Leveraging Technology to Make Parent Training More Accessible: Randomized Trial of In-person Versus Online Executive Function Training for Parents of Autistic Children

Scientific Abstract

This randomized trial compared the first online parent training program for an evidence-supported executive function (EF) intervention for autism to in-person parent training with the same dose and content. Parents of autistic children (8–12-years-old; Full Scale IQ above 70) were randomized to in-person (n = 51) or online (n = 46) training. Training acceptability and feasibility were rated highly by parents, without significant differences between groups. The completion rate was lower for parents in the online versus the in-person training condition, but neither the total time spent with training materials, nor parent and child outcomes differed by condition. Parents reported that training resulted in a reduction in caregiver strain (Cohen’s d=0.66) and modest improvements in child flexibility, emotional control and global EF, but not planning and organization. Dose of parent training had a significant positive effect on child planning and organization problems. These findings did not support the hypothesized superiority of online to in-person training, but they did indicate online is as effective as in-person training at helping parents learn to improve their autistic children’s EF abilities and reduce their own experience of parenting strain. Implications included increased access to training for parents who experience barriers to receiving in-person care.

Lay Abstract

This study compared the first online parent training program for executive function intervention for autism to in-person parent training on the same content. Participants were parents of autistic children, who were between 8–12 years of age and did not have intellectual disability. Parents were randomized to the in-person (n = 51) or online (n = 46) training conditions. Both trainings were developed with stakeholder (parents and autistic people) guidance. In this trial, most parents reported that they liked both trainings and that they were able to implement what they learned with their children. Parents in both groups spent equivalent amounts of time (about eight hours) with the training materials, but while 94% of parents in the in-person training attended both parent trainings, only 59% of parents in the online group completed all 10 online modules. Parents reported that it was difficult to stay motivated to complete the online trainings over the 10-week trial. Parent and child outcomes did not differ significantly between the groups. Overall, parents reported that the trainings resulted in a reduction in their own parenting strain and improvements in their child’s flexibility, emotional control, and global executive function, but not planning and organization. These findings indicated brief in-person and online training can help parents learn to support and improve their autistic children’s executive function abilities, reducing their own experience of parenting strain. The finding that the online training was equivalent to the in-person trainings is important because it is accessible to parents who encounter barriers to in-person care.

An enduring challenge for autism spectrum disorders (ASD) is that learned skills do not consistently generalize to new contexts (Brown & Bebko, 2012). This leads to poor outcomes (Charman, et al, 2011) and wasted resources on treatments (National Resource Council, 2001) that do not change behavior in real world settings. Parents, as well as teachers have special power to support generalization by virtue of their presence in the daily lives of children. Parent-mediated and parent-implemented interventions are evidence-based techniques for autistic children (Bearss et al., 2015a; Wong, 2015; Althoff et al., 2019), and can yield a secondary benefit for the parents’ wellbeing. A recent meta-analysis (Tarver et al., 2019) of parent interventions targeting autistic children’s emotional and behavioral problems found a small postive effect on parent stress from a variety of interventions, including brief parent trainings (Tellegen & Sanders; 2014). Reduction in parent stress is especially important because parents of autistic children report elevated levels of stress, even when compared to parents of children with other developmental disabilities (Schieve et al., 2007).

Executive dysfunction is common in autism and is related to academic and quality of life outcomes (Pugliese et al., 2020). It also drives negative child outcomes that put extra strain on parents, such as poor adaptive skills (Pugliese et al., 2016), and co-occuring psychopathology (Lawson et al., 2015). Executive Function (EF) merits interventions which are widely accessible and generalize to daily living. Unstuck and On Target (Unstuck; Cannon, et al., 2011; 2018) is a school-based intervention for 7–11-year-olds that has been shown in two randomized controlled trials (RCTs) to improve EF and related behaviors in autistic children (Anthony, et al., 2020; Kenworthy, et al., 2014a). In both trials, it was administered with good fidelity by school staff, indicating its promise as a low-cost, community-based treatment that can overcome disparities in access to clinic-based care. The Unstuck school trials included a low intensity, complementary parent-mediated intervention (Bearss et al. 2015b; Kenworthy et al., 2014b) that was adapted for culturally and economically diverse families (Ratto et al., 2017). The parent intervention provided training regarding EF, and direct instruction in the use of evidence-based child intervention techniques, such as modeling self-regulatory behaviors and using self-regulatory “scripts”, or vocabulary.

