The reported prevalence of autism has risen dramatically in recent years, with worldwide estimates suggesting that 1 in 100 children are autistic (Zeidan et al., 2022). Autism is a neurodevelopmental difference characterized by persistent difficulties in social communication and reciprocity, and restricted/repetitive patterns of behavior, activities, or interests (American Psychiatric Association, 2013). In addition, many autistic children experience difficulty with daily living skills (DLS), defined as a set of activities undertaken regularly (e.g., daily) that promote optimal health or support adaptive functioning (i.e., eating, sleeping, taking medication, exercising, toileting, personal hygiene, and home-care activities). Such difficulties may result from differences in communication, fine and gross motor skills, and sensory processing, all commonly associated with autism (Hannant et al., 2016).

Difficulties with DLS can have significant and lifelong implications for an individual’s social inclusion, learning, and physical health; thus, significantly impacting their independence. Moreover, when these skills are not attained, children may become reliant on their caregivers for support (Boutain et al., 2020), contributing significantly to the personal, social, and economic cost of autism. For instance, in an Australian longitudinal study conducted by Gray et al. (2014), 89 autistic children, with or without intellectual disability, were followed into adulthood. Results indicated that as adults the majority (89%) of the autistic individuals lived with their families or in an environment that provided support with DLS, with just 11% living independently or in university housing (Gray et al., 2014). This finding is widely reflected in research, demonstrating a strong connection between DLS and autonomy for autistic individuals (Ayres et al., 2011; Dollar et al., 2012; Reyes et al., 2021). Thus, effective treatment is necessary for the short-term and long-term wellbeing of autistic children.

There is an emerging body of research demonstrating the effectiveness of therapist and parent delivered behavioral interventions for DLS in autistic children (Estes et al., 2019; Leaf et al., 2011). This includes, but is not limited to, toilet training (McLay et al., 2015), brushing teeth (Du et al., 2018; Melati et al., 2019), personal hygiene (Dowdy et al., 2018; McLay et al., 2021; Veazey et al., 2015), sleep (Pattison et al., 2020; Vriend et al., 2011), and home-care activities (Byra et al., 2018). However, the rise in prevalence of autism coupled with a lack of trained service providers, high cost of support, and rural location, can limit access to vital services for many families. Furthermore, the current COVID-19 public health crisis has highlighted the limitations of traditional face-to-face service delivery, with many caregivers of autistic children experiencing a breakdown in intervention services.

Telehealth, broadly defined as a modality of healthcare service that utilizes information communication technology (e.g., internet/website, mobile applications, video conferencing, telephone, messaging/emails) to deliver formal assessment, education, and treatment (Sutherland et al., 2018), may offer an alternative or adjunct to face-to-face service delivery. Research suggests that telehealth can reduce the costs associated with behavioral interventions by up to half for service users and providers (Horn et al., 2016; Lindgren et al., 2016). Furthermore, parents living in rural and urban communities around the world report favoring telehealth approaches for providing support to autistic and non-autistic children (Campbell et al., 2019; Cheung et al., 2022; Wallisch et al., 2019). This is reportedly due to the convenience and accessibility of intervention, decreased need for travel, the ability to fit intervention into family routines and situations (Campbell et al., 2019; Cheung et al., 2022; Wallisch et al., 2019), and increased privacy (Campbell et al., 2019). Further, some parents note that during the COVID-19 pandemic, telehealth has provided a safe alternative for accessing services (Cheung et al., 2022).

Notwithstanding the benefits of telehealth, it is important to note that there are limitations associated with this model of care, including expenses associated with installing and setting up telecommunication equipment, the potential for equipment malfunction during consultations, and user difficulties with technology. Difficulties with maintaining client engagement have also been reported by service providers (Traube et al., 2021). As this is an emerging area of clinical practice and research, there is also limited evidence of the effectiveness of telehealth-delivered interventions.

Over the past five years, a number of literature reviews have emerged examining telehealth-delivered assessment, diagnosis, and intervention for autistic individuals (Alfuraydan et al., 2020; de Nocker & Toolan, 2021; Ellison et al., 2021; Ferguson et al., 2019; Kane & DeBar, 2023; Meimei & Zenghui, 2022; Sutherland et al., 2018). Collectively, results suggest that telehealth-delivered services can be effective in assessment and diagnosis and can improve the social skills, language skills, challenging behaviors, and internalizing behaviors (e.g., anxiety) of autistic children. Moreover, reviews have provided evidence of the effectiveness of telehealth-delivered caregiver training in behavioral procedures, suggesting that parents and caregivers of autistic children can successfully implement behavioral procedures that they have learned from therapists via telehealth (Unholz-Bowden et al., 2020). However, while previous reviews have broadly addressed telehealth-delivered supports for autistic individuals, to our knowledge, no review has systematically appraised the literature examining the effectiveness of telehealth-delivered interventions for DLS for autistic children.

