Introduction

Autism Spectrum Disorder (ASD) is characterised by poor social communication, repetitive behavior or stereotypes, and sensory issues (APA, 2013). There has been an increase in the prevalence of ASD over the last few decades which could be related to a variety of factors (Rice et al., 2012). Global prevalence is estimated to be 0.5–1% (Baxter et al., 2015; Elsabbagh et al., 2012). Early intensive behavioral intervention by a therapist has been found to be beneficial (Reichow et al., 2014). However, a majority of children with ASD live in resource-poor settings in South Asia and other low and middle-income countries (Rahman et al., 2016) where there is a lack of evidence-based interventions tailored for those settings and a huge shortage of therapists (Patel et al., 2013). Access to care for childhood mental health disorders has been recognised as one of the major challenges (Collins et al., 2011). Parent training programs are one way of increasing access to interventions for ASD (Wainer & Ingersoll, 2015). Parent-mediated intervention has also been found to improve outcomes in ASD (Pickles et al., 2016). Cochrane review by Oono and colleagues indicated that parent-mediated intervention can lead to positive changes in parent–child interaction, improvement in child language comprehension, reduction in ASD characteristics, maternal knowledge about ASD, and maternal communication (Oono et al., 2013). A high-quality Randomised Control Trial (RCT) for parent-mediated intervention for ASD in south Asia, (Rahman et al., 2016) adapted the Preschool Autism Communication Trial (PACT) (Pickles et al., 2016) for local culture and delivery by trained lay health workers. They found that it improved parent–child synchrony and child-initiated communication with the parent but reduced mutual attention at the end of six months compared to treatment as usual. However, parents living far away from the main centres can have difficulty accessing 12 sessions for PASS interventions. Other researchers have also found a shortage of professionals, limitations in finances, transportation, child care, waitlist, and time commitment as important barriers for widespread dissemination of parent-mediated interventions (Meadan & Daczewitz, 2015; Stahmer & Gist, 2001; Symon, 2001; Wainer & Ingersoll, 2015).

Literature Review

Technology-based interventions can potentially improve access to evidence-based treatments, at convenient times, with reduced costs (Baggett et al., 2010; Gros et al., 2013). Therefore, it is worth exploring technology-assisted evidence-based parent-mediated interventions that can support parents remotely. A variety of technological aids have been used to assist parents of children with ASD. They have varied from training programs for parents that are done remotely delivered using technology to others that used digital resources such as mobile apps, computer programs, DVDs, and robotic interventions that can be used by parents to engage their children with ASD (Aresti-Bartolome & Garcia-Zapirain, 2014; Grynszpan et al., 2014).

A review of tele-practice, for assessment and treatment of individuals with ASD, included eight studies with case series or ABAB designs (Boisvert et al., 2010). A meta-analysis included the use of technology-based interventions such as desktop computers, interactive DVD, and virtual reality for children with ASD. It included 22 studies that had group pre-post design, control studies, and RCTs. The ten RCTs included in this review had patients from the age of three to twenty-nine and focused on facial recognition, affect recognition, or emotional vocabulary. The individual studies in this study had different comparisons such as ASD patients with and without treatment, patients with ASD and no ASD or pre & post scores without any control arm. Combining the results for all of these studies, an effect size d = 0.47 with no significant difference related to IQ or age has been reported (Grynszpan et al., 2014). Knutsen et.al, included 36 articles predominantly pilot studies, single-case designs, case reports, surveys and, one RCT in their systematic review on telemedicine for ASD (Knutsen et al., 2016). They found that technology use was feasible and acceptable but recommended larger RCTs to better evaluate impact. In the systematic review on remotely delivered training for parent-mediated intervention for ASD, only interventions outside urban areas were included. They included seven trials (two pre-post cohorts, three that used multiple baselines and, two RCTs) that used self-guided websites with or without the support of therapists. Improvement in parent knowledge, intervention fidelity and child social communication skills was reported (Parsons et al., 2017).

Reviews focusing on a computer, app or robot-assisted interventions have reported promising results in a variety of developmental disorders and have shown to improve social, academic and intellectual functioning in ASD (DiPietro et al., 2019; Kokol et al., 2019). Ferguson et.al, focused exclusively on the use of telehealth training, supervision, or consultation of interventionists (professionals or parents) for delivery of Applied Behavior Analysis (ABA) therapy to individuals with ASD. They had 28 studies (where at least one person had ASD) of various designs of which only eight (28%) used a group design, all of which had a weak quality rating. More than 60% (n = 17) of studies rated positive improvements for all participants across all variables studied while 32% (n = 9) showed mixed efficacy. Some studies found improvement in the fidelity of delivery by the interventionists but no significant improvement in social communication behaviors in the participants (Ferguson et al., 2019).

