.

In step 3, the IPT was further revised through a literature review. Six⁶ papers were included in the final review. The papers were reviewed, and relevant data related to the context, mechanism and outcomes that supported the IPTs as framed or could inform further refinement were extracted. The IPT was framed as shown in box 1 below.

Box 1. Initial programme theory.

IF the principal investigator (PI) has a network of potential collaborators who have complementary research equipment and expertise for transdisciplinary research (C) and the PI has the financial resources and the autonomy to implement collaborative health research and research capacity strengthening (intervention).

THEN equitable research collaborations will be established (O).

BECAUSE the PI will be empowered and motivated to leverage their collaborators’ equipment and expertise (M).

Our goal in this paper is to test and refine the ARISE IPT to establish how and why the research collaborations were working to strengthen the research capacity of the ARISE stakeholders and beneficiaries, and under what conditions. Online supplemental table 2 describes the nature of partners across the three cases; these details will help the reader to operationalise the results section.

Methods

Koenig²⁷ confirmed that realist evaluation aligns with the case study research design. To this end, we adopted a multicase realist evaluation design with three case studies. The selection of the three cases was based on the size of the host research institution, the economic status of the host country (which was hypothesised to influence budgetary allocation for health research at the national level) and the nature of the research. Our three case studies constituted innovation-based, policy-based and laboratory-based research projects. The case studies were sequentially conducted to ensure that insights from one case study were applied in the subsequent one—see online supplemental figure 2. Each case formed a unit of analysis, and both within-case and cross-case syntheses were conducted in line with the guidance provided by Yin.²⁸

Data collection and analysis

Only realist-informed qualitative methods were employed in the study. The choice of methods was based on the understanding that Indigenous research principles lend themselves well to qualitative methods, given their reliance on indigenous data collection methods such as storytelling.^{29 30} Participants were purposively selected based on their experience with the programme and the specific programme activities and component(s) they were engaged in.

Data collection in realist evaluation allows for formulating and interrogating CMO hypotheses to verify, validate or refute the programme theories.³¹ We used the realist interviewing technique whereby IPTs are the subject of discussion and the basis for gathering additional information to clarify, modify, approve or discredit the IPTs.³² We specifically found the realist interviewing ‘starter set’ of questions useful.³³ Specifically, realist interviews, storytelling and observations were used.^{24 34 35}

Our data collection goal was to test our theorised inner workings of the programme by eliciting information from the stakeholders’ based on their experience of how the programme works. Probing questions sought to elicit deeper information about relevant context conditions, generative mechanisms or emerging outcomes.

Data were collected through (1) in-depth interviews with the PIs and research collaborators, PIs’ mentors or supervisors and research support staff, (2) observation sessions with the PIs, (3) storytelling with the masters/PhD trainees and (4) key informant interviews with programme-level partners including the AAS, the EC and the ARISE’s scientific steering committee representatives—see table 1.

Table 1. Summary of data collection and participants.

Data collection method	Total number
In-depth interviews with PIs, collaborators, PIs’ mentors or supervisors and research support staff	24
Participant observation sessions with the PIs	3
Storytelling sessions with masters/PhD students	8
Key informant interviews with the programme partners	4

PIs, principal investigators.

In line with Manzano’s³⁶ suggestion, all the data collection tools were structured around causal links within programme theories rather than the separate context (Cs), mechanism (Ms) and outcomes (Os). Before engaging the participants, we shared the participant information sheet with them, which provided the details of the study and the consent forms for their signature if they decided to participate. This approach ensured that their participation was voluntary and that they had a choice not to participate if they wanted to.

Our data analysis was informed by Wiltshire and Ronkainen’s³⁷ realist approach to thematic analysis, which was achieved by formulating experiential, inferential and dispositional themes. The experiential, inferential and dispositional thematic analysis in the realist-informed thematic analysis is aligned with the realist philosophical underpinnings of real, actual and empirical domains.³⁷ Data analysis began at the theory development stage, identifying and placing information within wider CMO configurational maps.³⁸ The analysis helped identify emerging data gaps that must be addressed in subsequent interviews, observations or storytelling sessions. Consistent with realist approaches, retroductive theorising informed the overall data collection approach guided by the CMO heuristic, and abductive reasoning (both inductive and deductive logic) was applied to the various data sources.³⁹ We also drew on our understanding/experience (acquired through the experience of supporting research capacity-strengthening programmes in Africa) to formulate mechanism-based theories. Since the elicitation of IPTs and the refinement of theories partially depend on the evaluator’s judgement and knowledge,⁴⁰ reflective journaling was used to document the decisions made during the analysis, thus ensuring transparency. Using the CMO heuristic tool, we identified excerpts denoting mechanisms (eg, incentives, motivations, skills, knowledge, etc) triggered by the ARISE across the three project contexts to generate intended or unintended outcomes.

