Abstract
Many autistic young adults (AYAs) want to build healthy relationship skills that apply across romantic relationships, friendships, work, school, and community life. Yet, few have access to neurodiversity-affirming resources that are tailored toward adults and accessible asynchronously. To address this need, we developed and pilot-tested a video library of short, targeted lessons designed to support AYAs in building the skills they need to have successful, healthy relationships. Working with an advisory board of experts in autism and relationships (N = 5) and AYAs (N = 5), we conducted formative focus group research (N = 14 AYAs) to inform the development of the relationship skills video library. We created 20 videos on the skills Apologies and Addressing Hurtful Comments. Next, we pilot-tested the relationship skills video library with AYAs with reciprocal communication skills (N = 17) to determine whether exposure to the video library improved participants self-rated and observed skills regarding apologies and responding to hurtful comments. We found that participants’ self-rated skill levels increased in two of the three evaluated skill areas. In addition, observed skill usage significantly improved between the preintervention and postintervention evaluations. We collected extensive participant feedback on desired additions to the web platform and priority skills and topics for future videos.
Community Brief
Why was this program developed?
Many autistic adults would like stronger friendships and romantic relationships, more job opportunities, and the chance to live separately from family or caregivers. It can be hard for autistic adults to find help to build their relationships and meet these goals. Most programs to support healthy relationship skills are for children. Our team, dfusion, created and tested a collection of videos called SkillFlix AYA for autistic adults to build healthy relationship skills.
What does the program do?
The new program is a collection of 20 videos focused on two specific relationship skills. The skills are “Apologies” and “Addressing Hurtful Comments.” The videos teach each skill using short videos that show the steps people can use to apologize or address hurtful comments. Autistic adults and other autism experts helped us create the video content.
How did researchers assess the program?
Seventeen autistic young adults tested SkillFlix AYA. First, they took a survey and rated their relationship skills on a scale of 1–7. Next, they did a role-play with project staff in a video call to show two relationship skills. The skills were “Giving an apology” and “Responding to a hurtful comment.” After the role-play, the youth had a week to watch 20 videos. Then, they took the same survey and did another role-play. We recorded all the role-plays. Project staff reviewed the recordings to count when the young adults used relationship skills. Staff also talked to the young adults about their experience using SkillFlix AYA.
What were the findings?
Young adults showed more relationship skills and Microskills after using SkillFlix AYA. The number of skill steps used in each role-play went up from about three to nine skills. On the survey, young adults rated themselves much higher on their ability to give an apology and ask for an apology. They also rated themselves a little higher on how to respond to a hurtful comment. The young adults said that SkillFlix AYA was useful and gave our team ideas on how to make it better.
What were the weaknesses?
A small number of autistic young adults who can have back-and-forth conversations tested SkillFlix AYA. This means we cannot be sure the program will help all autistic young adults. Another weakness is that the role-plays did not happen in person.
All role-play skills were demonstrated over Zoom video chat to ensure safety from COVID-19.
What are the next steps?
We were given funding to make and test about 150 more videos for autistic young adults. The new videos also teach neurotypical families and friends how to communicate with autistic people. We made these videos to watch and discuss together. We will test the new videos in a research study.
How will this help autistic adults?
SkillFlix AYA can help autistic young adults who want to learn new relationship skills. It can help people deal with complicated social situations. People can search the large set of videos and find written resources, helpful links, and discussions.
Introduction
We created a video-based resource to support relationship skill-building for autistic young adults (AYAs), and we engaged autistic people throughout the development of this resource. In this article, we describe the participatory research process involving community advisors, focus groups, scriptwriters, actors, usability testers, and pilot participants for the development and pilot evaluation of the resource. We discuss each phase of formative research, content development, and technology testing, as well as the establishment of an inclusive and innovative evaluation methodology. Furthermore, we provide insights into the feasibility and acceptability findings from our pilot study and share preliminary efficacy results. Finally, we outline our plans for the continued development and evaluation of this resource.
Autism in Transition-Age Youth
The 2017 National Autism Indicators Report 1 estimated that there are 450,000 autistic youth aged 16–24 years in the United States, and approximately 50,000 autistic youth turn 18 yearly. In the next decade, over half a million autistic youth will age out of school-based services, and many will need support transitioning to adulthood. Navigating the transition to adulthood can be particularly challenging for AYAs. 2 Research suggests many will have difficulty accessing support to succeed in interpersonal relationships and life skills, as services for autistic adults are limited.3,4
Recently, organizations have developed programs to address the high rates of unemployment, 5 limited participation in postsecondary education, 6 and low rates of independent living among autistic adults. 7 However, few interventions focus on social relationships critical to school, work, and everyday life. Many autistic people report having few friendships8,9 and difficulties sustaining romantic relationships. 10 While not all autistic adults share the same friendship or relationship goals, all deserve support in developing the skills necessary to meet their personal goals in these critical domains of adulthood.
