Eligibility criteria

To be included, studies had to report quantitative and/or qualitative outcomes of educational interventions aimed at building the capacity of health professionals to work effectively with individuals with intellectual or developmental disabilities (including autistic individuals) of any age. Outcomes included clinician knowledge, attitudes, confidence (a broad term in earlier research) and/or self-efficacy (context-specific confidence). The focus was on qualified health professionals (e.g. medical, nursing, allied health), not undergraduate students; studies with mixed cohorts required >50% healthcare professionals to be eligible. This distinction reflects differing educational needs and motivations between practising clinicians and students. Eligible interventions used any method or mode of training, but needed a clear goal of enhancing clinical care capacity not solely diagnostic- or treatment-specific training.

Studies were included if they met all the following:

1. (a) participants were health professionals working in mainstream (non-specialist intellectual disability, intellectual and developmental disabilities, or autism health) settings, including community and hospital settings;

2. (b) interventions aimed to build clinician capacity to work effectively with individuals with intellectual or developmental disabilities (including autistic individuals) and included an educational component;

3. (c) quantitative and/or qualitative within or between-individual comparisons were reported;

4. (d) outcomes included at least one of the following: clinician knowledge, attitudes, confidence or self-efficacy.

Studies were excluded if they met one or more of the following:

1. (a) >50% participants were students, trainees (including professionals in training) or non-health professionals (e.g. disability support workers);

2. (b) interventions focused solely on diagnosis or screening, (e.g. diagnosis of foetal alcohol spectrum disorder, conducting cervical screening);

3. (c) no relevant quantitative or qualitative findings were reported (e.g. studies reporting only diagnostic outcomes without assessing health professional knowledge, attitudes, confidence or self-efficacy).

Search strategy

Studies were identified by searching PubMed, PsycINFO, ProQuest, Web of Science, CINAHL, EMBASE and the grey literature. The Cochrane Library and PROSPERO database were also searched to ensure no other systematic reviews were planned or had been published in this area. A University of Queensland librarian reviewed the search strategy. The search was not restricted to date, thus a variety of search terms that included historical terms for intellectual disability and autism were included. Broad terms for clinicians were used to cover the wide variety of health professionals, and possibility that interventions were only conducted on one professional group. Search terms are summarised in Table 1.

Table 1 Search terms

Each search term was limited to title/abstract, each search string was connected using the Boolean operator ‘AND’, and the * indicates that the search is broadened to include any words starting with these letters preceding the *.

Study selection

Authors C.F. and K.B. undertook dual independent blinded reviews of the titles and abstracts of articles remaining after removal of duplicates, using the Covidence platform on Windows and macOS (Veritas Health Innovation, Melbourne, Australia; see https://www.covidence.org). C.F. and K.B. then undertook dual independent blinded review of the full text of included articles. Further information was sought from authors of two included publications to clarify published information relevant to inclusion. Each article described a unique study, and there were no studies that were reported in more than one article. Disagreements between reviewers were resolved at each stage through discussion, and consensus was able to be reached without involvement of a third reviewer.

Data collection, extraction, synthesis and analysis

Data was extracted on the Covidence platform, using a form designed for this study by C.F. Data extraction was dually independently undertaken by C.F. and S.G. Data extraction commenced on 31 August 2024 and was completed on 17 December 2024. Conflicts were resolved through discussion and consensus achieved at each stage of the above process.

Data items that were extracted for each study were: reference details; country (where intervention was delivered); consumer involvement (co-design (yes/no) and co-delivery (yes/no)); delivery mode (face to face, online, video, written); format (didactic, interactive, mixed (noting if Extension for Community Healthcare Outcomes (ECHO)); duration of intervention; setting (primary care, community health, community mental health, paediatric hospital, mental health in-patient care); study design (qualitative, single group pre-test/post-test, prospective waitlist control, non-randomised two-group, partial stepped wedge randomised controlled trial); outcome measure area (knowledge, attitudes, confidence, self-efficacy, skills, behaviour) and name of measure; quantitative methods; and results. For studies that included qualitative methods, the aim, approach and main themes were summarised. Missing data were denoted by ‘NS’ (not stated). Each data point was extracted by two independent reviewers, C.F. and S.G. Any differences in data were discussed, and consensus was reached.

