We conducted a modified version of the Nominal Group Technique (NGT) during the two-day Conference of the Nigeria Implementation Science Alliance (NISA) held from August 17th to 18th, 2023, in Abuja, Nigeria, titled under the theme “Bridging the Implementation Science Gap: Strategies for Effectively Translating Evidence into Practice.” The NGT, a structured, face-to-face group brainstorming method, was selected for its efficacy in collecting and establishing consensus on diverse issues, challenges, and solutions among a heterogenous group of stakeholders (24, 25). Its suitability for time-limited, in-person engagement with diverse groups, and its ability to combine the strengths of both quantitative and qualitative approaches by allowing for collective prioritization through consensus and ranking and the inclusion of rich contexts, made it an ideal choice. This study utilized a modified NGT, incorporating hybrid voting platforms, purposive assignment of participants to tables, virtual documentation of barriers and facilitators using Microsoft Forms (as opposed to traditional paper) and a two-day phased process for consolidation and plenary voting (24, 25). The NGT process is particularly advantageous as it ensures equitable participation, thereby mitigating the dominance of more vocal individuals (17). This advantage was particularly important given the diversity of expertise among implementation science researchers and the need to surface and rank the most context-relevant barriers and strategies. Compared to the Delphi method, which requires multiple rounds of remote input over time, and Focus group discussions (FGDs) and interviews that limit the scope of perspectives, the NGT methodology was deemed to be the most suitable approach as it would result in the elicitation of practical, context-specific solutions from conference attendees, relying on their firsthand experiences, to identify and rate key challenges and actionable strategies in a time-limited group session.

Eligible participants were adult attendees of the 2023 NISA conference who identified as researchers, policymakers, implementing partners, civil society actors, or affiliated stakeholders in implementation science or public policy. Inclusion criteria included self-identification with one of these professional categories during the conference registration process and willingness to participate in the group session. There were no exclusion criteria aside from non-attendance or withdrawal of consent. Participants were purposively allocated to 32 tables using registration data collected prior to the conference. Allocation was stratified by stakeholder category (e.g., policymaker, researcher, implementer) as self-identified at registration, in order to ensure heterogeneity within each table and to maximize cross-sectoral perspectives. The assignment process aimed to ensure a balanced and representative mix of disciplines at each table, reduce the risk of stakeholder dominance and increase diversity of perspectives. Thus, by integrating a wide spectrum of voices (across government, academia, and practice) into each group, we aimed to minimize sample bias thus ensuring a more comprehensive consensus-building.

The NGT session was led by four trained facilitators and followed a hybrid model, combining in-person and virtual tools. Each group had a designated leader who was provided with a Samsung Galaxy Tablet® (10.1″) and standardized instructions. The session began with a brief presentation on the purpose and structure of the NGT. Participants were at their assigned tables (Table numbers 1–32), which had been pre-arranged to ensure stakeholder diversity as described above. The NGT unfolded in four phases across two days and comprised four distinct phases as illustrated in Figure 1. Phase 1 was dedicated to identifying barriers that hinder successful engagement between policymakers and implementation researchers in knowledge translation. This phase encompassed three stages. The first stage was idea generation, where individual participants listed their top three barriers to successful engagement between policymakers and implementation researchers for knowledge translation. In the group discussion stage, the barriers generated by individual participants were thoroughly discussed within each group. In the voting stage, group members subsequently identified and voted for the top three barriers that had emerged from their discussions.

Moving on to Phase 2 of the NGT, participants focused on pinpointing the three most promising strategies to address each of the top three barriers identified within their respective groups. As in Phase 1, this phase followed a structured approach, comprising idea generation, discussion, and voting. In Phase 3, the facilitators amalgamated the top three barriers identified by all 32 groups, de-duplicating before consolidating them into cohesive themes. Proposed strategies for each group's top three barriers were also merged into unique strategy themes following de-duplication. All three phases happened on Day 1. Phase 4 involved a plenary session in which participants first voted for the top three barriers identified from the themes generated in Phase 3. They then voted for the top three strategies to address each of these barriers. Phases 1 and 2 of the NGT together lasted approximately 1 h and 20 min. During Phase 4, voting on barriers lasted about 7 min, while voting on corresponding strategies took about 20 min. Both voting activities were conducted on Day 2 of the conference (see Supplementary Table). Minimal attrition (<3%) occurred during Phase 4 plenary voting, due to a few participants leaving the conference early and a brief period of technical/connectivity issues. Only complete voting records were analysed, consistent with NGT data handling standards, with no partial votes included.