Relationships and Social Isolation
Social isolation is a pervasive problem for autistic adults, 11 with up to two-thirds having no close friendships.8,12–14 This lack of healthy relationships compounds social isolation and loneliness, which are associated with higher rates of depression and anxiety among autistic adults.15–17 Mazurek (2014) found that loneliness in autistic adults was associated with increased depression, anxiety, decreased life satisfaction, and self-esteem, while greater quantity and quality of friendships were associated with decreased loneliness. 17 The number of friends an autistic adult had showed unique independent effects in predicting self-esteem, depression, and anxiety. Mental health difficulties such as depression may negatively impact autistic adults’ ability to meet other personal goals, such as maintaining meaningful employment, achieving a postsecondary education, or taking care of one’s health and home. 18
For many young people transitioning into adulthood, achieving a stable job is a significant milestone, but only 58% of autistic people aged 21–25 report ever having been employed. 19 Research suggests that misalignment with culturally neurotypical communication and relationship expectations has a significant negative influence on employment for AYAs.20–22 For some, communication and social differences with supervisors and coworkers impact job performance and can lead to termination.23–26
To have fulfilling and healthy lives, some AYAs have requested support in building skills for healthy relationships across the relationship spectrum, including friendships, colleagues, and romantic relationships. Highlighting the importance of this area, organizations have recently developed several resources, including written resources 27 and group-based healthy relationship interventions. 28 In addition, programs have used video modeling and prompting to develop vocational skills for autistic adults. 29 Unfortunately, there are still few existing autism resources that address complex adult relationship skills, and most do not address skills prioritized by AYAs. In addition, the majority require in-person attendance rather than being accessible online. 30 To address this critical need, dfusion, Inc., a company dedicated to advancing health and well-being with innovative strategies, developed and pilot-tested a relationship skill-building video library, SkillFlix for Autistic Young Adults. We designed the video library to build and support skills for AYAs to have successful, healthy relationships.
Emerging Practice
The SkillFlix for AYA streaming video library
We designed the SkillFlix AYA streaming video library so individuals can develop relationship-focused skills at their own pace. We divided the video library into skill sets (e.g., apologies), which are further broken down into 8 to 15 subskill videos (e.g., giving a serious apology that is not accepted). We refer to these subskills as Microskills®, natural “bite-sized” skills, ideal for quick observational learning that can be modeled in two to eight steps over 30–120 seconds of video. Each video models these specific Microskills are modeled accompanied by synchronized supporting text. We developed and tested two skill sets in our pilot test of the video library, 20 videos focusing on Apologies (Giving and Asking for) and Addressing Hurtful Comments. See Table 1 for a complete list of Microskills associated with these skill sets.
SkillFlix for Autistic Young Adults—Healthy Relationship Skills Pilot-Tested
Theoretical basis
The SkillFlix model is grounded in the social learning theory 31 (SLT) and adult learning theory32–34 (ALT). The SkillFlix model’s key feature is learning through repeated observation of realistic modeling of Microskills in the context of the larger skill and topical information for skill acquisition (SLT principle). The SkillFlix model facilitates user-determined customized learning paths (ALT principle) by offering multiple methods of accessing the material, including tailored recommendations, selection of skills and Microskills, content of interest, and a “next video” feature. Users can also decide what content is most relevant, including searching for specific streaming content when needed—for example, before a work function or date.
The SkillFlix model is also informed by the neuroscience framework of mirror neurons, 35 which explains how the visual information received when watching someone perform a skill or action is mapped on the equivalent motor representation in the brain by mirror neurons. This permits an individual to participate in another’s actions without imitating or practicing directly. Experimental research with mirror neurons has validated this approach for participants to learn a variety of skills. 36 We believe this model will be effective with AYAs because neural response has also been shown to play a role in skill improvement for autistic individuals.37,38 Research has found that video modeling has been found to be effective among AYAs, and short online videos are a preferred learning method.39–41
Development of the SkillFlix AYA video library
Years of experience training educators in different skills inspired the creation and patenting of the Microskills approach. Three studies 40 found the SkillFlix method to be effective across different populations: educators, parents, and pharmacists. Participants showed significant improvement in demonstrated skills in these studies. Based on the success of this approach in building communication skills, particularly among parents of autistic teens, the team sought to build a similar resource for AYAs.
