Healthcare professionals' knowledge, experience, and training in caring for autistic adults: A systematic review

2.1. Research design

A comprehensive systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2020) for reporting of the findings (See Figure 1). The protocol of this systematic review has been registered with PROSPERO (CRD420251120773).

2.2. Search methods

Preliminary searches were conducted using PubMed by F.R. using a list of search terms to identify appropriate terminology related to healthcare access and utilisation among autistic adults. The findings from the preliminary search informed the refinement of the final search terms. A structured search was conducted by F.R. in Scopus, PsycINFO, PubMed, MEDLINE, and Web of Science for studies published from 2013 to 2026. The search strategy combined terms relating to (1) autism, (2) adults, (3) healthcare professionals, and (4) knowledge and/or experience. Search terms included autism (``ASD'', ``ASD symptoms'', ``autism spectrum disorder'', ``Aspergers'', ``neurodevelopmental disorders''), adult (``adults''), healthcare professional (``allied healthcare professionals''), knowledge (``knowledge about autism'', ``knowledge about ASD'') and experience (``experience of autism''). Boolean operators were applied, where similar terms with concepts were combined using Or, and concepts were combined using AND (Table 1). The searches were limited to full text, English-language publications. Reference lists of included studies were screened for further eligible studies. Eligibility criteria were pre-defined, and studies were included if they reported on HCPs from any discipline or clinical setting who worked with autistic adults. Discipline/healthcare settings included were ``medicine, nursing, physiotherapy, speech and language therapy and radiography''. Grey literature (e.g. theses, dissertations, reports and pre-prints) was excluded. Study designs included qualitative, quantitative or mixed-methods, peer-reviewed publications from 2013 - 2026, ensuring that the evidence reflected the updated, Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5) which was published in 2013 (American Psychiatric Association, 2013).

2.3. Study selection

Identified articles from the search were exported into Covidence (Covidence, 2023), and duplicates were removed. Dual screening was carried out, and all voting was blinded. Titles and abstracts were independently screened by F.R. and M.G. against the inclusion criteria and labelled as ``yes'', ``no'', or ``maybe''. Inter-rater agreement at the title and abstract stage was 82%. Where there were conflicts or disagreements that could not be resolved, S.K. and M.M. acted as independent reviewers to make the final screening decision. F.R. then conducted a full-text review of all included studies against the inclusion criteria which was verified by a second reviewer (M.G.). Studies that partially met the criteria were discussed with the review team, and the final decisions were reached by group consensus. Data extraction for the final sample of studies was completed using a modified data extraction table created following guidance outlined in the Covidence data extraction template (Covidence, 2023). Data extraction was conducted by F.R. and cross-checked for accuracy and completeness by M.G. Extracted data included from the studies specified the name of the author(s), publication year, country, study design, study aim, data collection methods, sample and setting, and summary of key findings (see Table 2).

2.4. Data synthesis

A narrative synthesis was conducted to summarise the key characteristics of the included studies. F.R. carried out the analysis independently to gain an understanding of the themes that emerged. A meta-analysis was not feasible due to considerable variation in the outcome measures and assessment tools used to evaluate knowledge across studies. Therefore, an adapted CanMEDS competency framework was used to guide the data synthesis and organisation of findings. This framework enabled the categorisation of results relating to HCPs' knowledge and experiences of caring for autistic adults across multiple domains of healthcare competence.

2.5. Operational definitions

For the purpose of this review, key terms were operationally defined to ensure consistency in study selection and interpretation. Knowledge: HCPs' knowledge was defined as factual and self-perceived understanding of autism, including core characteristics, associations, communication needs and co-occurring conditions relevant to healthcare. Experience: Experience was defined as HCPs' frequency of direct clinical contact and self-reported exposure in caring for autistic adults during their professional practice. Training: To allow for the systematic examination of both the existing education provision and reported gaps in education, training was conceptualised as ``training received'' and ``training needs''. Training received was defined as any autism-specific education previously undertaken, including formal and informal education. Training needs were defined as HCPs' self-identified areas where additional autism-specific education or skill development was required to support effective care. While presenting the constructs as analytically distinct, it is acknowledged that the studies selected demonstrate conceptual overlap between knowledge, experience, and training. Clinical experience can function as an informal source of knowledge acquisition and training interventions are designed to influence both knowledge and practical competence. To support an integrated interpretation of the findings, the CanMEDS competency framework was used as an interpretive lens, viewing knowledge, experience and training as inter-related components of clinical competence.

