a) Identification of digital repositories and other infrastructures for health policy

The study employed a comprehensive search strategy to identify digital repositories and other relevant infrastructure. This approach included generalised search engines such as Google, specialised search engines and targeted databases such as PubMed, Scopus, and JSTOR to ensure broader and more nuanced coverage of relevant resources. Additionally, artificial intelligence tools, including ChatGPT, ⁵ were utilised to uncover resources not readily accessible through conventional search methods. We further identified digital repositories and infrastructure by visiting directly the websites of relevant global and regional health organisations, academic repositories from leading public health universities, commercial databases, proprietary platforms, and resources from international health non-governmental organisations. This multifaceted approach ensured a thorough and diverse identification of resources necessary to support our landscape analysis and evaluation objectives.

Electronic searches were conducted using keywords and Medical Subject Headings (MeSH) across the abovementioned platforms. The primary search terms included “health policies,” “health laws,” “health strategies,” “health guidelines,” “health action plans,” “digital repositories,” “digital infrastructure,” and “global health.” To ensure comprehensiveness, synonyms for each term were identified and integrated into a robust search strategy, detailed in Search terms .

In addition to database searches, reference lists and associated links from identified repositories and infrastructures were screened to discover other potentially relevant sources. This iterative process ensured the inclusion of a diverse and comprehensive selection of repositories and infrastructures pertinent to health policy documents.

b) Inclusion and exclusion criteria

We included digital repositories that systematically collate health policy documents, as defined in Table 1. The review encompassed only platforms where the content was either available in English or offered an English translation as one of their languages.

Platforms functioning primarily as search engines, analytical databases, hubs for academic literature, or think pieces—rather than dedicated repositories for health policy documents—were excluded.

Furthermore, repositories were dropped from the review if they lacked an accessible link, had broken or missing links, or required access through login or subscription to access any information, rendering them inaccessible to the public or our research team. However, repositories with partial or limited access requiring a subscription to access more information or information after a set period were reviewed. This carefully defined inclusion and exclusion process ensured a focused and targeted analysis of publicly accessible digital infrastructures systematically collating health policy documentation. It is important to note that a few repositories were identified after the data collection window had closed and were thus not included for that reason (listed in List of excluded repositories ).

a) Screening of identified repositories

We employed a two-step screening process to ensure rigorous verification of the repositories and digital infrastructure to be evaluated in detail ( PRISMA flow chart ). In the first step, two reviewers (TM and BM) independently screened all repositories identified through the electronic database searches. During this stage, the reviewers conducted a first assessment of all inclusion/exclusion criteria (see below) while eliminating duplicate records. This initial screening ensured a more focused and relevant set of repositories for further review.

In the second step, two reviewers conducted an additional review of the above criteria to finalise the list of repositories to be taken forward for detailed evaluation. Any disagreements between the two reviewers were carefully examined and discussed. When consensus could not be reached, a third-party expert (MGC, MK, TN, or CSM) was consulted to resolve the discrepancies. This collaborative approach ensured a thorough process for including relevant repositories in the evaluation and a consistent and more objective assessment across the repositories.

b) Data extraction and evaluation of the scope, content, and functionality of the repositories

We developed a set of assessment and review criteria based on which data was extracted from the repositories for evaluation. These criteria were piloted on three repositories before the actual data extraction took place to enhance the consistency of the results. The set of evaluation criteria guided specific data extraction. We briefly employed a structured approach by extracting the following data, where possible. 1.

Repository name – The official legal name of the repository.

The information extracted from each of the included repositories and digital infrastructures was collated in a Microsoft Excel (Microsoft Corporation, Redmond, Washington, United States) database (for a condensed extract, see Summary of Digital Repositories of Health Policy included in our evaluation ). A list of repositories considered for evaluation but not included in the final list is also provided ( List of excluded repositories ).

Eligible repositories were evaluated based on four key dimensions: the health policy focus area, data content, adherence to FAIR (Findable, Accessible, Interoperable, and Reusable) principles, and UI/UX features supporting content discovery. Where available, information on sustainability and funding models of the health policy document repositories was also assessed. 1.

Health policy focus area

We examined the scope and comprehensiveness of health policy documentation curated within these repositories. We analysed the range and specificity of health or disease areas covered in each repository and evaluated the format of health policy documents, focusing on their relevance to specific policy domains.

Adherence to FAIR principles varied widely among the repositories. Many appeared to have been established without explicitly considering these principles, leading to significant gaps in the data’s findability, accessibility, interoperability, and reusability (see Table 2). This inconsistency poses challenges for the broader health community, particularly for policymakers, researchers or others interested in health policy who may employ these repositories, generally set up to make health policy data widely available, as a basis for comparative analyses, to promote accountability and ultimately to inform national responses in each health area.