We developed the SkillFlix AYA video library based on the input of two advisory panels and focus groups with AYAs. The advisory panels included (1) a panel of five professionals in autism support and research and (2) a panel of five AYAs. We recruited the advisors via professional networks and word of mouth. The advisory panels helped establish content and target skills for the SkillFlix AYA video library, identified questions and topics for focus groups, and played an influential role in the iterative development of video scripts.
The research team identified interpersonal conflict as a primary skill area to explore in the pilot. From this broad skill area, we narrowed down to addressing Apologies and Addressing Hurtful Comments as the two skill sets in the pilot evaluation. The advisory panels and the research team collaborated to develop scripts using ideas described by focus group participants. Both advisory panels reviewed script drafts and had their feedback incorporated. The research team prioritized hiring a diverse cast for video filming. Five of the eight actors were autistic, and three were transgender/nonbinary. Due to COVID-19, we recorded the videos virtually. We uploaded the 20 videos to the SkillFlix AYA website, along with informational fact sheets and links to resources.
This 20-video library depicts culturally neurotypical social expectations that many AYA find difficult to navigate and often create challenges in different types of relationships. This study explored ways to support AYAs in building healthy relationship skills and tested the usability, acceptability, and effectiveness of a streaming video library, Skillflix AYA, to strengthen two specific relationship skills.
Evaluation Methods
Study design
The mixed-methods study included focus groups (n = 15), usability testing (n = 9) with an observation protocol and a brief survey, and pilot testing (n = 15) with a pre–post intervention survey and observation protocol involving role-play scenarios. Throughout the project, the research team intentionally prioritized the needs, priorities, and opinions of AYAs. The expert advisors, focus groups, scriptwriters, actors, usability testers, and pilot participants informed the instrument development, interpretation of results, and conclusions. The Institutional Review Board from ETR reviewed and approved the study procedures, through an agreement with dfusion, Inc.
Focus groups
Participants
The focus group participants were 15 young adults with formal autism diagnoses aged 18 to 28 years. They were recruited via word of mouth and referral through organizational partnerships. See Table 2 for demographics.
Demographics for SkillFlix for Autistic Young Adults (N = 17)—Focus Groups, Advisors, and Pilot Participants
Participants selected all responses that applied for gender identity and race/ethnicity. Range of participants who completed items: n = 16 to n = 17.
Instrument
We used a semistructured focus group guide to learn about what participants desired in a video-based relationship skills resource. The protocol began with open-ended questions about relationship skills and challenges, such as: “What are common relationship challenges that AYAs may have” and “what relationship skills would you find helpful to have modeled by video?” We prepared probing questions and examples to support participant engagement. The protocol delved into detail on specific relationship skills, such as apologies, as well as factors impacting willingness to watch the SkillFlix AYA videos, reactions to content, and website features.
Data collection
The research team conducted two 90-minute online focus groups via Zoom. An autistic research team member moderated the focus group and they were attended by the principal investigator (PI). An assistant moderator took field notes. For accessibility, we encouraged participants to respond via either the chat or vocally, depending on their preference. We read chat comments aloud and incorporated them into the discussion.
Data analysis
After the focus groups, the moderator, PI, and assistant moderator debriefed themes and interpretations of the data. They refined the field notes and updated them to include observations from the research team. We used a content analysis approach to analyze the data, which allowed flexibility in the use of both deductive (e.g., coding based on the questions asked) and inductive coding (e.g., identifying themes that emerged from the data). First, the research team digitally recorded, transcribed, and then read each focus group. Then the research team created and reviewed a codebook, and two independent coders analyzed each transcript and met to discuss differences in coding and achieve consensus.
Usability testing
Once we created the Skillflix AYA pilot library, we conducted usability testing to inform and improve website functionality.
Participants
Nine adults with formal autism diagnoses aged 18– 27 years, participated in usability testing. They were uninvolved in any other aspect of the project. We recruited them via word of mouth and referral through organizational partnerships. See Table 2 for demographics.
Instrument
The structured usability data collection protocol included open-ended qualitative questions covering four website task areas (i.e., navigating the home screen, searching for a video, and playing a video) and a survey of site impressions, including expectations for the functionality of various icons and buttons.
Data collection
Usability test participants joined 1:1 virtual sessions with trained facilitators. During the session, participants shared their computer screens while completing tasks on the SkillFlix AYA site. The facilitator guided participants to use a think-aloud protocol to articulate their thoughts and respond to prompts while using the resource. Facilitators observed the participant’s process using the site and noted any challenges or questions. Once the tester completed all activities, we asked them to share their likes and dislikes about the prototype and how it could be improved.
