Introduction

The adoption of electronic medical records (EMRs) across healthcare systems is rapidly accelerating, particularly in Asia and Europe1. EMRs play a crucial role in capturing and storing a wide range of health data, encompassing medical history, clinical information, and personal details2,3. This wealth of health data has the potential to improve patient care and generate value within healthcare organisations4. The value of EMRs within healthcare organisations is created by enabling clinicians to access patient records from within the same system at any time, streamlining care and facilitating innovation5. Further work is needed to ensure that EMR systems prioritise patient-centeredness, delivering equitable benefits at a population health level while enabling seamless data sharing across multiple agencies. Moreover, enhancing the utilisation of this data in technologies like Artificial Intelligence and Machine Learning is essential. Despite the considerable advantages presented by the electronic collection and sharing of patient information between service providers and clinicians, there are still challenges, particularly regarding privacy, security and governance6. These challenges are further exacerbated for many Indigenous people7, for whom the willingness to embrace new technology may be tainted by past experiences of unethical data collection and management, including through research, stemming from inherent racism biases and failure to recognise and respect the rights of Indigenous peoples8.

The Declaration on the Rights of Indigenous Peoples (UNDRIP) was adopted by the United Nations in 2007 to establish universal minimum standards for the rights of Indigenous Peoples9. The UNDRIP Article 31 specifically includes standards for Indigenous Peoples to exercise control over intellectual property pertaining to their communities, lands, and resources10. In addition, Article 18 addresses the data rights of Indigenous Peoples, emphasising their inclusion in decision-making processes that impact their rights in alignment with their own established procedures8. These standards offer comprehensive approaches to managing Indigenous Peoples’ data beyond mainstream notions of research processes, knowledge generation and intellectual property11.

Indigenous Data Governance (IDG) and Indigenous Data Sovereignty (IDS) are relatively new methodologies increasingly advocated for in Indigenous communities to be able to govern the collection, analysis and interpretation of data that relates to their sovereign rights. These principles have been developed largely from standards contained within the UNDRIP, and generally, they reaffirm the rights of Indigenous Peoples to control the collection, access, analysis, interpretation, management, dissemination, and reuse of data relating to their communities12. The implementation of Indigenous data sovereignty revolves around two fundamental principles: (i) the sovereignty of Indigenous People concerning data pertaining to them, regardless of its location or custodian, and (ii) the entitlement to access the data necessary for Indigenous Peoples’ nation-building efforts13. While standards exist to advocate for Indigenous data sovereignty, the practical application of these standards in research activities involving data from Indigenous communities remains unclear.

This prompts the research question: What are the current practices used in research for governing Indigenous Peoples’ routinely collected health data? The primary objective of this study is to systematically review the data governance approaches employed when using routinely collected health data for Indigenous Peoples for research purposes. The secondary objective was to understand the advantages and challenges of using this data for research, which is particularly relevant for Indigenous Peoples given the burden of research on Indigenous Peoples, who are some of the most researched groups in the world14.

Methods

This scoping review was guided by Arksey and O’Malley’s framework for scoping studies15. In addition, the study selection and presentation followed the PRISMA extension for scoping reviews (PRISMA-ScR) guideline16. The PRISMA-ScR checklist is available in Supplementary Table 1. The scoping review methodology was selected for this study because, unlike systematic reviews, it is particularly effective in synthesising research and mapping literature in areas that were either not extensively reviewed or are complex and diverse in nature17.

Search strategy

We searched five electronic databases (PubMed, EMBASE, CINAHL, Web of Science, ATSIHealth), including one database which focuses on Aboriginal and Torres Strait Islander health studies (ATSIHealth), for materials published from 2013 to 6 December 2022. A professional librarian provided help to develop the search strategy; full search terms are available in Supplementary Table 2.

The search strategy was designed to identify papers that included: (1) Indigenous Peoples across various countries worldwide and (2) Access to routinely collected health data. To identify studies which included Indigenous Peoples, we used subject headings such as ‘Health Services, Indigenous’, ‘Indigenous People’s’, ‘United States Indian Health Services’ and related free text searches. Similarly, studies which accessed routinely collected health data were found using subject headings such as ‘Medical Record Linkage’, ‘Routinely Collected Health Data’ and related free text searches.

Study selection

Title and abstract review inclusion and exclusion criteria were drafted, and a sample of 50 papers were reviewed by two researchers (T.E. and J.W.) to refine and agree on the final criteria. The same sample of 50 papers was reviewed by the other researchers (H.W. and S.K.), and conflicts were discussed to ensure all reviewers had a consistent understanding of the criteria. Inclusion and exclusion criteria are described in Table 1.