The delivery method for the Unstuck parent training was in-person sessions provided by clinician-researchers, however, which limited the dissemination of Unstuck outside of research or clinic supported settings. Some parents also find in-person trainings burdensome and encounter disparities in access related to cost, geography and time constraints. Rural and poor families with autistic children are at a disadvantage when training or treatment is provided in-person (Mandell, Novak & Zubrinsky, 2005), and distance to any kind of clinical service has been identified as a barrier to receiving diagnosis and care by underserved families of children with ASD (Elder, Brasher & Alexander, 2016).

Internet-based training has the potential to overcome barriers in access to training for parents of disabled children (Vismara et al., 2013; Meadan & Daczewitz, 2015). Leveraging technology to bring parent interactive training tools into the home has an added benefit of allowing parents repeated access to the materials as they learn new intervention skills, which is necessary for generalization of skills (Ylvisaker, 1998, 2008). A review (Hall & Bierman, 2015) of 48 studies of technology-assisted parent education and interventions described a burgeoning field which demonstrates the potential of parent training through technological devices but found that evidence for such trainings is not yet conclusive. Such trainings are generally feasible, although concerns remain regarding low completion rates in some cases. Hall and Bierman also noted that only a few studies compared the impact of modality (technology versus in-person) to determine relative efficacy. For this reason, they found that “it is premature to draw strong conclusions about the utility of technology-assisted interventions, but important to encourage further rigorous study of their potential” (p.3). Regarding parent-mediated intervention for ASD specifically, a systematic review (Parsons et al., 2017) found preliminary evidence that remote training delivery methods may improve parent intervention fidelity and child skills, but of seven studies reviewed, only two were RCTs.

In this study, we investigated whether an online parent training, e-Unstuck, was superior to in-person trainings and tested the following hypotheses:

1. Both in-person and online parent training delivery methods are acceptable and feasible for parents;

2. Online training provides greater benefits than in-person training in terms of:

   1. Reducing parent strain
   2. Increasing parents’ sense of competence and self-efficacy
   3. Reducing parent reported EF problems in children;

3. The amount of time spent interacting with training materials (i.e., “dose”) and implementing the Unstuck techniques with their children (i.e., degree of implementation) will be positively related to improvements for parents and their autistic children in both online and in-person conditions.

Methods

Participants

Ninety-seven parents of autistic children were randomized to one of two parent training conditions, in-person or online, in this IRB-approved trial (Children’s National Institutional Review Board, Protocol 8129). Parents who brought children in for inclusion assessment signed a written consent and children provided verbal assent for assessment. If parents had prior assessments available for their children, an in-person assessment for inclusion was not necessary, and parents provided online consent only. One parent withdrew immediately after randomization, leaving a final study sample of 96 parents and their children (Figure 1). Each parent was an English speaker with an autistic child between 8–12 years of age, whose Full-Scale IQ score (FSIQ; as assessed by a standardized IQ test, see Supplemental Materials for details) was ≥ 70, and who had parent-reported flexibility or planning/organization problems (as indicated by a Behavior Rating Inventory of Executive Functions-2 Shift or Plan/Organize scale T score >60). Autism diagnosis reflected Diagnostic and Statistical Manual-5 (DSM-5) criteria as determined by a clinical psychologist and cutoff criteria on the Autism Diagnostic Observation Schedule, Module 3, first or second edition (ADOS; ADOS-2; Lord et al., 2000, 2012). Community administration of the ADOS was accepted when a report was available for review and DSM-5 autism symptoms were clearly described. Research reliable staff conducted the ADOS (n=47) with any child who did not have an ADOS administered in our clinic at the Children’s National Center for Autism Spectrum Disorders (n=42) or in the community (n=7) from a trained, experienced autism clinician. In all cases, ADOS research-reliable staff determined if the child met DSM-5 autism criteria based on all available information, including the ADOS. During screening, parents were asked up to 10 knowledge questions to gauge familiarity with training content. Parents who answered all 10 questions correctly were excluded. See Table 1 for demographics.