Another meta-analysis included seven RCTs that used mobile apps for children with ASD. Their results suggested improved visual and fine motor skills in the intervention arm based on data from two RCTs but no significant benefits in social communication (Moon et al., 2019). Another systematic review evaluated the evidence for assessment, monitoring and treatment of all neurodevelopment disorders including ASD and ADHD. They included forty-seven trials of various designs and were interested in clinical effectiveness, economic impact and user impact while using various types of devices such as mobile apps/tablets, robots, gaming, computerized tests, videos, and virtual reality. About half of the studies reported positive effects (Valentine et al., 2020).

Though previous reviews have suggested that technology-assisted interventions for ASD are feasible and acceptable, their effectiveness has not been fully established (Ferguson et al., 2019; Grynszpan et al., 2014; Moon et al., 2019; Parsons et al., 2017). Previous reviews included fewer studies (Boisvert et al., 2010), poor quality studies (Knutsen et al., 2016), only studies from remote areas (Parsons et al., 2017), included all developmental disorders (Kokol et al., 2019; Valentine et al., 2020), based on only one type of intervention e.g. ABA (Ferguson et al., 2019), or only mobile apps (Moon et al., 2019) or are old & combined results from studies with different designs, comparisons and outcomes together in their meta-analysis (Grynszpan et al., 2014).

Study Aim

This meta-analysis aimed to examine the effectiveness of technology-based interventions in assisting parents to deliver interventions for their children with ASD based only on RCTs. The focus was on social communication and interaction as they are generally targeted and improve with parent-mediated interventions (Oono et al., 2013; Pickles et al., 2016). Parent-mediated interventions generally do not focus on restricted and repetitive behaviors and hence that was not an outcome of interest for this review. As parental involvement and effectiveness are likely to be lesser with adolescents than with children, this study was limited to children 12 years or less. To our knowledge, there was no meta-analysis exclusively on RCTs of technology-assisted interventions supporting parents in delivering ASD-based interventions in children and focusing on social communication outcomes. This study aimed to explore the effectiveness of various evidence-based (RCT) technology-assisted, parent-mediated interventions for children with ASD on social and communication-related outcomes.

Method

The protocol was registered in the International Prospective Register of Systematic Review Protocols, PROSPERO number: CRD42020162825.

We conducted a systematic search in 6 databases: (1) MEDLINE, (2) EMBASE, (3) Cumulative Index to Nursing and Allied Health Literature (CINAHL), (4) Education Resources Information Center (ERIC), (5) PsycINFO, and (6) Pubmed in February 2020 and updated the search in January 2021. No date range was used for search.

A systematic search strategy was created for the followingterms “Autism”, “ASD”, “Telehealth”, “Telemedicine” and “Randomised Controlled Trial” firstly for MEDLINE and then adapted it accordingly to each database. Search terms for parent or caregivers were left out as including them significantly reduced the number of studies and left out several studies that met the inclusion criteria for this review. The detailed search strategy used for different databases is presented in appendix 1. After the removal of duplicates, identified studies were shortlisted through screening of the title and abstract. Full texts of shortlisted studies were independently reviewed by two reviewers KK and HJ based on the inclusion and exclusion criteria mentioned below. References from included studies and past systematic reviews on this topic were hand searched to check if any additional studies met the inclusion criteria for this review.

Inclusion Criteria

  1. (1)

    Randomized Controlled Trials with either a control intervention or waiting-list control group;

  2. (2)

    Children were diagnosed with ASD according to DSM-5; with autism, Asperger's syndrome, or pervasive developmental disorders not otherwise specified using DSM-IV criteria; or diagnosed with Childhood autism, atypical autism or Asperger syndrome under ICD-10 criteria.

  3. (3)

    Studies were included if:

    1. (i)

      the children with ASD were aged twelve years or less;

    2. (ii)

      where studies included children over twelve years, the proportion of such children was under 50% of cases.