Where too many contexts and mechanisms were identified, those elements were assessed for their pragmatic utility⁴⁰ to prioritise CMO elements with greater (but not final) epistemic credibility and greater explanatory potential. This approach entailed applying the notion of judgemental rationality to evaluate and compare the explanatory power of different CMO configurations and to select theories that most accurately capture research collaborations in each of our three cases. To do this, data fragments were abductively chosen to build CMO configurations, especially where multiple mechanisms were simultaneously present, as characterised by a long causal chain leading to outcomes.

The process of retrodiction was applied to build the overarching or refined programme theory explicating the nature of research collaborations in the ARISE programme consistent with Mukumbang.⁴¹ In retroduction, we identify individual causal pathways and the mechanisms that produce them, and in retrodiction, we investigate what mechanism bundles interacted in what way to produce any particular event.⁴¹ For each programme theory, we constructed ‘if…then…because… statements’ and figures representing ‘what works, for whom, how and why, and under what conditions’ to improve the evidence translation for policymakers and funding agencies.⁴²

Patient and public involvement

Neither patients nor the public were involved in the study’s design, implementation or dissemination.

Results

Three case-based theories and a refined programme theory are presented. A reflexivity statement has been included to help the reader situate the authors’ positionality and how it might influence the research—see online supplemental file 1.

Theory 1: trust, motivation, inspiration and sense of responsibility (case A)

In case A, multiple contexts were identified as key in triggering specific mechanisms. The PI reported that it was her first time leading a research project (C1) of the ARISE magnitude. As a first-time PI (C1), the researcher was keen not to fail in her first research project, and this triggered motivation (M2) and a sense of responsibility (M4) to effectively work with and get the most out of the collaboration relationship. The PI and the collaborator reported that they had previously worked together on specific research initiatives (C2), and this longstanding relationship triggered trust (M1). Importantly, the PI reported that the collaborators were committed and distinguished researchers (C3) in their fields who had generated impactful research. Working with committed and authoritative research collaborators (C3) inspired (M3) and motivated (M2) the PI and activated a sense of responsibility (M4) needed to optimise the benefits/value of the collaboration relationship. This theory is captured in the quotes below.

…the other thing is trust and confidence [M1] in your team members; you need a reliable collaborator who will do the job, follow through, and deliver high-quality outputs. …it demands a lot of commitment, and you need to, you know, be there whenever the team needs you. …It motivates the PI [M2] to know that they have people who have their back. I’ve been there myself. …the PI and I have previously collaborated on several small projects, which has helped us build a trusting relationship. [Interview, Collaborator, Case A]

I think one of the key aspects about doing the project is the current collaboration we have with the [Global North partner]. I am here in [Europe] to advance the research I was doing at [institution], so I’m working at their lab. …so, I’ve learned so much from the team, particularly their experience in deploying 5G in digital health, but I’ve also shared with them the results from my research and the contextual challenges we have faced deploying such technologies in some of the remote areas in Africa. …there is incredible learning and knowledge exchange that has happened between us (O), thanks to the ARISE programme and the collaboration between the Dr. [PI] and [Global North institution]. [Storytelling, Masters Student, Case A]

I think that by working with such distinguished researchers in your field and having mutual trust and respect [M1] among you, you will feel that you cannot fail the team but show leadership. It not only motivates [M2] you are working with such a resourceful team of collaborators, but I think it somewhat inspires you [M3], as the PI, to take responsibility and show leadership [M4]. …the stakes are even higher if it’s your first time [PI] leading a research project because failure is not in the equation. [Interview, PI, Case A]

Highlighting the complexity of the collaboration, the PI shared that being an ‘authoritative researcher’ was not enough and, therefore, commitment was a key trait. The PI pointed out that working with an authoritative/distinguished research collaborator in a previous research project was problematic because the collaborator was extremely busy and lacked commitment. Notably, the PIs and their local (individual) collaborators highlighted that because of their history of working together, they did not need to draw a Memoranda of Understanding (MoU) or Agreement because the collaboration was based on trust and respect. However, each party knew their role and the allocated project activities. The PI, however, highlighted that working with institutional partners was different, and an MOU stipulating the responsibilities of each party was required. The MOUs, according to participants, further cemented the trust.