Figure 1.

Consort Diagram

Table 1.

Participant Demographics and Mean Child EF and Parent Strain and Competence Scores at Baseline

	In-Person	Online	Group Comparison
Total N	50	46
Age, years, M (SD)	9.8 (1.43)	9.89 (1.52)	t94 = −0.30, p = .76
FSIQ1, Standard Score, M (SD)	104.76 (17.01)	104.09 (17.47)	t94 = 0.19, p = .85
Child N male (% male)	38 (76%)	40 (87%)	χ2 (1) = 1.89, p = .17
Parent N female (% female)	45 (90%)	44 (96%)	χ2 (1) = 1.13, p = .29
Parent Education, years, M (SD)	16.68 (2.03)	17.04 (2.28)	t94 = −0.83, p = .41
Race, N			χ2 (3) = 2.32, p = .51
Asian	1	4
Black/African-American	4	4
White	38	33
Multiracial	7	5
Ethnicity, N			χ2 (1) = 1.44, p = .23
Hispanic or Latino, N (%)	7 (14%)	3 (6%)
Child Outcomes: BRIEF2 (T-scores)
Shift	74.52 (10.23)	76.33 (8.73)	t(94) = −0.93, p = .36
Planning and Organization	68.24 (7.29)	67.83 (7.65)	t(94) = 0.27, p = .79
Emotional Control	67.4 (9.55)	68.35 (9.41)	t(94) = −0.49, p = .63
Global Executive Composite	73.18 (7.38)	73.78 (7.44)	t(94) = −0.40, p = .69
Parent Outcomes
Child-related strain (CSQ)	21.5 (59.2)	22.26 (6.10)	t(94) = −0.62, p = .54
Caregiver competence	28.34 (3.80)	28.09 (3.58)	t(94) = 0.34, p = .74
Caregiver self-efficacy FES	23.78 (3.26)	23.54 (2.67)	t(94) = 0.39, p = .70
Training Evaluation
Acceptability of Training	16.42 (3.25)	16.49 (3.55)	t = −.091, p = .928
Satisfaction with Training	19.30 (3.36)	19.97 (4.04)	t = −.821, p = .414
Feasibility of use of Training Techniques	19.88 (4.90)	19.56 (4.30)	t = .313, p = .755

¹FSIQ: Full Scale Intelligence Quotient

²BRIEF: Behavior Rating of Executive Function-2, Parent Report

Study design

Parents were randomized using a random number generator followed by even/odd segregation into the in-person or online condition. The in-person training included two, two-hour sessions, which occurred at the beginning and midpoint of the ten-week trial. The sessions were led by an Unstuck author (LK or MW) at two different locations and on two different dates in order to facilitate attendance. The online training occurred virtually and asynchronously. Participants were asked to independently review one 20–30-minute module weekly for ten weeks. All parents were given a published parent manual (Kenworthy et al., 2014b).

Parents completed outcome measures online via RedCap or the PARiconnect system at baseline (Pre), and at the conclusion of the trial (Post). Outcomes were all parent-rated and included self-ratings of sense of strain, self-efficacy, and competence, and ratings of child’s EF problems. Less than 10% (N = 5) of the whole sample was lost to follow up (see Figure 1). They did not differ from the sample as a whole on demographic variables. One parent stated that a new job prevented her from continuing the training, otherwise reasons for disengagement from the study are unknown. An additional 16 parents (6 from the in-person training, 10 from the online training) provided only partial outcome data at post-assessment (see Figure 1 Consort Diagram), but the completion rate for the outcome data at post assessment was not significantly different between the two groups as determined by a chi-square analysis (X² = 1.56, p = 0.21).

