Neurodevelopmental conditions (including intellectual disability, autism, cerebral palsy, attention deficit hyperactivity disorder, and some genetic syndromes) are usually diagnosed in childhood, and are associated with varying levels of support needs [1–3]. Siblings of children with these conditions can provide important companionship and care, sometimes having informal carer roles as children and more formal care roles as adults [4–7]. Siblings are often described as having positive character traits, and develop rich relationships as a result of growing up alongside someone with a disability [8, 9], yet some may also be at risk of poorer health outcomes compared to their peers, which may be the result of shared genetics, environmental factors, family factors, or a combination of these [10–12]. Neurocognitive difficulties, psychiatric diagnoses, and self-harm have been observed in some siblings during childhood and young adulthood [13–17].

Health outcomes for siblings are often reported at the individual level, using interviews [18–20] or surveys [21, 22], and are sometimes subject to convenience sampling of available siblings or families of children who have specific neurodevelopmental conditions [23, 24]. The evidence base around the health outcomes of siblings at a population level is relatively sparse. Compared to surveys and clinical samples, population-based data can help obtain adequate sibling sample sizes and representation for minority groups and rare conditions [25–31]. Population data also allow the selection of comparison groups from the same time period and data collection parameters, which strengthens the context and interpretation of the findings. Further, it can also reduce the need for follow-up if enough data are available. Existing data minimises recall bias, attrition, or limitations on resources to engage participants, that can occur when collecting data from individuals directly [32].

Population data relating to individuals with neurodevelopmental conditions are usually collected and maintained in specific administrative databases [33–36], or in disability-specific population databases [26, 37–40], and these rarely include data on other family members. Using data linkage methodologies, disability data can be connected to health and other datasets, including genealogical connections, to follow individuals or families across wide-ranging domains. Linkage allows for large-scale analyses to occur for specific sibling health outcomes and wellbeing trajectories, and provides a unique evidence-base from linking together information that would otherwise remain siloed or fragmented [32, 41–43]. The use of high quality linkage methods and databases can minimise linkage error bias that can occur when linking population data [44].

The benefits of applying data linkage to independent data sources for sibling research are significant [45, 46]. Linkage can update or extend existing data repositories and can rapidly merge relevant information for faster interpretation and response, especially important for mental health concerns. Life course experiences of siblings can be tracked [47] across a wide variety of contacts with health (e.g. hospital visits, medication prescriptions, mental health contacts) or other registries (e.g. education, census data, crime) until the end of the follow up period, or death [28, 41]. Data linkage can be used to improve visibility of siblings of individuals with rarer neurodevelopmental conditions for whom understanding health patterns and circumstances may require data extensions beyond a single administrative system. The data linkage process itself is able to preserve confidentiality [48] and reduces the need to approach individuals to collect information that may already exist, or be difficult to collect.

The objective of this scoping review was to explore the application of population data linkage as a research method to capture health outcomes of siblings of children and young adults with neurodevelopmental conditions, and ways data linkage can enhance knowledge in this field. This review describes the population characteristics and methodology of studies meeting the eligibility criteria, the neurodevelopmental conditions of index children, sibling health outcomes reported, and the infrastructure used.

Before commencement, scoping review databases (Open Science Framework, Figshare, JBI Evidence Synthesis, and BMJ Open) were searched for similar registered reviews, but no protocols involving the key terms for this review were identified.

What are the characteristics of population data linkage studies that capture health outcomes of siblings of children and young adults with neurodevelopmental conditions?

How has population data linkage been used in these studies to capture health outcomes that originate in childhood or young adulthood, what are the capabilities of data linkage to explore wide scale health outcomes for siblings, and what could be enhanced to improve future sibling research involving linked data?

Studies using population data linkage (administrative data, registry data, and research databases) in a defined geographical area (a country, state or territory, or region bounded by government or administration) to capture health outcomes (physical and/or mental health) of siblings of children and young adults with neurodevelopmental conditions were considered. All study designs were included e.g. quantitative, population based, longitudinal or cohort studies, if they met the inclusion criteria.

Studies were included if both index children with neurodevelopmental conditions and their siblings were children or young adults, defined as being aged up to 25 years [49]. Siblings did not need to be the primary group of interest as long as sibling data (full or half siblings) were presented. In studies where sibling ages were not stated or were combined with cases/controls, the estimated mean age for index children could not exceed 25 years. Twin studies were included if twins were identifiable in subgroups, and if other singleton sibling data were available. Where possible, comparators comprising of siblings of children and young adults without neurodevelopmental conditions were included.