2.6. Quality review

The Mixed Methods Appraisal Tool (MMAT; Hong et al., 2018) was used to appraise the quality of all included studies (see Table 3). Quality appraisal was conducted independently by F.R., with a subset (30%) reviewed by M.G. Agreement was high, and discrepancies resolved through discussion. Across the quantitative studies, research questions were generally clearly stated and aligned with the data collected. Most studies employed appropriate measurement tools and statistical analyses relevant to their stated objectives (Alaghband-rad et al., 2023; Cashin et al., 2022; Eray & Murat, 2017; Gerçek et al., 2024; Malow et al., 2025; Micai et al., 2021; Ryan et al., 2022; Tromans et al., 2019; Shawahna et al., 2021;Watson-Grace et al., 2026; Zeleke et al., 2021). However, methodological limitations identified through quality appraisal reduce confidence in study conclusions, particularly in relation to training interventions where evidence should be interpreted as potential rather than definitive effectiveness due small sample size and inconsistent measurement tools. Several methodological limitations were identified, in particular, sample representativeness and potential response bias were frequently unclear or insufficiently reported in 21% of the studies (Alaghband-rad et al., 2023; Carlier et al., 2023; Gerçek et al., 2924; Zeleke et al., 2021). Additionally, aspects of sampling strategy were not consistently addressed in 10.54% of the studies, limiting the extent to which findings may be generalised (Alaghband-rad et al.; 2024, Zeleke et al. 2021). The mixed-methods studies included typically provided a clear rationale for the use of a mixed methods approach, although in 10.53% it was difficult to ascertain if there was adequate rationale for a mixed-methods approach (Mazurek et al. 2020; Stogiannos et al., 2021). Qualitative components were generally appropriate to the study aims, with data collection methods judged to be suitable in most cases; however, in 21% of the studies it was difficult to determine if inconstancies were addressed (Dillenburger et al., 2016; Edwards et al., 2025; Mazurek et al., 2020). One qualitative study met all MMAT criteria, demonstrating strong methodological coherence with appropriate data collection and analytical approaches clearly aligned to the study aims (Watson-Grace et al., 2026).

