Inclusion, Diversity, Equity, and Accessibility (IDEA) are increasingly recognised as foundational principles for advancing health research and addressing structural inequities. A proliferation of resources and best practices reflects a growing movement to integrate IDEA across research environments [1–4]. IDEA-related strategic planning can help organisations interpret best practices and IDEA concepts in the context of a particular field or domain and articulate tailored approaches the organisation will take to promote equitable, accessible, and inclusive research. However, there is limited published detail describing how organisations develop IDEA strategies and translate broad principles into contextually relevant priorities and actions [5–7]. This lack of practical guidance limits opportunities for other organisations to learn from and adapt established approaches, particularly in specialised research settings such as population data science.

As the science of data about people, population data science has long emphasised the responsible use of data for public good and the ethical challenges associated with handling sensitive data [8]. At the same time, there is growing recognition that data, systems, and research practices underpinning this field can reproduce systemic biases and inequities and contribute to harm among systematically marginalised populations [9–12]. Addressing these challenges requires more than technical improvements. It also necessitates examining how power operates within research structures and processes, including who defines research priorities, how expertise is valued, and the extent to which communities are involved in governance and decision making [11, 13]. This imperative is especially salient in multi-jurisdictional research environments, such as distributed research networks, where researchers and research support organisations must navigate complex administrative processes, diverse data systems, varying data governance requirements, and differing expectations regarding public and community involvement [9, 10, 14–16].

Health Data Research Network Canada (HDRN Canada) is a not-for-profit corporation consisting of a pan-Canadian distributed network of organisations collectively focused on supporting use of population-level data to advance health and health equity. Established in 2019 with 18 member organisations, HDRN Canada is a distributed network where most member organisations hold linkable population health and health-related data or have mandates relating to the access or use of those data. HDRN Canada is organised into several cross-cutting Teams that support the network’s strategic priorities, including initiatives related to data access, public engagement, Indigenous data governance, training and capacity building, and equity-oriented data practices. Network members share knowledge and identify opportunities to collaborate and support researchers conducting multi-regional research, share and build advanced analytics, expand types and sources of data linkage, create infrastructures to improve access to and collection of data, and establish partnerships with relevant parties, Indigenous communities, and other knowledge users [17]. In this context, advancing Inclusion, Diversity, Equity, and Accessibility (IDEA) is closely connected to, while distinct from, commitments to the principles of OCAP® and broader Indigenous Data Sovereignty, which establish specific rights and responsibilities related to the governance and use of Indigenous data.

In its 2022–2026 Strategic Plan, HDRN Canada formally included a strategic objective related to IDEA and committed to developing a comprehensive IDEA strategy and action plan that would embed IDEA across the organisation’s five strategic goals [18]. Although many publications describe why IDEA is important and identify promising practices, few provide detailed accounts of how organisations can develop tailored strategies in complex research environments. To address this gap, this paper describes the structured, participatory process HDRN Canada followed to co-create its IDEA Strategy. Our approach was informed by principles of co-creation and participatory strategic planning, which emphasize engaging diverse stakeholders throughout the planning process to build shared ownership and ensure that resulting priorities reflect the needs and perspectives of those involved [19–21]. By focusing on strategy development itself rather than implementation, this paper offers a practical roadmap to other organisations seeking to translate IDEA principles into contextually relevant priorities and actions within population data science.

Population-based data research using administrative data has significant potential to improve population health outcomes. However, the systems and practices underpinning this field can also reproduce inequities and contribute to harm [22, 23]. Some of these harms are related to the ways different populations are represented and described in research. For example, administrative data sets used in population data research often rely on binary or inconsistently defined sex and gender variables, limiting researchers’ ability to accurately identify and describe transgender, non-binary, and gender diverse populations [24]. Inadequate gender measures also increase the risk of research using outdated or stigmatizing terminology when describing this community. Similar challenges exist when data are linked across systems that collect demographic information in different ways or when important variables such as race and ethnicity are unavailable or inconsistently captured [25, 26].

Other harms arise through the invisibility of populations within data systems and research. The absence of sociodemographic indicators related to social exclusion, such as housing status, substance use disorders, engagement in legal sex work, and incarceration experiences, makes it difficult to identify and address the health needs of groups experiencing marginalization [27]. Individuals may also be missing from administrative data altogether because of barriers to service access, stigma, or concerns about safety [28]. As a result, the benefits of population health research may not be equitably distributed, and groups that are underrepresented in data may be excluded from decisions regarding health and social service planning and resource allocation [29–31].

There is also increasing reliance on artificial intelligence and machine learning methods in population health research. While these approaches may improve efficiency and analytic capacity, they are susceptible to biases arising from missing data, historical inequities, and assumptions embedded in model development [32]. Studies have identified algorithmic biases, including the omission of social context and reliance on potentially biased clinical decisions, that may reinforce rather than reduce inequities [33]. Similar challenges arise when researchers develop computational phenotypes to identify specific populations in large datasets. These methods require defining inclusion criteria that are shaped by both human assumptions and data limitations. For example, Rich et al. developed a computational phenotype to identify transgender individuals using diagnostic codes and exogenous sex hormone prescriptions [34]. Although innovative, this approach may inadvertently reinforce the assumption that all transgender people undergo some form of medical transition.