Data analysis
We gathered insights into an Excel file and noted ones that emerged from more than one usability tester or that severely impacted participant experience for the website developer. The team and developer discussed insights and possible solutions in detail based on what was possible to achieve within the project time frame, budget, and web development platform.
Pilot testing
Participants
We recruited AYAs (N = 17) by asking 121 organizations that work with autistic adults to share our emailed study flyer with their networks. Interested individuals emailed the research team and completed an online eligibility survey.
The eligibility criteria included the following: (1) being between the ages of 18 and 28 years; (2) being English-speaking because the videos were in English; (3) self-reporting a formal autism diagnosis or believing themselves to be autistic (i.e., self-diagnosis); (4) self-reporting comfort with spoken back-and-forth (reciprocal) conversation to complete the role-play tasks; and (5) the ability to complete the pre- and postintervention surveys independently. See Table 2 for demographic information. Based on feedback from advisors, focus groups, and our commitment that SkillFlix AYA be effective for those who have struggled to access autism resources (e.g., late-diagnosed and self-diagnosed individuals who had barriers to diagnosis such as women, people of color, people from low-income backgrounds, or who live in rural areas), we included individuals with either self- or formally diagnosed autism in the pilot. We chose the 18– 28-year age range to broadly include AYAs who may transition to adulthood or reach adult milestones at different ages while still attempting to ensure that most were generally in the same stage of life.
Intervention
We provided participants with unlimited access to SkillFlix AYA for a week and encouraged them to watch all 20 videos and explore the resources.
Instruments
Pre- and Postintervention Surveys
Participants completed a preintervention survey with demographic questions and three questions on self-perceived skill (e.g., “How would you rate your skill level for….”). We evaluated the skills (1) Giving an Apology, (2) Asking for an Apology, and (3) Addressing Hurtful Comments. Participants rated their self-perceived skill for each of these using a 7-point Likert scale (1, “very unskilled”; 7, “very skilled”). The postintervention survey included the same questions about self-perceived skill. In addition, participants responded to eight acceptability questions (e.g., ease of use, how much they liked the site contents) using a 7-point Likert scale (1, “not at all”; 7 “very much”). Finally, participants responded to open-ended qualitative questions about what they did and did not like about the SkillFlix AYA video library, any challenges in the use of library, and suggestions for site features and video topics. (see Supplementary File S4 for the pre- and postintervention survey).
Pre- and Postintervention Role-Play Scenarios
To measure participant use of targeted skills before and after exposure to the SkillFlix AYA video library, the research team developed four naturalistic role-play scenarios, an accompanying structured observation protocol, and a detailed evaluation rubric for each role-play scenario (see Supplementary File S1 for role-play protocols, Supplementary File S2 for role-play flowcharts, and Supplementary File S3 for the role-play ruberic). We used the semiscripted role-plays to observe real-time skill demonstration in realistic scenarios that could not have been captured through ratings of self-perceived skills alone. We selected Giving an Apology and Addressing Hurtful Comments from the three skills. The role-play scenarios were brief, 2–4-minute conversations. The research team designed role-play scenarios to have similar premises while covering different topics, giving participants the opportunity to demonstrate the same skills before and after exposure to SkillFlix AYA in a new conversation.
The scenario we depicted for Giving an Apology represented a friend feeling let down and frustrated after the participant’s character (1) did not attend a friend’s band performance they had agreed to attend or (2) ignored a friend’s request for support after a breakup.
For Addressing Hurtful Comments, the role-play scenario represented a friend insulting the participant’s ability after the participant’s character (1) made a mistake while cooking or (2) made an error putting together a piece of furniture.
To provide a semistructured framework that would allow facilitators to react to a spectrum of potential responses throughout the role-play scenarios, the research team delineated a flowchart of responses. We extensively tested conversation flow options with staff not working on the study to account for possible responses, define conversation parameters and comments that would fall outside role-play scenarios. For example, through testing, we established that if a participant responded with irrelevant factors such as, “I didn’t text back because my cat was in the hospital,” the role-play facilitator would remind the participant that we provided all the information for the role-play scenario and they were not expected to generate further background. The facilitator would inform participants that they were stepping back into character and ask participants to return to the same point in conversation, “redoing” their last response.
