Introduction

There is growing interest in the role of non-governmental organisations (NGOs) in international health research [1–3]. Defined as formally constituted organisations that are largely independent from government and aimed at promoting welfare rather than making a profit [4], NGOs are typically known for their work as service providers or advocates [5–7]. However, research has grown as an area of NGO activity [1,8–11], and there have been calls for more NGO involvement in research from both NGOs and the wider research community [2,12–15]. The World Health Organization Strategy on Research for Health, for example, urges governments to promote NGO participation in national health research systems [16]. This interest in NGO involvement in research is based partly on the value of research for strengthening NGOs’ own work, particularly with increased attention to evidence-based practice in development [2,8,9,17–19]. NGO involvement also has potential value for strengthening health research: partly due to their service delivery and advocacy activities, NGOs are sometimes well-placed to identify relevant research questions, access field sites, and disseminate research findings [2,18,20–24]. Those calling for NGO involvement in research suggest NGOs are and should be active throughout this research cycle, including identifying priorities, conducting research, dissemination and using research findings [2,12,13]. In this article, we focus on NGOs conducting research.

To conduct research, NGOs need research capacity. We aim to contribute to the understanding of NGO involvement in research by providing information on their capacity to conduct research, and to the understanding of health research capacity by bringing experience from NGOs. Based on the experience of three NGOs in Malawi, we examine factors within and outside the organisations that affect key elements of capacity to conduct research.

There is currently limited evidence on capacity to conduct research among NGOs. Although an extensive literature examines capacity for health research [25–31], this literature focuses on academic institutes and think tanks, rather than on NGOs where research is an additional function alongside service delivery and advocacy. Within reports on NGO involvement in research, limited capacity to conduct research is frequently noted [2,8,9,18,22,32–35]. However, these reports provide little detail on factors that limit capacity or indication of how and why capacity varies between NGOs. Existing reports also focus largely on international NGOs (INGOs), with less attention to research capacity among NGOs based in developing countries. Supporting NGO involvement in research requires more understanding of current capacity gaps and their causes.

A complete definition of research capacity encompasses the entire research cycle, involving the “ability and resources of individuals, organisations and systems to undertake, communicate and use high quality research” [36]. In line with our focus, we concentrate on ability and resources to conduct research. In conceptualising capacity, we draw on analyses of both research and wider organisational capacity. Both sets of literature emphasise that capacity is multi-faceted, involving tangible elements such as technical skills or material resources, and intangible elements such as leadership and commitment [26,28,31,37–43]. The literature also highlights the need for capacity at individual, organisational and institutional or environmental levels, for example individual skills, effective organisational structures, and enabling funding streams [26,29,31,36,38,41,43,44]. Different elements of capacity and different levels of capacity are interdependent, with interactions between elements such as skills and leadership, and between individual, organisational and environmental levels [26,38]. A further insight is the influence of external relationships on capacity, with power differences between organisations affecting scope for action [28,40,42]. In line with this, discussions of “unleashing” research capacity highlight the significance of externally imposed constraints [26]. Finally, required capacity depends on the task, and different kinds of research need different resources [26,45].

Alongside these conceptual insights about the nature of capacity, the literature indicates specific resources and abilities needed to conduct research. Numerous elements are identified within existing analyses, for example infrastructure, library access, and systems for peer review [29,44]. Three resources consistently discussed are funding, staff time and research skills [25,26,29,31,41,46]. These three resources of funds, time and skills are also frequently noted in discussions of NGO research [2,8,9,18,22,32–35], and they emerged as priorities for the NGOs in our research. Consequently, while recognising that capacity to conduct research involves many more components, including wider organisational abilities such as strategic vision or networking [42], we focus on research funding, staff time and skills to reflect primary concerns in the literature and among our research participants.

NGOs play an important role in Malawi’s health system [47]. The NGO sector has grown since 1994, when Malawi moved to a multi-party system [48]. The number of NGOs is unknown, but one estimate suggests over 500 [49]. Early NGO work focused on service delivery, but with increasing political space, NGO involvement in advocacy increased [48]. Most Malawian NGOs rely on funding from foreign donors, often via INGOs. Following the global recession of the late 2000s and cuts to aid in donor countries, funding declined and became increasingly competitive [50]. The new funding environment also involved increasing donor emphasis on evidence to demonstrate results [50], contributing to a growing interest in conducting research among NGOs in Malawi.