The Findability guideline in the FAIR principles ensures that data (here, health policy documents) and/or metadata stored as digital assets can be easily located by both humans and machines. Metadata, also known as (meta) data, refers to data about data—in our case, data that provides information about health policy documents. This requires unique identifiers and well-structured, searchable metadata. Our assessment examined four indicators of findability (F1–F4, Table 2). Results show a range of performance: while only three repositories utilised unique identifiers, 14 repositories were indexed in searchable resources, significantly enhancing discoverability.

The Accessibility guideline emphasises using standardised protocols for data retrieval and clear access conditions to guarantee data availability. This includes ensuring that metadata remains accessible even when the associated health policies are unavailable. Our evaluation found that four ⁴ repositories provided accessible metadata, while nine ⁹ implemented open, free, and universally applicable protocols, reflecting strong adherence to accessibility standards ( Table 2).

The Interoperability guideline ensures that data can integrate seamlessly with other datasets and systems using standardised vocabularies and frameworks. We evaluated three indicators of interoperability (I1–I3, Table 2). The findings revealed that 13 out of 15 platforms used formal, accessible, and widely applicable metadata languages. Furthermore, 11 platforms adhered to FAIR vocabulary standards and included qualified references to other metadata, demonstrating high compatibility and integration ( Table 2).

The Reusability guideline evaluates whether data is provided with clear usage licenses and comprehensive documentation, enabling users to effectively understand, reuse, and build upon it. As shown in Summary of Digital Repositories of Health Policy included in our evaluation , thirteen ¹³ repositories offered richly described metadata with relevant attributes, and fourteen ¹⁴ included metadata with detailed provenance information. However, only six repositories provided a clear and accessible data usage license, and ten ¹⁰ adhered to domain-relevant community standards. These findings underscore the need for improvements in licensing clarity to enhance data reusability. 3.

User interface (UI) and User experience (UX)

To evaluate the UI/UX features of the health policy repositories, we focused on the availability and functionality of navigation tools, such as search bars and filter options, which help users locate specific information and navigate large databases/datasets. Additionally, we assessed the presence of analytical and interactive features, including data visualisation tools such as maps, tables, and graphs (all of these being either interactive or static), which enhance the usability and interpretability of the data provided by these repositories. These elements collectively determine how accessible, functional, and user-friendly a repository is for its intended audience.

Almost all 15 evaluated repositories include basic search and filter functions. These tools generally fall into two categories: search boxes and filtering options. The search box feature allows users to perform keyword-based searches within the repository. This function enables users to input specific terms or phrases—such as the name of a country, a health condition, or a policy title—to quickly retrieve relevant results. This free-text search capability is particularly useful for targeted queries where users know what they are looking for. On the other hand, the filter functions allow users to narrow down data based on predefined criteria.

In the 15 repositories we evaluated, common filtering options allow users to isolate health policy documents specific to a particular country or region, making it easier to explore country-specific data. Similarly, narrowing down by health focus area is common, where data (health policy documents in this case) are subsequently organised by thematic disease categories or broader public health topics. These filter options help users refine their search according to specific health priorities or geographical interests. While the implementation of these features was generally consistent across repositories, their effectiveness often depends on the quality of data and tagging (metadata) within the repository, which can influence the precision and relevance of search results.

The availability of functionality and visualisation features varies across repositories. Figure 1 highlights the functionality (such as search and filtering) and visualisation (static or interactive maps, graphs, and tables) features based on the focus areas of the repositories. Repositories focused on HIV/AIDS and broader health scopes offer a wide range of functionality and visualisation options. In contrast, repositories dedicated to public health and MNCH provide fewer features, with MNCH-focused repositories having the least. Notably, advanced analytical features were rare across all repositories, and none included AI-enabled tools.

While our assessment of UI/UX may be subject to observer bias, it is clear that a certain level of computational competence is required to navigate and explore these repositories, let alone extract data. In many cases, expert knowledge in the field may be necessary to effectively interact with a given repository to access, use and interpret the data contained. As a result, many repositories may unintentionally or intentionally make it difficult for non-experts—such as citizens, patients, or even non-health researchers—to maximise the use of the information. Indeed, given the often relatively basic search functions to interrogate vast amounts of data, it is difficult, even for experts (policymakers, health professionals, etc) to succeed in finding what they might be looking for. Some repositories (7/15), however, do provide guides on how to access and use the data (e.g., via how-to pages or videos).

