Abstract
Objective: An umbrella review to synthesize the fragmented evidence on the experiences and outcomes of transitions out of school for autistic young people, including those with intellectual disability (ID), and to identify recommendations for practice, research, and policy. Methods: An integrated narrative synthesis of 25 reviews covering 435 primary studies. Findings: Reviews were generally well-conducted, but primary evidence quality was variable. Early individualized planning and collaboration were consistent facilitators. Barriers included exclusion from decision-making, fragmented coordination, and uneven access. Long-term outcomes or adverse events were rarely considered. Few reviews included ID-specific analyses or examined intersectionality. Conclusions: The evidence base remains conceptually narrow, with limited guidance on equitable transition outcomes. Recommendations: Plan with, not for, autistic young people, including those with ID and other intersecting identities. Practice, research, and policy should align to prioritize, measure, and report meaningful outcomes.
Keywords
The transition out of school marks a pivotal period of change and development for all young people. As they leave the structured school environment, they navigate less defined pathways to adulthood, including postsecondary education, vocational training, employment, or community living. This transition typically occurs between the ages of 16 and 25 and overlaps with “emerging adulthood,” a distinct developmental period characterized by instability, with experiences shaped by social and structural contexts rather than age alone (Arnett, 2000, 2024; Nelson, 2020).
For young people who require additional support, the transition introduces further challenges. Autistic young people may experience unique and compounded demands, as the expectation of increased independence coupled with a reduction in structured support may result in complexities around coping with change, navigating social differences, developing practical life skills, and encountering stigma and discrimination (Crompton & Bond, 2022; Liptak et al., 2011; Toor et al., 2016). This abrupt shift in support has been likened to falling off a “services cliff” (Roux et al., 2015). It has been shown to impact immediate psychological well-being (Cooper et al., 2024) and longer-term outcomes, including problems sustaining employment, ongoing reliance on family support, prolonged disengagement from health and care services, and elevated mental health difficulties (Roux et al., 2023; Wehman et al., 2014).
Although there is debate over what constitutes “good” adult outcomes for autistic young people (Lounds Taylor, 2017; Pellicano & Heyworth, 2023), research has typically described outcomes using objective indicators, such as participation in education or employment, relationships, independent living, and mental health (Howlin & Magiati, 2017). These indicators are often reported as poorer for autistic adults compared with same-age peers (Howlin & Magiati, 2017), although they do not necessarily capture autistic people’s own priorities or life satisfaction. Further, they reflect disparities in access to resources, opportunities, and supports, shaped by systemic and environmental barriers (e.g., workplace discrimination and inflexible employment practices) rather than individual impairment (Dreaver et al., 2020; Scott et al., 2019). These inequities are often amplified for the approximately one-third of autistic young people who have co-occurring intellectual disability (ID) (Roman-Urrestarazu & van Kessel, 2022; Zeidan et al., 2022), for whom such indicators may be unrealistic or misaligned with individual priorities (De Paor et al., 2025; Young-Southward et al., 2017).
In this review, the transition out of school is conceptualized as holistic, relational, and context-dependent, rather than a discrete event. Holistically, transition spans key areas, including education, health, and care. Relationally, experiences are shaped through interactions between young people, families, and professionals. Contextually, interactions are embedded within wider service, policy, and societal conditions. This conceptualization aligns with socioecological accounts of transition (Crompton & Bond, 2022; Kranzler et al., 2020), grounded in ecological systems theory (Bronfenbrenner, 1979), situating interactions at a micro level (individual and family), meso level (services), and macro level (policy and broader societal structures).
This conceptual framework is interpreted through a human rights-based lens grounded in the UN Convention on the Rights of Persons with Disabilities (CRPD; United Nations, 2006). Under this lens, disabled young people and their families are rights holders, and services and institutions (e.g., schools and colleges, local government authorities, health and social care, employers) are duty bearers with legal obligations to remove barriers and enable participation (Bayefsky, 2000).
In practice, these rights and duties are enacted through formal transition planning. Transition responsibilities, eligibility criteria, and planning processes vary internationally. In some countries, school-based transition planning may begin as early as age 12 for those with the highest levels of need and is delivered through established planning processes—for example, the US Individualized Education Program (IEP); England’s Education, Health and Care Plans (EHCPs); and Australia’s requirements for reasonable adjustments and equitable access. Despite these frameworks, many young people, including those who are autistic, face gaps in provision, underpinned by variation in planning quality and timing, eligibility thresholds, support access, and discontinuities between child and adult systems (Roux et al., 2023). These gaps occur between settings (school to post-school) and between agencies responsible for planning and support.
Against this backdrop, policy and services not only shape practice but also influence the research questions investigated. Consequently, the literature is weighted toward transition intervention effectiveness and outcome measurement (Roux et al., 2023); areas misaligned with the priorities of the autistic community, which emphasize well-being, participation, and self-determination (Nicolaidis et al., 2025; Pellicano et al., 2014b). Alongside this imbalance in focus, representation is uneven. Autism-focused studies frequently exclude young people with ID (Russell et al., 2019), whereas researchers focusing on young people with ID do not always report findings separately for autistic participants (Foley et al., 2012; Jacobs et al., 2020). Proxy reports are not equivalent to first-person accounts (Krämer et al., 2021), raising questions about whose perspectives are represented in the evidence and whose are missing (Akl et al., 2024).
The growing body of transition-related primary research for autistic young people has been accompanied by numerous systematic reviews. However, these systematic reviews exist in silos, focusing on discrete diagnostic groups such as autism (e.g., Chancel et al., 2022) and ID (e.g., Young-Southward et al., 2017), or domains where the transition is most visible, such as education (e.g., Chandroo et al., 2018) or health (e.g., Levy et al., 2020). Umbrella reviews offer a methodological approach to synthesize findings systematically from multiple reviews and meta-analyses, aggregating and summarizing evidence from diverse sources (Aromataris et al., 2015). An umbrella review of transition research significantly advances the field through consolidation of fragmented evidence, advancing cross-domain understanding and generating a foundation for evidence-informed recommendations for future practice, research, and policy.
Only one prior umbrella review has addressed a related area. Tincani et al. (2023) synthesized evidence on vocational interventions for autistic adolescents and adults across 31 systematic reviews and meta-analyses covering 287 primary intervention studies. Although not framed around the transition out of school, the research questions focused on the vocational intervention evidence base (targeted skills and outcomes, settings, service delivery types, effectiveness, and populations), and the majority of reviews included primary studies focused on transition-aged young people. This review highlighted that vocational evidence is dominated by interventions targeting discrete job-task skills, with comparatively limited attention to broader employment outcomes (e.g., job attainment, wages, hours worked, benefits, job satisfaction) and limited reporting of demographic characteristics across primary studies.
The current umbrella review extends the only existing synthesis in four key ways. First, it considers the transition as a holistic, relational, and context-bound process rather than a discrete event. Second, it broadens the scope by reviewing literature across education, health, and social care domains. Third, it includes the perspectives of all those involved in the transition process to consider who is heard, how participation is enacted, and where perspectives are absent. Fourth, it aims to capture research, including or meaningfully describing the transition process for autistic young people with ID, to reflect the diversity of autistic young people’s experiences.