To assess fidelity to the intervention and dosage of intervention received, participants were asked to keep a daily diary tracking the amount of time spent with Unstuck materials, including the manual. Parents were reminded weekly to log in online to record this data. 99% of parents completed the diary at least once, and 70% completed entries each week.

At the completion of the trial parents were asked to report on psychoactive therapies or medicines that their children received in the community during this trial, see Supplemental Materials for a full description by group. Overall, the online and in-person training groups did not differ significantly in terms of the frequency of any intervention received (all X² values <1.5, all p’s > 0.48), but more than half of both groups (in-person = 58%; online = 52%) received psychotherapy during the trial. Forty-eight percent of children in the in-person group and 54% in the online group received psychotropic medication during the trial. A smaller proportion of children received tutoring or EF coaching during the trial (in-person = 18%; online = 11%).

Content of Parent Trainings

In-person and online trainings presented the same content, which was drawn from the parent manual (Kenworthy et al., 2014b). In-person trainings included PowerPoint presentations with embedded video and other engaging visuals. Key topics in both parent trainings were empirically supported:

1. Distinguishing volitional behaviors that children chose to demonstrate from behaviors that result from brain-based EF deficits (Schopler, Kunce, & Mesibov, 1998; Pennington, 2002);

2. Providing specific supports and accommodations prior to teaching new skills (Schopler, Mesibov & Hearsey, 1995);

3. Maximizing motivation through positive reinforcement and collaboration (Wong, 2015, Ylvisaker et al., 2008);

4. Teaching strategies to recognize and cope with intense feelings (Samson, et al., 2012);

5. Modeling, scaffolding, and supporting the use of specific self-regulatory vocabulary (e.g., Flexible, Big Deal/Little Deal, Goal-Why-Plan-Do-Check; Cannon et al., 2018) in order to create a common language at home (Wong, 2015, Ylvisaker, 1998).

The online training (e-Unstuck) delivered instruction through a combination of video presentations (modeling skills, interviews with parents, experts, and autistic self-advocates) and interactive exercises with individualized feedback and reports based on user responses. The e-Unstuck program’s software was built on the 3C Institute’s proprietary Dynamic e-Learning Platform (DeLP; 3cisd.com/e-learning-for-behavior-change). DeLP implements the evidence-based cognitive theory of multimedia learning to effectively engage diverse types of online learners (Mayer & Moreno, 1998), and is highly rated in usability, acceptability and efficacy (Bruzzese et al., 2020, Jaycox et al., 2019, Sanchez & Bartel, 2015). Each online module intermixed didactic instruction, demonstration/modeling, and description of concepts with self-assessments, and interactive practice opportunities for exploration, practice, and application of concepts and skills (Clark & Mayer, 2008, Shams & Seitz, 2008). See Figure 2 for a list of the online training modules.

Figure 2.

Online Training Modules

Community Stakeholder Driven Development Process

The parent training content and modalities were developed iteratively with formal stakeholder input at each step of development:

1. Training content was originally developed in response to stakeholder requests for EF interventions and guidance regarding the methods used and topics addressed. Stakeholders included autistic youth, autistic self-advocates, and parents and teachers of autistic children (Kenworthy et al., 2014a, Cannon et al., 2018). Two autistic self-advocates reviewed the content of the school curriculum and made suggestions, such as the importance of recognizing the utility of inflexibility for autistic people, that were incorporated into the parent training. The published parent manual has a parent author in order to ensure that the parent perspective guided the material. It was further reviewed by a parent advisory committee at Children’s National Hospital. Recommendations from the committee were implemented prior to publication and included reducing the complexity of the strategies that parents were trained to model.

2. Translation of the in-person parent training to an online platform followed a survey of 24 parents of autistic children who had previously completed or were currently completing in-person Unstuck trainings. They strongly endorsed the need for an interactive, online training for parents (M = 4.38 on a 5-point scale).

3. Prototype online training modules were reviewed by 56 parents of autistic children (without intellectual disability; 8–12 years old). They recommended revisions to interactive instructions and feedback and emphasized the importance of the integrated personalized reports and printables. These suggestions were incorporated into the final online training modules.