  4. (4)

    Studies included technology-assisted parent-mediated interventions;

  5. (5)

    The technology employed could include mobile apps, DVD, video conferencing, web-based interventions, virtual reality, robots, or others;

  6. (6)

    Social behavior and communication outcomes for the child were assessed;

  7. (7)

    Studies were published in English.

Exclusion Criteria

  1. (1)

    Interventions delivered by clinicians or others (except parents or caregivers);

  2. (2)

    Studies with only face to face interventions;

  3. (3)

    Studies using a single-subject multiple baselines;

  4. (4)

    Studies that do not report any social or communication outcomes for the child;

  5. (5)

    Interventions delivered in schools or specialist centers.

Disagreements on the inclusion between the two reviewers (KK and HJP) were resolved by discussions with co-author PK. Data from the included studies were extracted independently by one of the authors (KK, HJ, SM) using a prepared proforma (Appendix 2) and was then verified by another author. The differences were resolved by further discussion to reach a consensus. Mean (SD) and total sample size in each arm from validated measures and subscales were extracted. To determine the treatment effect Mean Difference (MD) and Standardized Mean Difference (SMD) along with 95% Confidence Interval (95% CI) were applied depending on whether the outcome measurements were made with the same assessment tool or different assessment tools respectively. For dichotomous data number of events and number randomized in each group were extracted. Risk Ratio (RR) along with 95% Confidence Interval (95% CI) was used as the effect measure. Analysis was conducted by following the guidance from the Cochrane handbook of systematic review (Higgins JPT, 2019). Data analysis was conducted using Revman 5.3 (2014).

Mantel Hansel method for dichotomous and inverse variance method for continuous data was employed with random effects model in calculating the pooled estimate. Heterogeneity was assessed with the Cochrane Q test along with I2 statistics. An I2 statistics range from 0 to 100%. An I2 index less than 25% is indicative of low heterogeneity, between 25 and 75% represents average heterogeneity, and more than 75% means that considerable heterogeneity is present (Higgins et al., 2003). Subgroup analyses were undertaken for studies reporting endpoint score (final measurement outcomes) and change in score values (changes from baseline) separately when different scales were used for measuring the same outcome.

Quality Assessment

The quality of studies included in this review was assessed based on the standards set by Reichow et al. (Reichow et al., 2008). They set up primary indicators (factors that relate to the validity of the study) and secondary indicators (other important factors). We followed their guidelines to rate each of the primary indicators as high (H), adequate (A), or inadequate (U) based on defined criteria. The secondary indicators were either positive ( +) or negative (−) (Reichow, 2011). Furthermore, based on the number of primary and secondary indicators applicable, the studies were categorized into strong, adequate, or weak in overall quality. Strong ratings were offered for those studies if they scored high for all the primary indicators and at least four of the secondary indicators. Those studies with a high rating for a minimum of four primary indicators and two secondary indicators were rated as having adequate quality.

The weak rating was reserved for studies that had less than four high ratings for primary indicators and two secondary indicators. One of the three authors (KK, HJ, or SM) rated the quality of each of the included studies, and then it was verified by one of the other authors. Whenever there was a difference of opinion it was resolved by consensus or further discussion with PK. Only 6.25% (N = 1) of the articles required verification by another author to resolve differences in opinion. As we had less than ten studies in each of the analyses we were unable to check for publication bias using a funnel plot.

Results

A total of 16 studies (thus K = 16; n = 786) that fulfilled the inclusion criteria were included. The PRISMA flowchart is depicted in Fig. 1. The studies were mostly conducted in the developed world, eight in the United States, and four in Australia, one each in United Kingdom, Italy, Macedonia, and Israel. There was variability in the age ranges included in the studies, eight studies included participants under the age of six (Esposito et al., 2017; Fletcher-Watson et al., 2016a; Ibañez et al., 2018; Ingersoll et al., 2016; Lindgren et al., 2020; Parsons et al., 2019; Vismara et al., 2016; Whitehouse et al., 2017) another five included those 12 or under (Beaumont, 2018; Gev et al., 2017; Voss et al., 2019; Williams et al., 2012; Young & Posselt, 2012). We also included three other studies, as more than 80% of their participants were under twelve or had a mean age less than twelve (Kelly, 2017; Vasilevska Petrovska & Trajkovski, 2019; Vasquez-Terry, 2014). Two of them were included in the analyses (Vasilevska Petrovska & Trajkovski, 2019; Vasquez-Terry, 2014). The details of the included studies are summarised in Table 1.