I think it’s a different dynamic when dealing with individual collaborators, like my colleagues at the University and the institutional partners. For individuals, and particularly when we have that history of working together, we already have built trust [M1], so we don’t need an MOU between us. We just needed a meeting where I shared the project implementation plan, highlighting where I need their help [M4] Sense of responsibility], and that was it. Everyone has a plan and knows what to do. …I can attest that the collaboration has allowed us to improve learning and knowledge exchange and share our research resources. [Interview, PI, Case A]

In this case, multiple contexts were necessary for the multiple mechanisms responsible for generating the two outcomes: shared research resources (outcome 1) and improved knowledge exchange and learning (outcome 2).

The programme theory has been summarised—see figure 1.

Figure 1. Programme theory 1. PI, principal investigator.

This theory can be represented using if…then…because the statement is shown in box 2 below.

Box 2. Programme theory 1.

IF a first-time PI (C1) has had ‘pre-African Research Initiative for Scientific Excellence engagement’ with a research collaborator (C2) and the collaborator is committed and renowned in their research field (C3), providing the principal investigator (PI) with the opportunity and resources to choose collaborators from anywhere (I).

THEN the research expertise and resources will be shared (O1) and knowledge exchange and learning improved (O2).

BECAUSE the PI forms trusting relationships (M1) and is motivated (M2), inspired (M3) and has a sense of responsibility (M4).

Theory 2: motivation and psychological readiness (case B)

In case B, the ‘pre-ARISE engagement’ with the national government ministries (C1) and the appreciation of the PI’s research contribution and leadership to health research by the policymakers (C2) were identified as critical contexts. The ARISE initiative has provided the resources and the opportunity for the PI to continue engaging and working with the national government ministries (ie, Ministry of Health and Ministry of Education), thus motivating (M1) the PI and enhancing their psychological readiness for policy leadership (M2). Participants highlighted that the championing of research uptake and translation by the PI among policymakers was crucial in demonstrating the value that research/evidence could add to the policy work. Some of the participants shared the following:

…we have been working with the [national ministry] in the pilot study that became the foundation for the ARISE project. …we have strengthened our collaboration for the last two years since we’ve worked together. …this [national ministry] is a partner you must have on board if you want your research to generate policy or sector level impact. …I think motivation has been a key driver [M1] because I know I can make a difference as a researcher in this field. [Interview, PI, Case B]

…as a result of her involvement and being vocal about [specialised] health work in [country], the PI was appointed as a member of the national taskforce on the [health field]. …I think that [collaboration] has given her the space or the platform for her to demonstrate the value and the contribution she can make to the [specialised] health issues in this country [psychological readiness for policy leadership role] [M2] and it has been recognised and appreciated by the government agencies we’ve been working with. [Interview, Project Manager, Case B]

The collaboration relationship has consequently opened new opportunities for the PI to be appointed to a policy-level leadership role (O1). The participants highlighted that the collaborative relationship has given the PI the platform to demonstrate the value she could add to government policy work and become a respected voice in the health (research) field. The appointment of the PI into a policy leadership role (O1), according to one of the participants, makes it possible for them to promote buy-in and ownership by policymakers, which is necessary for research uptake (O2). In this case, the PI’s appointment emerged as a positive but unanticipated outcome.

The programme theory two has been summarised—see figure 2.

Figure 2. Programme theory 2. PI, principal investigator.

This theory can be represented using if…then…because statements, as shown in box 3 below.

Box 3. Programme theory 2.

IF the principal investigator (PI) has had ‘pre-African Research Initiative for Scientific Excellence engagement’ with the national government partners (C1) and the government ministries (partners) value/ appreciate the PI’s contribution to health research (C2), by providing the PI with resources to continue working with/ engaging the government ministries in research (I).