Participant numbers were collected, along with their professional group (medical, nursing, allied health professionals – specific professional listed if identifiable). The number of participants who completed the intervention and post-intervention evaluation were recorded, compared with those who commenced and provided baseline data. The objective of the intervention was recorded, in addition to the area of focus (intellectual disability, intellectual and developmental disabilities or autism), whether it focused on paediatric or adult healthcare, and any specific education focus (the diagnosis of autism or intellectual disability, general clinical care).

Evaluation methods were a focus of this review. Data was collected to describe the study design, outcome measures and data points and results, in those studies where published or validated measures were used. In those studies where qualitative methods of evaluation were reported, the framework, analytic process and focus of investigation and themes were recorded for each study.

Study risk-of-bias assessment

The variation in methods of assessment and evaluation of health professional education made it difficult to meaningfully compare the quality and risk of bias of included studies. Therefore, the Mixed-Methods Appraisal Tool (MMAT) ^{Reference Hong, Gonzalez-Reyes and Pluye15} was used to assess qualitative methods, and the Medical Education Research Study Quality Instrument (MERSQI) was used to assess quality of quantitative studies. ^{Reference Reed, Cook, Beckman, Levine, Kern and Wright16} The MMAT is a critical appraisal checklist of the methods used. ^{Reference Hong, Gonzalez-Reyes and Pluye15} The MERSQI, on the other hand, is quantitatively focused and has a scoring system, but it does not specify specific thresholds of methodological quality. It has been validated as a reliable tool for this purpose. ^{Reference Cook and Reed11} The MERSQI consists of ten items, with total possible scores ranging from 5 to 18. One study suggested using a score of 14 to denote high-quality studies, ^{Reference Lin, Lin, Auditore and Fanning17} and another agreed and further proposed suggested ranges of 5–9 for low-quality studies and 9–14 for medium-quality studies. ^{Reference Clarke and Fung10} Authors C.F. and S.G. undertook dual blinded independent assessments, applying the MERSQI and the MMAT (where appropriate), using these thresholds. Disagreement in quality ratings were resolved through discussion between the two reviewers, until consensus was reached. Risk of bias in data extraction was minimised by using two independent blinded investigators (C.F. and S.G.) to assess each study via the Covidence platform. Disagreements in risk of bias assessments were resolved through discussion.

Selective reporting bias was assessed by comparing planned and reported outcome measures for each study via checking published trial protocols and, where not published, the methods specified in the publications.

Effect measures

P-values were selected to indicate statistical significance across studies, as they were commonly reported in this study and allow some comparison between studies.

Synthesis methods

Synthesis of mixed methods studies was initially conducted by analysing quantitative and qualitative findings separately. ^{Reference Voils, Sandelowski, Barroso and Hasselblad18} Data was extracted and recorded using the Covidence platform, and then collated into a table to record characteristics of individual study design (Tables 2 and 3). Qualitative findings were then collated in a separate table that included data on methods, themes and findings (Table 4). The frequency of each data field in Tables 2 and 3 were collated in Table 9 to synthesise findings.

Table 2 Intellectual and developmental disabilities study design characteristics

SGPP, single-group pre-test/post-test; ATTID, Attitudes to the Treatment of People with Intellectual Disabilities in Mainstream Services; TCS-ID, Therapy Confidence Scale–Intellectual Disabilities; GSE, General Self-Efficacy Scale; ECHO, Extension for Community Healthcare Outcomes. Study identifier: only the first author is noted, with year of publication. Bespoke refers to unpublished measures, constructed for purpose by authors.