Fig. 1
figure 1

PRISMA flow chart

Full size image
Table 1 Characteristics of included studies
Full size table

Parent Characteristics and Involvement

Amongst studies that used app-based interventions two reported basic English proficiency (Fletcher-Watson et al., 2016b; Vasquez-Terry, 2013), one study enrolled parents who had a university degree (Esposito et al., 2017) and one study had 20% parents in high socioeconomic status (SES), 60% in middle and 20% in the bottom (Fletcher-Watson et al., 2016b). Others did not report information regarding gender of the parent, age, education, or socio-economic status. Parent training varied from initial training at the beginning of the program (Esposito et al., 2017; Parsons et al., 2019; Whitehouse et al., 2017) to ongoing weekly training (Vasquez-Terry, 2013) or brief instruction document on how to run the activities (Fletcher-Watson et al., 2016a). The involvement of parents was mostly to facilitate the activities and play a supportive role.

In the online group, studies reported that parents with proficiency in English (Ibañez et al., 2018; Ingersoll et al., 2016; Vismara et al., 2018), three reported on level of education (Ibañez et al., 2018; Ingersoll et al., 2016; Vismara et al., 2018) and forty to sixty percent had either college or graduate education and in one trial that reported income levels about 40% earned an annual income of more than 75 thousand US dollars (Vismara et al., 2018). Most studies in this group offered substantial training to the parents in the form of web-based tutorials (Ibañez et al., 2018), coaching (Lindgren et al., 2020) technical support, or ongoing therapist assistance (Ingersoll et al., 2016), and additional website resources (Vismara et al., 2018). Parents also played a more active role in the implementation of the intervention, offered reinforcement for positive engagement of the child, and collected outcome data.

Of the three studies that used DVD-based interventions, one study investigated and found no additional benefit from parental involvement. They included English-speaking and non-English speaking parents (Gev et al., 2017). Generally, parents in the DVD and computer-based interventions groups only played a supportive role in helping their child’s engagement with the intervention.

Quality of Studies

The quality of studies was rated using the criteria by Reichow and detailed in Table 2 (Reichow, 2011). Six studies were rated strong, 10 were rated adequate and 1 study whose abstract only was available was rated weak in overall quality. Several studies had all primary indicators but only a few studies had blind raters (Fletcher-Watson et al., 2016a; Vasquez-Terry, 2014; Vismara et al., 2016; Voss et al., 2019) and only six studies had social validity (Ingersoll et al., 2016; Lindgren et al., 2020; Vasquez-Terry, 2014; Vismara et al., 2016; Williams et al., 2012; Young & Posselt, 2012).

Table 2 Quality of studies included
Full size table

Outcome Measures

The details of the various measures used in the trials have been summarized in Appendix 3.

Effectiveness

One of the studies (Ibañez et al., 2018) used a Bayesian distribution which led to skewed results compared to all the other studies. Including data from this study changed the direction of the results. Hence, we decided to exclude that from the analysis. Four studies that used app based interventions, one that used online intervention, and one with interactive DVD-based intervention reported social communication outcomes. Most studies used parent-rated measures such as the Autism Treatment Evaluation Checklist (ATEC), Communication Symbolic Behavior Scales (CSBS), frequency of social behaviors, and peer interest, however, one study used therapist rated measure Brief Observation of Social Communication Changes (BOSCC) (Fig. 2). The technology-assisted parent-mediated interventions did not offer significantly greater benefits in social communication compared to controls (MD 0.75, 95% CI − 0.16 to 1.68; participants = 282; studies = 6; I2 = 39%) (Fig. 2) moderate certainty using Grading of Recommendations, Assessment, Development and Evaluations (GRADE) criteria (Table 3). Two studies (one that evaluated app-based intervention and the other one compared interactive DVD based intervention) used therapist rated Vineland Adaptive Behavior Scale (VABS) social skills sub-scale to measure socialization as a functional outcome. Based on data from these two studies that included 129 participants there was no significant difference between the two arms (MD 1.83, 95% CI − 2.01 to 5.68; I2 = 0%) (Fig. 3) moderate certainty using GRADE criteria (Table 3).