More than half of the 15 repositories we evaluated dedicate their content to specific health focus areas, including HIV/AIDS, NCDs, or MNCH, while the rest cover a wide range of public health topics more generally. This pattern appears to be influenced by several factors. Many of the included repositories were established by NGOs and United Nations (UN) agencies, suggesting that the focus areas are shaped by the interests and mandates of the organisations setting them up, the specific needs they aim to address, and the funding available to support their work. For example, HIV/AIDS has historically benefited from significant funding, which likely contributed to the establishment of specialised repositories. Additionally, the methodology and processes used to collate data play a critical role. For instance, the WHO bases its repositories on information provided by member states, which, in turn, depends on the type of questionnaires and data requests sent by the organisation. If certain topics are not prioritised in these requests, there is little chance of developing repositories in those areas. This combination of organisational focus, resource availability, and data collection processes shapes the current landscape of health policy repositories.

Several critical health issues remain underrepresented and topics such as neglected tropical diseases, vaccine-preventable diseases, mental health, and cross-cutting themes like Universal Health Coverage (UHC) or health equity are notably absent. Emerging health challenges, such as pandemics and outbreaks of diseases like SARS, Marburg, and Ebola viruses, are particularly underrepresented. Only one repository (the COVID-19 Law Lab) was dedicated to collating health policy documents on rapidly evolving issues such as (re) emerging infectious diseases that pose a risk of developing into an epidemic or pandemic, including Ebola, Mpox, or other potential zoonotic outbreaks. It is important to note that the review did not evaluate this repository, since it was identified after our data collection period had closed. This gap in available policy resources for emerging and re-emerging health threats underscores the urgent need for dedicated repositories that can address rapidly evolving health crises and provide timely, actionable policy resources.

Arguably, repositories that are disease-agnostic and inclusive of all health areas should generally be more useful to a wider user base globally - provided that the content navigation functions are excellent. As outlined above, search functions tend to be too basic to successfully mine large data content. Repositories with a focus on a given health area may thus be more amenable to being accessed if users are interested in a specific health area or disease, risk factor or cross-cutting health area (e.g., health financing).

While many of the health policy repositories evaluated claim to offer comprehensive coverage of global health policy documents, a closer inspection reveals significant regional disparities, particularly in the representation of low- and middle-income countries (LMICs). These disparities stem from several interrelated factors, including limited institutional capacity for documentation and digitization, inadequate funding for repository maintenance, and a lack of technical expertise in these regions. Further, language plays a critical role and a limitation in our review. The mapping exercise focused on repositories where the content was either available in English or offered an English translation as one of their languages, limiting the breadth of data within repositories and the number of repositories we included in our analysis. As a result, health policies from LMICs are often either underrepresented or absent from global platforms, perpetuating a cycle of marginalisation where these regions cannot share their successful innovations and policies on a global stage.

The absence of LMIC health policy documentation and digital repositories presents critical challenges to achieving equity in global health governance. Without access to comprehensive repositories that include LMIC policy documents, researchers, policymakers, and practitioners risk making decisions that overlook region-specific health challenges, socioeconomic conditions, and cultural contexts. Policy documents addressing key issues, including infectious disease outbreaks, climate change impacts, or health system strengthening in resource-constrained settings, often provide valuable lessons for other countries facing similar challenges. However, the absence of such policies in accessible repositories prevents these lessons from being adapted or scaled in other contexts and limits opportunities for international learning and collaboration. Bridging this digital divide and gap in accessibility of health policy data between HIC and LMICs requires targeted efforts, such as fostering new and existing partnerships between South-South and North-South institutions and initiatives, investing in capacity-strengthening activities, and ensuring sustainable funding for the inclusion and maintenance of LMIC data in global health repositories. Addressing these disparities is essential for creating truly global and equitable health policy platforms that foster inclusivity, knowledge sharing, and collaboration across diverse contexts.

Adherence to the FAIR principles—Findability, Accessibility, Interoperability, and Reusability—is crucial for enhancing the usability and impact of data within health policy repositories. These principles emphasize “machine-actionability,” meaning that computational systems should be able to find, access, interoperate, and reuse data with minimal human intervention. This is particularly important as the volume, complexity, and speed of data creation increase, requiring more computational support. ¹⁴ Adherence to the FAIR principles varied widely across the repositories analysed, with many failing to explicitly incorporate these guidelines into their design and functionality. While some repositories excelled in findability and accessibility, others faced significant challenges with interoperability and reusability. ¹⁵ For example, few repositories employed unique identifiers for health policy documents or provided clear data usage licenses, which are essential for ensuring data can be easily located, accessed, integrated, and reused effectively.