This umbrella review is timely given the rapid expansion of transition-related research, evolving policy landscapes, and the growing emphasis on the inclusion of all voices. By synthesizing fragmented evidence across disciplines and centering the perspectives of all those involved in the transition process, this review addresses current research gaps and community priorities. We addressed the following research questions:
What are the experiences and outcomes of the transition out of school for autistic young people, with or without ID, from the perspectives of young people themselves and those supporting the transition process?
Where are the gaps in the existing research?
Method
We conducted the review according to the methodological guidance outlined in the Joanna Briggs Institute manual for evidence synthesis (Aromataris et al., 2024), with reporting aligned to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Page et al., 2021) and the Preferred Reporting Items for Overviews of Reviews (PRIOR) statement (Gates et al., 2022). The protocol was pre-registered on PROSPERO (CRD42024519301). Prespecified elements included the databases, date limits, eligibility criteria, screening procedures, quality appraisal using the JBI checklist, and the integrated narrative synthesis plan. An audit trail was maintained, and periodic checks showed no deviations from the protocol.
Positionality Statement
Our research team comprised five authors with diverse academic, lived, family, and professional experiences. Our team included academics at different career stages and a range of professional and disciplinary backgrounds (including education, psychology, and public health). Of relevance to this review, our team included autistic and non-autistic contributors. We foreground autism related positionality because this review examined autism-specific transition research and the related power dynamics in knowledge production. Although the team did not include an autistic contributor with ID, which we recognize as a limitation, several authors have professional links with autistic young people with ID. Additionally, we brought culturally and geographically diverse perspectives, which informed decisions at each stage (Krämer et al., 2021). We recognize that other identity dimensions shape perspectives in disability research and may not be fully represented in this work.
We aimed for transparency, making assumptions explicit and distinguishing delimitations (intentional scope choices) from limitations (constraints we could not control) (Alexander, 2020). This approach aligns with Cascio et al. (2019), who advocate for team-based open coding to strengthen intercoder consensus. It also reflects the participatory ethos advocated by Tan et al. (2024), in which authentic involvement of autistic researchers—a traditionally rare occurrence in autism research (Pellicano et al., 2014a)—improved research transparency and challenged power dynamics. We include a GRIPP2 short form checklist (Guidance for Reporting Involvement of Patients and the Public; Staniszewska et al., 2017) to support transparent reporting of community involvement (see Supplemental Table S1 in the online version of the journal).
Although this umbrella review used systematic methods to enhance transparency and reproducibility, we did not treat evidence as neutral or value-free. We located the synthesis within an interpretive, context-sensitive orientation, recognizing that transition outcomes are socially produced and shaped by relationships, environments, and policy structures. Consistent with the social model of disability (Oliver, 2013; Oliver & Barnes, 2012) and a neurodiversity-affirmative approach (Kapp, 2020), we interpreted findings as reflecting both individual experiences and systemic conditions. We treated gaps in autistic representation and voice as analytically meaningful rather than limitations. The synthesis offers an interpretative account of review-level evidence rather than as an estimate of effect or an objective account of transition.
Search Strategy
Our scope intentionally spanned education, health, and social care domains to reflect transition as a system-level process rather than a school-bound event. Two key decisions explicitly guided the breadth of the search strategy. First, given the recent umbrella review by Tincani et al. (2023), we excluded vocational skills interventions to avoid duplication. Second, a review on transition conducted within the field of autism research alone would likely miss autistic participants with ID (see Russell et al., 2019). To track who is represented and how (Alvarez, 2020; Krämer et al., 2021), we extended the scope of the search to the field of ID research to include autistic participants who are subsumed within broader samples of those with ID. We accepted that data relating solely to autistic participants may not be fully extracted.
We detail the data sources searched, date limits, inclusion and exclusion criteria, screening and appraisal procedures, and the approach to synthesis. A database coverage map (see Supplemental Table S2 in the online version of the journal), full search strings (see Supplemental Table S3 in the online version of the journal), decision trees used at title/abstract and full-text screening (see Supplemental Table S4 in the online version of the journal), and a list of excluded reviews with reasons for exclusion (see Supplemental Table S5 in the online version of the journal) are provided in the supplemental materials.
A subject-specialist (education) librarian guided database selection and peer-reviewed the search strategies. A literature search was conducted on March 7, 2024, using the following databases: Cochrane Database of Systematic Reviews, the EPPI Center Evidence Library, JBI Evidence Synthesis, the PROSPERO register, Education Resources Information Center (ERIC), MEDLINE (via Ovid), Web of Science, Scopus, PsycINFO (via Ovid), Social Services Abstracts, CINAHL, ASSIA, Elsevier Reference Collection in Neuroscience and Biobehavioral Psychology, and [end]. We searched the gray literature (theses and reports) to promote citational justice, widen representation beyond selective journals, and capture community-led or participatory work (Alvarez, 2020), using the following sources: Kings Fund Library, PsycExtra, Overton and ProQuest Dissertations, and Theses Global. Database searches used the terms contained in Table 1, combined with “AND.” Reference lists of eligible studies were searched between May 5 and May 23, 2024, with six additional references identified and five further studies included in this review. We judged additional hand searches unlikely to materially alter the included set given multi-database coverage, librarian input, and citation chasing. Searches were rerun on April 1 and April 2, 2025. No additional reviews were identified.
Search terms
Eligibility Criteria
The eligibility criteria reflected the scope of the review and were necessarily broad to capture the evidence addressing the transition out of school for autistic young people (with or without ID), ensuring the inclusion of reviews that captured: diverse transition experiences and outcomes; “how” or “why” interventions do or do not work; studies from different countries, including when and how transitions occur (which differ between, and sometimes within, countries); and research from different points in time, thus reflecting changing policies. Table 2 summarizes the inclusion and exclusion parameters.
Inclusion and exclusion criteria: Summary of eligibility criteria applied at title/abstract and full-text screening stages
Reviews were included if they examined the experiences or outcomes associated with the transition out of school for autistic young people, including those with ID, or reviews of young people with ID where autistic participants were separately reported. Eligible reviews focused on school-leaving or transitions into post-school life, including transition planning, support, and service coordination across education, health, or social care domains. Reviews that met inclusion criteria but yielded no autism primary studies (e.g., Watson et al., 2011) were retained to document evidence gaps but did not contribute findings.
Reviews were excluded if they focused solely on vocational or employment interventions, if the autistic or ID populations could not be distinguished, or where the transition context differed from the education-to-adulthood focus of this review. Reviews were excluded if they focused on outcomes framed around the transition but were not expressly focused on supporting the transition as a process. This included reviews of discrete skills interventions (e.g., daily living skills programs, Hong et al., 2017). Reviews that focused on post-transition support only were excluded, even if they included transition-aged young people (e.g., college-based reviews; Flegenheimer & Scherf, 2022). Transitions focused on health services only (e.g., Child and Adolescent Mental Health Services to Adult Mental Health Services) were excluded unless explicitly linked to leaving school or college. Lastly, reviews framed around adolescence in general were excluded (e.g., Cooper et al., 2024).