Table 3 Autism study design characteristics

SGPP, single-group pre-test/post-test; ECHO, Extension for Community Healthcare Outcomes; PCASE, Adapted, shortened version of the Primary Care Autism Self-Efficacy survey; AKQ-P, Autism Spectrum Disorder Knowledge Questionnaire – Physician Edition; CAMI, Dutch translation of the Community Attitudes to Mental Illness questionnaire; EBPAS, Evidence-Based Practice Attitude Scale; CCPAI, Confidence in Coping with Patient Aggression Instrument; CHABA, Challenging Behaviour Attributions Scale Short Form; KAP, Knowledge, Attitudes, Practice; ASD-KS, Autism Knowledge and Self-Efficacy Questionnaire. a. Delivered by a parent of an autistic child, no involvement by autistic person themselves. Study identifier: only the first author is noted, with year of publication. Bespoke refers to unpublished measures, constructed for purpose by authors.

Table 4 Qualitative design and outcome of mixed-method studies

Study identifier: only the first author is noted, with year of publication.

Study selection

A total of 6803 studies were identified on searching relevant databases. After duplicates were removed, 4602 studies remained for screening of title and abstract; 4498 of these did not meet inclusion criteria and were removed. After assessment for eligibility, another 69 studies were excluded. The main reasons for exclusion were that the participant group did not consist of predominantly health professionals (n = 28) or that there was no reporting of outcome measures or methods (n = 19). This left 34 studies that met the eligibility criteria for the review (Fig. 1).

Fig. 1 PRISMA flowchart for study identification, screening and inclusion. ^{Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann and Mulrow13}

Study characteristics

The design characteristics of the 34 studies are summarised in Tables 2 and 3, where they are arranged according to the specific focus of the education (intellectual disability, intellectual and developmental disabilities, or autism). All studies included quantitative measures, no studies were purely qualitative, and five of the 34 studies used mixed methods. Table 4 summarises the qualitative design of those studies where qualitative analysis was reported.

The earliest study was published in 1992, although the vast majority were published between 2014 and 2023. In four studies, the focus of the educational intervention was intellectual disability clinical care in adults. ^{Reference Dagnan, Masson, Thwaites, James and Hatton19–Reference Bessell, Kim, Chiem, McDonald, Thompson and Glozier23} There were two intellectual and developmental disabilities educational interventions focused on clinical care in children. ^{Reference McConkey, Hall, Soni, Macdonald, Sinclair and Veljkovik24,Reference Sadoo, Nalugya, Lassman, Kohli-Lynch, Chariot and Davies25} The remaining 27 studies were autism educational interventions: 19 focused on autism in children and eight focused on autism in adults. Of the 19 studies that focused on autism in children, nine focused on screening and diagnosis of autism ^{Reference Balogh, Wood, Lunsky, Isaacs, Ouellette-Kuntz and Sullivan26–Reference van ‘t Hof, van Nieuwenhuyzen, van Berckelaer-Onnes, Deen, Hoek and Ester34} and ten focused on clinical care. ^{Reference Ashburner, Ziviani, Rodger, Hinder, Cartmill and White35–Reference Sengupta, Lobo and Krishnamurthy44} The eight educational interventions that covered autism in adults all focused on clinical care. ^{Reference Ben-Sasson, Atun-Einy, Yahav-Jonas, Lev-On and Gev45–Reference McGonigle, Migyanka, Glor-Scheib, Cramer, Fratangeli and Hegde52}

Risk of bias in studies

The MERSQI was used to assess quality of quantitative data, with scores presented in; Table 5. Only two studies were rated high quality, 22 studies were rated moderate and ten were rated low. Notably, just 11 of the 34 studies used validated, published scales; the remainder relied on unvalidated, bespoke measures. The MMAT was applied to mixed-methods studies, with the lowest scoring domain determining overall quality. Two studies achieved 80% (four stars) and three scored 60% (three stars). A common limitation was the lack of rationale for using mixed methods. Higher-scoring studies reported a clear qualitative framework and methods, whereas lower-scoring ones did not.