Given the profound impact that DLS have on the independence, wellbeing, and functioning of an individual and their caregivers, the limitations and barriers that are associated with traditional face-to-face delivery of behavioral interventions, and the existing empirical support for telehealth-delivered behavioral interventions (TDBIs) for autistic children, it is imperative that the potential utility and effectiveness of TDBIs are further understood. The present review aimed to investigate the use of TDBIs for DLS for autistic children. TDBIs were assessed for feasibility, effectiveness, and social validity, and common components of telehealth-delivered interventions were identified. Finally, the quality of the evidence of the selected studies was appraised.

Method

This review was prospectively registered with PROSPERO prior to beginning the search, and conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Liberati et al., 2009).

Search Procedure

A systematic search was undertaken using the electronic databases PsycINFO, MEDLINE, EMBASE, Scopus, Cochran Library, and Cochran Register of Controlled Clinical Trials (CRCCT). The initial search was conducted in December 2021 by the first author, along with a specialist subject librarian with expertise in database searching for systematic reviews. An updated search occurred in June 2023. Search terms included subject headings appropriate to each database, and keywords related to autism (Group 1), telehealth (Group 2), DLS (Group 3), and age range (Group 4). Refer to Table 1 for a comprehensive list of search terms. A further search of Google Scholar, along with ancestral searches of included papers, and relevant systematic reviews and meta-analyses with a focus on telehealth delivery of interventions for autistic children (e.g., de Nocker and Toolan, 2021; Ferguson et al., 2019; Sutherland et al., 2018) did not result in any additional manuscripts that met inclusion criteria. See Fig. 1 for an overview of the search procedure and the number of studies included at each stage of the search.

Table 1 Summary of search terms across all databases
Full size table
Fig. 1
figure 1

Flowchart of search procedures and included studies. From: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71. For more information, visit: http://www.prisma-statement.org

Full size image

Inclusion and Exclusion Criteria

Included articles met the following criteria: (a) they reported quantitative outcome data on at least one participant aged 24 years or younger who had a diagnosis of autism, (b) the study focused on caregiver training or direct work with autistic children via telehealth, (c) the interventions were based on the principles of applied behavior analysis (ABA), and targeted at least one domain of DLS (e.g., sleeping, eating, personal hygiene), for autistic children, and (d) the articles were published in an academic, peer-reviewed, English language journal. Studies were excluded if they only reported qualitative data, or if outcome data for DLS was not reported separately from other unrelated domains (e.g., communication, academic achievement). Given the limited available research, no limit was placed on the date of publication or research design.

For the purpose of this review, telehealth was defined as a method of healthcare intervention that allowed specialists and care providers to deliver intervention remotely via the application of technology (as defined above) (Sutherland et al., 2018). Applied behaver analysis interventions are those that apply sometimes tentative principles of behavior to the improvement of target behaviors (i.e., participation in activities of daily living), while simultaneously evaluating whether the process caused any meaningful improvements (Baer et al., 1968). Examples of behavior analytic procedures include functional behavioral assessment and analysis, task analysis, prompting, modelling, discrimination training, and manipulating consequences (rewards and punishments). Daily living skills were defined as a set of activities that a person may undertake on a regular basis (e.g., daily) and which promote optimal health or support adaptive functioning. For example, engaging in leisure or recreational activities, taking medication, sleeping, food preparation, eating, drinking, bathing, showering, dressing, grooming, toileting, washing hands, completing household chores, or wearing appropriate personal protective equipment. The upper age limit of 24 years was chosen based on the World Health Organization’s definition of a young person (World Health Organization, 2018).

Initially, 1351 articles were identified. Once duplicates were removed, 897 remained. The titles and abstracts were screened by the first author. The majority of these studies (n = 848) were excluded because they did not utilize telehealth as a component of intervention delivery, the intervention did not target DLS, participants did not have a diagnosis of autism or were over the age of 24 years, researchers did not report treatment outcome data (e.g., they focused on parent treatment fidelity), or articles were not published within peer-reviewed journals. Authors MC and JvD conducted a full review of the remaining 49 articles to determine eligibility. Of these articles, 17 met all inclusion criteria. There was complete agreement (100%) on the final selection.

Data Extraction

Coding of Variables

Each included article was summarized according to: (a) participant characteristics (number, age, gender, diagnosis); (b) DLS targeted during intervention; (c) research design and follow-up assessment procedures; (d) intervention characteristics (agent, intervention type, intervention components, type and level of professional support); (e) dependent variables and data collection methods; (f) results; and (g) overall study quality (strong, adequate, weak).

Evaluation of Study Quality

In order to evaluate the methodological quality of the included articles, one of two rubrics developed by Reichow et al. (2008) were used: the single-case research design (SCRD) rubric, or the standard group design rubric. The Reichow et al. (2008) framework was developed in response to concerns that existing approaches to evaluating research methodology were not suitable for intervention research for autistic children. This framework has since been regularly used in systematic reviews of interventions for autism and other developmental disabilities (e.g., Ferguson et al., 2019; McLay et al., 2020; Yakubova et al., 2021). In a review of SCRD evaluation tools (Wendt & Miller, 2012), the Reichow et al. (2008) framework was rated as being of high quality due to its congruence with agreed standards for quality in SCRD, empirical evidence in support of its validation, and its ability to categorize studies based on their various levels of quality. Further, the framework was the only highly-rated tool that provided separate yet comparable guidelines for both SCRD and group design research (Wendt & Miller, 2012).