Study Selection
After initial conceptualization by Richards, Pellicano, and Crane, Han and Spiegler joined the team and contributed to study selection, coding, and interpretation. To begin, one researcher (Richards) independently reviewed the titles and abstracts of all articles to determine suitability for inclusion in the review using Covidence systematic review software (Veritas Health Innovation, 2024). Two researchers (Han and Spiegler) independently reviewed 50% each, ensuring all reviews were screened by two researchers. An article proceeded to full-text review if the title or abstract appeared to meet the criteria or if more information was needed. If the title/abstract of an article met any exclusion criterion, the article was excluded. In cases of disagreement, consensus was reached by discussion between two researchers (Richards and either Han or Spiegler). Next, at full-text review, one researcher (Richards) independently screened 100% of articles and two researchers (Han and Spiegler) each independently screened 50% of articles to enable consensus on whether articles met inclusion/exclusion criteria. Consensus was reached through discussion between two researchers (Richards and either Han or Spiegler) for all disagreements.
Cohen’s Kappa for title and abstract screening was .59 (moderate agreement). Raters reported that the initial screening was difficult, given the heterogeneity of review types, variability in vocabulary, and uncertainty about whether autistic participants were or could be involved. Titles and abstracts often inferred neurodivergent populations; at this stage, records were provisionally included to avoid false negatives and flagged for verification. This conservative stance produced more “not sure” decisions, resulting in moderate agreement. At full-text screening, supported by discussions with Pellicano and Crane, richer details clarified the study samples and enabled more precise application of criteria. Cohen’s Kappa reached .89 (almost perfect agreement). One hundred percent consensus was reached before proceeding to data extraction.
Data Extraction
Data were extracted by one researcher (Richards) for all articles (100%) using the Covidence software. To balance thoroughness with resource constraints, and following practices discussed in the systematic review literature—where partial independent data extraction supplemented by audits is considered acceptable (Mathes et al., 2017)—two researchers (Han and Spiegler) independently extracted data from a randomly selected subset of 24% of reviews (three reviews, 12% each). The remaining 76% were audited by one of two researchers (Han and Spiegler). Where data were incomplete or missing, consensus was reached by discussion. Where reported, results were extracted separately for autistic young people with and without ID. When multiple publications reported overlapping data from the same review or dataset, the most complete or most recent version was retained.
Data were extracted on the characteristics of the eligible systematic reviews and included: (a) review authors; (b) year of publication; (c) type of review; (d) publication type; (e) country of review; (f) context of review; (g) aims or objectives of the review; (h) search details; (i) total number of primary studies and participants reported; (j) aspect of the transition period; (k) appraisal tool and primary study quality; (l) method of synthesis/analysis; (m) summary of the main findings of the review; (n) community involvement in the review. Data relating to participants included in the primary studies (autistic young people and/or any other person supporting the transition process) were also extracted: (o) autistic participants’ age; (p) participants’ gender; (q) participants’ ethnicity; (r) autistic participants’ diagnosis(-es); and (s) any other reported participant information. In cases where a review included data from a doctoral thesis that was subsequently published, only data from the peer-reviewed journal article were included.
Reviews were coded as including autistic participants with ID where this was explicitly reported. Inclusion was not inferred from nonspecific categories. Where reporting was ambiguous, coding was recorded as “unclear.” Decisions were based on review-level descriptions and, where needed, were cross-checked against primary studies.
Quality Appraisal
Appraisal judgments were made at the review level (not at the level of primary studies), and ratings were weighted in the resulting synthesis. Uncertainty was indicated where a review showed methodological concerns, for example, regarding search strategies or reporting. Two researchers (Richards and either Han or Spiegler) independently assessed the reviews using an adapted version of the Critical Appraisal Checklist for Systematic Reviews and Research Syntheses (CACSRRS; Aromataris et al., 2024). The checklist considered 11 items to determine the quality of the following: (a) review question, (b) inclusion criteria, (c) search strategy, (d) sources/resources, (e) study appraisal criteria, (f) agreement between raters, (g) methods to minimize error in extraction, (h) methods to combine studies, (i) likelihood of publication bias (only for quantitative reviews and meta-analyses), (j) recommendations for policy and/or practice, and (k) directives for new research. Qualitative and scoping reviews were scored on 10 items rather than 11 because the primary risk posed by publication bias (i.e., providing a skewed view of the effectiveness of an intervention) was not applicable and was therefore excluded from the denominator.
Before independent assessment, a calibration exercise using three varied reviews aligned interpretations of positive, negative, or unclear appraisal. In calibration, agreement was reached that a replicable, multi-database search using subject headings and keywords satisfied the search strategy item; explicit independent critical appraisal (or documented consensus following dual review) satisfied the appraisal process item; and a clearly stated synthesis method appropriate to heterogeneity (for example, meta-analysis with suitable modeling or a transparent narrative approach) satisfied the methods to combine studies item. Once agreement in interpretation was established, the appraisal process was straightforward, with few disagreements, and none required escalation to a third team member. Each item was rated “yes” (to indicate low risk of bias), “no,” or “unclear” (to indicate high risk of bias). Following JBI guidance for umbrella reviews (Aromataris et al., 2024), a summary indicative percentage of “yes” scores was calculated for each review and converted to a rating (high > 80%, moderate 60–79%, and low <60%). Interpretation was based on review quality. Findings from low-quality reviews were included for completeness, but did not underpin conclusions. Each review’s appraisal of its included primary studies (tool used and summary judgement) was extracted (see Supplemental Table S7 in the online version of the journal). These appraisals were not re-rated here, but were considered in the narrative synthesis to indicate confidence in the underlying evidence.
Overlapping Primary Studies
One limitation in any umbrella review is double-counting or giving too much weight to the results of primary studies subsumed in one or more systematic reviews. Currently, there is no standardized methodological approach to visualize, quantify, and resolve overlap in primary studies across reviews. Due to the broad review question, there was an expectation that there would be many reviews captured by the search; the degree of overlap was therefore not considered a criterion for eligibility at the screening stages. To ensure no data were unnecessarily discounted, data were extracted from all eligible reviews. The risk of bias in reporting individual review results was mitigated through agreement on reported, synthesized findings among all authors; therefore, in this review, overlap was not managed until the synthesis and summary stage.
Available methods to account for overlap include mathematically smoothing results according to sample size and calculating the corrected covered area (Pieper et al., 2014). However, in this review, the heterogeneity of the review types, the wide range of sample sizes, and the prioritization of qualitative syntheses would undermine interpretability; therefore, it was not a useful calculation (Lunny et al., 2021). Instead, we displayed the overlap visually. A citation matrix was created, indicating which primary studies were included in which systematic reviews, using binary indicators (1 = included, 0 = not included) to reflect overlap.