Table 5 Quality assessment

MERSQI, Medical Education Research Study Quality Instrument; MMAT, Mixed-Methods Appraisal Tool. Study identifier: only the first author is noted, with year of publication.

Results of individual studies

The results of individual studies are represented in Tables 4 (qualitative), 6 (knowledge and attitudes), 7 (confidence and self-efficacy) and 8 (behaviour). Three studies reported item-level statistical analysis but not of total scores. ^{Reference Harper and Wadsworth21,Reference Johnson, Lashley, Stonek and Bonjour29,Reference Pasco, Clark, Dragan, Kalambayi, Slonims and Tarpan43} Three different studies reported scores, but did not perform statistical analyses. ^{Reference Swanson, Warren, Stone, Vehorn, Dohrmann and Humberd33,Reference Giachetto, Casuriaga, Santoro, Kanopa, Garrido and Fernández47,Reference Giarelli, Ruttenberg and Segal48} All 29 studies that reported statistical analyses reported statistically significant findings in at least one outcome area. Findings were similarly positive across each subject area of intellectual disability, intellectual and developmental disabilities, and autism. Qualitative data from the five mixed-methods studies also demonstrated positive findings and tended to focus on the effects of the training on clinical care (Table 4). Only the mixed-methods study by Sengupta et al ^{Reference Sengupta, Lobo and Krishnamurthy44} described participant feedback relating to the inclusion of people with lived experience in the delivery of the education, noting that participants found it very valuable (Table 4).

Results of syntheses

Table 9 shows that only eight out of 34 studies reported co-design or co-delivery, with just one incorporating both. Of the six co-delivered studies, three involved individuals with intellectual disability or autistic individuals, and three involved parents of autistic children. Interventions were delivered across 15 countries – 26 in high-income countries (including 15 in the USA, five in Australia, two in the UK) and eight in low- and middle-income countries. Delivery modes varied: 13 were fully online, 12 were face to face, seven used both and two combined written with either video or face-to-face formats. Twelve studies delivered education of 11–20 h, whereas nine were 1–5 h with positive findings. Most interventions were workplace-based, primarily in community health settings.

Most studies (30 out of 34) used a single-group pre-test/post-test design, with 15 collecting data only immediately pre- and post- intervention. Five studies included a qualitative component, although only three of these reported a guiding framework – all using thematic analysis. Because of varied outcome measures, effectiveness by delivery mode (online versus face to face) could not be reliably compared. However, all studies using statistical analyses reported at least one statistically significant positive finding in knowledge, attitudes, confidence and/or self-efficacy. Of the six studies assessing behaviour, five used objective measures, such as the number of cases referred for assessment or use of autism screening measures. Data were typically collected immediately pre- and post- intervention; just over half included a 3-month follow-up, with some extending further to assess for sustained impact. Knowledge was the most assessed outcome, followed by confidence, self-efficacy and attitudes. Behaviour was the least frequent outcome measured, but had some of the longest follow-up intervals, including 6 months, 12 months and 4 years (Tables 8 and 9). Across all studies, 15 studies only assessed for effects immediately following the education, 11 studies for up to 6 months and 2 studies for over 12 months (Table 9).

Table 6 Knowledge and attitudes results

PWP, psychological well-being practitioner; HIP, high-intensity practitioner; ATTID, Attitudes to the Treatment of People with Intellectual Disabilities in Mainstream Services; CAMI, Dutch translation of the Community Attitudes to Mental Illness questionnaire; EBPAS, Evidence-Based Practice Attitude Scale; KAP, Knowledge, Attitudes, Practice; ASD KSEQ, Autism Spectrum Disorder Knowledge and Self-Efficacy Questionnaire.

a. Follow-up was at 18 months. Study identifier: only the first author is noted, with year of publication. Bespoke refers to unpublished measures, constructed for purpose by authors. Data points: 0 = immediately pre- and post- the intervention. *P < 0.05, **P < 0.01, ***P < 0.001.