Studies evaluated using the Reichow et al. (2008) framework are assigned a final rating based on the primary and secondary indicators. Primary quality indicators are elements deemed critical for demonstrating the validity of a study and are defined on a trichotomous ordinal scale (high quality, acceptable quality, unacceptable quality; further detail of these definitions is included in Tomlinson et al., 2018). Primary quality indicators that are common to both the SCRD and group design rubrics include: (a) sufficient description of participant characteristics, including participant age, gender, and diagnosis; (b) independent variables (IV) described with sufficient detail to enable replication; and (c) dependent variables (DV) are operationalized, replicable, relevant, and appropriate. Additional primary indicators specific to SCRD include a minimum of three baseline measurement points which appear stable, have no trend or counter therapeutic trend, and are defined with replicable precision; all relevant data is graphed and visual inspection indicates stability, less than 25% of overlap in data points, and a significant change in trend in response to the manipulation of independent variables; and evidence of at least three demonstrations of the experimental effect at three points of time, following manipulation of independent variables. Primary indicators that are specific to group research designs include a link between the research question and data analysis, and appropriate statistical analysis with adequate power and sample size.

Secondary indicators are defined on a dichotomous scale (evidence or no evidence), and while important, are not considered necessary for the establishment of validity. Secondary quality indicators that represent both SCRD and group research designs include the collection of interobserver agreement (IOA) across all conditions, raters, and participants with interobserver agreement at or above 0.80 (80%); continuous measurement of treatment fidelity with measurement statistics at or above 0.80 (80%); collection of generalization and/or maintenance data; inclusion of blind raters; and a measure of social validity (e.g., the intervention was cost and time-efficient, the dependent variables were socially important, consumers were satisfied with the results). Additional secondary quality indicators for group research designs include random assignment to groups, comparable articulation (did not differ between groups more than 25%, and less than 30% at final outcome measures), and effect size of 0.40 or above for at least 75% of treatment outcomes.

Studies were categorized as ‘strong’ if they received high-quality ratings on all primary quality indicators and showed evidence of at least three (group design) or four (SCRD) secondary quality indicators; ‘adequate’ if they were rated high on four or more primary quality indicators, with no unacceptable ratings on any primary quality indicators, and showed evidence of at least two secondary quality indicators; and ‘weak’ if they received fewer than four high-quality ratings in primary indicators, and showed evidence of less than two secondary quality indicators. Author MC evaluated all 17 studies according to the Reichow et al. (2008) criteria. A second coder independently evaluated 35% (6/17) of the articles. The final agreement across indicators and ratings was 100%.

Results

The details of included studies are summarized in Table 2. Seventeen studies were included, all published between 2011 and 2023.

Table 2 Summary of research investigating telehealth-delivered behavioral interventions for daily living skills in autistic children
Full size table

Participants

The seventeen studies included a total of 536 participants (1-115 participants per study), ranging in age from 16 months to 19 years. The mean age could not be calculated as some studies only reported age ranges. One study (Corona et al., 2021) included approximately 25 (22%) participants without a diagnosis of autism. Therefore, the total number of autistic participants was 511. Of these, 395 participants aged between 16 months and 14 years were the recipients of parent-mediated intervention. The remaining 116 participants received intervention directly (i.e., not parent-mediated intervention).

Target Daily Living Skills

Eight out of the 17 studies included multiple intervention targets (Boutain et al., 2020; Corona et al., 2021; Gerow et al., 2021, 2023; Ibañez et al., 2018; Marino et al., 2020; Martin et al., 2023; Swaminathan & Pai, 2020). Seven studies targeted feeding or feeding-related behaviours such as using utensils and sitting at the dinner table (Alaimo et al., 2022; Bloomfield et al., 2021; Gerow et al., 2021; Ibañez et al., 2018; Marino et al., 2020; Martin et al., 2023; Swaminathan & Pai, 2020), six targeted sleep difficulties (Corona et al., 2021; Ibañez et al., 2018; Marino et al., 2020; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019), six targeted personal hygiene activities including washing and bathing, hand washing, and oral hygiene (Boutain et al., 2020; Gerow et al., 2021; Ibañez et al., 2018; Marino et al., 2020; Popple et al., 2016; Swaminathan & Pai, 2020), five targeted toileting (Corona et al., 2021; Dabney et al., 2023; Marino et al., 2020; Martin et al., 2023; Swaminathan & Pai, 2020), four targeted personal grooming activities such as applying lotion and dressing/undressing (Boutain et al., 2020; Gerow et al., 2021; Marino et al., 2020; Swaminathan & Pai, 2020), and three focused on home-care activities such as wiping counters, setting the dinner table, dish washing, doing laundry, and putting away personal items (Gerow et al., 2021; Marino et al., 2020; Swaminathan & Pai, 2020). Additional targeted DLS that were important for physical health and/or adaptive functioning were engagement in physical activity (Esentürk & Yarımkaya, 2021; Swaminathan & Pai, 2020), mask-wearing (Sivaraman et al., 2021), taking medication, using adaptive equipment, spending money, shopping, and using a telephone (Swaminathan & Pai, 2020).