Data Synthesis
Following the JBI Reviewer’s Manual for Umbrella Reviews (Aromataris et al., 2024), review-level findings were synthesized and primary studies were not reanalyzed. Given heterogeneity in populations (with or without ID), settings (for example, education or health), primary study designs, and review methods, a meta-analysis was neither feasible nor meaningful (Popay et al., 2006). Instead, we undertook a single integrated narrative synthesis (Popay et al., 2006; Wilson & Anagnostopoulos, 2021). Findings and contextual details were extracted as reported from each review; domain-specific summaries were prepared; and cross-domain convergences and divergences were identified. Where numbers were reported, simple descriptive summaries (direction of effect, counts, or proportions) were interpreted with equity and context in mind (Tabron & Thomas, 2023). An equity-sensitive reading was applied: subgroup findings (e.g., by race, ethnicity, gender, or socioeconomic status) were included where available, and no inferences were drawn where unreported. One reviewer (Richards) conducted the synthesis inductively, with coauthors discussing the work as it progressed. Interpretation was weighted by review quality, and conclusions were checked with lower-quality reviews excluded.
Results
This umbrella review synthesized findings from 25 reviews examining the transition out of school for autistic young people, with or without ID, across education, health, and social domains. The reviews varied in focus, method, and quality, and drew on partially overlapping primary studies. Results are presented in three stages: first, review characteristics and quality; second, the scope and distribution of the evidence; and third, an integrated narrative synthesis of experiential and outcomes findings across four themes.
Study Selection
Figure 1 shows the study selection results. Database searches yielded 4,284 records; after removing 2,380 duplicate references, 1,904 records were screened for title and abstract. In total, 1,754 records were excluded, leaving 149 studies for full-text screening. Of these, 124 were excluded due to study design, population, or focus; references are provided (see Table S5). Six additional reviews were identified through citation searches, resulting in 25 included reviews. One of the 25 was originally identified as a doctoral dissertation but later published as a review article; only the published version was extracted and counted (Kirby et al., 2016).
Study selection results (PRISMA; Page et al., 2021).
We excluded reviews focusing solely on vocational skills to avoid duplication (see Methods). Two systematic reviews overlapped with Tincani et al. (2023). Both reviews, Chancel et al. (2022) and Chandroo et al. (2018), were retained, as they explicitly framed the transition as a process. Chandroo et al. (2018) focus on autistic students’ involvement in IEP transition planning, including participation and voice within the transition process. Chancel et al. (2022) synthesized group-based educational interventions explicitly positioned within the transition to adulthood for autistic adolescents and young adults. As such, these reviews were more directly relevant to our transition-focused aims than Tincani et al. (2023), which is framed around vocational intervention evidence. Conclusions about vocational intervention effectiveness from these reviews were not taken forward into our synthesis.
Review Quality
Table 3 summarizes review quality appraisal (for further detail, see Supplemental Table S6 in the online version of the journal). Indicative appraisal percentages ranged from 20%–100%: 16 reviews were rated high (>80%), three moderate (60–79%), and six low (<60%). Quantitative reviews and meta-analysis ranked highest (82%–91%), reflecting their structured approach and methodology, predominantly losing points for limited gray literature or reference searching. Qualitative and mixed-methods reviews showed a wider spread (73%–100%). Scoping reviews were rated the lowest overall in terms of quality (ranging from 20%–80%) because few included quality appraisals, although this is not a standard requirement of scoping review methodology. Although lower-quality scores are noted, and appropriate weighting is given in the synthesis of results, scoping reviews contributed useful mapping and future research pointers.
Summary of quality appraisal of reviews
Note. High (≥80%): 16 reviews; moderate (60–79%): 3 reviews; low (<60%): 6 reviews. Mean % Yes: High = 88%; moderate = 69%; low = 37%.
Across the reviews, limitations in underlying studies were frequently identified (small samples, heterogeneity, risk of bias, and rare adverse-event reporting). Findings were tempered within the reviews themselves. Details of primary study quality appraisal are included in (see Supplemental Table S7 in the online version of the journal). In this review, those review-level judgments (both the quality appraisals and the cautions attached to findings) have been carried forward and used to weight the resulting synthesis.
Overlapping Primary Studies
Table 4 summarizes overlap among the 25 included reviews. Together they cited 790 studies, of which 500 citations were relevant to leaving education and mapped to 435 unique primary studies extracted in this review. Of these 435 primary studies, 384 studies (88%) were included once, and 51 studies (12%) more than once. The most frequently cited, Ruble et al. (2018), was included in five intervention-focused reviews (Bottema-Beutel et al., 2022; Hothi et al., 2022; Kwan et al., 2021; Levy et al., 2020; White et al., 2024). This study was a US-based randomized controlled trial (RCT) evaluating the COMPASS intervention, which promotes self-determination through active collaboration in planning. Outcomes were assessed using Goal Attainment Scaling, a predefined scale that is sensitive to change but captures proximal rather than long-term outcomes. Two studies appeared in four intervention-focused reviews. First, Hagner et al. (2012) evaluated a family-centered intervention supporting transition, which was included in Bottema-Beutel et al. (2022), Chandroo et al. (2018), Kwan et al. (2021), and Levy et al. (2020). Second, Nadig et al. (2018) reported an RCT focusing on self-determination and quality of life, which was included in Chancel et al. (2022), Hothi et al. (2022), Kwan et al. (2021), and Levy et al. (2020). One review, by Reale and Bonati (2015), had one relevant study for extraction, Cheak-Zamora et al. (2013). This study was in two other reviews: K. A. Anderson, Roux, et al. (2018) and Eilenberg et al. (2019).
Overlapping primary studies included within reviews
Overall, overlap was limited, but notable: reliance on a few well-known US trials of school-based programs means some positive effects reflect repeated citation rather than independent replication.
Review Context: Domains, Locations, and Participant Representation
To situate findings, the focus, geographical origin, and participant representation of the 25 reviews are presented in Table 5. Findings were included in full where all participants fit the inclusion criteria or partially extracted where autistic young people were subsumed within the population, and results could be extracted separately.
Characteristics of included reviews
Note. Aims, conclusions, and findings are reproduced verbatim to preserve author intent and minimize interpretative bias.
Domain: Education, Health, Socioecological
Most reviews focused on education (14 of 25, 56%), of which 10 were eligible for full extraction (71%), and four were partially extracted for findings relating to autistic young people only (Table 5). Nine reviews considered health transitions between child and adult mental health services (36%), with results relating to autistic young people being partially extracted. Two reviews focused on socioecological factors (8%). Despite a wide database search, no social care transition reviews were identified, highlighting an evidence gap. One review, Watson et al. (2011), met the inclusion criteria, but contained no autism primary studies, highlighting the absence of research at that time.
While most reviews were recent (post-2014), the underlying primary studies spanned from 1984 to 2023. This long timespan introduces policy lag, meaning older data may not reflect current systems.