Table 1 Title and abstract review inclusion and exclusion criteria
Full size table

The title and abstract review of each article was performed by two independent researchers, randomly assigned by Covidence to members of the research team (T.E., J.W., H.W., S.K.). Conflicts were resolved through group discussions with at least two researchers.

Full-text review was also conducted by two researchers independently, randomly allocated by Covidence to research team members (T.E., J.W., H.W., S.H., S.K., S.O.), with conflicts resolved through a group discussion with at least two researchers. Papers were excluded if: (1) could not locate a full-text article; (2) full text not available in English; or (3) not peer-reviewed original research article; or (4) not focused on Indigenous People (at least 90% of study participants); or (5) did not use routinely collected health data; or (6) study outcome was not a health outcome; or (7) did not use personal level health data.

Cohen’s Kappa was extracted from Covidence, and a weighted average was calculated to compare inter-rater reliability for both stages of the review.

Data extraction

Study characteristics, Indigenous data governance approaches and advantages and disadvantages of using routinely collected health data were extracted from the included papers. One reviewer (T.E.) developed a data extraction template in Covidence and tested it with four other reviewers (H.W., J.W., S.K., S.O.) independently extracting five articles each. Conflicts were discussed, and refinements were made to the data extraction template. Double data extraction was then completed by six reviewers (T.E., H.W., J.W., S.K., S.O., E.L.). One reviewer (T.E.) resolved conflicts for consistency.

A risk of bias assessment was not conducted as part of this scoping review, as the purpose is to examine Indigenous data governance practices reported, not to report on the outcomes of the studies.

Data analysis

Data extracted from Covidence was exported into a spreadsheet. The study characteristics were analysed using descriptive statistical techniques. A table was produced summarising the number of studies in each category. For the extraction of qualitative data, a thematic analysis approach was employed, following the methodology of Braun and Clarke18. This thematic analysis methodology involved two reviewers (E.L., H.W.) familiarising themselves with the data to generate coding elements and then iteratively comparing these coding elements to identify recurrent themes and subthemes.

The frequency of the main Indigenous data governance approach being reported in the included studies was summarised in a table. The Indigenous data governance approaches, advantages and disadvantages described in the studies were distilled into a checklist of considerations for using Indigenous Peoples’ routinely collected health data for research. The table was structured according to the horizons of digital transformation in health19, a commonly used framework in digital health. The horizon names were amended to focus on data selection, access and use.

Results

The combined searches identified a total of 1012 articles; after removing duplicates using EndNote and Covidence, 580 unique articles remained. After the title and abstract screening, 145 articles were included for full-text retrieval. Reviewer agreement was moderate for title and abstract screening (κ = 0.58). All full-text articles were found and assessed for eligibility, which resulted in 85 articles being included (Fig. 1). The reviewers had a substantial agreement on study inclusion (κ = 0.63).

Fig. 1: PRISMA study selection diagram.
figure 1

From: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71. For more information, visit http://www.prisma-statement.org/.

Full size image

Characteristics of included studies

The characteristics of the studies included in this review are summarised in Table 2, and the details of each study are available in Supplementary Table 3. The included studies were published between 2013 and 2022. Studies were carried out in four countries, including Australia (n = 38; 44.7%), the United States (n = 25; 29.4%), Canada (n = 19; 22.4%), and New Zealand (n = 3; 3.5%).

Table 2 Study characteristics of included studies
Full size table

Amongst the 85 articles included in this review, 82 articles reported on the number of participants, ranging from 8 to 138,551. One article considered the number of visits (i.e., 5373) of the target population, while another included 29 Aboriginal Community Controlled Health Services representing 34 individual clinics and 5 clinical hubs. One article did not describe the number of participants/visits. A majority of articles considered all genders (n = 70; 82.4%), while 10 studies focused on women only (11.8%). The articles included participants of one or more Indigenous backgrounds, with a majority being Aboriginal Australian (n = 38; 44.7%), Torres Strait Islander (n = 24; 28.2%), Alaska Native (n = 19; 22.4%), American Indian (n = 18; 21.2%) and First Nations living in Canada (n = 14; 16.5%). More than half (n = 45; 52.9%) of the studies focused on populations in rural or regional populations, 14.1% considered urban areas only, and 28.2% considered both.

The studies examined one or more health outcomes including healthcare utilisation and access (n = 24; 28.2%), maternal and child health (n = 17; 20%), chronic diseases and comorbidities (n = 15; 17.6%), infectious diseases (n = 10; 11.8%), mental health and suicide risk (n = 8; 9.4%), public health and prevention (n = 6; 7.1%), dental health (n = 2; 2.4%), and other health conditions (n = 7; 8.2%).