Research Design

Nine of the 17 studies used a between-group design. Of these, three employed a randomized control trial (RCT) design (Ibañez et al., 2018; Marino et al., 2020; Popple et al., 2016), four used a pre/post-test design (Corona et al., 2021; Esentürk & Yarımkaya, 2021; Martin et al., 2023; McCrae et al., 2021), and two used a pre/post quasi-experimental design (Gerow et al., 2023; Roberts et al., 2019). Six studies used a single case design, employing concurrent multiple baseline (Gerow et al., 2021), nonconcurrent multiple baseline (Alaimo et al., 2022; Boutain et al., 2020; Dabney et al., 2023; Sivaraman et al., 2021), and changing criterion (Bloomfield et al., 2021) designs. The remaining two studies were case studies (Moon et al., 2011; Swaminathan & Pai, 2020) one of which included multiple cases (Moon et al., 2011).

Type and Level of Professional Support

Level of professional support varied significantly, ranging from 2 to 47 sessions (i.e., workshops or contact from a trained professional). In the majority of cases, sessions occurred at least once per week and had a fixed dosage (i.e., 50 min twice per week for 4 weeks). However, three studies (Boutain et al., 2020; Gerow et al., 2021; Sivaraman et al., 2021) continued intervention until the researcher or interventionist deemed behavioral outcomes satisfactory (i.e., until participants demonstrated at least 90% of the skill steps; Boutain et al., 2020). The longest reported period of participation was 8 months; however, the frequency and duration of sessions was not reported in this study (Boutain et al., 2020). One study that targeted bath time, snack time, and bedtime asked parents to independently work through an online tutorial in their own time over a period of one month, with the expectation that completion would take approximately six hours (Ibañez et al., 2018).

Intervention Characteristics

Parent-Mediated Interventions

Fifteen of the 17 studies were parent-mediated (Alaimo et al., 2022; Bloomfield et al., 2021; Boutain et al., 2020; Corona et al., 2021; Dabney et al., 2023; Esentürk & Yarımkaya, 2021; Gerow et al., 2021, 2023; Ibañez et al., 2018; Marino et al., 2020; Martin et al., 2023; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019; Sivaraman et al., 2021), meaning that interventions targeted parent behavior change as a means to modify targeted behavior in their child.

Intervention Components and Delivery

In most cases, trainers provided the opportunity for parents to receive individualized coaching and/or feedback via regular video conferencing (Alaimo et al., 2022; Bloomfield et al., 2021; Boutain et al., 2020; Corona et al., 2021; Dabney et al., 2023; Gerow et al., 2021, 2023; Marino et al., 2020; Martin et al., 2023; McCrae et al., 2021; Sivaraman et al., 2021) or telephone contact (Moon et al., 2011). Individualized coaching was often paired with additional resources such as written material (Bloomfield et al., 2021; Corona et al., 2021; Dabney et al., 2023; Gerow et al., 2021; Sivaraman et al., 2021), videos (Alaimo et al., 2022; Corona et al., 2021), and workbooks or planning guides (Corona et al., 2021; Dabney et al., 2023; McCrae et al., 2021). In addition to direct training/coaching, some parents were asked to complete homework, or practice skills in their own time in between sessions (Bloomfield et al., 2021; Martin et al., 2023; Moon et al., 2011). In five studies, parents worked through online materials autonomously, including PowerPoints™ (Roberts et al., 2019), interactive modules (Ibañez et al., 2018) and instructional videos (Esentürk & Yarımkaya, 2021). An additional study provided parents with detailed written instructions and no feedback. However, after five sessions of no improvement in targeted behaviors, parents received full dyadic training via videoconferencing (Boutain et al., 2020).

Intervention Types

In all parent-mediated studies, training focused on psychoeducation and/or behavioral skills training. Studies that targeted sleep provided education on sleep physiology, types and causes of insomnia, sleep hygiene, specific strategies for responding to sleep-related problems (e.g., bedtime refusal, delayed sleep onset, night wakings), and the use of specific behavioral procedures such as stimulus control, bedtime fading, fading parental presence, visual supports, and positive reinforcement (Corona et al., 2021; Ibañez et al., 2018; Marino et al., 2020; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019). McCrae et al. (2021) also included content on relaxation, night-time worries, anxieties, and nightmares. Studies that focused on feeding and mask-wearing typically included modeling, prompting, graduated exposure, and reinforcement (Alaimo et al., 2022; Bloomfield et al., 2021; Sivaraman et al., 2021); and psychoeducation focused on the reinforcing nature of escape and avoidance, and the management of challenging behavior (e.g., refusal, aggression). Studies relevant to personal hygiene, grooming, toilet training, and home care normally involved parent training in the use of behavioral strategies such as prompting and reinforcement, visual supports, visual schedules, and modeling (Boutain et al., 2020; Dabney et al., 2023; Gerow et al., 2021; Ibañez et al., 2018; Marino et al., 2020).