Location
Figure 2 depicts the reviews by the originating country and review type. Most reviews were conducted in the United States (two quantitative, two qualitative, two mixed-methods, and three scoping reviews), followed by Canada (three quantitative, one qualitative, one meta-ethnography, and one scoping review). The UK conducted no quantitative reviews (three qualitative, one mixed-methods, and one scoping review), whereas Australia conducted one quantitative review, one qualitative synthesis, and one mixed-methods review. Two further reviews were captured from within Europe, but no reviews or meta-analyses were found originating elsewhere. Table 5 shows the origin of the primary studies included within these reviews. While there are primary studies originating from the rest of Europe and one in China (White et al., 2024), no systematic reviews from other regions were found. This Global North concentration constrains generalizability to settings with different policy architectures, resources, and services.
Origin and type of review.
Demographic Information
Participant reporting at the review level was limited and skewed. Detailed demographics in the underlying primary studies were not consistent; a summary of participant information, as reported in the included reviews, is provided (see Supplemental Table S8 in the online version of the journal). The gender of the autistic young person was reported in 17 of 25 reviews (68%), and all of those (100%) described predominantly male samples. Ethnicity was reported in 8 of 25 reviews (32%), all indicating predominantly White participant samples. The role of parent respondents was reported in 6 of 25 reviews (24%) with mothers most frequently represented (5 of 6 reviews, 83%). Socioeconomic indicators were reported in 6 of 25 reviews (24%), with middle- or higher-income skew described in 3 of 6 reviews (50%). Professionals’ characteristics were infrequently represented, and where reported, educators were often female and held a university degree or higher. Six reviews (24%) reported participants with co-occurring intellectual disability (ID). Only two reviews (8%) identified autistic young people with ID as a defined subgroup. No reviews reported disaggregated findings for autistic young people with ID. This pattern suggests sampling and publication biases in the available literature; therefore, the evidence does not support intersectional inferences.
Adverse Events
Only one review, Bottema-Beutel et al. (2022), systematically assessed adverse events, or the unintended negative effects of an intervention. They found reports of adverse events in just two out of 193 primary studies, highlighting this omission as a fundamental flaw. Levy et al. (2020) planned an analysis of adverse events, but rare reporting meant no synthesis was possible. No other reviews reported adverse events.
Overall, the evidence is skewed toward education- and health-based systems in high-income countries, with underrepresentation of autistic young people with ID and minoritized backgrounds, and minimal adverse-event reporting. This bias shapes what can be concluded about transition experiences and outcomes. The integrated narrative synthesis that follows interprets findings within these constraints.
Review Findings
Across the 25 reviews, evidence clustered into two strands: experiential research capturing what transition feels like for autistic young people, families, and practitioners; and outcome-focused studies, measuring indicators of transition (progress, participation, and well-being) and/or factors that correlate with successful transitions. We distinguished between proximal outcomes, changes that are observable during or soon after support (e.g., transition readiness, self-advocacy, adaptive skills), and distal outcomes, which reflect later life circumstances (e.g., sustained participation in education or employment, interdependent or independent living, community participation, well-being, and quality of life). These strands are interwoven. Together, they illustrate how systems and relationships shape what the transition offers, and for whom. Table 5 summarizes the findings by review. We identified four nested themes from the synthesis, shown diagrammatically in Figure 3: (1) voice and choice shape transition pathways; (2) relationships and trust enable progress; (3) joined-up systems create continuity; and (4) access and advantage shape who benefits.
Thematic map.
Theme 1: Voice and Choice Shape Transition Pathways
This central theme privileges the young person’s voice and illustrates how authentic early involvement aligns planning with what matters and builds agency and confidence. The two subthemes reflect parallel perspectives of young people and parents.
Subtheme 1: Autistic Young People’s Voices Need to Be Central to the Transition Process
Experiential findings were reported in 11 reviews (44%), with 10 (40%) analyzing autistic young people’s own accounts, typically through interview, focus group, or self-report. Autistic first-person perspectives were absent in 15 reviews (60%), with proxy, practitioner, or dataset-based perspectives predominating. The underlying primary studies were mostly qualitative (with some secondary data review in healthcare studies), using longitudinal and cross-sectional designs. Counts reflect review-level coding of autism-specific extracted findings and therefore signal presence, not prevalence. Most reviews did not disaggregate results, so fine- grained quantification was not possible. Representation and primary study characteristics are provided in the supplemental materials (see Suppplemental Tables S8–10 in the online version of the journal).
What matters is not only including young people’s accounts, but also how directly they shaped the analysis. In most reviews where autistic young people’s views are included, they are mediated through authors’ interpretations or combined with other evidence. For example, a meta-ethnography of peer-mentoring programs by Nguyen et al. (2020) translated first-order accounts (young people’s descriptions of what mentoring felt like and what they found helpful) and second-order author interpretations (the primary authors’ explanations of why mentoring worked), into third-order themes that synthesized shared mechanisms across studies (e.g., support for connection, confidence, and navigating new contexts). A framework synthesis by Crompton and Bond (2022) kept first-person accounts analytically distinct from proxy perspectives (e.g., young people describing reliance on family support alongside parents describing advocacy and caregiving burden). Other reviews took a broader approach, presenting summary themes (K. A. Anderson, Sosnowy, et al., 2018; Broad et al., 2017; Lee et al., 2014; Toor et al., 2016; White et al., 2024) or integrating testimony alongside proxy or quantitative evidence (Chandroo et al., 2018; Nuske et al., 2019).
Although autistic young people’s accounts were often mediated through review-level interpretation, consistent priorities still emerged across the available first-person evidence. Overall, these studies indicated autistic young people valued early, meaningful involvement in decisions; accessible information that made planning understandable; and practical, supported opportunities to influence decisions (Chandroo et al., 2018; Lee et al., 2014; Nuske et al., 2019; Toor et al., 2016; White et al., 2024). In educational contexts, students wanted to be included as equal members of planning meetings rather than passive observers (K. A. Anderson, Sosnowy, et al., 2018; Chandroo et al., 2018; White et al., 2024). Regarding the transition to higher education, young people described wanting predictable environments, staff who understood autism, and support for disclosure and self-advocacy so they could access adjustments confidently (Crompton & Bond, 2022; Nuske et al., 2019). They valued practical help around navigating new environments, managing workload, and social expectations, understanding what would change (Toor et al., 2016), and who to turn to for support when problems arose (Crompton & Bond, 2022). In healthcare transitions, similar experiences were noted: young people valued planned-supported handovers, named coordinators, and clear information around how care would continue (J. K. Anderson et al., 2022; Broad et al., 2017).
However, nearly all reviews concluded that opportunities for genuine influence were rare, and participation was tokenistic. For example, young people reported feeling sidelined in education meetings, which felt bureaucratic, inaccessible, and dominated by paperwork and professional language (K. A. Anderson, Sosnowy, et al., 2018; Chandroo et al., 2018; White et al., 2024). Current conclusions reflect young people who are most able to self-report. Two reviews (one health, one education) explicitly included participants with ID (Broad et al., 2017; Nguyen et al., 2020); however, they did not disaggregate findings. Two other education reviews explicitly excluded this group (K. A. Anderson, Sosnowy, et al., 2018; Nuske et al., 2019).