4. The completed online training was evaluated by 33 parents (in four focus groups) who participated in the treatment trial in the online training condition to address the quality, value, usability, and feasibility of e-Unstuck. Two notetakers’ notes were then combined and organized into themes that related to recommendations. Rapid qualitative coding (RADaR; Watkins, 2017) allowed for feedback to facilitate changes before public launch. Overall, parents were positive about the online training’s accessibility:

“So much was so helpful…this is amazing that it is going to reach so many people that can’t afford [other services] or logistically make it happen and the timing how you can shut it down and come back to it… that was very powerful.”

Study Outcome Measures

Three types of parent-report outcome assessments were collected:

1. Training evaluation, completed at Post. (See Supplemental Material for full assessment). Acceptability, Satisfaction, and Feasibility of training was evaluated with 18 multiple choice items scored on a 5-point response scale. Higher scores indicated more positive experience or frequent use of strategies (e.g., 0 = 0 times; 5= more than 5 times).

   • Acceptability of Training, 4 items included the degree to which the training was valuable, engaging, innovative, and better than the alternatives. The internal consistency for this scale was good (Cronbach’s α = .88).
   • Satisfaction with Training, 7 items asked if the training improved their ability to understand and cope with their child’s EF problems, think differently about their child’s problems, and collaborate with their child. The internal consistency for this scale was excellent (Cronbach’s α = .90).
   • Feasibility of Use of Training Techniques, 7 items probed the frequency with which parents used specific Unstuck principles and techniques during the previous two weeks. The internal consistency for this scale was acceptable for a short scale (Cronbach’s α = .64).
2. Parent and child outcome measures were completed at Pre and Post.

   • The Caregiver Strain Questionnaire-Short Form 7 (CGSQ-SF7; Brannan, Athay, & de Andrade, 2012) measured child related-caregiver strain. Raw scores were totaled with higher scores indicating greater strain. The CGSQ-SF7 has acceptable psychometric properties, including internal consistency and construct validity (Brannan et al., 2012). The original CGSQ has been validated for use with parents of autistic children (Khanna et al., 2012). For this sample, Cronbach’s α at Pre was 0.88, and 0.90 at Post.
   • The Family Empowerment Scale (FES; Koren, DeChillo, & Friesen, 1992; Singh et al., 1995), assessed empowerment in families of children with mental, emotional and behavioral disorders and contains 4 subscales (Systems Advocacy, Knowledge, Competence, Self-efficacy). This study included the Competence and the Self-efficacy subscales. Raw scores were summed for a total score. Higher scores indicated greater competence or self-efficacy. Psychometric properties of the scale are acceptable (Singh et al., 1995). For this study, both subscales demonstrated good internal consistency at all timepoints (Competence, Cronbach’s α Pre = 0.79 and Post = 0.86; Self-efficacy, α Pre = 0.66 and Post = 0.76).
   • The Behavior Rating Inventory of Executive Function, Parent Report Form (BRIEF-2; Gioia et al., 2015) assessed behavioral manifestation of EF problems in children. It has demonstrated good test-retest reliability and validity (Gioia et al., 2015). BRIEF-2 scores include the Global Executive Composite (GEC), which is divided into eight scales, three of which were evaluated for this trial: Shift, Emotional Control, and Plan/Organize. Shift and Plan/Organize scales were selected because they were sensitive to intervention in a previous trial of Unstuck (Kenworthy et al., 2014a). GEC and Emotional Control were added in order to capture overall EF at home, and address a primary target of interest to parents: regulation of emotions. Scores are reported as T-scores, with higher scores indicating more problems and scores ≥ 65 falling in the suspected clinically elevated range.

3. Dosage of training received. Daily diaries documenting the number of minutes parents spent in any training activity including reading the manual were shared with study staff weekly during the 10-week intervention period.