Child-Focused Interventions

Two of the included studies worked directly with autistic children (Popple et al., 2016; Swaminathan & Pai, 2020). Swaminathan and Pai (2020) delivered intervention directly to the participant via videoconferencing. The intervention was based upon cognitive behavioral therapy (CBT) and self-determination theory (SDT), and included the use of exposure/response prevention, rewards, antecedent control, goal setting, and visual supports. Popple et al. (2016) delivered intervention directly to autistic participants through the delivery of videos via the internet, twice a day (morning and evening), for three weeks. Videos prompted and demonstrated the correct procedure for brushing teeth. Following the conclusion of the video, participants were asked to complete a brief survey that assessed whether the participant: (a) watched the video, and (b) brushed their teeth after watching the video.

Measures, Dependent Variables, and Data Collection Procedures

In most studies, dependent variables related to parent-reported child participation in activities of daily living, measured using pre- and post-treatment psychometric instruments or surveys (Corona et al., 2021; Gerow et al., 2023; Ibañez et al., 2018; Marino et al., 2020; Martin et al., 2023; Swaminathan & Pai, 2020). Other dependent variables included the number of steps correctly and independently completed by the child during intervention sessions (following prompting), as observed by the parent (Bloomfield et al., 2021) or the researcher (Alaimo et al., 2022; Boutain et al., 2020; Gerow et al., 2021; Sivaraman et al., 2021). Studies that targeted sleep were unique, typically measuring sleep onset latency (SOL; i.e., the time taken to fall asleep), frequency and duration of night wakings (NWs), total sleep time (TST), sleep efficiency (SE; i.e., percentage of time asleep while in bed) and time awake after sleep onset (WASO) (McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019). Outcomes were measured using parent-reported sleep diaries, the Children’s Sleep Habits Questionnaire (CSHQ), and actigraphy (McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019).

In Popple et al. (2016), dental evaluators assessed the presence of plaque on each participant’s teeth pre-intervention, post-intervention, and three weeks following the intervention. During each check-up, parents also completed a survey evaluating their perception of their child’s oral hygiene. Throughout the intervention, survey data was gathered twice a day, assessing whether each child participant watched the intervention video, and/or brushed their teeth after watching the video (Popple et al., 2016). In Esentürk and Yarımkaya (2021), the Leisure Time Exercise Questionnaire (Godin & Shephard, 1985) was used to evaluate the physical activity levels of participants before and after intervention.

Intervention Outcomes

Intervention Effectiveness

Overall, researchers reported improved outcomes across dependent variables and measures (Alaimo et al., 2022; Bloomfield et al., 2021; Boutain et al., 2020; Corona et al., 2021; Dabney et al., 2023; Esentürk & Yarımkaya, 2021; Gerow et al., 2021, 2023; Ibañez et al., 2018; Marino et al., 2020; Martin et al., 2023; McCrae et al., 2021; Popple et al., 2016; Sivaraman et al., 2021; Swaminathan & Pai, 2020). There were however two notable exceptions in which treatment was less effective. Both were studies evaluating the effectiveness of telehealth-delivered intervention for sleep disturbance (Moon et al., 2011; Roberts et al., 2019). Although Roberts et al. (2019) reported subjective improvements in the parent-rated CSHQ for both groups, actigraphs did not indicate that there was any significant improvement in TST, SE, or NWs. Further, actigraphy indicated that both groups (telehealth and face to face) displayed a significant increase in SOL following intervention, and WASO increased for the telehealth group. Moon et al. (2011) reported a decrease in SOL as measured by both actigraphy and sleep diaries; however, actigraphs, sleep diaries, and the CSHQ all indicated that there was no difference in TST. Changes in SE were also variable, with one participant showing a reduction, one participant showing an increase, and one showing no significant difference.

Interestingly, in one study (Boutain et al., 2020), significant improvements in the targeted skills (handwashing, face washing, applying lotion), were not achieved by providing parents with detailed written instructions only. However, after parents received behavioral skills training via video conferencing, all children completed the targeted skills with high accuracy and independence (Boutain et al., 2020).

Three studies compared telehealth delivery with face-to-face methods and produced mixed results. Marino et al. (2020) reported a significant improvement in the demand-specific subscale of the Home Situation Questionnaire (HSQ), which measures compliance with daily living tasks (i.e., toileting, showering, dressing, etc.) in the telehealth group, but not in the face-to-face group. In Roberts et al. (2019) neither the telehealth nor the face-to-face group showed significant improvements in objectively measured (i.e., actigraphy) sleep variables, and while the telehealth group produced a significant decrease in NWs, as measured by the CSHQ, the face-to-face group did not (Roberts et al., 2019). Conversely, Corona et al. (2021) reported that the face-to-face and hybrid delivery groups showed greater improvement in Clinical Global Impressions of Improvement scores (measuring child functioning in daily routines), compared to those in the telehealth-only group. This suggests that improvement in child participation in daily routines was less evident for those who received intervention via telehealth only (Corona et al., 2021). The two studies that compared telehealth to no-treatment control groups reported better outcomes for the telehealth groups, compared to those who did not receive intervention (Ibañez et al., 2018; Popple et al., 2016).