The neurodevelopmental conditions of index children must have been diagnosed clinically and/or recorded as an entry in either a population administrative dataset(s), diagnosis specific registry or project database using standardised classification systems (e.g. Diagnostic and Statistical Manual of Mental Disorders (DSM) [50], International Classification of Diseases (ICD) [51]). For example, identification of a diagnosis of intellectual disability using DSM in a disability specific registry, a record of autism using an ICD code in registry-based patient records, or identification of attention deficit hyperactivity disorder using a medication code in a prescribed drug registry.

Sibling health outcomes captured must have been presented as clinical diagnoses and recorded as a standardised entry in a relevant database. For example, identification of a diagnosis of obesity using a record of ICD code in registry-based hospital records, or a record of death using a cause of death code in a mortality registry. No inclusion criteria were applied to social, cultural and gender constructs.

This scoping review followed the JBI methodology for scoping reviews [52], and a JBI checklist was completed as an assessment of conformity [53]. A protocol was published a priori in Open Science Framework [54] which outlined the objectives, inclusion criteria, and methods planned [55].

The identification and inclusion of papers were made according to the guidelines of Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping reviews (PRISMA-Scr) [53] (Figure 1). A total of 31 papers met the criteria for inclusion.

CINAHL, Ovid (MEDLINE, Embase, PsycINFO), Scopus, and Web of Science were the final search sources of evidence for papers that captured sibling health outcomes of children and young adults with neurodevelopmental conditions using data linkage methods. Papers were published in a variety of journals, most of which related to the neurodevelopmental condition of index children or sibling health outcome observed (e.g. papers capturing data on psychiatric disorders were published in psychiatric journals). Some journals had a broader focus (e.g. paediatrics, open medical journals), and a few journals related to population-based methodology (e.g. epidemiology, population health) [59, 60]. A summary of paper characteristics of the final sources of evidence are reported in the Summary table (Table 1). The full results of this scoping review are reported in the Data extraction table (Supplementary Material 3). The following findings are presented as a descriptive summary of the paper characteristics, methodologies, neurodevelopmental condition of index children, and sibling health outcomes presented.

Print publication dates of included papers ranged from 2013 to 2024, most of these since 2019 (Table 1). Many papers referenced familial or first-degree relatives in the title and included siblings within this broad category (Table 1) [61–64]. Around one third of papers referenced siblings in the title of the paper [65–67], and in most of these, siblings were the primary interest group. The remainder of papers had no reference to siblings or family in the title [68–70]; however, sibling data were included in the abstract and body of the papers.

The majority of papers related to studies from Norden countries. The total papers were represented by the following countries in order of Sweden, Taiwan, Finland, Norway, Denmark, and Canada (Table 1). Thirty-nine unique datasets were used for data linkage across all included studies, of which 20 were databases used in the 20 Swedish studies (Supplementary Material 3). On average, each paper reported on data from five linked datasets, up to 11 in some instances. The most commonly reported datasets were those relating to births, family connections, hospital, medication, health insurance, and deaths.

Half of the included papers reported a whole population study design, and the remainder a matched case-control study design (ratios ranged from 1:4 to 1:50 matched controls) (Table 1). Sibling cohort size for whole population studies ranged from 3,410 to 4,289,186, with a mean of 1,270,764 individuals (Table 1). As expected, sibling population sizes were smaller among matched control designs, ranging from 1,304 to 665,994, with a mean of 82,793 individuals. One third of included papers reported on paternal and maternal half siblings, in addition to full siblings. Almost half of all papers reported sibling characteristics such as birth order and/or sibling size.

Across all papers, duration of follow-up ranged from 15 to 77 years with a mean of 31 years, and sibling ages ranged from 0 to 44 years (Table 1). Cohort birth years ranged from 1932 to 2013, with around one-third of studies having birth years beginning between 1980 to 1989.

Siblings were the primary group of interest in almost one third of papers, while the remaining two-thirds of papers included sibling data for comparison to the index cases and/or controls (Table 1). Each paper’s aim has been provided in relation to the sibling data captured within the paper (Supplementary Material 3).

The neurodevelopmental conditions of index children included autism, attention deficit hyperactivity disorder, intellectual disability, cerebral palsy, and developmental delay (Table 1). Most studies on autism included subgrouping for the presence of intellectual disability [71], and over half captured sibling outcomes for these subgroups [72, 73]. Two studies on intellectual disability included subgrouping by severity levels, and one of these reported sibling outcomes for the subgroups (Supplementary Material 3) [74].

Sibling health outcomes captured were wide ranging, from broad outcome categories to specific conditions (Supplementary Material 3) [6, 75–79]. The most common were psychiatric or mental health conditions, followed by physical health conditions, and neurodevelopmental conditions (Table 1).