In addition to these challenges, inadequate metadata standards ¹⁴ and poor data interoperability further undermine the effectiveness of many repositories. These shortcomings reduce their ability to function as comprehensive, reliable sources of health policy information and limit their potential to support evidence-based decision-making, knowledge sharing, and collaboration. ¹⁶ Restricted access, including requirements for subscriptions, organisational affiliations, or specific credentials, also creates barriers, particularly for stakeholders in low-resource settings. This exclusivity undermines the inclusivity of these platforms and leaves critical users—such as community leaders, advocates, and smaller health organisations—unable to benefit from the resources available. Notably, we excluded seven repositories from our evaluation due to such restrictions, though we included repositories with limited or partial access.

The absence of primary data source linkage is a substantial limitation in numerous repositories we evaluated, a void that is closely associated with deficiencies in the FAIR principles. This problem considerably impedes users’ capacity to verify the content of policies, trace their origins, or access supporting evidence. Researchers, policymakers, and other stakeholders who depend on a comprehensive comprehension of the policy landscape to inform decision-making are impeded by the absence of such linkages. It is difficult to contextualise or compare policies or evaluate their alignment with evidence-based practices in the absence of direct connections to underlying data sources, such as national health databases, program reports, or epidemiological studies. For instance, users who are endeavouring to analyse a national health strategy or assess its efficacy may encounter challenges in identifying the data that influenced its development. This can result in the inability to validate assumptions or make meaningful comparisons between regions or time periods. Similarly, identifying existing health policy documents as exemplars for potential adaptation to other settings is hampered by the inaccessibility of much of these data nationally and globally. Furthermore, repositories’ effectiveness as instruments for promoting transparency and accountability are restricted by their absence of linked data sources. It becomes more difficult to evaluate the development, implementation, and reliability of policies in the absence of accessible primary source documents. This may erode the trust of stakeholders in the repository as a dependable source of information. The absence of linked data sources also diminishes interoperability with other digital health platforms from a technological standpoint. Richer analyses, such as modelling the impact of policy decisions or identifying voids in resource allocation, could be facilitated by the integration of datasets from health systems, demographic surveys, or financial resources. Repositories that lack these linkages are unable to accommodate such sophisticated use cases, which restricts their capacity to inform the comprehensive strengthening of health systems. As such, addressing these challenges through more robust adherence to the FAIR principles will be essential for optimizing the role of digital repositories in advancing health systems and policy development, ensuring they can be used as comprehensive and reliable resources for all stakeholders.

The user interface and experience (UI/UX) of health policy document repositories play a crucial role in determining how effectively users can access, navigate, and utilize the information. While many repositories offer basic search and filtering options, the functionality and quality of these tools are often constrained by the underlying data and metadata. This may be due to a lack of funding or technical expertise to develop advanced UI/UX features, such as data visualization tools or AI-enabled analytics, which are rare in most of repositories we assessed. These advanced features could significantly enhance user engagement and analytical capabilities, but they often require specialized knowledge to interpret, potentially limiting accessibility for non-experts, such as community advocates or policymakers without technical backgrounds. As a result, these platforms may fall short in supporting inclusive, evidence-based decision-making.

Furthermore, outdated technological infrastructure and limited capacity for adaptation can significantly hinder the effectiveness of digital health repositories. Many platforms lack regular updates to incorporate modern functionalities, such as AI-enabled analytics or mobile-responsive designs, increasing the risk of obsolescence. The absence of these functionalities complicates the process of analysing and understanding policy information, limiting their utility. Although AI holds great potential to enhance these repositories, its application is not without challenges. AI tools may perpetuate biases present in their underpinning algorithms or training data, leading to skewed insights that can misguide the interpretation of outputs and if applicable decision-making. Additionally, the high costs and technical expertise required to implement and maintain AI-driven solutions create significant barriers, particularly in resource-limited settings. Unless these technological and practical challenges are addressed, repositories may struggle to meet the evolving needs of diverse users.

Assessing UI/UX is inherently subjective, influenced by users’ familiarity with digital platforms, professional expertise, and personal preferences. Therefore, perceptions of ease of use, navigation efficiency, and aesthetic appeal can vary widely. Our evaluation acknowledges this limitation and suggests that a more comprehensive assessment would involve engaging a diverse group of end-users in testing and providing feedback. This approach would offer a broader perspective on how repositories meet user needs and identify areas for improvement, ensuring these platforms are accessible and functional for all stakeholders.

Lastly, sustainability is another pressing concern for many repositories, especially those dependent on short-term or project-based funding. Funder-driven platforms often face uncertainty once the funding cycle ends, leading to stagnation, reduced updates, or even closure. This reliance on external support undermines the long-term viability of these platforms, compromising their capacity to remain relevant and effective in providing up-to-date policy information and responding to evolving health policy needs.