The research team scored the role-play scenario observations using rubrics designed specifically for each role-play scenario. Scoring rubrics for each role-play scenario included all of the applicable Microskill steps demonstrated in the SkillFlix videos. We categorized skills as either (1) skills that are practiced throughout the conversation, such as “using a calm tone of voice,” which would earn one point per role-play scenario, or (2) skills that could be utilized multiple times in conversation, such as “using an I-Feel statement,” which we gave one point per usage. Behaviors explicitly discouraged in the videos, such as “using dismissive language,” we assigned negative point values to subtract from the total conversation score per use; we decided a priori that scores could not be negative.
The Giving an Apology role-play scenario rubric consists of 13 possible Microskill steps and six discouraged behaviors that participants could demonstrate during role-plays. Summed role-play scores could range from 0 to 13. The Addressing Hurtful Comments role-play scenario rubric consists of 17 possible Microskill steps with the same six discouraged behaviors. Summed scores could range from 0 to 17.
Before use, the research team tested and refined the rubrics on recorded tests of the role-play scenarios with nonstudy staff. We discussed skill steps not consistently scored by multiple reviewers. If we could not reach a consistent criterion, we eliminated those skill steps for being too subjective. Trained research team members watched and scored recorded videos based on rubrics until inter-rater reliability reached 85%, whereupon one research team member coded the remaining videos. We applied the rubric to each role-play scenario and derived a summed skill score for each participant’s four role-play scenario demonstrations (Giving an Apology range = 0–13; Addressing Hurtful Comments range = 0–17).
Data collection
Pre- and Postintervention Surveys
Once participants were enrolled in the study, the research team emailed them a link to complete the online preintervention survey. At the end of the study period, after participants completed the intervention and their role-play scenarios, study staff sent the postintervention survey via email. All participants completed the postintervention role-plays and surveys.
Pre- and Postintervention Role-Play Scenarios
We designed the role-play protocol to decrease potential participant anxiety and increase comfort with the process. We described the evaluation procedures as role-play scenarios and avoided references to tests or assessments to describe the process in an approachable way. The research team trained three role-play facilitators of varying ages, racial and ethnic backgrounds, and genders to reduce potential anxiety about participating in a role-play with a stranger. The team shared the facilitator’s photographs and personal bios with participants, who then selected their preferred role-play partner. The young adults’ advisors assisted in making this process accessible.
Participants scheduled the preintervention recorded video call with their chosen role-play facilitator and role-played two scenarios. The evaluation team used an online random number generator to assign participants two of the four role-play scenarios (one in each skill set) for preintervention and assigned the other for postintervention. After the intervention, the research team emailed participants a link to schedule the one-on-one video call for the two postintervention role-plays.
Data analyses
Pre- and Postintervention Surveys
We used a nonparametric test for paired data, the Wilcoxon signed-rank test, to compare preintervention with postintervention self-ratings of skills. We have presented responses to acceptability questions as descriptives. For the open-ended qualitative questions, we used content analysis to analyze responses. As described in focus group analyses, we used inductive and deductive coding to analyze participant feedback about the SkillFlix AYA resource. The research team created and reviewed a codebook, and a single research team member applied codes to data.
Pre- and Postintervention Role-Play Scenarios
For role-play scenarios, we generated summed scores based on applying predefined scoring rubrics to video clips for each role-play scenario. We compared preintervention role-play scenario summed scores with postintervention role-play scenario summed scores using the Wilcoxon signed-rank test.
Qualitative Synthesis
Similar themes emerged across feedback from focus groups, advisors, and participants. To synthesize qualitative findings from this project and present them in a concise and digestible form, we created a data display that included themes from each separate qualitative analysis to view them together. We then sorted those themes into overarching categories. We compared the data (quotes) from project-specific themes to thematic categories as a validity check. We present the thematic categories below.
Results and Lessons Learned
Usability
The website developer used the usability testing results to improve website functionality, making changes such as improving the search functionality to be responsive to a broader array of search terms, adding ways to navigate to the home screen, adding a dark mode for people with light sensitivities, and changing icon images to be more intuitive.
Pilot testing
Pre- and Postintervention surveys
For the skill Giving an Apology, on average, participants showed a significant increase from their preintervention self-rating (Mdn = 5) to their postintervention self-rating (Mdn = 6; z = 2.994, p = 0.003). In the skill Asking for an Apology, on average, participants showed a significant increase from their preintervention self-rating (Mdn = 4) to their postintervention self-rating (Mdn = 5; z = 2.810, p = 0.005). The mean increase for participants’ self-rated skill for Responding to a Hurtful Comment was not significant (preintervention Mdn = 5, postintervention Mdn = 6; z = 1.189, p = 0.235).