Taken together, the autistic voice is present, but not predominant, in the review-level evidence. This constitutes a central evidence gap.
Subtheme 2: Parents Are Essential Partners
In parallel, parents were described as essential partners in planning: families sought to keep plans focused on what mattered to their child while scaffolding and promoting autonomy (Crompton & Bond, 2022; Toor et al., 2016; White et al., 2024). Yet they often lacked support to shift from decision-maker to supporter, uncertain about when and how to step back (K. A. Anderson, Sosnowy, et al., 2018; Lee et al., 2014). In practice, parents frequently carried coordination work across services when formal processes faltered (K. A. Anderson, Sosnowy, et al., 2018; Chandroo et al., 2018; Lee et al., 2014).
Outcome findings aligned with experiences: where there was student-led planning or individualized appointments, reviews reported proximal improvements in self-advocacy, decision-making, and transition readiness, suggesting a young person’s role can shift from presence to influence (Chandroo et al., 2018; Levy et al., 2020). However, these benefits were short-term and context-bound with little evidence of translation to the broader life outcomes valued by young people and their families, such as increased wayfinding around new environments (Lounds Taylor et al., 2012), social interactions (Hothi et al., 2022), well-being, and quality of life (Levy et al., 2020).
Theme 2: Relationships and Trust Enable Progress
This theme highlights that no single intervention or technique reliably improves outcomes: more effective approaches combine person-centered skill building with relational, context-sensitive support.
Subtheme 1: Relationally Anchored Programs Show Proximal Gains, Not Generalizable Outcomes
Young people reported that they valued a support network that was knowledgeable about autism and knew them well (White et al., 2024). Relational continuity, along with parents as advocates, kept plans actionable and centered on their priorities (K. A. Anderson, Sosnowy, et al., 2018; Toor et al., 2016; White et al., 2024). Formal peer-mentoring programs provide valuable support and facilitate smoother transitions through trained peers and structured sessions (K. A. Anderson, Sosnowy, et al., 2018; Chandroo et al., 2018; Nguyen et al., 2020; Toor et al., 2016; White et al., 2024; see also Hadley and Mapondera, 2023).
Intervention reviews reinforced experiences to some extent. Structured programs, for example, CBT approaches centered on emotional regulation and resilience building, performed best when combined with family-inclusive approaches (Kwan et al., 2021). The strongest positive effects were seen in multi-component programs combining skills training with therapeutic or relational elements, such as the US-based UCLA PEERS Program (Chancel et al., 2022), COMPASS program (Ruble et al., 2018) and the ACCESS program, which were presented as cost-effective, accessible, and a scalable way to deliver interventions for autistic young people without ID (Chancel et al., 2022). However, across reviews, effects were reported as modest for proximal skills and minimal for broader, distal outcomes (Chancel et al., 2022). Further, evidence overlapped substantially, so gains should be interpreted cautiously (and as US-specific).
Overall, despite person-centered rhetoric, the literature remains behaviorally framed, rather than relational in nature (Bottema-Beutel et al., 2022; Kwan et al., 2021). Most outcomes stemmed from taught programs and were measured in the same settings. The evidence was dominated by small, single-site studies with heterogeneous designs and outcome measures (Bottema-Beutel et al., 2022; Chandroo et al., 2018; Hothi et al., 2022; Kwan et al., 2021; Levy et al., 2020; Munsell & Coster, 2021; White et al., 2024) and group programs lacked assessor blinding (Chancel et al., 2022). Overall, the intervention evidence is narrow and inconsistent.
Subtheme 2: Relationships Are Fragile in Practice
High staff turnover and inconsistent relationships feed instability. School staff’s advocacy role was often unclear (K. A. Anderson, Sosnowy, et al., 2018; White et al., 2024), and gaps in knowledge and understanding were highlighted by educators and health professionals (K. A. Anderson, Sosnowy, et al., 2018; White et al., 2024). Similarly, many parents reported knowledge gaps, including how to navigate planning, support systems, and available options (Lee et al., 2014).
Theme 3: Joined-up Systems Create Continuity
This theme highlights that transitions were reported to be most effective when services collaborate and relationships continue across boundaries; when they are not, support collapses. Across experiential reviews, continuity depended on planned handovers, collaborative planning, and consistent communication, supported by stronger appraised evidence (J. K. Anderson et al., 2022; K. A. Anderson, Sosnowy, et al., 2018; Crompton & Bond, 2022; Eilenberg et al., 2019). Few reviews considered how service coordination works in practice. Transitional Care Interventions (TCIs) offered one example of joined-up working: individualized health transition plans and guided appointments were associated with moderate-to-strong improvements in short-term transition readiness and self-advocacy.
Most reviews highlighted fragile or absent mechanisms. For example, healthcare transitions broke down due to limited transition services, reduced adult provision, and abrupt service transfer, which reduced service use at transition age and connection to adult care (J. K. Anderson et al., 2022; K. A. Anderson, Sosnowy, et al., 2018; Eilenberg et al., 2019). Within education, siloed responsibility between school and post-school systems led to inconsistency and loss of momentum, leaving families bridging gaps (White et al., 2024). In general, the evidence on the mechanisms of coordination remains thin.
Theme 4: Access and Advantage Shape Who Benefits
This theme illustrates that transition outcomes reflect not only individual readiness but also differential access to social, economic, and institutional resources. The uneven evidence reflects and reinforces inequity in practice, particularly for autistic young people with ID and those who are multiply marginalized.
Subtheme 1: Transition Outcomes Are Patterned By Race, Gender, and Socioeconomic Status
Participant demographics, where reported, were skewed toward White, male, and middle-income (see Supplemental Table S8 in the online version of the journal). Eilenberg et al. (2019) conducted the only included review to directly analyze disparities based on race, ethnicity, and socioeconomic status. They found that both low-income and racial/ethnic minority autistic young people were less likely to participate in transition planning meetings, continue to postsecondary education, live independently, participate in social activities, and receive healthcare transition services compared to their White, higher-income peers. However, these categories often overlap within datasets, and analyses did not separate the effects of race and income.
Subtheme 2: Autistic Young People With ID Are Routinely Excluded
Autistic young people with ID were particularly underrepresented across both experiential and intervention reviews. In several cases, they were explicitly excluded (K. A. Anderson, Roux, et al., 2018; Hadley & Mapondera, 2023; Nuske et al., 2019) and, where included, findings were rarely disaggregated (Broad et al., 2017; Nguyen et al., 2020). Among the nine intervention reviews, two autism-specific reviews explicitly excluded participants with ID (Chancel et al., 2022; Munsell & Coster, 2021), and none reported outcomes separately for this group. Broader neurodevelopmental reviews also failed to disaggregate autistic young people with ID in their analyses. When inclusion was reported, their perspectives and outcomes were absent, limiting insight into the relevance and effectiveness of existing supports for this group.