Indigenous data governance

The frequency of Indigenous data governance approaches described in the studies is included in Table 3. A checklist of considerations for using Indigenous Peoples routinely collected health data for research synthesised from the included studies is shown in Table 4, structured by the horizons of digital transformation19.

Table 3 Frequency of Indigenous Data Governance Approaches Described in Included Studies
Full size table
Table 4 Considerations for selecting, accessing, and using Indigenous Peoples Routinely Collected Health Data for Research
Full size table

Indigenous data sovereignty

Data sovereignty aspects were described in 34 (40%) of the studies. Eighteen studies outlined the requirement of the state health services to maintain control of the data20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36, while 15 studies outlined Indigenous Peoples or Communities sovereignty over their own data27,31,32,37,38,39,40,41,42,43,44,45,46,47,48. Fourteen (16.5%) studies described the researchers’ inability to share data publicly24,26,33,34,35,41,45,46,47,48,49,50,51,52 due to privacy and ethical restrictions31,35,41,47,49,51,52. However, in 11 (12.9%) studies, the researchers described the data can be obtained upon reasonable request30,31,35,37,41,47,48,49,50,51,52, subject to additional institutional23,26,27,28,29,30,31,32,33,34,35,36,44,45,49,52,53 or Tribal27,31,32,37,41,42,43,44,45,46,48 approvals, and/or compliance with privacy policies26,27,34,46,47. Furthermore, researchers had considered the use of de-identified data to analyse and present information with the intention to promote the anonymity of the Indigenous People whose data was accessed in the research study20,21,22,23,24,25,26,27,29,30,31,32,33,36,38,41,46,48,50,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72.

In addition to the approvals, studies highlight that data collected should be securely stored in various repositories21,31,32,50 and regulated by organisations such as Tribal health Organisations40, healthcare/government departments38,39, and data custodians25,30. Access to the data is restricted to research investigators21 or people who meet prespecified criteria for data access27,31,35,47,50,51. This information was most commonly included in a Data Availability Statement, which is increasingly being required by journals, and hence was more prevalent in recent studies (10% of studies published 2013–2015 vs. 63% of studies 2020–2022).

In terms of data sovereignty principles and access, the procedures vary with individual context. For example, in Manitoba, approval from specific entities is required to access data27,31,32,33,34. The Navajo Nation41 and Western Australia24,35,47 also have specific processes for data access, while at the Sioux Lookout, access and management of data need to be conducted in accordance with the principles of ownership, control, access and possession (OCAP)40.

Approach to consent

The majority of studies (n = 57, 67.1%) did not describe how consent was approached. Of the 28 (32.9%) studies which did include this information, nine studies obtained consent from individual participants28,31,61,65,73,74,75,76,77. Fifteen studies employed a waiver of consent to access participants' health data21,23,24,33,36,37,41,42,43,47,49,55,63,66,71. Another approach to consent was to obtain permission from Community Leaders46 or Community Organisations (Aboriginal Community Controlled Health Services) involved48. Furthermore, in studies that required follow-up care78 or further review of specific individuals’ documents79, additional consent was sought from the participants at that stage.

Involvement of indigenous community and people in research

The researchers in 65 (76.5%) studies described various measures undertaken to ensure their research was conducted with the involvement and approval of the Communities they worked with. In several studies, researchers obtained approvals from Indigenous Leaders32,33,55,74,80, institutional organisations34,37,39,44,52,56,59,61,62,63,74,81,82,83, and Tribal health Organisations21,23,31,32,33,35,37,39,42,43,44,45,47,50,51,52,53,56,59,61,62,63,68,72,81,82,83,84,85,86,87,88,89 to commence, undertake and/or disseminate findings at various stages of the research study. In addition, partnerships were established with various Indigenous Leaders34,55,61,74, Communities22,30,31,34,44,49,52,62,66,69,73,75,79,80,90 and Organisations21,23,26,29,31,33,35,39,40,41,46,47,48,50,53,55,56,59,66,67,69,71,73,74,75,77,79,80,85,87,88,91,92,93 to incorporate their perspectives and ensure cultural relevance23,34,50,51,56,62,73,79.

In one study by Struck, et al.56, the researchers described that the research needs to be conducted in the spirit of truth and reconciliation with recognition of the harms conducted to Indigenous People. By focusing on transparency, mutual respect, and maintaining a shared understanding of Indigenous data,56,81 it may be possible to achieve deep trust81 and respectful collaboration with Indigenous People61. While in some studies21,23,24,33,36,37,41,42,43,47,49,55,63,66,71,81 researchers received approval for a waiver of informed consent, efforts were made to maintain transparency and trust between researchers and Indigenous Communities81.