Data used in population health research are typically collected by governments, healthcare organisations, and other institutions for administrative rather than research purposes [35]. Because researchers are often removed from the original data collection context, they may overlook important historical, social, and structural factors that shape how data are generated and interpreted [13, 36, 37]. This can lead to analyses that decontextualize communities and perpetuate harmful narratives. For example, interpreting the overrepresentation of Indigenous peoples in justice or child welfare systems without acknowledging the impacts of colonialism, racism, and intergenerational trauma can reinforce deficit-based stereotypes [38, 39].

Together, the absence of meaningful sociodemographic data, the invisibility of marginalised populations, and research practices that overlook structural context can contribute to epistemic harm by privileging dominant ways of knowing while dismissing community knowledge and lived experience [13, 40]. These practices shape assumptions about who is visible, who counts, and whose experiences are considered legitimate within research, ultimately limiting the ability of population health research to advance health equity [41].

There is growing recognition of these harms and ongoing efforts to address them. Recommendations include improving sociodemographic data collection, engaging communities affected by research, diversifying evaluation methods, and critically examining how systems of oppression and intersectionality shape both data and research practices [16, 42–44]. HDRN Canada’s IDEA Strategy responds to this imperative by providing a pathway to embed IDEA considerations within research practices and across organisational structures.

HDRN Canada created its IDEA Strategy through a structured five-phase, organisation-wide process designed to support shared leadership, collaborative planning, and inclusive engagement (Figure 1). The process began with preparatory work to assess organisational readiness and identify priorities, followed by activities to establish governance structures, create a shared understanding of the strategy’s purpose and approach, engage stakeholders in strategic thinking and feedback, and iteratively draft and refine the strategy. Although implementation activities are underway, the focus of this paper is on the process used to co-create the strategy. The implementation component shown in Figure 1 refers to the strategic directions described in the following section rather than an additional phase of work.

As part of the co-creation process, participants identified four interrelated strategic priorities that together provide a framework for embedding IDEA across HDRN Canada’s governance, operations, and research infrastructure. These priorities reflect areas where participants believed coordinated action was needed to strengthen equitable data practices, build organisational capacity, and influence the broader population health research ecosystem.

The first lesson is to ensure processes exist for acknowledging differences across the group, with the aim of preventing and sorting out conflicts. Co-creation and collaborative strategy development requires fostering psychological safety, naming and identifying power dynamics, and agreeing on shared rules of engagement [19, 20]. Shortly after embarking on this project, we realised that while all Project Team members value IDEA, collective buy-in was not guaranteed. Individual team members’ experiences with IDEA and the values and priorities they held influenced their decision to participate and desire to create something meaningful, especially in a context where IDEA-related outputs do not always result in expected change [55]. At times, this resulted in a difference of opinion on how to proceed with strategy development. Establishing the project charter at the outset and clearly defining the consensus-based approach to decision making was critical for naming power as an essential factor, creating space for people to share their opinions, values, beliefs, and providing a reference point for team commitment and accountability.

Our second lesson is not to underestimate the importance of facilitation. Co-creation requires significant skills and expertise to manage the complexity and conflict that can emerge in group processes [56]. Recognising that HDRN Canada members had the technical expertise in population health research and conceptual expertise in IDEA, we ultimately decided hiring a skilled facilitator who could navigate the challenges of bringing people together across an extensive, geographically dispersed network would yield a better return on investment. Noting that the quality of facilitation has a profound impact on an initiative’s success [57], it was important that a key selection criterion would be to have a clear process that would address concerns about how much time and resources would be needed to complete the strategy. The facilitation kept the focus on possibilities rather than barriers. This was not just for the pragmatic reason of ensuring we could complete the strategy activities, but also recognised that the complexity of IDEA itself could end up being a significant barrier to adopting the IDEA strategy [58]. Relying on an expert to help balance power among participants and enable different voices to influence discussions and decisions is necessary for fostering inclusion and shared ownership needed for co-creation [20].

Our final and perhaps most fundamental lesson was that visible and sustained leadership commitment creates the conditions necessary for meaningful organisational change. Despite having identified a clear imperative for HDRN Canada to focus efforts on embedding IDEA within and across the network, efforts to develop and implement the IDEA strategy would not be possible without support from the Executive Team and Board of Directors. Implementation and organisational change research consistently reports that visible and sustained executive sponsorship is pivotal to success in complex, equity-oriented initiatives [59–61]. By having executive and director-level participation in strategy sessions, roundtable discussions, and town hall presentations, we were not only able to signal that IDEA was an organisational priority but also secure the necessary resources for strategy development and implementation. Leadership participation also modelled an openness to the change and transformation needed to achieve the IDEA Strategy’s objectives and goals. Ultimately, executive sponsorship, engagement, and visibility throughout the strategy development process provided the authority, resources, trust, and shared ownership necessary to translate the IDEA strategy from a concept into something meaningful and actionable for HDRN Canada.