Findings from this review demonstrate the utility of population data linkage to capture health outcomes of siblings of children and young adults with neurodevelopmental conditions, including when siblings are not the primary focus. The sibling data captured in the reviewed papers, and the final number of included papers, was shaped by the data linkage capabilities in the area where the research was being conducted, and therefore narrowed the findings accordingly. Despite this, the search was able to generate an informative understanding of the application of data linkage as a method to deliver population perspective in this field. The included papers demonstrated a wide scope in target populations and sample sizes, a range of diagnostic observations, and availability of sibling data spanning decades of data collection. These qualities have significant benefits for sibling research by facilitating the exploration of sibling wellbeing in adequate samples across disability types, developmental ages and a broad range of outcomes, allowing representation of siblings in a variety of contexts.

Recent increases in the number of eligible papers indicates a general growth in sibling-based research. The countries represented in this scoping review reflects the availability of population data infrastructure and capabilities in those regions, the linkages that readily occur between datasets, and locations where genealogical connections may be more available [80–84]. Sweden, for example, has established two population registers, the population register (maintained by the Swedish National tax agency), and the Total Population Register (TPR) (maintained by the government agency Statistics Sweden). Both registers contain large amounts of health and other data, and family links, and updates are transmitted daily [81]. Whereas, in Taiwan, a data repository site, the Health and Welfare Data Science Center (HWDC), was established to centralise the National Health Insurance Research Database (NHIRD) and around 70 other health databases, to streamline research data management [82]. The power of data aggregation from both types of infrastructure is significant, and combined with the availability of family linkages resulted in the inclusion of many papers into this review. The absence of papers from other geographical areas with established data infrastructure and linkage capabilities [85, 86] reflect capability or methodological differences that prevented their inclusion in the current review, and highlights avenues for future research in this field to be explored.

The types of registries used in the included studies varied, depending on the diagnostic groups of index children, sibling health outcomes observed, and the availability and suitability of data sources [87]. Many studies used government registries e.g. birth registers, which support broad population data collection and increase the feasibility to conduct linkage to other datasets. The use of longitudinal government-maintained statutory databases assumes a consistency and quality of data and data linkage methods, providing confidence to the findings of studies that use these resources. Studies typically linked multiple registries, demonstrating the versatility and broad application of data linkage methods for sibling research, where data may need to be captured across different domains. This allows considerable power and breadth in the types of analyses that can be made at the population level, and demonstrates the benefits of combining data from sources that would otherwise remain independent [88, 89].

The majority of studies used ICD codes extracted from government-based records (e.g. hospital, prescription, or mortality databases) to identify diagnostic groups and/or outcomes [81, 82], enhancing consistency and replication of the studies. However, validation issues surrounding adequate identification of neurodevelopmental conditions such as intellectual disability in these databases have been raised and should be acknowledged [90]. Other studies used DSM classification systems to identify diagnostic groups and/or outcomes, which may be partial to similar limitations. In addition, diagnostic trends and the ability to co-diagnose some neurodevelopmental conditions (e.g. autism and attention deficit hyperactivity disorder [91]) have evolved over time, and comparability across a long time period may be limited. Knowledge has also increased of the presentation of some conditions, such as attention deficit hyperactivity disorder in individuals with a learning or intellectual disability [92], or autism in females [93], which may also impact on comparisons across different studies or time cohorts. It also warrants caution in using coded data for inferring genetic risk. In this review, studies often used non-government registries (i.e. Stockholm youth cohort, psychiatric registers), which typically housed data dedicated to specific conditions, not purely for administrative purposes [94]. Such registries can provide unique data and perspective, greater diagnostic accuracy, inclusion of rarer conditions, and finer details such as severity level, which are vital to assessing sibling outcomes [14, 95]. However, for condition-specific registries, availability of family links may not be possible or sustainable due to the resources required to build and maintain them. Resources devoted to developing these data sources, and enhancing linkage capabilities between government-based datasets and disability specific/non-government datasets would help improve the availability of data capture in the future.

The lengthy follow-up times of the included studies highlight a significant benefit of using linked data for sibling research, enabling longitudinal data analyses, spanning different developmental and life stages within a relatively short analytical timeframe. In contrast, establishing a prospective study to collect equivalent data would require sustained data collection and resources over the entire follow up period. It is vital to consider life-course approaches to sibling mental health, as many siblings adopt formal carer roles later in life, and experience challenges as adults [96], but for some, mental health concerns are already observed in childhood [21]. Data linkage allows the analysis of health and wellbeing indicators from birth to understand trajectories of development and change, covering boundaries of childhood, where siblings live with their families, to early adulthood where independence is reached, and formal caregiving may occur within the enduring sibling relationship. Data linkage allows bridging of childhood and young adulthood data with health outcomes observed later in adulthood to support a lifespan approach to monitoring and improving sibling wellbeing.