Table 7 Confidence and self-efficacy quantitative results

TCS-ID, Therapy Confidence Scale–Intellectual Disabilities; GSE, General Self-Efficacy Scale; PCASE, Adapted, shortened version of the Primary Care Autism Self-Efficacy survey; KAP, Knowledge, Attitudes, Practice; AKQ-P, Autism Spectrum Disorder Knowledge Questionnaire – Physician edition; CAMI, Dutch translation of the Community Attitudes to Mental Illness questionnaire; CCPAI, Confidence in Coping with Patient Aggression Instrument. Study identifier: only the first author is noted, with year of publication. Bespoke refers to unpublished measures, constructed for purpose by authors. Data points: 0 = immediately pre- and post- the intervention. *P < 0.05, **P < 0.01, ***P < 0.001.

Table 8 Behaviour results

M-CHAT, Modified Checklist for Autism in Toddlers; ASD, autism spectrum disorder. Study identifier: only the first author is noted, with year of publication. Data points: 0 = immediately pre- and post- the intervention. *P < 0.05, **P < 0.01, ***P < 0.001.

Table 9 Combined design characteristics of included studies

Table 10 describes the validated scales used in these studies in further detail. The majority (25 out of 34) of studies used bespoke measures that were designed for purpose and not published or validated. Of the nine studies that used outcome measures that were published and validated, ten different scales were used, and none stood out as more commonly used. Some scales were developed specifically for use in this area, whereas others were not. The ASD Knowledge and Self-Efficacy Questionnaire ^{Reference Atun-Einy and Ben-Sasson53} and Autism Spectrum Disorder Knowledge Questionnaire – Physician’s Edition (AKQ-P), ^{Reference van ‘t Hof, van Berckelaer-Onnes, Deen, Neukerk, Bannink and Daniels54} were designed to measure knowledge of autism. By contrast, only one scale measuring attitudes was specifically designed for use in relation to intellectual disability or autism, the Attitudes Towards Intellectual Disability (ATTID) scale, ^{Reference Morin, Crocker, Beaulieu-Bergeron and Caron55} whereas the other two published scales were designed for assessing attitudes to evidence-based practice (Evidence-Based Practice Attitude Scale, EBPAS) ^{Reference Aarons, Glisson, Hoagwood, Kelleher, Landsverk and Cafri56} or mental illness (Community Attitudes Towards the Mentally Ill, CAMI). ^{Reference Taylor and Dear57} Confidence was measured using the targeted Therapy Confidence Scale – Intellectual Disabilities (TCS-ID) ^{Reference Dagnan, Masson, Cavagin, Thwaites and Hatton58} and the non-disability specific Confidence in Coping with Patient Aggression Instrument. ^{Reference Thackrey59} Self-efficacy scales were specific for intellectual disability (General Self-Efficacy Scale, GSE) ^{Reference Schwarzer, Jerusalem, Weinman, Wright and Johnston60} or autism (Primary Care Autism Self-Efficacy survey, PCASE). ^{Reference Mazurek, Brown, Curran and Sohl61} The Challenging Behaviour Attributions Scale (CHABA) was used to assess beliefs around aggression. ^{Reference Hastings62}

Table 10 Outcome measures

a. abbreviated form. Scale author and year, Study using this scale: only first author of publication is noted, with year of publication.