THEN the PI is appointed into a policy leadership role (O1) and there is improved knowledge uptake and translation of research output in policymaking (O2).

BECAUSE the PI is motivated (M1) and psychological readiness (M2) for policy leadership role is activated.

Theory 3: activated sense of agency and empowerment (case C)

Based on our observations, interviews and storytelling, when faced with the challenge of locally unavailable research expertise and equipment (C1), the PI (case C) had to work with a collaborator with the requisite expertise and equipment. The PI highlighted her wish to work with an African partner, but the potential African partners charged significantly higher fees for the data analysis services. As the research project leader, the PI had the autonomy (C2) to identify a collaborator providing complementary research expertise and equipment. These contexts triggered the agency (M1) needed to establish research collaboration with a Global North partner.

…the beauty is that if a partnership, or collaboration as you call it, is not working, I can look for someone else who can do the job without tolerating non-cooperative collaborators. …at the end of the day, you just want to deliver your research project successfully, and, as a leader, you just need to work with the right people. …I think that independence activates a sense of agency [M1] that a PI needs to do what needs to be done. [Interview, PI, Case C]

…as part of the capacity-strengthening arrangement between Dr. [PI] and [Global North partner], I travelled to [Europe] to learn some specific skills in bioinformatics. Given their advanced, highly powered analytical technologies, I could complete the analysis of my data over the three months I was there. These are technologies that I couldn’t access here in [country] or even in Africa [C1]. [Storytelling, PhD Student, Case C]

The quote below shows some collaboration-related decisions the PI (case C) had to contend with, such as those involving African versus Global North-based collaborators and the availability of critically needed research expertise and equipment. While lack of choices can be disempowering for the PI, the ARISE programme availed resources that empowered (M2) the PI with options, thus enabling the PI to access the needed research expertise and equipment outside (if not available in) Africa (outcome). The PI pointed out that although the data processing and analysis were outsourced in her ARISE project, the services were paid for, the roles of both parties were clearly stipulated, and the control of the resources remained solely with the PI.

…but of course, before considering the collaborator, I looked around for potential centres that could do the analyses [agency] [M1], but the challenge is the prices from the African countries were almost double what our collaborator [in Global North] was quoting. In addition, even though there are [specialised field] facilities in Africa, they are dedicated to pathogens and viruses. Very few are dedicated to analysing human samples. …that’s why Dr. [name] from the University of [name] was my best option [empowerment] [M2]. [Interview, PI, Case C]

The programme theory three has been summarised—see figure 3.

Figure 3. Programme theory 3.

This theory can be represented using the ‘if…then…because’ statements as shown in box 4 below.

Box 4. Programme theory 3.

IF research expertise and equipment is not locally available (C1) and the African principal investigator (PI) has autonomy over a research project (C2), by providing African PIs the resources (funding and support) and the opportunity to choose collaborators from anywhere (I).

THEN the PI will likely form meaningful research collaborations with relevant Global North partners who possess complementary research expertise and research resources (O).

BECAUSE their sense of agency is enacted (M1), and they feel empowered (M2).

Cross-case refined theory

After identifying three within-case programme theories, we systematically conducted a cross-case analysis to generate a more refined theory for understanding the dynamics of research collaborations in global health. The three theories were placed in a juxtaposition, and then, as Mukumbang⁴¹ suggests, we searched for the variations in the contexts accounting for the differences between them and aimed to ‘generalise’ across by looking at how the important outcomes may be achieved. The different CMO elements across the three within-case programme theories were synthesised and an illustration connecting between/among contexts, mechanisms and outcomes was created—see figure 4.

Figure 4. Cross-case theory. PI, principal investigator.