3.1. Knowledge

Across six pre-post intervention studies, training was consistently associated with improvements in autism-related knowledge, although the extent of improvement varied depending on whether knowledge was measured objectively or by self-report Eray & Murat, 2012; Havercamp et al., 2016; Mallow et al., 2023; Mazurek et al., 2020; Ryan et al., 2022; Zeleke et al., 2021;). Improvements were observed in both objective knowledge scores, for example, absolute increases +1.7 [18.9 to 20.6] after a 12-week programme (Malow et al., 2023) and +0.8 points (28.2 - 29.0) among interdisciplinary trainees as well as self - perceived knowledge increases such as +20.8 points (58.9 - 79.7) (Ryan et al., 2022). Collectively, these findings suggest that structured training may improve autism-related knowledge across a range of HCPs and healthcare setting; however, direct comparison across studies is limited by heterogeneity in knowledge assessments and reporting, which prevented the calculation of pooled effect estimates. In contrast, four non-intervention studies that assessed knowledge at a single time point demonstrated a consistent pattern of low-to-moderate perceived autism knowledge among HCPs, with substantial proportions rating their knowledge as poor/fair and few reporting high confidence (Alaghband-rad et al., 2023; Gerçek et al.,2024; Stogiannos et al., 2021; Zerbo et al., 2015). Where objective tools were used, such as the Knowledge about Childhood Autism among Health Workers (KCAHW), average scores suggested only moderate knowledge (Gerçek et al., 2024). Overall, the cross-sectional evidence indicates that baseline autism knowledge remains variable and frequently limited in routine HCP practice (Corden et al., 2022). These deficits are clinically consequential as limited understanding of co-occurring conditions may contribute to misattribution of mental health or medical symptoms, consistent with diagnostic overshadowing (Hamdan et al., 2024; Strömberg et al., 2022), while communication-related gaps may underpin a broader mismatch between HCPs and autistic adult patients (Malik-Soni et al., 2022; Mason et al., 2021; Nicolaidis et al., 2015). By profession, perceived knowledge tended to be higher among HCPs reporting more frequent contact with autistic adult patients (Cashin et al., 2022; Urbanowicz et al., 2020). This is evidenced by physicians and radiographers frequently reporting knowledge gaps despite clinical experience (Carlier et al., 2023; Zerbo et al., 2015), while nurses and interdisciplinary trainees commonly reported limited formal preparation and relied on informal learning (Cashin et al. 2022; Ryan et al., 2022). Overall, the knowledge evidence supports a need for training that moves beyond general awareness toward competency-based content. When interpreted through the CanMEDS framework, these findings relate primarily to gaps within the ``medical expert role'' (autism-specific clinical knowledge, including co-occurring conditions) and ``communicator role'' (knowledge to support effective communication accommodation). Additionally, the variable awareness of evolving evidence and limited indication of ongoing professional development also aligns with challenges within the ``scholar role''.

3.2. Experience

Across studies, clinical experience with autistic adult patients varied greatly by role and country. Reported experience ranged from very limited experience (7.1% among specialist intellectual disability nurses in Australia) to widespread experience in other settings (93.7% for UK radiographers) (Cashin et al., 2022; Stogiannos et al., 2021). By profession, experience was most consistently reported among radiographers and primary care physicians, reflecting routine clinical contact in imaging and general practice (Carlier et al., 2023; Stogiannos et al., 2021; Zerbo et al., 2015), whereas specialist intellectual disability nursing cohorts reported lower proportions of routine experience in some settings (Cashin et al., 2022). Across professions, greater experience was generally associated with higher perceived self-efficacy, suggesting experience may support practical adaptation of care processes in healthcare settings. By country, experience patterns differed even within the same profession. For example, radiographers reported greater experience in the UK compared with Switzerland, indicating that local service pathways (Carlier et al., 2023; Stogiannos et al., 2021). Another cross-study pattern was that limited experience itself was frequently cited as a barrier to delivering appropriate care, whereas greater experience with autistic adults was associated with higher perceived self-efficacy (Edwards et al., 2025; Urbanowicz et al., 2020). In Urbanowicz et al., (2020), HCPs working in disability or autism-specific settings reported higher perceived competence in providing care than those working outside these settings, suggesting that routine experience may shape confidence as well as skill development. Edwards et al., (2025) similarly reported that more frequent clinical contact with autistic patients predicted higher clinician self-efficacy, indicating that repeated experience may support the development of strategies for reasonable adjustments and adaptations. However, as these outcomes are generally self-reported and often resulting from cross-sectional designs, these should be interpreted as indicative of an association rather than evidence of an effect. Taken together, these findings suggest that experience functions as a potential mechanism through which HCPs develop practical competence in adapting assessment, communication, and procedures for autistic adult patients, particularly in time-pressured healthcare settings. Interpreted through CanMEDS, these patterns align with the ``medical expert role'' (practice-based development of clinical judgement and procedural adaptation) and have implications for the ``communicator role'' as experience may support the development of tailored communication strategies and reasonable accommodations.

