Child health research helps uncover how biological and contextual environmental factors interact to influence child health outcomes. These factors are strongly interconnected; for example, socioeconomic adversity can negatively affect physical and mental health [1, 2], while compromised health in childhood may limit educational achievement [3]. Population-based studies allow researchers to capture these complex relationships through large-scale evidence [4]. Multi-domain administrative data offer a valuable source for studying child health, including physical, mental, and social well-being, by linking data on health and non-health domains [5–7].

Administrative data are routinely collected records to manage health care, education, social programs, and other sectors, rather than for research purposes; yet they capture large, representative populations that reflect real-world settings, making them valuable for research [4, 8, 9]. Population-based administrative data also offer key advantages as they are relatively low-cost, allow for long follow-up periods, and reduce challenges such as attrition that commonly affect primary data collection [10], which is often constrained by the need for parental consent, child participation, and long-term developmental follow-up [11–14]. Moreover, these barriers are compounded by imbalanced recruitment across socioeconomic and cultural groups, as well as loss to follow-up when families relocate, change schools, or transition between paediatric and adult care systems, all of which can lead to underrepresentation and biased results [15, 16]. Therefore, linked administrative data are useful in child health research by enabling long-term follow-up within large population-wide cohorts.

Linked multi-domain administrative data integrate information from different sectors (e.g., health, education, justice, social services) at the individual or family level [7]. They allow researchers to combine health datasets, such as physician claims, prescriptions, and immunisation records, with non-health datasets, such as education, child protection, and social services [4, 17]. In this review, a domain refers to a data sector representing a distinct aspect of people’s lives or service systems. For example, household income reflects socioeconomic status, typically obtained from provincial income support or social assistance records, while child protection involvement refers to contact with child welfare services. These non-health domains capture many of the social and structural determinants of health, such as poverty, family structure, and justice involvement, that influence children’s life trajectories but are often underrepresented in clinical data [18, 19]. Another key advantage of administrative data for child health research is the ability to link children’s records with those of parents and grandparents, enabling intergenerational analyses of family exposures and related child health risks [20, 21]. These advantages make administrative data particularly useful for studying child populations.

Canada’s universal systems and data infrastructures provide unique opportunities for linking health and non-health datasets for child health research [22]. Several Canadian provinces have established research data centres, such as the Manitoba Centre for Health Policy (MCHP), the Institute for Clinical Evaluative Sciences (ICES) in Ontario, and the Population Data BC (PopData) in British Columbia, which provide secure access to linked health and non-health datasets [23–25]. Despite Canada’s unique strengths in multi-domain administrative data, there is an unclear overview of how the data sources have been used to study children’s health. Examining how existing studies link health and non-health datasets is crucial to strengthen evidence on the influence of social, educational, and family contexts on child health outcomes and to identify future research opportunities. Although some reviews have examined how administrative data have been used to address specific topics, such as child maltreatment [26] or adult-focused mental health and addictions [27], no review has comprehensively summarised Canadian population-based child health studies that integrate both health and non-health domains from administrative data. Addressing this gap will help identify how health and non-health information are being linked across Canada and guide future use of administrative data in child health policy and research. Therefore, the purpose of this scoping review is to describe Canadian population-based child health studies that used linked multi-domain administrative data.

This review identified 42 Canadian articles using multi-domain linked administrative data in population-based child health research, all conducted within a single province or territory, reflecting Canada’s decentralised data systems [66, 67]. This decentralised concentration may also reflect differences in data governance frameworks, availability of established data repositories, researcher capacity, and funding across provinces or territories. Similar systems exist in other countries, such as Australia and the United States, where multi-jurisdictional collaboration requires complex coordination [68, 69]. However, emerging strategies are beginning to facilitate coordinated multi-jurisdictional data access. Initiatives such as Health Data Research Network Canada (HDRN) aim to improve and coordinate access to administrative data across multiple jurisdictions [70]. Similar national data coordination exists internationally, such as Health Data Research UK (HDR UK) in the United Kingdom and the Population Health Research Network (PHRN) in Australia, which provide secure, federated infrastructures to support cross-jurisdictional data linkage [71, 72]. In situations where direct coordination is not feasible, meta-analysis offers an alternative approach by pooling region-specific estimates to generate overall national estimates while maintaining privacy protections that prevent the sharing of data between regions [73]. As a result, multi-jurisdictional research is necessary to improve generalizability in child health research findings.

Of the 42 included articles, nine did not include any parental linkage, and most linked only to mothers. This limitation constrains the ability to examine intergenerational risk factors and family level associations; for example, linking both parents allows for comprehensive analyses of how parental justice involvement and household income influence child health outcomes [74–76]. It also supports advanced analytical designs, including sibling and cousin comparisons, which can strengthen causal inference [77]. Expanding both-parent linkage is therefore essential to capture the full family context and advance robust intergenerational research on child health.

Across studies, the commonly reported non-health variables were immigration status and household income, while limited studies reported on information related to Indigenous identity and parental justice involvement. Studying child health requires linking health and non-health data, because child health outcomes are influenced by education, family, and social factors [4, 5]. Linking these datasets at the individual or family level allows exposures (e.g., poverty, child protection involvement) and outcomes (e.g., vaccination coverage, academic achievement) to be measured within the same population, particularly in vulnerable groups who may be underrepresented in health data alone [18, 19]. Other important non-health variables were largely absent or limited in the 42 included articles, such as housing instability and childhood adversities. This likely reflects how administrative data are generated and structured, as these variables are not routinely captured in administrative data and are often confined to sector-specific systems, such as social services, where access and linkage are more complex, limiting their inclusion in population-based research. These variables are essential for understanding inequities rooted in socioeconomic and cultural contexts, and their inclusion would allow researchers to examine disparities in child health outcomes to better support marginalised populations. Addressing this limitation will require better integration of data across sectors, more consistent collection of key social determinants, and transparent processes to ensure equitable data collection.

This review has strengths and limitations. First, it provides a comprehensive review of Canadian population-based child health studies using linked multi-domain administrative data. Second, the review followed PRISMA-ScR guidelines and applied predefined eligibility criteria to ensure transparency and reproducibility. Third, two reviewers independently screened all articles at the titles/abstracts and full-text stages, resolving disagreements through discussion to maintain review quality. Fourth, the search strategy was developed in collaboration with a health sciences librarian, strengthening its comprehensiveness through database-specific adaptations. However, the search strategy might not have captured all relevant articles, as some publications may have used alternative terminology (e.g., medical claims) that did not align with our searching strategy. In addition, this review did not search grey literature sources, such as government or institutional reports, which often use administrative data but are not published in peer-reviewed journals. Future studies could expand the current search strategy by incorporating additional generic or alternative terms and grey literature searches to optimise the retrieval of relevant articles.

This scoping review highlights the growing use of multi-domain administrative data in Canadian population-based child health research. To advance the field, future research should strengthen multi-jurisdictional collaboration, enhance parental linkage and expand linkages to additional non-health domains to better capture the wide range of social determinants that shape child health outcomes and to produce comparable evidence that informs child health policy and services across Canada.