Indigenous organisation author affiliations

The inclusion of Indigenous Organisations in the research study was evident in 60 (70.6%) studies where one or more co-authors were affiliated with Indigenous Health, Research or Community Organisations. These co-authors participated in the design, development, data collection and analysis of the research study24,35,42,60,65,71.

Indigenous ethics approval

Thirty-five (41.2%) studies reported receiving ethics approval from an Indigenous-specific ethics committee for their study23,24,29,35,36,37,39,40,41,42,43,44,45,46,47,48,56,59,61,63,64,68,69,80,82,83,84,85,86,87,90,92,93,94,95. Forty-three (50.6%) studies detailed receiving ethical approval to conduct their study from a non-Indigenous ethics committee. Six studies did not describe whether ethical approval was received, and one study stated ethical approval was not required.

Indigenous guiding principles

Eighteen (20%) studies described using Indigenous guiding principles to inform their research22,31,32,34,35,40,46,48,50,52,53,56,72,73,78,89,92. For example, in Canada, several studies focused on the use of OCAP (ownership, control, access, and possession)31,32,40,46,50,92 and OCAS (ownership, control, access, and stewardship)56 principles in Indigenous health research. These principles were followed throughout the study31,32 to ensure governance of Indigenous data46. One study also described the inclusion of Chiefs of Ontario First Nations Data Governance Committee and the Grand Council Treaty towards the review of the study’s compliance with the OCAP principle50, while another study was supervised, and the data were maintained by the Sioux Lookout First Nations Health Authority in accordance with the OCAP principle40. Other studies focused on including several ethical and scientific standards from the various Canadian Institutes (Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada, and Social Sciences and Humanities Research Council of Canada)73. In particular, Section 6 of the Tri-Council Policy Statement regarding the Ethical Conduct for Research Involving Humans, that involves First Nation, Métis or Inuit People26,89. Moreover, one study by Pena-Sanchez, et al.22 utilised the Indigenous medicine wheel as its foundational framework, supported by cultural safety and patient-oriented research principles. It was guided by two specific Calls to Action from the Truth and Reconciliation Commission (TRC) of Canada. TRC 18 emphasised acknowledging the then-current state of Aboriginal health and implementing the healthcare rights of Aboriginal Peoples, and TRC 19 called for establishing goals in consultation with Aboriginal Communities to identify and address health outcome disparities. Additionally, one study co-developed its protocol with the Isumataiit Sivuliuqtii, ensuring a foundation grounded in Inuit ways of knowing (Inuit Qaujimajatuqangit)34.

In the United States, one study focused on promoting trust and respectful collaboration with the Indigenous People concerning research participation and patient confidentiality61. While, in Australia, four studies emphasised compliance with the National Health and Medical Research Council (NHMRC) guidelines for ethical conduct in Aboriginal and Torres Strait Islander health research35,52,72,78. Another Australian study72 was driven by shared values such as spirit and integrity, reciprocity, respect, equity, cultural continuity, and responsibility in all network activities.

Advantages and challenges of using routinely collected Indigenous Peoples’ health data

Advantages of using routinely collected Indigenous Peoples’ health data

Out of the 85 articles, twenty-eight (32.9%) of them discussed the benefit of using routinely collected Indigenous data in research in terms of enhanced efficiency and inclusivity while minimising biases and participant burden. One of the key benefits is that by leveraging existing data sources30,37 and linking them together23,29,34,36,49,70,96, researchers are able to access a wealth of information23,28,30,34,36,37,56,62,70,75,95,96,97 without requiring additional input from participants. This minimises the participant burden72 and reduces reporting and recall bias33,56, while also making the research process more efficient87. Moreover, it can also provide data that is broadly representative of the Indigenous Communities75. As a result, the studies can achieve more robust and representative findings.

Another significant benefit lies in the ability to extract comprehensive and detailed information on patients’ diagnoses, treatments, follow-up care, and relevant outcomes45,59,97. Researchers can utilise this data, which is often underutilised69, not only to examine high-risk populations49,59,96 and health trends of Indigenous People28,29,34,36,37,48,56,70,72,75,85,88,96,97 but also to monitor service utilisation34, interventions20 and outcomes69. These are essential for strategic planning and operational decision-making in healthcare services29. In addition, EMRs can allow for improved data validity and reliability87 while also automating data collection and analysis tasks48,65, which enhances the sustainability of surveillance systems48. The automation of these processes provides a significant advantage to researchers over the use of manual procedures65. This increases the efficiency and longevity of research projects, allowing them to have a lasting impact even beyond their initial funding period48.