Table 11 demonstrates knowledge and confidence had the most statistically significant improvements, although fewer studies measured outcomes other than knowledge. Qualitative findings revealed various knowledge and skills gains across interventions. One study discussed that although participants thought they already knew about autism, the intervention helped them ‘see that there was much more to learn’. ^{Reference Sengupta, Lobo and Krishnamurthy44} An intervention that focused on therapists working with people intellectual disability reported increased awareness of the needs of people with intellectual disability and how they could improve their practice. ^{Reference Dagnan, Masson, Thwaites, James and Hatton19} Examples of changes that were identified included adaptation of materials, nature of communication and the interventions used in therapy. ^{Reference Dagnan, Masson, Thwaites, James and Hatton19,Reference Sadoo, Nalugya, Lassman, Kohli-Lynch, Chariot and Davies25} Changes in attitudes were also highlighted, and related to the involvement of a parent of an autistic child as part of the expert hub providing education, which gave greater awareness and insights into the experience of the autistic child and their family. ^{Reference Sengupta, Lobo and Krishnamurthy44} One paper evaluated the use of co-mentoring as a technique that was chosen to improve the uptake of learnings in the workplace following a didactic training course. Benefits included the opportunity to discuss learnings and how they might be implemented, challenges were the logistic challenges in a busy workplace. ^{Reference Ashburner, Ziviani, Rodger, Hinder, Cartmill and White35}

Table 11 Synthesis of quantitative findings

Reporting biases

Few studies performed statistical analyses on collected data or analysed only single questionnaire items. Although no results were missing, reporting bias likely exists because all studies with statistical analysis reported at least one positive finding, and it would be expected that there would be some studies in this area that showed no change in measured parameters.

Improving knowledge, attitudes, confidence and self-efficacy

Recent calls for improved health professional education evaluation ^{Reference Reed, Cook, Beckman, Levine, Kern and Wright16} prompted development of the Standards for Quality Improvement Reporting Excellence in Education (SQUIRE-EDU) ^{Reference Ogrinc, Armstrong, Dolansky, Singh and Davies65} and measures to assess quality of medical education research. ^{Reference Cook and Reed11,Reference Reed, Cook, Beckman, Levine, Kern and Wright16} The MERSQI measure, used to assess quality in this review, ranks outcomes in increasing order of quality, with lowest ranking scores for subjective self-report measures of satisfaction, attitudes, perceptions and general facts, followed by knowledge and skills, followed by behaviours; and highest ranking scores for the use of patient/healthcare outcomes. Most studies in this systematic review scored the lowest scores. There was some measurement of behaviours in six studies (Table 8), such as health professionals performing periodic health assessments, or blinded patient assessments of skills to assess diagnostic accuracy in one study (Tables 2 and 3). ^{Reference Eagleson, Weise, Cvejic and Trollor20} The challenge in the area of intellectual and developmental disabilities is that, apart from screening and diagnosis in children, there are few behavioural or patient outcome parameters that can be measured in adults. One potential objective measure could be the number of patients identified with intellectual disability in records pre- and post-intervention.

This review’s synthesis of qualitative and quantitative results suggests that education is especially effective in improving knowledge in this area. However, it is of concern that attitudes of health professionals, a major source for concern in this area, was relatively under-investigated quantitatively, reported in only seven papers. This is despite the availability of several published and validated scales to measure attitudes in intellectual disability, ^{Reference Morin, Valois, Crocker and Robitaille66–Reference Akrami, Ekehammar, Claesson and Sonnander68} and some in autism. ^{Reference Cage, Di Monaco and Newell69,Reference Flood, Bulgrin and Morgan70} It is also noteworthy that the synthesis of quantitative knowledge outcomes was problematic, as most scales were unpublished and situation specific. It was not clear what type of knowledge was improved in most of the quantitative results, and how generalisable that might be to healthcare for people with intellectual disability and those on the autism spectrum. Qualitative findings did suggest improvements in knowledge and provided some examples of the ways that this knowledge improved practice. Confidence, and the closely related concept of self-efficacy, were measured, with confidence showing particularly impressive improvements in eight of the ten studies that measured it.

