Positionality Statement

This survey was conducted as part of the first author’s (IC) master’s thesis. IC is a non-autistic researcher focusing on autism health services delivery. CW was one of the Autistic advisors in the Project Advisory Group. She has an M.A. in Critical Disability Studies and has consistently been carrying out a variety of professional roles in autism and accessibility since graduating with her M.A. in 2017. The other authors are all non-autistic scientists who also have clinical training (medicine, psychology, occupational therapy). Together, we acknowledge that discussions about therapies accessed by autistic children are nuanced and can be emotionally laden, and we aim to use this work to respectfully engage with that nuance. Our aim is to identify key values that will inform dialogue about the acceptability of therapies and promote wellbeing among autistic children and their families.

Advisory Group

A Project Advisory Group (PAG) advised on the development, dissemination and analysis of the survey. This committee was composed of four autistic people, three parents of autistic children and youth (recruited from larger advisory groups at the investigators’ institutions) and three clinicians (recruited from the investigators’ networks), all within the province of Ontario. All members received an honorarium for their participation in the PAG. Due time constraints associated with completing a master’s thesis, fully participatory research, where researchers and community members shared equal decision-making, was not feasible. In such cases, advisory groups are an important alternative [28]; however, their use requires clear communication about their role and expectations. The expectations and timelines of the advisory group were communicated to members of the PAG, in addition to potential limitations of the project (e.g., deadlines for the master’s thesis) [2]. Four meetings were conducted with the PAG: two focused on survey development and two on contextualizing results. At the first meeting, PAG members were familiarized with each of the seven constructs of the TFA and led through a discussion about what they believed was important to capture through survey questions for each of the constructs. Following the first meeting, a draft survey was created and distributed to PAG members for their asynchronous review. The second survey meeting focused on a discussion related to these draft questions, and how they could be improved to better capture lived experiences. The survey questions were updated following discussion, and the final draft of the survey, including the informed consent form, was emailed to PAG members to elicit their feedback on survey content, accessibility, ease of understanding, and language. At the two meetings focused on contextualizing the results, a high-level overview of both qualitative and quantitative findings was shared with members to ascertain the extent to which the results aligned with or diverged from their lived experiences. Discussions at the PAG prompted iterative analysis, and returning to qualitative analysis informed by members’ experiences and perspectives. Throughout the project, a text-based meeting platform, one-on-one meetings, and asynchronous participation (i.e., emails rather than meetings) were offered to facilitate participation.

Survey Design

As no validated questionnaires relating to the acceptability of autism services have been published, a novel survey was constructed with input from the PAG. The seven constructs of the TFA were used to structure the survey. The survey was created using REDCap, a secure electronic data capture tool [29], and was pilot tested with an autistic person, a parent, and a clinician for language use and clarity. Ethics approval was obtained through a tertiary pediatric hospital and university, and all methods were carried out following institutional guidelines.

Demographic information (age, gender, race/ethnicity, income, and province/territory of residence) was captured. For each therapy type, a broad definition was provided, and respondents answered a set of questions on a Likert-type scale and free-text questions (survey and related definitions available in Appendix A). For each of the seven constructs of the TFA there were 2–3 questions, with Likert-style response options of Strongly Agree, Agree, Neutral/No Opinion, Disagree, and Strongly Disagree. At the end of each section related to one of the therapy types, respondents were asked two free-text questions: 1) “What do you think the purpose of [therapy type] is for autistic children? and 2) “In your opinion, is [therapy type] an acceptable therapy for autistic children? Why or why not?”. Free-text questions were suggested by the PAG to allow respondents further explain their perspectives if desired. The survey collected information on four therapeutic approaches: SLP, OT, ABA and physiotherapy. Respondents reported low engagement and familiarity with physiotherapy; therefore, this approach is excluded from reporting.

Participants

Parents of autistic children were recruited to complete the survey using social media, snowball sampling, listservs, and newsletters of national and provincial autism organizations. Respondents were asked if their child(ren) received a diagnosis of autism spectrum disorder. Other inclusion criteria were that respondents needed to live in Canada, have access to the Internet, and be able to read and write in English, or access supports (i.e., another individual) that could facilitate survey completion. Importantly, the TFA allows for the evaluation of perceived acceptability from those who have not engaged in the intervention, and we sought parental perspectives irrespective of whether their child had participated in a specific therapy. Informed consent was obtained and documented through the online survey, as respondents had to read and agree to the consent page to participate and navigate to the main survey page. From pilot testing, the survey was estimated to take 20–30 minutes to complete. At the end of the survey, respondents had the option entering their email address to be entered into a raffle for one of 20 $50 gift cards, and their email addresses were not linked to their responses. Recruitment took place between January 27^th 2022 and August 25^th 2022. No measures were included to prevent potential fraudulent responses to the survey [30].