Subtheme 3: Predictors Reflect Individual and Family-Level Advantage
Reviews identifying predictors of positive transitions (K. A. Anderson, Sosnowy, et al., 2018; Barr et al., 2017; Eilenberg et al., 2019; Kirby et al., 2016; White et al., 2024) consistently pointed to individual capacities, notably adaptive functioning, and parental engagement, as foundations for effective support strategies (K. A. Anderson, Roux, et al., 2018; Kirby et al., 2016; White et al., 2024). The quality of relationships and social inclusion were also considered critical (K. A. Anderson, Roux, et al., 2018; Barr et al., 2017; Eilenberg et al., 2019; Kirby et al., 2016; Toor et al., 2016).
At the family level, parent education and income were the most commonly measured socioecological correlates. Higher levels of household income, education, and involvement in decision-making generally (not always) were associated with more positive transition outcomes (K. A. Anderson, Roux, et al., 2018). Acknowledging financial pressures improved the realism of plans (Barr et al., 2017).
Overall, the strength of the evidence in this area is constrained. Across reviews, there was an overemphasis on individual factors without considering broader social contexts (K. A. Anderson, Roux, et al., 2018; Eilenberg et al., 2019). The reliance on secondary data, overlapping primary studies that relied solely on large US datasets (the National Longitudinal Transition Study-2, NLTS2; and the Rehabilitation Services Administration 911, RSA-911), limits demographic and contextual diversity. Methodological variation constrains comparability. Outcome definitions vary widely, and policy influences were rarely examined. This limits generalizability (K. A. Anderson, Roux, et al., 2018; Eilenberg et al., 2019; White et al., 2024).
Discussion
Transitions out of school are pivotal, yet evidence remains fragmented across disciplines and populations. Existing reviews have examined aspects of the transition, yet none synthesize across domains while considering autistic young people with and without ID. This umbrella review addressed this gap by synthesizing evidence from cross-disciplinary reviews and by framing transition as a relational process rather than a single event. Findings aligned with policy aspirations for early, personalized collaborative planning. Four nested themes were identified: (1) voice and choice shape transition pathways; (2) relationships and trust enable progress; (3) joined-up systems create continuity; and (4) access and advantage shape who benefits. To interpret these themes, we adopt a bioecological multilevel lens grounded in Bronfenbrenner’s ecological systems theory (Bronfenbrenner, 1979) as subsequently applied in autism (Crompton & Bond, 2022; Kranzler et al., 2020), with three levels: the macro level reflects system conditions; the meso level considers service coordination and relationships; and the micro level encompasses lived experience. Interpreted through this lens, the four themes interrelate: access and advantage operate at a macro level to shape who benefits; joined-up systems and relationships at the meso level to create continuity and stability; and these factors, in turn, enable voice and choice at the micro level to exert influence.
Macro-Level Gap: The Defined Population Versus the Actual Population
Access and advantage set the starting conditions by shaping who is reached and resourced. Research remains predominantly concentrated in the Global North, with no reviews or primary studies originating in low- or middle-income countries (LMIC). This finding maps onto jurisdictions where transition planning is statutorily required, focusing attention and resources on transition and steering the agenda of transition-based research (Roux et al., 2023). Therefore, our understanding of transition experiences is shaped by a narrow subset of the global autistic population, raising concerns about the applicability of findings across diverse sociocultural and economic contexts. Approaches developed in high-income contexts require cultural and system-level adaptation before transfer to lower-resource systems (de Leeuw et al., 2020), for example, co-designing with local communities, simplifying resource-intensive components, embedding reasonable adjustments within existing services, and setting realistic implementation standards for staffing, training, and data sharing (de Leeuw et al., 2020).
Beyond geography, the evidence rests on a narrow participant profile. Across reviews, samples were US-based, metropolitan White males, leaving significant gaps in understanding the experiences of autistic females, those identifying outside the gender binary, and those from racially or socioeconomically marginalized communities. This pattern likely reflects structural inequalities and sampling and publication biases, thus limiting generalizability. Consequently, interpretations are least certain for underrepresented groups, and intersectional conclusions are not advanced here. By basing best practices on such a narrow segment of the autistic population, transition supports fail to capture the full range of needs.
Location and inclusion combine at the macro level to shape the evidence, offering a starting point, not a template, for other settings and populations.
Meso-Level Gap: Planning “With” Young People and Planning “For” Young People
The meso level considers participation and coordination: where joined-up systems are fragile, relationships and trust do not persist across boundaries, and plans lose continuity, muting voice and choice in decisions that matter.
Across the results, advocacy (securing access, convening agencies, and keeping plans centered on the young person) was valued but rarely realized. Participation was tokenistic, coordination unclear, and families frequently carried the coordinating role. This pattern is consistent with wider analyses showing autistic self-advocates (including those with ID) are frequently marginalized within decision-making and advocacy contexts (Petri et al., 2021). Broader evidence corroborates these findings, highlighting how frustrations differ by perspective. Autistic young people feel excluded from decisions (Burke et al., 2024; Crane et al., 2022), and find navigating the transition with limited or inconsistent support overwhelming (Burke et al., 2025). Families report feeling disempowered and unsupported (Crane et al., 2023), and professionals report limited time and training (Boesley & Crane, 2018; Crane et al., 2021).
Even if advocacy were strengthened, a misalignment of aims exists. Young people seek participation while parents balance advocacy with service constraints. Without explicit effort to reconcile these aims, plans drift from what matters to autistic young people. Tyldesley-Marshall et al. (2025) bring these concerns together in a five-element model of multi-agency collaboration emphasizing legitimate participation in partnership, personalized and consultative planning, respectful communication, preparation for effective partnership roles, and adequately resourced cross-organizational working. However, while coordination frameworks exist (on paper), the culture and the required resources rarely do.
Compounding these issues, participation was uneven. For autistic young people with ID, our results showed sparse or non-disaggregated evidence. This evidence gap makes it difficult to judge whether current practices are appropriate and risks obscuring different routes to participation. A pragmatic response is to design for access by default, embedding accessible features that make involvement possible and influential: accessible information, reasonable adjustments (to timing, format, location, and sensory environment), and supports for language and communication. These principles align with calls for clearer coordination, stronger participation, and more integrated person-centered planning in ID research (Best & Burke, 2024; De Paor et al., 2025). Coproducing transition tools will center valued goals, communication styles, and specific support needs (Crompton & Bond, 2022).
Although these approaches demonstrate pockets of success, scalability is challenging. Good practice is often developed in small-scale pilot projects, with limited pathways for embedding these approaches into mainstream practice, reinforcing disparities in support. To move forward, systems need investment and policies that embed inclusive practice.
Micro-Level Gap: Measured Outcomes Versus Meaningful Outcomes
The micro level considers the young person’s immediate supports, relationships, and environment, and the outcomes closest to them. Without stable relationships and system continuity, voice and choice at the point of care become tokenistic. Gains remain proximal, patterned by access and advantage.