Fig. 2
figure 2

Forest plot of comparison 1 Tech assisted Versus Control outcome 1.1 social communication

Full size image
Table 3 Summary of findings—Tech assisted parent mediated intervention compared to Control for ASD
Full size table
Fig. 3
figure 3

Forest plot of comparison: 1 Tech assisted Versus Control outcome 1.2 social skills

Full size image

Emotion recognition was reported in three studies, one study that used an online intervention and two that used interactive DVD intervention, using two different parent-rated measures (Emotion Comprehension Test & NEPSY-II affect recognition). Though the treatment arm was significantly more effective than control (SMD 1.25, 95% CI 0.54–1.96; participants = 112; studies = 3; I2 = 63%) (Fig. 4) lack of blinded outcome assessment, significant heterogeneity and high risk of publication bias led to downgrading to very low certainty according to GRADE criteria (Table 3).

Fig. 4
figure 4

Forest plot of comparison 1 Tech assisted Versus Control outcome 1.3 emotion recognition

Full size image

Four studies that used app-based interventions provided language outcomes as total scores, receptive language, gestures and expressive language scores. Across all four outcomes, language total score (MD − 0.06, 95% CI − 2.76 to 2.64; participants = 179; studies = 3; I2 = 43%); receptive language (MD 10.49, 95% CI − 13.11 to 34.09; participants = 177; studies = 3; I2 = 59%); gestures (MD 1.71, 95% CI − 1.24 to 4.66; participants = 129; studies = 2; I2 = 0%) and expressive speech (SMD 0.03, 95% CI − 0.36 to 0.42; participants = 102; studies = 2; I2 = 0%) there were no significant differences between the two arms with moderate certainty according to GRADE criteria (Table 3).

Publication Bias

Due to a small number of trials in each comparison it was not possible to check for publication bias.

Discussion

The focus of this review was to evaluate if technology-assisted parent-mediated interventions were effective in improving social communication outcomes for ASD. Unlike other systematic reviews (DiPietro et al., 2019; Ferguson et al., 2019; Grynszpan et al., 2014; Parsons et al., 2017) in this study, only randomized controlled trials were included. Sixteen studies with 748 participants were included in the narrative synthesis of this review. An increasing number of RCTs published on this topic in the last few years indicates the growing interest and importance of this area.

Similar to previous reviews (Aresti-Bartolome & Garcia-Zapirain, 2014), studies in this review used different technologies such as mobile apps, computer games, interactive DVD applications, online web-based interventions and superpower glass intervention with support of an app. Though a previous review (DiPietro et al., 2019) included robotic interventions, none of those studies had any parent involvement and many were conducted in schools, therapist’s centre or research labs and hence not included in this review. Unlike previous meta-analyses that were interested in the impact on academics, the main focus of this review was the effect on social communication and interaction (Aspiranti et al., 2020). Hence RCTs that focused only on behavior (Hanrahan et al., 2020; Kuravackel et al., 2018; Turgeon & Lanovaz, 2019) anxiety (Conaughton et al., 2017) parental knowledge (Jang et al., 2012), parental stress (Marino et al., 2020), parent satisfaction (Fisher et al., 2014) or executive function (De Vries, 2015) were excluded.

This review supports previous research that suggests the feasibility and acceptability of technology-based interventions for ASD (Ferguson et al., 2019; Moon et al., 2019; Parsons et al., 2017). Further, studies included in our review showed that parents reported high levels of satisfaction with technology (Fletcher-Watson et al., 2016a; Ingersoll et al., 2016; Vismara et al., 2016). However, one of the app-based studies had significantly high rates of dropouts which merits further exploration (Parsons et al., 2019).

Data from a total of eight comparable studies could be combined in a meta-analysis. Analysis of six studies reporting social communication outcomes revealed no significant difference between intervention and control arms. A previous review that included only app-based interventions that combined data from two RCTs also had similar results. They found that fine motor and visual skills were improved in the intervention arm but no difference in speech, gestures, social communication, and symbolic play (Moon et al., 2019). Results from our analysis contrast with a previous meta-analysis where the effect size from 14 controlled trials was d = 0.47 (95% CI 0.008–0.86) (Grynszpan et al., 2014). The results were significant even when they redid the analysis with only the 10 RCTs in their review. Unlike this review where there was separate analysis for each outcome, their meta-analysis averaged the effect sizes of all outcome measures in each study. They included trials with different research designs such as comparing ASD patients with non-ASD patients, pre-post design, ASD patients with and without treatment) delivered in different settings with parent or therapist mediated delivery modes. These differences could explain why they found completely different results to our analyses.