Data analytic plan

Baseline characteristics (demographics, direct child assessments, and behavior rating scales prior to the trainings) were compared between the treatment groups to ensure comparability of groups after randomization. Acceptability, satisfaction, and feasibility of training were evaluated with descriptive analyses and compared across the training conditions with independent samples t-tests. Feasibility was further assessed through description of training completion rates. Qualitative parent report regarding the Unstuck content, skills, and materials they used most were coded using Atlas.ti to generate response frequencies which are presented in rank order (ATLAS.ti, 2016–2018). Change in parent and child outcome scores as a result of treatment type scores were compared through a series of ANCOVAs with pre-training score, age, gender, Full Scale IQ, and parent education included as covariates to isolate the treatment effects as much as possible. Follow-up paired t-tests comparing pre- to post outcome scores were used to evaluate the magnitude of each of the treatment’s effects (Cohen’s d). Dosage (time spent with training materials) was tallied from the daily diaries and Pearson correlations probed the relationship between total minutes of training and change on each of the outcome scores.

In order to insure that all the data from participants who were randomized were analyzed, multiple imputation was conducted with all randomized participants who did not withdraw prior to starting treatment (n=1), using the parcel summaries multiple imputation method. Missing data was imputed with 20 imputations for each item level variable. All analyses were rerun with the imputed data, and all findings were maintained, therefore results reported below are derived from the unimputed data.

Results

Quantitative data was analyzed using SPSS v26, Version 9.2 of the SAS systems for Windows, and RStudio (IBM Corp., 2019, SAS Institute, Inc., 2010, RStudio Team, 2019). The two groups did not differ on any baseline characteristic measured. Acceptability, Satisfaction, and Feasibility of Use of Training Techniques were rated positively by parents in both conditions, with no significant differences between the groups (ts < 1; ps > 0.4), see Table 1. Eighty-eight percent of parents rated the trainings as “very much” better than the alternatives. Parents reported regular use of strategies taught in the trainings, especially use of the self-regulatory vocabulary (89% reported using the vocabulary sometimes or often) and accommodations (98% reported using accommodations sometimes or often). Ninety-four and 98% percent of parents in the in-person condition attended both, or one of the trainings, respectively. Fifty-nine percent of parents completed all of the online modules, while 96% completed some of the online modules. Overall, parents completed an average of 75% of the online training content.

ANCOVAs controlling for pre-training baseline scores, child IQ, and parent education showed no differences between the in-person and online training conditions for any of the outcome variables (all ps > 0.38). With the exception of the BRIEF GEC post-score, for which there were no significant predictors, the only significant predictor for each outcome variable at post-test was the pre-score for that variable (all ps < .05). Significant improvement in all parent self-ratings (ts between 2.64 and 6.15; ps < .01) and parent ratings of their child’s flexibility, emotional control and global EF problems (ts between 2.51 and 3.88; ps < .02), but not planning/organization, indicated that parents and their children in both conditions benefitted from the training. See Figure 3. The effect sizes for these improvements were small (Cohen’s ds between 0.27 – 0.42), with the exception of improvement in Caregiver Strain which showed a medium effect size (d = 0.66). See Table 2. In order to ensure that the inclusion of the covariates was not influencing these findings and because the two groups did not significantly differ on any variables at baseline, the analyses were rerun as repeated measured ANOVAs and the results were the same.

Figure 3.

Change in Parent (a) and Child (b) outcomes following in-person or online training

¹BRIEF-2: Behavior Rating of Executive Function-2, Parent Report

²GEC: Global Executive Composite

Table 2.

ANCOVAs Measuring Change in Outcome Variables, Accounting for Pre-intervention Score, Child IQ, and Parent Education (BRIEF-2 t-scores reported, all others raw scores)

Outcome measure	Mean Change-Pre - Post (SD)	Effect size (Cohen’s d)
Child-related strain (n = 87)	3.24 (4.91)	0.66**
Caregiver competence (n = 87)	−1.44 (3.96)	−0.36**
Caregiver self-efficacy (n = 87)	−0.93 (3.28)	−0.28**
BRIEF-2 Shift (n = 85)	1.74 (6.39)	0.27*
BRIEF-2 Emotional Control (n = 85)	2.51 (6.48)	0.39*
BRIEF-2 Plan/Organize (n = 85)	0.47 (5.89)	0.08
BRIEF-2 Global Executive Composite (n = 85)	2.06 (4.90)	0.42*