The length of exposure, based on the youngest age at which outcome conditions were diagnosed in siblings, can be used to determine at what ages sibling supports are most needed [25, 45, 83]. Some studies in this review used a narrow age range, e.g. nine calendar years, and reported specific sibling outcomes for that age period [97, 98], showcasing more precise capabilities of data linkage methods. However, the majority of studies grouped sibling data, regardless of age and length of follow up, sometimes spanning 77 years [75], thereby limiting the interpretation by sibling age due to the lengthy time periods of mixed-age data aggregation. The use of linked longitudinal population data to advocate for sibling supports at more specific life stages is an important capability, and is a recommended focus in future work.

Reliable integration of genealogical data can be difficult to incorporate at scale using other non-linkage-based study designs. The papers in this review all used population-based genealogy to establish sibling relationships (full siblings, maternal/paternal half siblings etc.), of up to 4.3 million individuals in the largest sample [42]. Sibling characteristics such as sibling size, spacing and birth order are considered important risk or protective factors for various health outcomes [99, 100] and reporting of these characteristics is readily facilitated by linked population data. In this review, some papers included frequencies of sibling characteristics as demographic content [67, 68, 76, 101–103], while others included sibling characteristics as covariates to control for confounding effects [104], often combined with other socio-economic variables, thus limiting the context of the findings for siblings. In papers that reported these characteristics more comprehensively, additional benefits of population data for sibling outcomes were demonstrated [60, 79]. For example, presenting outcome rates based on the sex of the sibling, sibling size, and age distance to the index child [60], and using birth order to define sibling dyads and degree of risk [79]. Incorporating and reporting detailed sibling characteristics allows a greater understanding of how sibling relationships and family environments shape their outcomes across multiple domains.

Siblings can have unique life experiences, often shaped by the type and severity of neurodevelopmental conditions among children in the family [9]. Study designs that allow for comparisons to be made between condition types and severity levels can provide greater accuracy about sibling experiences and outcomes. Data linkage methodologies can facilitate this design, particularly for rare or lower prevalence conditions, and the ability to link to registries containing finer diagnostic detail. All papers in this review focused on index children with the same neurodevelopmental condition, thus limiting the ability to compare outcomes between different diagnostic groups in the one study design. One paper that focused on intellectual disability, reported outcomes for different intellectual disability severity levels, allowing enhanced sibling data to be captured, for example, showing stronger sibling risk of attention deficit hyperactivity disorder for mild and moderate intellectual disability than for severe intellectual disability [74]. Most papers did not report results by level of intellectual disability, citing potential overlapping genetic risk for neurodevelopmental conditions [105, 106] or the lower prevalence levels in intellectual disability subgroups [94], it may also be due to inadequate identification of intellectual disability in the databases used [90]. Studies including index children with autism, often subgrouped by the presence or absence of intellectual disability, reflecting the recognised independent effects of both autism and intellectual disability on outcomes [102, 107–109]. Studies where sibling data can be captured for different severity levels allow interpretation of the impact these differences may have on sibling health outcomes, and consequentially, understanding how to best support siblings according to their family circumstances.

The neurodevelopmental conditions of index children and sibling outcomes among studies included in this scoping review reflected population prevalence and publication trends, whereby neurodevelopmental conditions such as autism, and sibling health outcomes such as mental health conditions were collectively the most commonly explored [25, 45]. All studies captured sibling data showing some degree of increased risk or higher prevalence of the assessed outcomes among siblings of children and young adults with, compared to those without neurodevelopmental conditions [110]. The poorer outcomes observed for siblings at the population level supports the use of data linkage as a research tool to explore and report health disparities across a diverse range of outcomes. Most studies concluded the shared causal or risk factors prompt the need for further studies [67, 111] to assess genetic and environmental associations in the same study design. Some extended this to recommend preventive and clinical interventions be explored, and early identification and assessment for siblings [65, 66, 79, 103, 107, 112, 113]. These recommendations indicate a real need to increase the focus on sibling data across psychosocial, educational, financial, environmental and other domains.

This scoping review demonstrated the application of data linkage as an effective method to capture data for siblings of children and young adults with neurodevelopmental conditions at a population level. While a narrow inclusion of geographical areas was represented, increasing data capabilities and more vigilant reporting in the future will help improve the amount and visibility of research conducted with ongoing benefits for addressing the wellbeing of siblings.

Ethical approval for this scoping review was not required.

This research is funded by the Australian National Health and Medical Research Council (NHRMC)(GNT1184770), and a Senior Research Fellowship (1117105); and a UTP/RTA Scholarship from The University of Western Australia. The contents of the published material are solely the responsibility of the individual authors and do not reflect the views of NHMRC. DR is supported by the NIHR Biomedical Research Centre at the University of Bristol and University Hospitals Bristol and the Weston NHS Foundation Trust and a member of The MRC Integrative Epidemiology Unit (MC_UU_00011/1).