Maintenance of Treatment Effects

Five out of 17 studies included a follow-up measure. This was gathered between 4 weeks and 13 weeks post-intervention (Dabney et al., 2023; Ibañez et al., 2018; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019). Improvements were maintained at follow-up in all four studies (Ibañez et al., 2018; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019).

Treatment Fidelity

Treatment fidelity (i.e., the reliability of the administration of intervention procedures) was measured in 11 out of the 17 studies (Bloomfield et al., 2021; Boutain et al., 2020; Corona et al., 2021; Gerow et al., 2021; Ibañez et al., 2018; McCrae et al., 2021; Sivaraman et al., 2021). Treatment fidelity was typically measured using questionnaires (e.g., parental efficacy scale, Ibañez et al., 2018), parent-completed compliance logs (Dabney et al., 2023; McCrae et al., 2021), or through researcher observation of parental fidelity (Alaimo et al., 2022; Bloomfield et al., 2021; Boutain et al., 2020; Corona et al., 2021; Gerow et al., 2021, 2023; Martin et al., 2023; Sivaraman et al., 2021). Boutain et al. (2020) reported that parents did not correctly implement graduated guidance procedures after receiving written instructions only, however, after these parents received behavioral skills training via video conferencing, all parents implemented graduated guidance procedures with near-perfect fidelity (Boutain et al., 2020). Treatment fidelity was reportedly high across all other studies (Alaimo et al., 2022; Bloomfield et al., 2021; Corona et al., 2021; Dabney et al., 2023; Gerow et al., 2021, 2023; Ibañez et al., 2018; Martin et al., 2023; McCrae et al., 2021; Sivaraman et al., 2021).

Social Validity

The social validity of an intervention considers the social significance of the intervention goals, the social acceptability of the procedures, and the social importance of the intervention’s outcomes (Wolf, 1978). Social validity was measured in 12 out of the 17 studies (Alaimo et al., 2022; Boutain et al., 2020; Corona et al., 2021; Dabney et al., 2023; Esentürk & Yarımkaya, 2021; Gerow et al., 2023; Ibañez et al., 2018; Martin et al., 2023; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019; Sivaraman et al., 2021). This was typically measured with the Treatment Acceptability Rating Form-Revised (TARF-R), or surveys developed by the researchers to assess treatment usability, feasibility, and acceptability factors (Alaimo et al., 2022; Boutain et al., 2020; Corona et al., 2021; Dabney et al., 2023; Esentürk & Yarımkaya, 2021; Gerow et al., 2023; Ibañez et al., 2018; Martin et al., 2023; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019; Sivaraman et al., 2021). Within the limits of the measurement tools applied, all eight studies reported high social validity (Alaimo et al., 2022; Boutain et al., 2020; Corona et al., 2021; Dabney et al., 2023; Esentürk & Yarımkaya, 2021; Gerow et al., 2023; Ibañez et al., 2018; Martin et al., 2023; McCrae et al., 2021; Moon et al., 2011; Roberts et al., 2019; Sivaraman et al., 2021).

Roberts et al. (2019) was the only included study that provided qualitative social validity responses. In their comparison of face-to-face and telehealth delivery, there was no significant difference between groups for satisfaction ratings except for “talking/blogging with other parents”. This item received a significantly higher satisfaction rating for parents in the face-to-face group. Qualitative survey responses in this study also revealed that parents in the face-to-face group felt that the most valuable component of the sessions was meeting other parents and learning about the strategies that they had tried, and many suggested that time for parent discussion was built into the program. Parents in the face-to-face group noted constraints such as session timing, location, or childcare. Some parents from the telehealth group reported that the convenience and ease of use of the telehealth website were the program’s best features, while others cited difficulty with the use of technology.

Methodological Quality

Of the six studies that employed a SCRD, one met criteria for a ‘strong’ quality rating, four met criteria for an ‘adequate’ quality rating, and one met criteria for an overall ‘weak’ quality rating. Of the two case studies, both received a ‘weak’ rating. Of the nine studies that employed a group design, eight studies were classified as being of ‘adequate’ quality, and one received a ‘strong’ quality rating. An overall rating for each study can be found in Table 2.

The SCRD studies were rated highly on several primary indicators, with 100% of the studies meeting criteria for ‘high’ quality ratings for DV and IV, and all but one study (Moon et al., 2011) was rated highly according to the participants’ criteria. By contrast, only five studies were rated highly according to baseline criteria (Alaimo et al., 2022; Bloomfield et al., 2021; Boutain et al., 2020; Dabney et al., 2023; Sivaraman et al., 2021), and five studies met ‘high’ quality ratings for visual analysis and experimental control.

Of the studies that employed a group research design, all were rated highly on participants, comparison group, DV, and link to research criteria, and all but one study (Marino et al., 2020) was rated highly according to IV criteria. Only two group-design studies were rated highly according to the statistics criteria (McCrae et al., 2021; Marino et al., 2020), with the remaining five studies receiving an ‘acceptable’ rating. None of the group design studies were rated as unacceptable on any of the primary indicators.