^*p < .05

^**p < .01

^***p <.001

Both before and after removal of two outlier values, parents in both training groups combined reported spending more than 8 hours with the training materials (including the parent book), and time spent with training materials was not different between groups (t = − 0.68; p = .50). Because there were no significant differences in outcome across the two training groups, but there was improvement in outcomes for the two conditions combined, dosage effects were probed for the two conditions combined with the outliers removed. The Pre to Post change in parent reported problems on the BRIEF Plan/Organize Scale was significantly correlated (r = −0.26; p = .016) with the time parents spent with the training materials, such that the more time parents spent with the materials, the greater the reduction in their child’s problems with planning and organization. Information regarding how to support planning and organization skills comes at the end of both the book and online training modules. Time interacting with training materials was not significantly related to change from Pre to Post assessment on any other outcome variable (all rs < 0.14, all ps > 0.23).

Discussion

We report on a randomized trial comparing the first online parent training program for an empirically-supported EF intervention for autism to an equivalent dose of in-person parent training. Both the online and in-person trainings were acceptable and feasible for most parents to implement with their child. Contrary to our hypothesis regarding the superiority of online training compared to in-person training, both formats were equally effective at reducing parent strain and increasing parent’s sense of competence and self-efficacy. Parents also reported that in-person and online trainings were equally effective at reducing their child’s EF problems. Significantly fewer parents completed the online training modules than attended both in-person training sessions, although the amount of time spent with training materials, including the parent book, did not differ across the groups. There was a significant positive effect of the dose of parent training on their child’s planning and organization problems, but not on the other outcome variables.

In-person and online parent trainings were both developed with extensive stakeholder review and responsive revision of materials. Perhaps as a result, both training modalities were generally acceptable to parents in this trial, and they described high levels of satisfaction with the skills they learned. They also reported regular use of the strategies taught in the trainings, especially use of self-regulatory vocabulary (89% reported using the vocabulary sometimes or often) and accommodations (98% reported using accommodations sometimes or often). Completion rates differed across the groups, with 94% of parents in the in-person training attending both parent trainings and 59% of parents in the online group completing all 10 online modules. Completion rates for online training in this trial are within the range observed in a recent systematic review of 14 technology-assisted parent training programs (68.6%±13.1%; Baumel et al., 2017). Parents in the online training condition completed an average of 75% of the training content and spent an equivalent amount of time with the training materials as did the parents in the in-person group. On average, the groups spent more than eight hours with training materials. Care was taken to make the in-person trainings temporally and geographically accessible (repeat trainings offered in two locations at two different times based on parent preferences) and therefore the attendance rate at the in-person training sessions does not reflect typical clinical practice. The fact that less than two-thirds of the parents finished the online training was addressed by the parents in the focus groups, who suggested that spreading the trainings across 10 weeks was not ideal (“when I watched it all at once I was able to see the how they are all related”). Parents also asked for more support from experts and fellow parents to make implementation easier: “I would like an in-person session. I think the feedback like this [focus group] is nice throughout the course of the program. Hearing other people’s opinions and thoughts is very helpful.” And “Two things I wish I had more of: contact with parents who are going through the same thing and having an online webinar with some of the authors at a few points throughout the class would be wonderful. I had a lot of questions that I wished I could have talked with one of you about. I would totally pay for that.”