Barriers and strategies were recorded via Microsoft Forms with QR codes for rapid consolidation.. During Phase 1 and 2, the voting procedure took a digital form, utilizing OpinionX© (https://www.opinionx.co) for collating group barriers and strategies while Phase 4 utilized Mentimeter© (https://www.mentimeter.com), an external online platform tailored for virtual interactive polling and ranking, to deliberate on strategies. OpinionX© was incorporated for barrier ranking due to its ability to handle a high number of options, and Mentimeter© for strategy ranking, which allowed for interactive visualisation and quick consensus (both being a deviation from traditional in-room card or flipchart voting). Facilitators generated QR codes linked to the polls that were then projected onto a screen for participants to access. Responses were only unveiled after all participants had completed their voting.

The analysis process adhered to the prescribed methodology for handling NGT data, as outlined in the guidelines proposed by McMillan and colleagues (24). To examine the barriers and potential strategies facilitating productive collaboration between policymakers and implementation researchers in diverse contexts, a thematic analysis approach was applied. This involved the aggregation and categorization of all valid responses, which were then grouped according to common themes. Descriptive statistics, including frequencies and percentages, were employed to provide a concise summary of the voting outcomes.

This study involved human subjects. However, the need for ethical approval for this study was waived by the College of Medicine Research and Ethics Committee (COMREC), University of Nigeria, as the research did not involve interventions or collection of sensitive personal data that required review, according to institutional guidelines. All participants were provided with clear information about the purpose, procedures, and voluntary nature of the study both before and during the NGT session. Informed consent was obtained verbally from all participants, and participation in the NGT activities implied consent to contribute data anonymously to the study. No identifying information was collected during idea generation, discussion, or voting phases. Data were stored in password-protected files and accessible only to the research team to ensure confidentiality and data protection. Given that this activity was conducted during a professional conference, and responses were aggregated at the group level, the risk to participants was minimal. Participants had the right to decline participation at any time during the process without consequence.

Following de-duplication, one hundred and eighty-one (181) unique responses were received from the groups during Phase 1 of the NGT, with each group contributing a minimum of between three to five barriers. These barriers were categorized under themes during the collation process. The themes identified include “knowledge/capacity gaps in stakeholder engagement”, “lack of an engagement framework between policymakers and implementation researchers”, “effects of the political environment and bureaucracy”, “competing interests”, “lack of resources”, “communication gaps”, “lack of collaboration”, “the lack of relevance of research findings to the local context”, “absence of dissemination opportunities”, etc. Eighteen (18) unique barrier themes were identified and voted during Phase 4 of the NGT process as seen in Figure 2. “Competing interests between policymakers and implementation researchers” emerged as the highest-ranked barrier, accounting for 28% of collated votes. This was followed by “knowledge/capacity gap in stakeholder engagement”, and “absence of an engagement framework between policymakers and implementation researchers”, in that order. Conversely, “lack of time and availability of stakeholders”, “complexity of the research”, and “leadership tussle in implementation” were ranked as the least relevant barriers.

Eighty-one (81) responses were received from the groups during Phase 1 of the NGT, for each of the top three listed barriers. Participants highlighted creating “a platform for researchers and policymakers to engage”, “stakeholder consultation meetings on research projects” and “aligning research with national priorities” as the top three strategies for successful engagement between policymakers and implementation researchers. Strategies to overcome the “knowledge/capacity gap in stakeholder engagement” from the NGT process include “training and education of stakeholders”, “participatory engagements”, and “multi-level and complete stakeholder mapping”. To overcome the “absence of an engagement framework between policymakers and implementation researchers”, participants listed “continuous engagement to harmonize the interests and priorities of researchers and policymakers”, “developing an engagement framework between policymakers and researchers”, and “creating or sustaining collaborative platforms between researchers and policymakers” as the top three most effective strategies (Figure 3). It should be noted that outputs presented in Figures 2, 3 reflect relative ranking rather than scale-based measurements.