Discussion

The present review sought to provide an overview of the current empirical research evaluating the effectiveness and social validity of TDBIs for DLS for autistic children, and to appraise the quality of this research. Seventeen studies that met inclusion criteria were identified, 13 of which were published in the last three years, suggesting that this is a new and rapidly advancing area of research. The findings of this review provide preliminary evidence of the effectiveness of TDBIs for toilet training, eating (i.e., number of bites consumed, using utensils, sitting at the dinner table), sleeping (i.e., SOL, bedtime resistance, WASO, TST, SE), personal grooming (i.e., dressing, applying lotion), personal hygiene (i.e., showering, bathing, washing hands, brushing teeth), wearing personal protective equipment (PPE; i.e., facemasks), and engaging in physical activity. Overall, telehealth-delivered intervention produced better outcomes in studies that compared telehealth to control conditions, and in all 12 studies which measured social validity, caregivers rated TDBIs as highly feasible and acceptable.

In general, telehealth provided caregivers with psychoeducation and training in ABA-based interventions, including visual supports, modeling, prompting, graduated exposure, praise, and rewards. In most studies (n = 12/17), intervention involved individualized coaching delivered via videoconferencing, or in one case, via the telephone. Individualized coaching was typically paired with additional resources such as written material, instructional videos, workbooks, and planning guides. In the remaining studies, TDBIs were self-directed.

A notable finding in the reviewed literature relates to self-directed TDBIs. Differing from clinician-delivered content (i.e., via video conferencing or telephone), self-directed programs (also known as technology-based) are primarily delivered via the internet or mobile technology, with little or no clinician contact. While clinician guided TDBIs have many advantages compared to traditional face-to-face intervention, particularly for increasing accessibility and reach (Campbell et al., 2019; Cheung et al., 2022; Wallisch et al., 2019), self-directed approaches do not require support from a trained specialist. Self-directed TDBIs, therefore, have greater potential for dissemination and may broaden access to care for the autism community (Glenn et al., 2022). Of the five studies within the current review that included a self-directed component, treatment effects were promising. Four studies noted parent-reported child improvements, including participation in daily routines (bedtime, bath time, eating) (Ibañez et al., 2018) and exercise (Esentürk & Yarımkaya, 2021), and improved oral hygiene (Popple et al., 2016) and sleep (reduced SOL, NWs, sleep-related anxiety, total sleep severity score) (Roberts et al., 2019). On the contrary, Boutain et al. (2020) reported that significant improvements in the targeted skills (handwashing, face washing, applying lotion) were not achieved by providing parents with detailed written instructions only. However, after parents received behavioral skills training via video conferencing, all children completed the targeted skills with high accuracy and independence (Boutain et al., 2020).

These findings suggest that parents and caregivers can successfully implement behavioral procedures that they have learned via technology-based training, and that such training can result in improved DLS for autistic children (Ibañez et al., 2018; Popple et al., 2016; Roberts et al., 2019). Nonetheless, in some cases, coaching may be required to achieve clinically significant improvements (e.g., as found in Boutain et al., 2020). Future research may aim to evaluate the relative contribution that different components of intervention delivery make to treatment effectiveness (i.e., written material, instructional videos, coaching via videoconferencing), and the conditions under which technology-based training alone may be effective.

Although these initial findings are encouraging, results should be interpreted with caution. According to the quality rating criteria applied, only two of the 17 studies (evaluating TDBIs for sleep and feeding) met criteria for a ‘strong’ classification. The remaining studies were classified as ‘adequate’ (12/17 studies) or ‘weak’ (3/17 studies). Identified limitations within the research included a shortage of reliability (IOA) and fidelity (treatment integrity) checks, limited assessment of generalization or maintenance of treatment effects, and only two out of the 17 studies were able to include blinded data collection. In addition, issues with outcome measurement, such as reliance on subjective measures (i.e., parent-report questionnaires and surveys), may have compromised data reliability. These findings are largely consistent with previous research which has reported similar methodological flaws associated with telehealth research (e.g., de Nocker and Toolan, 2021; Ferguson et al., 2019; McLay et al., 2020).

Given that telehealth is a relatively new and emerging field, these limitations may be expected. Notably, 52% of the studies (n = 9/17) included in this review were pilot or feasibility trials, suggesting that study protocols and procedures in the telehealth field are still under trial. Further, in response to the COVID-19 pandemic, it is likely that many providers pivoted rapidly to online delivery (i.e., by necessity rather than design). Thus, systematic and careful planning of study procedures may not have been feasible. Markedly, more than 40% (n = 7/17) of the studies included in the present review cited the COVID-19 pandemic in the rationale for their research. Further, given the nature of telehealth (i.e., services delivered via distance), researchers may have difficulty collecting data (especially direct/observational), making IOA, treatment fidelity, and reliability checks challenging. To strengthen our understanding of the effectiveness of TDBIs, future researchers should attempt to pre-empt and manage these issues.