As shown in figure 4, multiple mechanisms were triggered to generate varied outcomes across the three contexts. For instance, empowerment (M) is triggered when the African PI has autonomy over research resources. In a context where the PI faces the challenge of locally unavailable research expertise and equipment, the PI’s liberty to choose collaborators from anywhere (I) can enact a sense of agency (M) needed to establish equitable collaboration with a Global North partner. In a context where the PI has had a longstanding relationship with the collaborator and the collaborator is committed and renowned in their research field, trust (M) will likely be triggered, and a sense of responsibility (M) will be enacted. Importantly, in the context where the African PI is leading a research initiative for the first time, inspiration (M) will be triggered. The multiple mechanisms triggered in varied contextual conditions will enable ‘improved knowledge exchange and learning’. In a context where the PI has had previous engagement with the government ministries, and the government ministries value and appreciate the PI’s contribution to health research, this will not only trigger motivation but the PI’s demonstration of ‘psychological readiness for policy leadership’ will open up new opportunities for the PI, that is, appointment to the policy leadership position and also improved uptake of research outputs by policymakers. Therefore, the programme resources will make it possible for the PI to continue working with and engaging government ministries in research.

Discussion

This study confirms that each research partnership is unique, and the experiences of the local (African) PIs will differ across contexts. This uniqueness is demonstrated by multiple and varied mechanisms (ie, sense of agency, inspiration, trust, sense of responsibility, motivation, empowerment and psychological readiness for policy role) that were triggered across the three contexts to generate varied outcomes (ie, equitable research collaboration established, improved knowledge exchange and learning and appointment of the PI to policy leadership role and improved uptake of research outputs by policymakers). Notably, the study has shown that by providing financial resources to the PIs, guaranteeing them autonomy, and giving them the liberty and space to define their research needs and choose collaborators from anywhere, the PIs were thus able to establish effective and equitable research collaborations.

In a context where the PI is at liberty to choose a collaborator from anywhere, a sense of agency is enacted, and the PI will choose a collaborator who brings in complementary research resources and adds value to the research project. Ngongalah et al⁴³ highlighted the centrality of the researchers’ need for collaborative research and the complementarity of research equipment and expertise as a critical factor in health research capacity-strengthening efforts. Working with a Global North partner was advantageous due to the complementary research expertise and resources that the African PI could leverage in both implementing their research and strengthening the capacities of their masters/PhD trainees. Giving the African PIs the liberty to choose collaborators from anywhere and guaranteeing their autonomy over research resources can empower and give them a sense of agency needed to establish effective and equitable collaborations through which they can access critically needed research expertise and equipment that is not locally available. As Ngongalah et al⁴³ argue, this makes it possible for the research capacity-strengthening programmes in the South to balance issues of epistemic and normative power in global health research with the desire to make strides in research fields dominated by Northern players.

Motivation (M) was triggered in a context where the government ministries valued and appreciated the PI’s contribution to health research (C). Though not explicit, we could deduce that the PI’s motivation was both intrinsically and extrinsically driven, as defined by Ryan and Deci.⁴⁴ It is possible that they are extrinsically motivated by the resources provided by the programme, but also intrinsically by the fulfilment they derive from carrying out collaborative research work that has real value to their country. This is consistent with Svartefoss et al⁴⁵ who posit that researchers are motivated by curiosity and the practical application of their work. For instance, the provision of resources and support through the ARISE and the PI’s continued engagement with the national government ministries (ie, Ministry of Health and Ministry of Education) could have extrinsically motivated (M) them and activated their psychological readiness (M) for a policy leadership role. It is not surprising that, as Ommering et al⁴⁶ argue, the social aspects of collaborative research can be a source of motivation in themselves. Consistent with Ryan and Deci’s⁴⁴ description of self-determination theory, the PI’s intrinsic motivation could be attributed to their psychological need for autonomy (sense of self-control), competence (sense of capability) and relatedness (sense of social connectedness with government actors), which were either part of the context or mechanism activated by the ARISE programme—see figure 4 above. The motivation and the psychological readiness activated by the programme, in this case, resulted in the appointment of the PI to a policy leadership role (O) and improved uptake of research outputs (O). It is reasonable to hypothesise that the appointment of the PI to a policy leadership role would mean, as an unintended outcome, spending less time on actual research.