Pre- and Postintervention role-play scenarios
Participants increased their demonstration of Microskills between the pre-and postintervention role-play scenarios. In the Giving an Apology role-play scenario, on average, participants showed a significant increase in the number of skills demonstrated from preintervention (Mdn = 2) to postintervention (Mdn = 5, z = 3.046; p = 0.002). In the Responding to a Hurtful Comment role-play scenario, participants showed a significant mean increase in skills demonstrated from preintervention (Mdn = 1) to postintervention (Mdn = 5; z = 3.298; p < 0.001). Both role-play scenarios showed a statistically significant increase in total Microskill steps demonstrated, improving from Mdn = 3 to Mdn = 10 of 26 total possible skills (z = 3.257; p = 0.001).
Acceptability
Participants rated the SkillFlix AYA website as easy to use, professionally presented, educational, and useful (Table 3). In response to the qualitative questions, participants offered that they liked the focus on examples instead of providing a single script. They appreciated the short videos with specifics, the multiple examples of how a conversation could happen, and the ability to watch and think about how they would approach the situation, noting that the content was “helpful and thought-provoking.” Participants appreciated accessibility features, such as the skill steps appearing in writing during the video. One participant noted, “The scenarios I appreciate as examples of others’ thoughts. It was eye-opening to see what people get upset over.”
Acceptability Scores for SkillFlix for Autistic Young Adults (N = 17)
Participants responded using Likert-style scale with range 1–7, 1 being least positive and 7 being most positive. All participants did not complete all items, range n = 16 to n = 17.
Observations from practice and lessons from initial implementation efforts
Through the interviews, focus groups, usability testing, and pilot participant feedback, advisors and participants in every study phase described ways to build and improve resources for AYAs. Six cross-cutting themes emerged from the qualitative data:
Center autistic voices
An important theme expressed repeatedly across all groups was the centering of autistic voices. Participants appreciated the inclusion of autistic advisors, staff, scriptwriters, and actors in developing the SkillFlix AYA video library. One pilot study participant said in their post-test user-satisfaction survey, “Representation is important; you need to keep and grow autistic staff.” An advisor shared, “It’d be important for the scripts to be by people with those actual life experiences.”
Use identity-first language
Focus group participants stressed the importance of using identity-first language and concepts developed by autism self-advocates to celebrate autistic identity. One pilot study participant shared, “There is a lot of content out there where autism has a negative connotation, so when we go to therapy, we talk about things we don’t like about ourselves, or others don’t like—always highlighting differences and difficulties. It’s important to highlight good traits; it can’t be just negative reinforcement.” Another stated that they liked “…that it was very autism-focused, I really appreciated the use of autism as an identity. Using identity-first language is great.” Another participant stated that researchers should “avoid stereotyping. Put humanity first, have message be for the individual rather than autism!”
Neurotypical behavior is not expected
Pilot study participants appreciated the lack of pressure to behave neurotypically within the SkillFlix AYA content. While participants were interested in learning skills to help them better communicate, most were uninterested in materials that encourage suppressing autistic traits. One study participant shared, “I like that there’s nothing to do with Autism Speaks, nothing ABA, or act neurotypical. I don’t want to be told to hide myself.” One study advisor stated, “To me, autism acceptance is when I don’t have to camouflage myself to those around me to fit in. It is when I don’t have to make significantly more eye contact than I am comfortable with or hide my stimming behaviors. It is when I can just be me.”
Expansive view of relationships
Participants requested for more topics and relationship types beyond friends and romantic partners, including classmates, professors, family members, potential romantic interests, and acquaintances. Pilot study and focus group participants expressed interest in transitional phases of relationships that arise when moving from one kind of relationship to another and a desire for understanding relationship dynamics in the workplace. They suggested topics including navigating power dynamics, seeking accommodations, engaging with coworkers, and becoming friends with coworkers. Study participants also suggested a section on skills supporting self-advocacy and community activism. Several participants shared an interest in receiving support for leadership and advocacy skills to help navigate community relationships.
Communication skills for neurotypical friends and family
The young adult advisors were interested in the creation of resources to help the non-autistic people in their lives work on their communication skills with autistic people. They suggested that SkillFlix AYA include videos intended to be viewed and discussed with neurotypical peers/family covering communication needs and ways to “meet in the middle.”
Additional site content
In the postintervention survey, participants made several content suggestions. Participants mentioned topics such as dating, relationships, flirting, how to introduce yourself, maintaining friendships and romantic relationships, expectations about friendships, meeting a partner’s needs, active listening, and talking about “serious matters.” Some requested videos addressing issues at work, or specific skill areas such as interview skills. One participant requested content on navigating power dynamics: “Videos about power and control relationships: teacher, parent, boss. Nonpeer. Negotiating and making compromises.” Some wanted to manage more aggressive or conflictual interactions, suggesting that content about personal safety and dating violence may also be important to include: “…how to deal with someone who is overtly aggressive and doesn’t want to resolve the conflict, other options for aggression when someone doesn’t want to resolve a conflict.”