Data Analysis and Integration

A participant’s responses were included in the final analysis if they completed the Likert questions for at least one of the included therapies (16 questions). Descriptive statistics were calculated for the demographic data. For each therapy type, a composite sub-score was created based on responses to the scale questions; the highest possible sub-score (most acceptable) was 64 and the lowest sub-score (least acceptable) was zero. Creating a composite score from Likert data is supported when measuring more abstract constructs like acceptability [31]. A measure of internal consistency, Cronbach’s alpha, was calculated for each therapy sub-score, which is recommended when producing a composite score from survey data [31]. The alpha coefficient is between zero and one, and describes the interrelatedness of items of a survey [32]. Coefficients closer to one are indicative of higher internal consistency. The Cronbach’s alpha for OT was 0.87, for SLP was 0.85, and for ABA 0.88, all representing a satisfactory measure of internal consistency.

Guided by the seven TFA constructs, free text responses were analyzed following deductive content analysis [33]. NVivo was used to collate and organize responses. Deductive content analysis is useful when seeking to extend or validate an existing framework like the TFA. Content analysis includes three phases: preparation, organizing and reporting [34]. In the preparation phase, the free-text responses were read and annotated by the researcher (IC). In the organizing phase, responses were categorized based on the the TFA, with subcategories created based on the included quantitative survey questions. IC and MP met biweekly to discuss coding and the analysis. As previously described, qualitative analyses were reported to the PAG for discussion and member checking, which prompted further engagement and analysis of the data in order to enhance credibility [35].

Following a mixed methods approach, the qualitative and quantitative strands were collected concurrently, and their results integrated through triangulation [36,37]. Narrative synthesis (weaving together qualitative and quantitative findings) [38] and joint displays, which are visual representations of qualitative and quantitative data integration, were used to integrate qualitative and quantitative strands [38]. Illustrative quotes from qualitative analysis are included within the joint displays.

Participant Characteristics

The survey was opened by 101 potential respondents; of those, 87 parents completed the demographic information and at least one set of the scale questions and were included in the analysis (attrition rate: 13.9%). As multiple sampling methods were used, the response rate cannot be calculated. There was no systemic screening completed for evidence of fraudulent responses; however, there were no incoherent responses nor responses submitted in close succession. Nearly 20% of parent respondents (n = 16) identified that they had more than one autistic child, and 12 parents self-identified as autistic (13.8%) (Table 1). Most parent respondents had children who engaged with OT (85.1%) or SLP (82.8%). Fewer had engaged in ABA (47.1%). The ages of parent respondents are visualized using a histogram (S1 Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive Statistics of Parent Respondents (n = 87).

https://doi.org/10.1371/journal.pone.0353678.t001

Acceptability Scores

S2 Fig provides a boxplot of the composite perceived acceptability sub-scores for respondents. The median SLP acceptability sub-score was 41.5 (Inter-quartile range (IQR) = 10.5, n = 78), the median OT acceptability sub-score was 40 (IQR = 13, n = 87), and the median ABA sub-score was 28 (IQR = 14, n = 77). Significant between group differences were identified using one-way analysis of variance tests [F (2, 239) = 43.4, p < 0.01]. Among those who identified accessing a therapy, the median acceptability scores were 42 (IQR = 10.25, n = 64) for SLP, 40.5 (IQR = 11.75, n = 74) for OT, and 30 (IQR = 18, n = 35) for ABA. There were no significant group differences between participants who had access to therapy and those who did not for both OT and SLP; however, this difference was significant for ABA (t = 3.36, p < 0.01). The median acceptability score for parents whose children had not accessed ABA was 25.5 (IQR = 12.75, n = 42).

Narrative Synthesis

Qualitative analysis identified codes predominantly in three TFA constructs (Ethicality, Perceived Effectiveness, and Self-Efficacy), with few to no codes identified under other TFA constructs. Greater analytic attention was paid to these three constructs as they generated richer responses with greater engagement of respondents, thus suggesting their strong resonance. These are described in more detail below along with Likert responses. Likert responses for other TFA domains are available in Appendix A.