In comparison with other low-income countries, Malawi’s health research sector is strong by some standards. An assessment of health research output based on number of publications ranked Malawi as ninth in the WHO African Region [51]. However, human resources, infrastructure and funding for research all remain limited [47,52–55]. There are efforts to develop research capacity, and programmes such as the internationally-funded Health Research Capacity Strengthening Initiative have contributed to a growth in the number of skilled researchers and stronger institutional structures [47,54,55]. NGOs have been part of some initiatives to strengthen research capacity [55], and they are recognised as a stakeholder within national research policies [53,56].

Methods

We used comparative case studies, involving in-depth examination of experiences in contrasting settings to enhance explanation and understanding of context [57,58]. We focus on three organisations: a Malawian NGO working on issues affecting women and young people (MALN); the Malawi country office of a large INGO working to strengthen health service delivery (INTA); and the Malawi country office of a medium-sized INGO working in several sectors including health (INTB). Key characteristics at the time of fieldwork are summarised in Table 1, using approximate figures to protect confidentiality.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Case NGO characteristics.

https://doi.org/10.1371/journal.pone.0198721.t001

Selection of NGOs was purposeful [59], based on identifying organisations with specific features that enabled relevant insight. Our focus was NGOs involved in service delivery and advocacy as well as research, not, for example, NGOs such as think tanks. We chose NGOs that provided contrasting organisational contexts, such as international and national structures and varied research experience. They were all NGOs that undertook research and saw conducting research as part of their organisational strategy; this stated commitment to research is not typical of all health NGOs. Case NGOs were initially approached by email, with further discussions about their participation through telephone and face-to-face meetings.

Data were collected during six months’ fieldwork in Malawi followed by email, Skype and telephone conversations. Fieldwork was undertaken by an experienced researcher who previously worked in NGOs (KG). This background contributed to rapport with participants, but familiarity with NGO language and settings also created a risk that aspects of NGO practice might be taken for granted and so unquestioned. This risk was managed through critical discussion with co-authors and other colleagues to provide alternative viewpoints and challenge assumptions. A research diary was also kept throughout fieldwork and analysis to reflect on relationships with participants and how these might affect the data.

In-depth interviews were conducted with 26 staff or former staff in the three NGOs, including staff leading on research, directors and those involved in service delivery and advocacy. Ten repeat interviews were conducted, split between seven staff with particularly extensive knowledge of each NGO’s research capacity. Most interviews lasted 1–1.5 hours and took place in private offices. Semi-structured topic guides were used, tailored to each participant’s role and adapted during fieldwork to pursue emerging themes. Interviews asked broadly about each NGO’s research experience and included specific questions related to capacity, such as strengths and challenges in conducting research, variation in capacity between district or country offices, examples of research that could or could not be taken forward, training or capacity building for research, and support from other organisations or donors. Further insights came from numerous informal conversations with these and other staff.

Focus groups were held with four to five staff in MALN and INTB, lasting around two hours. Participants overlapped with those interviewed, leading to similar findings, but group discussions brought some additional perspectives through interaction and cross-checking between participants.

Focus groups and interview participants were selected based on their involvement in and knowledge of the NGO’s research and to give a diversity of organisational positions. Participants were initially approached by email or in person, depending on logistics and whether they had already met the researcher. No one refused to participate, although work schedules meant some staff could not attend focus groups; views from these staff were sought through interviews instead.

Observation of NGO meetings and activities provided additional information on research processes and organisational contexts. The extent of observation varied between cases depending on logistics such as desk space in NGO offices, transport access, and timing; with MALN over five weeks were spent working from their office, whereas observation was limited to attending organisational research meetings in the INGOs. During some meetings in MALN, observation involved a high degree of participation, with KG actively involved in discussion, but in most cases meetings were attended primarily as an onlooker or guest.

Document review provided background information on each NGO. Relevant material was identified through internet searches and discussion with NGO staff.

To provide information on NGOs’ relationships with other actors and the wider context, and some indication of whether issues raised in case organisations were shared more widely, further interviews were conducted with donors (2), government (2), academics with experience of NGO collaboration (3) and other NGOs (4). These participants were selected based on suggestions from knowledgeable contacts in Malawi and other participants.

Methods are summarised in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of methods.

https://doi.org/10.1371/journal.pone.0198721.t002

All interviews and focus groups were conducted in English. 39 of the 46 interviews and both focus groups were audio-recorded and transcribed verbatim by KG. Where recording interviews was impractical and during observation, detailed notes were taken and expanded the same day.