Outcome evidence came largely from quantitative intervention reviews and emphasized proximal outcomes (e.g., readiness, context-specific self-advocacy, goal attainment). Long-term trajectories (e.g., education or employment participation, living arrangements, and quality of life) and harms were rarely reported. Some programs showed proximal improvements, but design limits, sampling limits (including small single-site samples and overly heterogeneous samples), and the disregard of system factors constrain inference. Ethical concerns arise where interventions focus on prioritizing normative outcomes (Bottema-Beutel et al., 2022). Reliance on large pre-existing datasets masks complexity, particularly the challenges faced by autistic young people with ID (Eilenberg et al., 2019; White et al., 2024). This explains why “what works” often fails to travel: the evidence gives limited attention to joined-up systems, the relationships that scaffold change, and voice and choice are limited when defining outcomes.
Consistent with these findings, and despite the recognized impact of contextual factors, research continues to focus on individual and biological factors, overlooking the influence of social, environmental, and broader contextual elements that significantly contribute to individual variation and development (Mailick et al., 2025).
Key questions arise: what counts as success, and for whom? Since proximal gains do not necessarily translate into longer-term outcomes valued by young people, conventional success metrics (e.g., employment, independent living, or other normative milestones) require re-evaluation to reflect diverse experiences. In evaluating meaningful (as opposed to simply measurable) outcomes, this review found limited evidence. Pellicano and Heyworth (2023) proposed that research should focus on individual well-being rather than impose narrow and normative definitions of success, calling for a shift toward outcome measures that reflect lived experiences. Likewise, Roche et al. (2020) showed autistic people and families prioritize research on access to services, long-term stability, and meaningful social inclusion (see also Nicolaidis et al., 2025). This focus extends beyond transition research, with a recent systematic review considering quality of life measurement in mental health interventions for autistic adults also advocating for increased use of holistic quality of life measures, particularly emphasizing the need for autistic community involvement in the definition and validation of such measures (Timmerman et al., 2024).
Co-designed goals require sustained relational support, often beginning with parents. Findings suggested parents are seen as partners rather than as obstacles to independence. Within wider evidence, Moser et al. (2024) extend this to propose a view of adulthood as requiring both independence and interdependence: balancing support and autonomy tailored to individual desires and capabilities. When professional support is consistent, parents can step back, but the reality of fluctuating professional involvement shifts the burden back to the parents (Crane et al., 2023; Crompton & Bond, 2022).
Progress is also constrained by who sets priorities. Although there is growing momentum toward participatory, coproduced research, ensuring autistic people are actively involved in shaping research agendas from the outset (Fletcher-Watson et al., 2019; Pellicano et al., 2014a), engagement varies across countries and institutions, with many community research partnerships forming late in the process rather than from the beginning (Roche et al., 2020; Tan et al., 2024). If autistic young people, their families, and allies are not meaningfully engaged from the beginning, research risks failing to inform either policy or practice, reinforcing the disconnect between evidence and real-world impact.
Implications for Practice, Research, and Policy
This review highlights a clear agenda for future practice, research, and policy. For practice, transition planning should be undertaken with, not for, young people by co-designing goals that matter with all young people, including those with ID. Planning should be structured and accessible, linking short-term goals to explicit future goals. Trusted relationships should be maintained throughout changes in support.
For research, it is essential to identify what works, for whom, how, and why, using participatory methods from the outset so that research approaches remain relevant, accountable, and ethically grounded in the priorities of those they are intended to support. Representation across race, gender, and class, including young people with ID, is essential. Research must test across contexts, with longitudinal follow-up, reporting benefits, and harms. Continuity (collaboration and relational stability) should be examined as a mechanism for access and equity.
For policy, it is crucial to ensure existing systems work equitably in practice by guaranteeing access and participation across all groups and defining and measuring success in alignment with what matters to autistic young people and their families.
Limitations
An inherent limitation in any umbrella review is its reliance on the secondary analysis of underlying primary studies without direct re-evaluation of the original data. As a result, factors such as primary study quality, heterogeneity, and overlapping data may have influenced the findings from our review. To address this risk, all eligible reviews were critically appraised independently, and the effect of overlapping studies was considered and discussed.
We acknowledge the limitation of using the same quality appraisal protocol to consider both systematic and scoping reviews, resulting in scoping reviews scoring lower due to their methodology. Likewise, we acknowledge that our search strategies may not have been exhaustive, particularly regarding the exclusion of non-English-language studies.
A recurring limitation of the reviews themselves was the sparse or absent reporting by intersecting identities (e.g., race and ethnicity, gender, income), which constrains transferability, a limitation we emphasize in the synthesis.
Finally, we included some health transitions where autistic participants were subsumed within larger participant groups in the broader realm of transfer between health services. This inclusion refers only to those reviews capturing participants explicitly described as transitioning out of education, meaning findings related to healthcare transitions are not guaranteed to be exhaustive.
Conclusion
In this umbrella review on the transition out of school for autistic young people with or without ID, we drew together existing syntheses of primary evidence on transition experiences and outcomes. Despite the widespread recognition of the importance of transition planning, our findings indicate challenges intersect at the systemic (macro), institutional (meso), and personal (micro) levels. System-level inequities determine who is reached; fragile coordination and relationships affect how support is delivered, and narrow, individualized outcome measures limit what counts as success. The implications drawn from this review therefore extend beyond any single setting: effective transition planning requires inclusive relational and equitable approaches, co-designed with young people and their families, measured against outcomes that they value, and supported by structures that enable consistent access across systems. Strengthening these dimensions will not only improve the evidence base but also make transition policy and practice more responsive to the realities of autistic young people’s lives.
Supplemental Material
sj-docx-1-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-1-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-10-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-10-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-2-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-2-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-3-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-3-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-4-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-4-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-5-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-5-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-6-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-6-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-7-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-7-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Supplemental Material
sj-docx-9-rer-10.3102_00346543261442123 – Supplemental material for Transition Out of School for Autistic Young People: An Umbrella Review
Supplemental material, sj-docx-9-rer-10.3102_00346543261442123 for Transition Out of School for Autistic Young People: An Umbrella Review by Norah Richards, Elizabeth Pellicano, Emeline Han, Jessica Spiegler and Laura Crane in Review of Educational Research
Footnotes
Author Contributions
Richards: Conceptualization, methodology, validation, writing—original draft, writing – review and editing, visualization, project administration. Pellicano: Conceptualization, writing—review and editing, supervision. Han: Investigation, writing—review and editing. Spiegler: Investigation, writing—review and editing. Crane: Conceptualization, writing—review and editing, supervision.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Authors
NORAH RICHARDS is a doctoral student at the Centre for Research in Autism and Education (CRAE) at the University College London. Email:
ELIZABETH PELLICANO is a professor of autism research in the Department of Clinical, Educational and Health Psychology at the University College London, having recently been a professor at Macquarie University, Sydney, Australia. She is a developmental cognitive scientist and educational psychologist. Email:
EMELINE HAN is a Postdoctoral Research Fellow in the School of Psychology at the University of Birmingham and an Honorary Research Fellow in the Department of Behavioral Science and Health at the University College London. Email:
JESSICA SPIEGLER is a doctoral student in the Department of Clinical, Educational and Health Psychology at the University College London. Email:
LAURA CRANE is professor of autism studies at the University of Birmingham, where she is director of the Autism Centre for Education and Research (ACER). Email:
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