Parent and child outcomes did not differ by training modality. Parents reported equivalent improvements in self-efficacy and competence following each training as well as equivalent reductions in their child’s EF problems and their experience of child related strain. Parents reported that training resulted in modest improvements with small effect sizes in their child’s flexibility, emotional control and global EF, but not planning and organization. This pattern of greater impact on flexibility than planning and organization was also observed in trials of the Unstuck school-based intervention, combined with in-person parent training (Kenworthy et al., 2014a). More broadly these findings indicate that online parent training in EF can contribute to behavior change in children. They contrast with reports regarding other web-based parent trainings that found it was equivalent to in-person training for promoting parent knowledge, but not for its capacity to lead to decreased problem behaviors (Kable, Coles, Strickland & Taddeo, 2012). On the other hand, parent-reported improvement in their children’s EF in this study did not represent a tightly-controlled test for several reasons. The data was not provided by a masked observer. Parents were aware of, and likely highly invested in, the training they were receiving. In addition, other community-based interventions occurred concurrently with the parent training trial, and they could be the source of some of the positive affects reported by parents.

Though significantly reduced, parent-reported child EF problems remained high after training completion, indicating that brief parent training alone is insufficient to fully address EF deficits in autism. A metanalysis of in-person parent training to address disruptive behaviors in autism identified an overall medium effect on disruptive behaviors but noted that the shortest program studied had the smallest impact on disruptive behaviors (Postorino et al., 2017; Tellegen and Sanders, 2014). In the case of Unstuck, parent training is conceived of as complimentary to the school-based Unstuck intervention. Although parents reported spending more time with Unstuck training materials than expected, their eight hours total average training time and the lack of any personalized check-ins or parent support indicated that this was a low-intensity parent training. Alternative technologically-driven modalities could increase the intensity of the parent training. Parent intervention programs delivered through telehealth sessions have been found acceptable for parents of autistic children (Bearss et al., 2018) and have led to reductions in autistic child behavior problems (Kuravackel et al., 2018). e-Unstuck has been delivered through a hybrid of synchronous and asynchronous sessions, as suggested by the parents in the focus group, with preliminary findings of effectiveness (Wilkinson et al., 2021a, Wilkinson et al., 2021b).

Parent report of the impact of the trainings on their own wellbeing showed a similar pattern to that observed in a recent Unstuck school intervention trial (Swain et al., 2021), such that reductions in child-related parent strain were of the greatest magnitude (medium effect size). Parents reported small increases in their sense of competence and self-efficacy. This pattern was consistent with that reported in Tarver and colleagues’ (2019) metanalysis of parent outcomes following parent interventions for emotional and behavior problems in their autistic children, in which they found positive effects on parent strain, but not on parent efficacy. Even small changes in parenting strain were important in this high-risk group (Swain et al., 2021), and were impressive in such a “low dose” intervention, as indicated by parent feedback such as this: “Having something like this and giving me something that I can use to immediately work with my son and see improvement is life changing”; “For me honestly this has been life changing. I was starting to get worried because the playdates were not ending well and to be able to get my son to change his thinking…he is starting to have more playdates.”

This study was limited by its reliance on parent report for all outcomes. As such, outcomes were reported by individuals who were not masked to intervention, and that may have led to overestimation of the impact of the trainings (Sonuga-Barke et al., 2013). This is particularly problematic because we did not find the hypothesized differences between the online and in-person training groups. As a result the positive changes in parent sense of competence and self-efficacy and child EF, could simply be the result of time, the extra attention provided as part of this research trial, or other treatments that were received in community during the trial. The lack of a measure of gains in knowledge of EF and EF intervention techniques prevented us from investigating the association between increased knowledge and outcomes. Finally, the vast majority of the participants were white and not Latinx. Parent education levels were high, and parents were overwhelmingly female, while their children were predominantly male. Further investigations with representative samples across racial, ethnic, and gender groups are needed.

Conclusion

While it is not a superior training modality, eUnstuck is a viable alternative to in-person parent training. It leverages technology to increase access to training materials for families seeking to support generalization of EF skills. It is as effective as in-person trainings and will be more widely available for those who experience geographic, financial or schedule-related barriers to care. The relatively small effect of parent training on child outcomes indicates that this training should remain a complimentary intervention to more intensive direct child interventions or be combined with direct support for parents. Nonetheless, if replicated, these findings indicate that both brief in-person and online training tools can help parents learn to support and improve their autistic children’s EF abilities, and reduce their own experience of parenting strain.