Analysis was ongoing throughout fieldwork, using new data to inform future methods [59,60]. More focused analysis involved familiarisation with the data as a whole through re-reading transcripts and field notes from interviews, focus groups and observation, followed by thematic coding of both notes and transcripts by hand and in NVivo. Codes included both deductive and emerging themes, and comprised organisational categories (such as ‘funding and donors’ or ‘research skills’); substantive categories (such as ‘research seen as irrelevant’ or ‘focusing on own work areas’); and more theoretical categories about relationships (such as ‘project funding determining research’) [61]. Matrices and network diagrams were used to explore relationships between themes and variations between cases [61–63], for example to identify factors affecting funding and the effect of funding on other elements of capacity. Initial explanations were refined by looking for conflicting evidence and considering alternative interpretations [59,62,64].

All stages of analysis were undertaken by KG in discussion with co-authors. A summary of findings was shared with case NGO participants for their feedback. Rigour was also supported through triangulation (within interviews and focus groups, between interviews, focus groups, observation and document review, and between research participants), through prolonged engagement with each NGO to develop rapport, and through steps previously mentioned such as reflexivity, audio recording and seeking negative cases [59,64–66]. The study was approved by the University of Leeds ethics committee (reference number HSLTLM/11/004) and the National Commission for Science and Technology (RTT/2/20) and Centre for Social Research (CSR/11/11/05) in Malawi. Informed consent was based on discussions with NGO directors at organisational level and with individual participants, alongside provision of information sheets. Consent was verbal because NGO participants saw written procedures as unnecessary and inefficient. Written procedures also risk damaging rapport and hence data quality, and they were impractical given numerous informal conversations [67,68]. Where research involves ongoing interaction with participants, valid consent is often achieved through continual open communication [67].

Results

The case NGOs vary in their capacity to conduct the quantity and type of research they would like. We describe a range of factors that influence their funding, time and skills to conduct research, the three important elements of capacity highlighted in the introduction. Quotes are drawn from interviews, focus groups and informal discussions during participant observation.

Discussion

Our findings indicate variations in funds, time and skills to conduct research among NGOs, and internal and external factors that affect availability of these resources. As shown in the introduction, existing literature points to gaps in NGOs’ research capacity (e.g. [2,18,22]). Our findings help to explain both these capacity gaps and variations in research capacity among NGOs. Table 3 summarises the findings, indicating key factors that affect NGOs’ funding, skills and time to conduct research.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Factors affecting availability of funds, time and skills to conduct research in NGOs.

https://doi.org/10.1371/journal.pone.0198721.t003

Several constraints identified among the case NGOs are indicated in other literature. For example, concerns about time to conduct research alongside service delivery [18,20,22] and about including research budgets in proposals [22] are also reported for other NGOs. Literature from Malawi and elsewhere also discusses the broader processes found to affect research capacity in the case NGOs, including a lack of core funding, limited donor support for staff salaries, unpredictable and short-term grants tied to specific activities, and the move of skilled staff from national to international organisations [39,69–77].

The NGOs’ experiences also reflect more conceptual aspects of the nature of research capacity identified in other contexts, including interdependence between different elements of capacity and between individual, organisational and environmental levels [26,38], the effect of external relationships [40,42], the role of intangible resources [26,28,31,37–42], and fitting capacity to needs [26,45]. In relation to interdependence between elements of capacity, there are interactions between skills, time and funding to conduct research within the case NGOs. For example, securing research grants requires time and research experience, while flexible core funds or donor support enable access to time and skills by allowing NGOs to hire research staff and provide training. Conversely, insecure, small grants make it harder to allocate time for research by increasing service delivery and fundraising workloads and hindering planning. Links between different levels of capacity are also clear, for example the international funding environment affects organisational access to funding for research training, and consequent scope to strengthen individual research skills. These effects of funding highlight the influence of external relationships, particularly with donors. The role of intangible resources is seen in the significance of prioritisation for research capacity, something noted for NGOs [8] and other research organisations [29]. For example, prioritisation of service delivery over research can mean NGO staff do not include research in project proposals, make time for research in their schedules, or use existing research skills. However, increased prioritisation of research would not overcome all resource constraints. For example, when NGOs need to implement contractually obliged activities and cannot afford to hire extra staff, making time for research is harder than when research staff are available. Finally, required skills, time and funding vary between and within the case NGOs depending on research aims and designs. NGOs that conduct multi-year research trials need different types and levels of skills and funding to NGOs interested in short-term qualitative research.