Co-design and delivery

The value of involving consumers in research has gained increasing recognition, with supporting policies and resources emerging in countries such as the UK, Canada and Australia. ^71–73 Inclusion of individuals with lived experience in the design and delivery of intellectual and developmental disabilities health education and research is now widely acknowledged as a priority. ^{Reference Frankena, Naaldenberg, Cardol, Garcia Iriarte, Buchner and Brooker74} However, this review found such involvement to be rare – only one study reported both co-design and co-delivery. Although seven different studies involved people with lived experience in the delivery of the education, only one reported feedback from participants on how this affected their educational experience. There can be challenges involving individuals with intellectual disability or those on the autism spectrum, but it is not a new practice. As early as 2008, Tracy and Iacono included tutors with intellectual disability in teaching medical students on communication with people with intellectual disability. ^{Reference Tracy and Iacono75} This approach requires careful planning and support, and should be embedded early in project development, and considered in the evaluation of the education.

Effective dosing and mode of delivery

This review also sought to examine whether there was a minimum dose of education necessary to improve outcomes. Education of paid employees is a financial burden to the employer, so this finding could help encourage efficient use of educational interventions in this area. In this review, approximately a third of interventions were 5 h or less in duration, and about a third were over 11 h. Unfortunately, the variety of measures used limits the strength of conclusions that can be drawn.

Guidance on the most effective modality can be considered by the finding of generally positive results across all modes of delivery, including the 13 out of 34 studies that used online delivery only. This shows promise for the effectiveness of online education, which has greater flexibility in remote areas, and in ease of access and administration. However, further research is required to allow more robust consideration of these variables.

Limitations

One of the main limitations of this review is its focus on restricting inclusion of papers that described either qualitative or quantitative outcome measures. This was necessary to ask the key question in relation to effectiveness of education. However, it means that the findings relate only to this group of evaluated educational interventions, and these may not be representative of all educational interventions in this area. The requirement for published outcomes also means that any papers with narratively discussed outcomes would also have been excluded. The heterogeneous focus of included studies (e.g. diagnosis versus clinical care) may also have made comparisons between education with different types of approaches less reliable or valid.

Implications and future directions

This systematic review found that educational interventions consistently improved healthcare professionals’ knowledge, attitudes, self-efficacy, confidence and skills. However, many studies lacked validated or comparable outcome measures, limiting the strength of findings. Co-design and co-delivery were uncommon and should be prioritised in future research and incorporated into evaluation. There is a clear need for a validated knowledge scale tailored to health professionals to support comparisons between studies. Qualitative components provided valuable insights into intervention impact, but overall, more robust outcome measures are needed to strengthen future study designs.

In conclusion, this mixed-methods systematic review found that educational interventions had positive effects on health professionals’ knowledge, skills, confidence and or self-efficacy. Health professional participants often expressed surprise at how much there was to learn, and reported that training improved the quality of the care they provided to people with intellectual disability and those on the autism spectrum. Involving people with intellectual disability and those on the autism spectrum in the design and delivery of interventions was noticeably scarce, and should be prioritised in future efforts. Further research is needed to determine the most effective delivery modes and duration of education.

Lived experience summary (by R.d.G. and C.F.)

Education is important for health professionals to help them give better care to people with intellectual disability and autistic people. Some need to learn more about how to communicate with people with intellectual disability and autistic people, and how to listen better. Otherwise, people with disabilities don’t get the care they need.

It would be good for this education to be mandatory everywhere, like it is in England, because then all health staff would learn more about people with disabilities and how to care for them better.

It’s important that people with disabilities are involved in the design, so that they can say what people with disabilities want health professionals to hear, so they are teaching the right things. Involving people with disability in giving the education is a good way to show health professionals what it’s like to have a disability, so that they can understand better and be more familiar with disability.

Measuring if the education works is important because people with a disability need to be able to see health professionals who can offer equal service to people with a disability to the general population. It’s important that health professionals learn from the education and that it can be shown that they have learnt, to offer the best healthcare to people with a disability.