Trust (mechanism) is triggered when a first-time African PI has a longstanding collaboration relationship with the collaborator(s). To effectively collaborate, Lasker et al⁴⁷ argue that people and organisations need to be confident that other partners will follow through on their responsibilities and obligations and will not take advantage of them. The first-time PIs highlighted that leading a research initiative was a high stake and that trust was key in effectively undertaking collaborative research. This understanding is consistent with Dean et al,⁴⁸ who reported that participants believed pre-existing relationships between partners, such as a former PhD supervisor-student relationship, often resulted in more successful project outcomes. This study also established that the African PIs aspired to mutual trust and accountability in the partnerships—consistent with Matenga et al.⁴⁹ It is not surprising that Parker and Kingori⁵⁰ conclude that ethical and moral dimensions such as trust are critical in judging the merits, sustainability and effectiveness of research collaborations besides the researchers’ competence and commitment to good scientific practices.

One complexity the PIs and individual collaborators highlighted was that they did not see the need to draw a formal MOU and felt confident with this arrangement since they had previously worked with each other. Building an effective collaborative relationship can be more challenging among parties without a history of cooperation and trust.⁴⁷ The pre-existing collaboration is, therefore, essential for building reciprocity, which is central to establishing mutual trust.⁵¹ As Ngongalah et al⁴³ argue, this informal collaboration relationship plays a key role in mitigating power imbalances by diminishing perceived hierarchies between or among individual researchers. Hierarchy in research expertise (working with committed and renowned collaborators) was deemed a critical aspect associated with high-quality research outputs and leadership skills rather than a source of inequity.⁵² While the notion of ‘authoritative’ or ‘renowned’ collaborator can insinuate a ’transactional’ relationship, in this study, it has a transformational angle, which allows for mechanisms that generate desirable effects to be activated. Establishing a collaborative relationship built on trust necessitated a voluntary abrogation of power in equity characteristic of PIs and collaborator relationships, based on the assumption that neither party would be exploited.1553,55 Concerning institutional partners, however, the PIs reported that they had established formal MOUs at the start of the project. Discussing and agreeing on the collaboration and what responsibilities each collaborator will undertake is critical in building trust.⁵⁰ Ensuring clarity and alignment on expectations and how to distribute them among the partners can promote equity.⁵⁶ If partners enter any project without agreeing on where the funds will be managed and by whom, what research priorities are to be addressed, and how the research benefits are perceived and shared among partners, the effect on the partnership can be debilitating.⁴⁸ It is not surprising that Nielsen⁵⁷ cautions that, given the complex social-psychological processes, interpersonal types of trust should not be applied to interinstitutional kinds of collaborations.

The PI’s autonomy over research resources, which has a power implication, determines how effective and equitable collaborations can be established. The findings are consistent with the study by Dean et al,⁴⁸ who found out that the project teams perceived inequity in partnerships as underpinned by where (between Global North and South partner) the award/funds were held and who had decision-making power regarding the spending of resources. The authors established that despite the bulk of the funding (50–70%) being spent in African countries, the African institutions had limited or no autonomy concerning when and how funds were spent.⁴⁸ It is not surprising, therefore, that the participants argued that “allowing the African partner to have more financial control would create more equitable partnerships” (p.7). Therefore, the authors concluded that for the research partnership to be equitable, the Northern partner should be prepared to relinquish control and accept considerable autonomy on the part of the Southern partner.⁴⁸ However, it is worth highlighting that funding partners inherently face a tension between ensuring accountability to taxpayers in their country and the need to guarantee their partners’ autonomy.⁵⁸ For instance, when it comes to managing funding resources, funding partners need assurance that there will be no wastage of funds or potential financial risk.⁵⁹ Therefore, it is unsurprising that under pressure from domestic players, funding partners may exert control over decision-making processes and implement initiatives unsuited to the local communities.⁵⁸

With the increasing call to decolonise global health research, ensuring justice and pursuing equity in research collaborations is critical.60,62 Inequities in partnerships reflect issues such as dominance by Northern partners in funding and its consequences for agenda-setting and power distribution within research partnerships.^{52 60 63 64} This study has established that if the African PI has autonomy and decision-making power, especially regarding administering the grant award and defining the research questions (research agenda setting), equity in partnership will likely be achieved. Similar studies have also found that African researchers were more likely to view partnerships as equitable if initiated by African researchers.^{16 52} To this end, Palmer et al⁶⁵ have suggested a revolutionary, transformative collaborative research framework that gives African researchers authority over research initiatives implemented in African countries. The framework seeks to shift the power to the Low- and- Middle Income Countries’ researchers by providing them with the funding/research resources, the freedom, and the space to conceptualise their realities and imagine better solutions to their research and development problems. One critical feature of the ARISE partnership was the bottom-up approach, in which the African research communities were allowed to define their research priorities and the research challenges they wanted to address through the funding.²³ The engagement of the local African research stakeholders in setting the research agenda and identifying their priorities was critical in achieving equity in partnerships, and it adheres to the Indigenous relationality principle.²³ Importantly, stakeholder engagement ensures that the research addresses the needs of the local African communities beyond the academic interests of researchers.^{23 66}