Several participants asked for informational-style videos to explain the reasoning behind critical skills. Currently, the videos exclusively demonstrate the skills, and many users desired more detailed information on the how and why of key skills. They made requests such as, “So, while the videos did a good job of showing an example of using the concepts, and the text under has a nice summary, I feel like it might be useful to have a version of the video where it pauses and talks about how they are using these skills. Connecting skills to bits in the video. Have another option analyzing how the skill is being used, how it reflects skill. Compare it with someone not succeeding at setting a boundary.” A number of participants called for more written resources and further explanations of skills.
Discussion
We created and tested the SkillFlix AYA streaming video library, designed to teach AYAs skills associated with healthy relationships. These pilot results show that participants exposed to the SkillFlix AYA video library demonstrated increased skills in apologizing and addressing hurtful comments. We also found increases in participant’s self-rated skill levels for giving and asking for an apology. This intervention and study incorporated novel methods: SkillFlix AYA is one of only three social or relationship skill interventions that we are aware of for autistic adults that incorporates video modeling and is one of the first to test the acquisition of skills via structured role-play scenarios rather than focusing on content knowledge or self- or proxy-assessment alone.42–44 The preliminary results for this pilot evaluation suggest that SkillFlix AYA may provide AYAs with easily accessible online support in building relevant, age-appropriate skills that align with their priorities.
Advancing the literature
Our study adds to emerging research supporting the efficacy of video modeling in helping AYAs develop and improve their communication and relationship skills. Researchers have used video modeling and prompting more extensively to develop vocational skills than friendship or romantic relationship skills. Campanero’s review of the vocational literature shows fairly positive results for video modeling and strong positive results for video-prompting interventions that break skills down into task steps (similar to Microskills). 29 However, we could find only two small-scale video modeling/prompting interventions focused on social or relationship skills for autistic adults, indicating a strong need for more resources and research. 43 The first, published in 2012, used a multiple-baseline single-case design (n = 2) and focused on skills such as eye contact, body orientation, and facial expression. They found a significant increase in social-communicative skills for one study participant and an increase in two out of three skills for the other, showing the efficiency with which video modeling can be used as an educational intervention. 44 The second, published in 2017 (n = 3), used a behavioral skills training package focused on cognitive coaching, including understanding thoughts, feelings, and behavior; social cognition; and generalizing these in a social setting. They found that all participants demonstrated an increase in the correct use of skill steps for all categories. 45
These studies incorporated feedback from trained clinicians to help participants improve their skills and foster healthy relationships. Our study is consistent with a body of literature suggesting that such feedback may not be necessary,46,47 meaning that it may be easier to disseminate the intervention since trained clinicians are not required for it to be effective. In addition, our study had input from AYA about the skills they would like to learn and focused on more sophisticated relationship skills rather than basic social skills. This study advances the literature by demonstrating that a set of videos modeling task steps, which AYAs can view in the convenience of their homes and based on their interests, can be effective for developing relationship skills for some autistic individuals.
The short online videos reflect the preferred learning methods for autistic adults, 48 a preference corroborated by our focus groups and study participants. Our participants expressed their desire to be able to search for topics (e.g., going on a date) and find fast-paced, relevant solutions to their questions. Feedback from users indicated that it would be beneficial to expand the SkillFlix AYA video library to address more relationship skills and real-life situations and make user-suggested improvements to the SkillFlix AYA platform. The SkillFlix AYA video library fills a support gap for AYAs and can be widely implemented by therapists, residential facilities, and other support organizations at the individual and organizational levels.
Limitations
The objective of this pilot study was to ascertain feasibility and acceptability, with secondary objectives focused on preliminary efficacy. The main limitations were the small sample size and limited characterization of the sample (e.g., in terms of mental health conditions, educational attainment, employment, and relationship status), which limits our ability to know for whom these results may generalize. Autistic people have a variety of needs in the social, emotional, and communication domains and have different ways of learning, paying attention, or reacting to their surroundings. 49 Recognizing that it is not feasible to create a small pilot resource that meets all these varying needs, we incorporated various Microskills of differing difficulty, emotional intensity, and inter-relational awareness. Additional demographic information about participants’ prepilot social skills and experiences, communication and intellectual functioning, mental health diagnoses, and educational and vocational status would have strengthened our understanding of how the videos worked for this sample of AYAs and may have allowed for greater exploration of how baseline skills and experience correlated to postintervention outcomes.