NGOs provide a different organisational context to the universities and think tanks that form the focus of most literature on health research capacity. Comparing the case NGOs’ experiences with this literature suggests some commonality in the constraints faced by NGOs and other kinds of research organisation in developing countries. For example, limited ability to recruit and retain skilled research staff and lack of core funding are also discussed in relation to academic institutes [28,29,78]. The findings also point to differences. For example, NGOs may face challenges of balancing time for research with service delivery, rather than the balance with teaching responsibilities identified in universities [31]. Differences in scope and ambition of research agendas also mean different capacity requirements. Assessments of capacity in universities or other research-focused organisations highlight challenges around adequate numbers of research staff in different disciplines and with different levels of expertise [25,28,31,79]. In NGOs where research agendas are smaller, lower levels of staffing may be considered adequate. For example, INTB staff appreciated having one staff member whose work is partly focused on research.

Our research has limitations. We focused on NGOs that undertake research and that have a stated commitment to research. NGOs that are not conducting any research may face additional capacity constraints. We also focused on three key elements of capacity to conduct research, funding, skills and time. Complete research capacity includes many more components, including skills related to other aspects of the research cycle such as ability to identify research questions, and wider organisational capacities [26,28,31,44]. Finally, while NGO staff emphasised donor reluctance to fund research, the need to protect organisational confidentiality and study timeframes limited scope to verify their concerns through discussion with donors. The donors interviewed were known to be interested in research, and further investigation is needed to understand approaches to funding NGO research among a wider range of donors.

Conclusions

Our findings have implications for strategies that could support NGO capacity to conduct research. Above all, a first conclusion is the need to tailor capacity development to specific organisational contexts. Particular NGOs’ methods for developing research capacity are sometimes described as a model for other organisations (e.g. [80,81]). However, diverse research approaches, internal conditions and external relationships create differences in required and available capacity and in opportunities for capacity development. Some recommendations for developing NGOs’ research capacity may only be feasible in large organisations with flexible core funding. For example, advice to hire research staff and provide training [18] is more easily followed with adequate flexible funding or donor support. Similarly, recommendations to give staff dedicated time for research and to include research time and budgets in annual plans [18] are more achievable when NGOs can recruit additional staff and when they have secure funding that provides control over annual plans and budgets. Varied starting points among NGOs necessitate capacity assessments, something emphasised for research capacity development more widely [31,41,43].

Second, the importance of multiple components of capacity and links between funding, time and skills to conduct research mean that developing NGO research capacity requires more than skills training, again a message underlined in discussions of organisational and research capacity [28,39,41,82]. Research training courses have demonstrated an impact on production of research by NGO staff [83], but this impact seems likely to be diminished in NGOs without adequate staff time or funding to conduct research.

Third, the influence of prioritisation on availability of funds, time and skills and the influence of research design on requires resources both suggest there are options for NGOs wishing to conduct more research. In particular, NGOs could adapt research designs to suit available capacity: small-scale data collection or use of monitoring records may be feasible without additional resources. NGOs can also explore ways to strengthen capacity within existing resources, for example, there may be staff with research experience who can share skills with others. There may be opportunities for including research budgets within service delivery proposals, or for considering alternative donors if current funders do not support research. NGOs can also consider different models for conducting research, including partnership with other researchers. The potential value of collaboration with academics is often emphasised in discussions of NGOs’ research [2,3,22,35,84,85], and such collaboration is an important strategy for some case NGOs [24]. As INGB’s experience indicates, suitable academic partners can be hard to find, and collaboration faces challenges such as divergent organisational priorities. However, working with academics can help some NGOs that lack internal capacity to conduct research.

Fourth, the significance of external relationships and funding suggests recommendations for donors interested in supporting NGO research capacity. To support access to research funding, particularly for operational research, donors could provide options for including research budgets within service delivery grants, and clearly indicate willingness to fund research so that NGO staff recognise the opportunity and have confidence to apply. Donors could also consider how current funding approaches may constrain capacity. In particular, more predictable, longer-term grants that include support for staff costs might reduce time spent on securing funding and ease staff workloads, freeing up NGOs’ time for research. For some NGOs, a useful area for donor support may be covering salaries for skilled research staff who can undertake research, apply for funding, and develop research skills and prioritisation among other staff.

The feasibility and effectiveness of these strategies could usefully be explored through action research with interested NGOs and donors. Such an approach might strengthen research within participating organisations, and provide further understanding of constraints and enablers to NGO involvement in health research.

Acknowledgments

We are grateful to the NGO staff and other participants in Malawi who gave time to support this research through interviews and other assistance. We also appreciate the valuable input from the two peer reviewers.