While Hellowell and Nayna Schwerdtle⁶⁷ argue that as far as funding continues to flow from the ‘North’ to the ‘South’, ensuring proper representation of, respect for, and accountability to researchers in the Global South will remain a complex problem, we assert that a radical shift in funding models and relinquishing of power by Global North partners can foster equitable partnerships. In an ideal world where the Global North and South researchers have sufficient local funding, expertise and equipment to address their local health research needs and priorities, issues of power imbalance would not get in the way of global health research. Nevertheless, since we are far from an ideal world, the intentional pursuit of equity in research partnerships can help address local research needs, elevate the voices of the local people, and adhere to the values of rights, equity and justice68,70 even when the funding is flowing from the North to South. One way to do this is through, for instance, the formulation of partnership advisory boards that can oversee the operationalisation of a partnership and ensure mutual respect, value and benefit and that it is fit for purpose, considering the uniqueness of each partner ecosystem. Importantly, generating evidence on how and why collaborations work can help identify what makes collaboration equitable and effective, thus adopting and implementing evidence-based strategies.

This study has key strengths. First, it provides policy-relevant and practice-relevant evidence and insights on how research collaborations can lead to anticipated and unanticipated research capacity outcomes. It engages with the complexity of health research collaborations and how they contribute to research capacity strengthening in African settings. Second, the study employed an Indigenous realist evaluation approach, which provided an additional (Indigenous) lens to explore power dynamics, understand (in)equity in research collaboration/partnerships and give primacy to participants’ voices and experiences. Since this paper primarily focuses on the CMOs related to research capacity strengthening through research collaborations, the methodological insights on applying the Indigenous realist evaluation approach have been documented elsewhere.²³ Additionally, insights on how the storytelling approach was adapted to uncover ontologically deeper evidence on CMOs have also been reported.⁷¹ Notably, using the case study design and the sequential approach to the cases allowed a more nuanced, iterative and in-depth exploration of the research collaborations across the three cases. Experiences from one case informed the subsequent one(s), and follow-up interviews were conducted with case participants to expound on or clarify some aspects, thus increasing rigour.⁷² Importantly, the study adheres to the Realist and Meta-narrative Evidence Synthesis: Evolving Standards (RAMESES) II reporting checklist and meets the required standards—see online supplemental file 2.

The study was not limitation free. First, the study focused only on the individual-level psychological mechanism. The rationale for focusing on the individual-level psychological mechanism was that the ARISE programme primarily sought to strengthen the research capacities of the African PIs to develop them into research leaders. By engaging with institutional-level collaborations and related mechanisms, the study could potentially result in a nuanced understanding and uncover the complexity inherent in institutional research collaborations. Second, the study employed qualitative methods given the small population and diversity of programme participants/beneficiaries across the three cases (n=47), which could not allow for the administration of a survey. This meant that the outcomes could not be quantified. Notwithstanding, the qualitative methods employed in this study helped identify the broad range of outcomes generated by the ARISE programme across the three cases.

Conclusion

This paper has demonstrated that each research partnership is unique, and the experiences of the local (African) PIs will likely differ from one context to another. While the different contexts will shape the research collaboration mechanisms in varied and often unanticipated ways, key mechanisms such as the PIs’ motivation, trust, sense of agency and empowerment can be transformative, particularly if the PIs have the required research resources, autonomy over those resources and liberty to choose their collaborators. Therefore, there is a need to increase direct funding support to African PIs and guarantee their autonomy to define their research priorities and questions and implement their collaborative research initiatives. Importantly, we argue that establishing partnership advisory boards to oversee the operationalisation of research partnership(s) and ensure that they are based on mutual respect, value and benefit, and fit for purpose can be a transformative way of fostering equitable partnerships.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.