4.1. Training received by healthcare professionals

Across studies, autism-specific training was inconsistent and often limited, with clear variation across professions, countries and training pathways. By profession, training exposure appeared uneven. Physicians more commonly reported some prior autism-specific training, often distributed across medical school, residency and short courses, whereas radiographers and intellectual disability HCPs reported particularly low levels of autism-specific training, frequently relying on ``on-the-job'' learning. By country, access to training varied markedly, ranging from relatively higher levels (63%) among Turkish doctors (Gerçek et al.,2024) to very low levels (15%) among Ethiopian HCPs (Zeleke et al., 2021), indicating substantial inequity in access to autism-specific education. A recurring cross-study pattern was informal, experiential learning rather than structured education. Australian HCPs more commonly reported informal learning (58%) and comparatively few reported receiving (23%) autism-specific content during undergraduate or postgraduate education (Urbanowicz et al., 2020). This was similar in the United Kingdom (UK) / Northern Ireland (NI), with a minority (29%) of HCPs reporting any autism training, most of which were brief awareness sessions (Dillenburger et al., 2016), suggesting limited depth and coverage of applied competencies. Training exposure appeared higher during initial qualification for Palestinian nurses (54.6%), whereas formal postgraduate autism-specific training was notably low (7.6%) in Australian nurses (Cashin et al., 2022; Shawahna et al., 2021). Among US primary care physicians, over half (56.6%) reported previous autism training, although the extent varied over medical school, residency and short courses (Mazurek et al., 2020). Radiographers' training was also low with only 3% in Switzerland reporting specific autism training (Carlier et al.,2022), compared to higher but inconsistent levels among UK radiographers (Stogiannos et al., 2021), indicating that even within a single profession, training may depend heavily on local service priorities and opportunities (Carlier et al., 2022; Stogiannos et al., 2021). There were similar gaps in other professions, including only 33% of intellectual disability HCPs, reinforcing that autism-specific training is not routinely embedded across disciplines (Tromans et al., 2019). Within the CanMEDS framework, the pattern of limited and irregular training aligns with gaps in the ``scholar role'', specifically the acquisition and application of knowledge and suggests that key competencies may not be systematically developed across professional education or CPD pathways. Overall, the limited scope of training identified suggests these competencies may not be adequately addressed. Training needs of healthcare professionals The most frequently identified need across all countries was communication. By profession, communication and interaction adjustments were consistently prioritised, including among radiographers and primary care physicians (Alaghband-rad et al., 2023; Carlier et al., 2023; Zerbo et al., 2015). HCP needs also extended to recognising autism in adults, managing sensory sensitivities, and adapting routine clinical processes (Alaghband-rad et al., 2023; Carlier et al., 2023; Edwards et al., 2925; Gerçek et al., 2024; Micai et al., 2021; Tromans et al., 2019; Zerbo et al., 2015). By country, communication remained a dominant theme, including in countries such as Iran, Switzerland, the US, and the UK, however, emphasis varied with other priorities stated as screening tools and recognition, continuing education access and workplace-based training (Alaghband-rad et al., 2023; Carlier et al., 2023; Dillenburger et al., 2016; Tromans et al., 2019; Zerbo et al., 2015). In Iran, HCP prioritised training in communication strategies (64.8%) and screening tools (60.4%) (Alaghband-rad et al., 2023). Similarly, Swiss radiographers also emphasised communication with autistic adult patients and their carers (Carlier et al., 2023), along with 70.8% of US physicians (Zerbo et al., 2015). Radiographers in the UK reported substantial training needs in understanding (90%), communication (88%) and autism recognition (70%) (Stogiannos et al., 2021). Other studies highlighted broader educational gaps, including Palestinian nurses' need for continuing education (Shawahna et al., 2021) and HCPs' limited understanding of health risks associated with autistic adults (Micai et al., 2021). More recent findings also point to deficits in both entry-level education and in employment-related training (Watson-Grace et al., 2026). When mapped to the CanMEDS framework, these needs extend beyond the ``medical expert role'' and highlight gaps in the ``communicator role'', especially in relation to adapting interactions to meet the needs of autistic adults.