Consistent with results on social communication, analysis of studies that used Vineland adaptive behaviour scale (VABS) social skills measure also revealed negative results for the treatment arm. It is to be noted that previous meta-analysis had not evaluated social functioning (Grynszpan et al., 2014; Moon et al., 2019). Similarly, there were no significant differences between the two arms in any of the language outcomes (language total scores, receptive language, gestures, and expressive language).

The reasons for the lack of effectiveness are unclear. The duration of intervention in app-based studies varied from five min per day for two months (Fletcher-Watson et al., 2016a) to nineteen minutes (average) per day for six months (Whitehouse et al., 2017). It is possible that the apps were used as aids by parents to engage children and did not focus on training parents which may have resulted in greater improvements. It is also likely that the intensity and duration of interventions in these trials were inadequate to cause clinically meaningful benefits. Two studies with strong quality (Ingersoll et al., 2016; Vismara et al., 2016) involving a 12-week web-based parent training program using multimodality approach reported improvement in parental fidelity. These two small studies indicated a trend towards improving imitation (Vismara et al., 2016) and social functioning (Ingersoll et al., 2016) but did not reach statistical significance. It is worth evaluating if training with greater intensity, duration, and multimodal methods can improve effectiveness. It is worth noting that therapist assisted online training was superior to self-directed learning by parents (Ingersoll et al., 2016). Given that most of the studies included were conducted in developed countries, its generalizability in resource poor developing countries is limited. Studies included also used different classificatory systems and often excluded comorbid disorders which also limits the generalizability of the results.

In contrast, analysis of data on emotion recognition from three trials showed significant improvement in the intervention arms. Compared to the previous analysis in our review on social communication, social functioning and language outcomes in this comparison all three trials had non active controls. Thus, technology-assisted interventions may be better than no intervention in improving social communication outcomes, but this requires further study. However, in those studies that showed improved emotion recognition there was still no clinically meaningful improvement in functional outcomes. There was also a significant risk of bias leading to downgrading the certainty to very low according to GRADE criteria. Other studies have also reported that it is challenging to improve social skills using technology-related interventions (Kelly, 2017).

The quality of studies evaluating technology-based interventions in ASD is improving. In this regard, previous systematic reviews and meta-analyses comprised of studies of variable quality (Ferguson et al., 2019; Grynszpan et al., 2014; Knutsen et al., 2016) which may have impacted the outcomes. However, most of the studies in this review had strong or adequate quality according to criteria established by Reichow et.al (Reichow et al., 2008). Though other systematic reviews in this area had used Cochrane style risk of bias assessments (Griffith et al., 2020; Moon et al., 2019), criteria by Reichow et.al was chosen for this review because of its specificity to interventions for ASD.

Reichow and colleagues have also established criteria for evaluating and determining evidence-based practices in ASD. Two group design studies with strong quality conducted in different geographical locations or four group design studies with adequate quality conducted by two different teams were required to merit consideration as ‘established’ evidence-based program. Two group design studies of at least adequate quality could qualify for ‘promising’ evidence-based program (Reichow, 2011). Based on the above criteria none of interventions included in this review fulfilled criteria for either established or promising evidence-based program. Further research is required to understand the predictors for better outcomes while using technology-assisted parent-mediated interventions for ASD.

Limitations

The studies included in this review used different interventions, in different age groups, for varying durations, using diagnostic criteria from different classificatory systems, comparing different control arms and used different outcome measures. Almost all studies were conducted in the developed world impacting on the generalizability of the findings to developing countries. Further, it is possible that several other negative studies were not published and this may have resulted in publication bias. This could not be verified because of small number of studies in each analysis. Search strategy with terms including assistive technology could have resulted in the inclusion of other relevant articles.

Conclusion

There is burgeoning interest in technology-assisted parent-mediated interventions for ASD and quality of trials were mostly adequate or strong. Studies indicate these interventions are feasible, acceptable and users have reported high levels of satisfaction. While studies have shown some promising results in improving emotion recognition, they have not led to significant improvement in other social communication domains or more meaningful functional outcomes. There is currently insufficient data to either support or refute the effectiveness of technology-assisted parent-mediated interventions to improve social communication. At present technology-assisted parent-mediated interventions do not qualify for evidence-based programs for ASD. Greater intensity, duration, parental training, and active therapist involvement may improve outcomes, but these require further examination. There is a need for carefully planned controlled trials with greater consistency of methodology, implementation of standardized assessment tools, longer duration of intervention and protocols for follow-up.