An additional limitation of the current review relates to a lack of clear reporting of intervention outcomes and replication of treatment effects. Many studies (n = 8/17) within the current review included multiple domains of DLS. Of these, two (Boutain et al., 2020; Gerow et al., 2021) reported outcomes from each domain separately, while the remaining authors reported outcomes as total scores (i.e., increased overall participation in activities of daily living). While this builds a promising foundation for the use of TDBIs, the nature of associated change between domains of DLS remains unclear. To establish the effectiveness and generalizability of intervention effects across DLS domains, future research may employ further high-quality studies evaluating the effectiveness of TDBIs, with independent reporting of intervention outcomes for each domain. Specifically, domains that were underrepresented or produced mixed outcomes within this review, and therefore warrant further investigation, include sleep, dressing, home-care activities (i.e., wiping the counter, unloading the dishwasher, setting the table, putting away personal items), personal hygiene (i.e., showering, bathing), toileting, taking medication, preparing/obtaining food, using adaptive equipment, spending money, shopping, and using a telephone. Additional domains that have not yet been included in the literature and therefore warrant further investigation include transportation skills (i.e., using public transportation, road safety), time management and planning, technology use, and community participation (i.e., accessing community resources and participating in recreational or leisure activities).

Another direction for future research is to investigate the factors that contribute to attrition rates in telehealth research. Attrition rates were high in some of the studies included within the current review (Gerow et al., 2023; Marino et al., 2020; Popple et al., 2016; Roberts et al., 2019). For example, 61% of participants in Popple et al., 2016 did not complete the programme. Popple et al., 2016 reported that reliable transport to and from the clinic (for objective data collection) and a lack of access to technology were the primary causes for withdrawal. Interestingly, embedded in the inclusion criteria of Popple et al., 2016 were requirements for access to internet and email, and to have reliable transportation to and from the clinic. This suggests that preemptively excluding participants may not be sufficient for sustained engagement. Attrition over the course of a study presents a significant threat to the integrity of research. It also has implications for the real-word effectiveness of TDBIs. Therefore, it is important that the factors influencing participant retention and attrition are better understood.

Relatedly, to extend the potential scope and generalizability of TDBIs, future research may seek to include participants from diverse backgrounds. Few studies reported deliberate efforts to make telehealth attainable through the provision of telehealth-related hardware (Boutain et al., 2020; Gerow et al., 2021; McCrae et al., 2021), some studies noted that participants were excluded due to not having access to resources such as an internet connection (Boutain et al., 2020; Dabney et al., 2023; Ibañez et al., 2018; Martin et al., 2023; Popple et al., 2016), and just one study reported providing training or support with the use of the telehealth technology (Dabney et al., 2023). Further, participants included within the studies were predominantly White. Literature suggests that socioeconomic and geographical disparities can impede access to treatment services for autistic children, and that access to autism support services are limited for families from ethnic or racial minorities or lower-level education (Liptak et al., 2008; Murphy & Ruble, 2012; Thomas et al., 2007; Wilson et al., 2021). Therefore, providers must be deliberate about decisions related to accessibility and effectiveness of interventions to ensure that the needs of diverse families are met. Going forward, researchers should be intentional about working directly with community members from unrepresented communities to ensure they are developing interventions that are accessible and acceptable to all.

To ensure equitable support, policy makers and service providers may collaborate with local organizations and schools. Loaner devices may be offered to facilitate access to computers, tablets, or smartphones. Internet connectivity solutions can be explored, such as providing vouchers or subsidies for internet services, or leveraging public WI-FI hot spots in community centers, schools, or libraries. Where internet connectivity is not possible, offline resources and phone-based support options may be established (as in Moon et al., 2011). Additionally, providing training and technical support will assist families in effectively using telehealth platforms and resolve any technical issues that may arise.

An additional gap in the literature relates to the inclusion of autistic adolescents. Previous studies have almost exclusively focused on young children, with only four involving adolescents (Alaimo et al., 2022; Esentürk & Yarımkaya, 2021; Popple et al., 2016; Swaminathan & Pai, 2020). Each of these studies had relatively few participants, limiting generalization of treatment effects to this population. Further, only two studies (Esentürk & Yarımkaya, 2021; Popple et al., 2016) involved the young person directly (i.e., not a parent-mediated intervention) and no studies gathered social validity or outcome data from the young people involved. In research, preliminary findings advocate for including young people in the therapeutic process of intervention (Chu & Kendall, 2004; Monahan et al., 2021; van Deurs et al., 2019). Additionally, autistic individuals tend to be dissatisfied with the implications and recommendations of research when they are not included in the process (Pellicano et al., 2014; Walmsley et al., 2018). While this may not be feasible with younger children, future researchers should consider involving capable autistic adolescents in the intervention process, and consider their feedback when assessing the social validity of intervention procedures and outcomes.

Overall, there is a widening gap between demand for services and availability, posing a significant barrier for parents who strive to enhance the independence and wellbeing of their autistic children. The findings of this review present a possible solution, suggesting that TDBIs may be an effective means of broadening access to care for the autism community, and ultimately improving DLS in autistic children. Through ongoing evaluation, researchers and intervention providers will move closer to the goal of wide-spread dissemination of effective and socially valid services for autistic children.