The pilot also used a pre–post intervention implementation design with a cohort of AYAs rather than a randomized controlled trial (RCT) design, and therefore, we cannot conclude that our intervention caused changes in relationship skills. These results will inform further development of the SkillFlix AYA site and a future RCT to test efficacy.
Another limitation faced during study participant recruitment was the difficulty recruiting Black, Indigenous, or Other People of Color (BIPOC) who are autistic. One examination of demographics of various autistic interventions exposed that ethnic groups are not adequately represented in many studies, “including few Black (6.8%), Asian (6.8%), and Hispanic (2.5%) participants…[and] predominantly White male.” 50 White children have historically been more quickly and accurately diagnosed and utilize more services and organizations compared with Black, Latinx, Pacific Islander, and Native American children.51,52 Because our most reliable method of reaching the autistic community was through organizations and autism is primarily identified in White children, our pilot had limited racial diversity. Therefore, creating and testing a larger video library will require partnering with organizations that serve BIPOC autistic communities. We conducted the entire study during the COVID-19 pandemic, which shifted our approach for focus groups, video production, and usability testing. This impacted video production and required script rewrites and a revised filming strategy to include COVID-free actor pods for face-to-face communication. While we had intended to show a larger selection of in-person live-action videos, we found that some study participants reported liking the virtual format because it narrowed in on detailed facial expressions.
Next steps
Upon completion of the pilot study, the National Institute on Disability, Independent Living, and Rehabilitation Research awarded dfusion, Inc. a grant for Phase II funding for the SkillFlix AYA project to add video content and conduct an RCT evaluation of the video library with 110 participants. The enhanced content and website features reflect the user experience and pilot findings from Phase I participants. The video library resource for transition-age autistic youth addresses a comprehensive array of healthy relationship skills prioritized by autistic advisors. Based on advisor requests and literature detailing the double-empathy problem (i.e., that neurotypical people struggle to understand and communicate with autistic people just as much as autistic people have difficulty communicating with neurotypical people), we have added a collection of skill videos for neurotypical audiences that users of the SkillFlix AYA resource can share with neurotypical friends and family. Our aim is to recognize that when addressing the communication gap between neurotypes, it is important to not only focus on skills autistic people can use to navigate neurotypical social expectations, but also to provide skills for neurotypical individuals to learn about how they can better communicate with autistic people.
This study suggests several opportunities for future research. First, usability testing offered several practical recommendations for making the website more accessible for autistic people. This type of testing is not common in psychosocial intervention development, yet consultation with autistic people during development or before feasibility testing may be a useful step to improve accessibility of intervention materials. Second, expanded neurodiversity-embracing approaches can improve interventions for autistic people. Advisors, focus group members, and participants remarked on the importance of nonpathologizing content that provided examples of how conversations might go rather than emphasizing that there is one correct way to communicate and socialize. Finally, participant input also emphasized the wide range of needs for healthy relationship skills and that everyone begins from a different baseline skill level and has their own goals based on their values and what is currently happening in their lives. Offering a broad array of skills and allowing participants to select the resources that meet their needs are an approach that supports individual autonomy and self-determination and the power of motivated self-directed skill development. Clinicians and intervention developers can integrate more self-directed learning approaches (e.g., offering in-depth written materials on topics beyond what is covered in session).
Supplementary Material
Footnotes
Acknowledgments
The authors thank the advisors, focus group members, participants, usability testers, and actors for sharing their personal perspective and preferences to make this project as useful and applicable as possible. The authors would also like to express their appreciation to Erica Marsh for her work in developing the SkillFlix AYA platform.
Authorship Confirmation Statement
All authors have reviewed and approved the article. The article has been submitted solely to this journal and is not published, in press, or submitted elsewhere. L.E. (corresponding): Methodology (lead), investigation (lead), writing—original draft (lead), and writing—review and editing (support). R.F.-T.: Conceptualization (support), funding acquisition (support), supervision (lead), project administration(lead), and writing—review and editing (support). T.K.: Conceptualization (lead), formal analysis (equal), funding acquisition (lead), and writing—review and editing (support). R.F.: Writing—original draft (support), methodology (support), investigation (support), and writing—review and editing (support). L.G.H.: writing—review and editing (lead) and formal analysis (equal).
Author Disclosure Statement
No competing financial interests exist.
Funding Information
The project was funded by the National Institute on Disability, Independent Living, and Rehabilitation Research contract #90BISA0037-01-00.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