Challenges of using routinely collected Indigenous Peoples’ health data

Several reported challenges and potential biases in using healthcare data were identified in 39 (45.9%) studies included in this review. The most significant challenge reported in the included studies is regarding data completeness85. Incomplete health data, including demographics and family variables39, physiological and lifestyle measurements23,32,40, laboratory report31,50,74,87, disease history and severity58,71,92, costs40,50,97, socioeconomic status23, risks51,77, charts53, health service utilisation24,46, diagnosis and treatment39,40,49,55,80, discharge abstracts21, and critical social and cultural dimensions25, were reported as either missing, underestimated, not recorded, or inaccessible. In one study, the researchers reported missing data ranging from 0 to 15.8% depending on the database98, while in another study, 9% of age and sex distribution data was considered to be missing52. The incompleteness of data was reported to be a consequence of high population mobility54,65, unclear clinical catchments54, consults in other health services67,80, availability of other non-department of health services54,99, unclear definitions used in the storage of data45,47,53,90,99, the inability to contact participants69 and/or limited medical consultations24,36,44,100. For example, one study described that patients who seek care outside the Indian Health Service (IHS) would not have their data recorded in the National Patient Information Reporting System99. This would result in a small sample size85, potential coding errors75 and inability to provide accurate estimates regarding an outcome24,31,36,38,50,51,52,54,65,77,92,99.

Data quality was identified to be another significant challenge reported in the included studies. From the included studies, it is evident that the health data is entered by the clinicians and is reliant on the consistency and quality of clinician recording54,65,71; often criticised for its dependency on clinicians41. However, this is an ongoing challenge, especially for new staff, who need to quickly learn the system and perform the necessary actions, thus affecting the accuracy and comprehensiveness of data collection91. The challenge of data quality may also be because of the limited scale of data25,39, generalisability87 and misrepresentation or misclassification of data22,47,59,79 that could lead to bias22,33,45,62,88. To mitigate these issues, researchers have looked towards tracing individuals through the system and by manually verifying the data63; however, they have been unsuccessful due to limited access30.

Discussion

This review sheds light on Indigenous data governance approaches employed by researchers when accessing Indigenous Peoples’ routinely collected health data. The findings reflect that Indigenous Data Governance (IDG) is an emerging area with inconsistent reporting of these approaches. Some elements of IDG, such as ethical approval and the involvement of Indigenous Peoples in research, were widely reported, while others, such as how data sovereignty was maintained and the use of Indigenous guiding principles, were less often reported. We propose that reporting on IDG approaches provides readers with confidence that the research was conducted ethically. A reporting guideline for research using Indigenous Peoples’ routinely collected health data may be useful to encourage the explicit and consistent inclusion of IDG approaches.

The benefits of utilising routinely collected health data for research are widely recognised to enhance healthcare efficiency and delivery101. However, its use in research poses significant ethical challenges related to patient privacy and data access, especially for Indigenous Peoples102,103. Consequently, Indigenous data governance is crucial to ensure the power, authority, access to, ownership and use of data is maintained by Indigenous Populations104. While the implementation of such approaches requires time, resources, education, and planning, when properly executed, it can serve as a driver for Indigenous-led strategic planning and decision-making in public health105. These approaches can help develop deep trust81 and respectful collaboration with Indigenous People61 through transparent, mutual respect and shared understanding of Indigenous data56,81. Including Indigenous Community Leaders and People can ensure cultural appropriateness in the process of strategic planning and operational decision-making within healthcare services23,34,50,51,56,62,73,79.

Indigenous Peoples are considered to be some of the most researched groups in the world14, which has put a significant burden on these Communities to share information about their health and participate in trials. Utilising routinely collected health data provides an opportunity to conduct important research without the need to burden populations through additional data collection72. Routinely collected health data can make the research process more efficient and cost effective87, it can also enable comparison or follow-up across longer periods of time and access to more people than would otherwise be practical48. There are well-documented limitations in the quality and completeness of routinely collected health data, the most significant being the inaccurate identification of Indigenous People29,68. Researchers should consider these factors when deciding whether utilising Indigenous Peoples’ routinely collected health data is appropriate for their research.

Despite using rigorous methods to understand the approaches to Indigenous data governance in healthcare, this study has its limitations. The research study incorporated the ATSIHealth database which focuses on Aboriginal Australian and Torres Strait Islander, as well as other international databases, but did not include databases specific to other Indigenous groups. While it can be argued that the research may have potential biases, the authors included research assistants from Canada and the USA and ensured a comprehensive set of search terms to encompass the diverse Indigenous Communities. This ensured thoroughness in extracting data from various health databases.