This barrier mentioned by conference participants reflects a fragmented research–policy–practice interface, which, according to the EPIS framework, constitutes a failure of systems-level bridging mechanisms and organizational readiness and sustainment strategies (26). In the Nigerian context, which is similar to many low- and middle-income countries (LMICs), institutional fragility and limited absorptive capacity further complicate the seamless integration of research into policy. This challenge mirrors findings in the literature. For example, a study in the Benin Republic revealed a clear disconnect between research evidence and policy formulation in developing the user fee exemption policy for caesarean sections (CS), the latter being part of some electoral campaign promises made good by a prominent Beninois politician (27). Although public health evidence could have steered the policy toward reducing maternal mortality, its political goal was instead to end the detention of mothers and newborns unable to pay hospital fees by exempting them from a broad package of maternal health services., (the césarienne gratuite or free CS policy) (27). In the end, the policy was unable to achieve free CS for all and, unfortunately, made little or no impact in the reduction of maternal and neonatal morbidity and mortality.

This narrative above illustrates what the Knowledge-to-Action (KTA) framework identifies as a breakdown in adapting knowledge to the local context and a failure to assess barriers to its use (28–30). Within this model, stakeholder engagement is a dynamic and iterative process, requiring both capacity and mutual understanding, elements perceived by study participants to be lacking. The absence of such competencies on both sides (research literacy among policymakers and policy literacy among researchers) hampers the KTA cycle, effectively stalling knowledge translation into policy and practice. Bullock et al. (31) theorize that implementation science and policy implementation research have often evolved separately, creating theoretical gaps and limiting integration of policy perspectives in implementation frameworks. In other words, policy is often treated as context or a barrier rather than as an active process involving key actors, leading to insufficient integration of policymakers in implementation efforts. Crable et al. (2) further argue that most dissemination and implementation (D&I) frameworks lack explicit guidance for how and when to engage policymakers. In addition, limited awareness of D&I frameworks among policymakers often compounds the problem, further hindering rigorous evaluation and adaptation of health interventions (32). This gap is particularly problematic in LMICs like Nigeria, where fragile health systems demand not just innovative interventions but also timely and politically feasible implementation strategies. Consequently, scientific advances remain trapped within academic or pilot phases. Without intentional mechanisms to bridge this divide, engagement will remain fragmented and superficial, preventing evidence from influencing meaningful policy action (33).

The second major barrier, which was mentioned by conference participants to be competing interests, can arise because policymakers often prioritize political feasibility, resource constraints, and rapid decision-making, whereas researchers emphasize methodological rigor, evidence generation, and long-term outcomes, leading to potential misalignment in goals and timelines (2, 31). This divergence can be explained using the Stakeholder theory, which posits that different actors within a system hold distinct, and sometimes conflicting, stakes in decision-making processes (37, 38). Without deliberate mechanisms to reconcile these differences, collaboration becomes superficial or counterproductive. This barrier is also known to be exacerbated by asymmetry in capacity-building opportunities. For instance, a mapping study found most training and skills initiatives were skewed towards researchers (39). This challenge mirrors challenges of divergent interests and timelines previously identified in a study by Erismann et al. (40). Similarly, poor collaboration between researchers and practitioners has been reported at a similar gathering as a barrier to evidence-based policymaking in Nigeria (41).

Persistent gaps in policy literacy among researchers and research literacy among policymakers emerged as a major barrier mentioned by conference participants. The SPIRIT Action Framework (Supporting Policy In health with Research: an Intervention Trial) (51) is a conceptual model, first introduced in 2015, that helps guide and test strategies to increase the use of research in policymaking. At its core, the framework outlines a logical sequence of the following: a Catalyst which is a factor that triggers the demand or need for policy research (e.g., a pressing policy question or emerging public health issue); Capacity, which is the ability of an organization or individual to engage with research; Actions: which are the research-related behaviours that follow (e.g., accessing, appraising, generating, or interacting around research) and Outcomes which refers to the forms of research use in policymaking (e.g., agenda-setting) and longer term impacts. Within the SPIRIT Action Framework, this barrier is identified fall under the “capacity” domain, which are skills, knowledge, systems, and relationships required for effective research–policy engagement. Such capacity gaps limit progression from catalyst to action, stalling translation of evidence into policy. Similar emphasis on capacity appears in other Knowledge Translation frameworks, such as the Knowledge-to-Action (KTA) cycle (52) and the Capability, Opportunity and Motivation- Behaviour (COM-B) model (53), which frame capability as a prerequisite for behaviour change in evidence use. Without targeted strategies to strengthen mutual understanding and stakeholder engagement skills, interactions can become inefficient and burdensome, with researchers spending disproportionate time lobbying policymakers rather than co-producing policy-relevant evidence.