Ethicality

The TFA defines Ethicality as “the extent to which an intervention has good fit with an individual’s value system” [18]. For OT and SLP, most parent respondents selected ‘Agree’ or ‘Strongly Agree’ that the goals (OT: 78.2%, SLP: 85.9% ‘Agree’ or ‘Strongly Agree’) and methods (OT: 75.9%, SLP: 83.4% ‘Agree’ or ‘Strongly Agree’) of these services were in line with their value systems (Table 2). For ABA, 31.2% of parents selected ‘Agree’ or ‘Strongly Agree’ that the goals of this service were aligned with their value systems, and 29.9% selected ‘Agree’ or ‘Strongly Agree’ when asked if the methods of ABA aligned with their personal value systems.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Joint Display of Parent Perspectives on Ethicality, n (%).

https://doi.org/10.1371/journal.pone.0353678.t002

Analysis of free-text responses identified three main topics under the TFA construct of Ethicality: 1) concerns about potential harms or negative experiences (particularly related to behavioural goals and principles), 2) the importance of neurodiversity-affirming services, and 3) mixed views on behavioural principles.

The first topic, potential harms or negative experiences, included possible negative occurrences and outcomes from therapies. Across all three therapy types, parents provided examples of negative experiences they had witnessed (e.g., anxiety, trauma, distress) or wrote about harms they had learned about from autistic adults. One parent shared that for OT to be acceptable, it must draw on the child’s strengths and promote autonomy:

Yes if it is strength based and not looking to override a sense of self or autonomy/consent for the goal of compliance as a benchmark of success.

Nested within potential harms, some respondents specifically critiqued what they perceived to be the harms that could be caused by ‘behavioural goals or principles’. Some parent respondents were critical of goals that promoted neurotypical behaviours, such as making the child appear ‘normal’ or encouraging them to fit into neurotypical society. For example, one parent wrote that the purpose of ABA from their view was:

To train autistic children in socially acceptable behaviour so other people aren’t made uncomfortable by [them] and they fit in better.

This parent’s response suggests an ethical tension related to changing a child’s behaviour to make neurotypical people more comfortable. Other parent respondents described potential harms of aiming to change behavior through therapies, which included disempowerment of the child, removal of self-soothing or coping mechanisms, trauma, and loss of happiness or sense of self. Some parents identified concerns specific to use of behavioural approaches, such as rewarding a child for displaying a specific behaviour (external reinforcement). When asked about the purpose of ABA, one non-autistic parent wrote:

I believe the purpose of ABA is to train [a] child using extreme positive reinforcement. It attempts to change a child. Adults who underwent ABA often consider it abusive.

This parent was influenced by the reports of autistic people who experienced ABA therapy. Ideas of ‘training’ and ‘changing’ were perceived as harmful.

The second topic of the importance/value of neurodiversity-affirming services highlighted what affirming practices for neurodivergent children should look like. As one autistic parent wrote about OT:

Yes, it helps them accomplish goals by teaching new skills. It is not focused on suppressing who they are, coercing them or forcing them to change to make other people more comfortable even if there are huge mental health costs to the child.

About the acceptability of SLP, one non-autistic parent shared:

Hopefully this is done more in a neurodiversity affirming model which doesn’t go to remove autistic traits but aid in communication and well-being of [the] individual.

Other respondents described that incorporation of alternative and augmentative communication devices in SLP aligned with neurodiversity-affirming approaches by honouring different communication methods.

Many respondents viewed neurodiversity-affirming approaches to therapy as aligning with their personal ethics. Respondents valued accepting the child for who they are, providing accommodations, and not attempting to reduce observable autistic traits.

The last topic under the construct of Ethicality reflected mixed views on behavioural principles, particularly around the goal of ‘behaviour change’ and the use of behavioural approaches in therapies. Respondents had differing opinions on whether these goals and approaches aligned with their personal values and beliefs. Some viewed behavioural approaches, often discussed in relation to ABA, as effective for teaching new skills and reducing ‘negative’ behaviours, such as task refusal, self-harm/injury, and elopement, whereas others questioned their appropriateness or ethicality. One non-autistic parent described that the purpose of ABA was:

To help mitigate negative behaviours and building new skills, to gain the skills to lead an independent life.

Respondents also identified a broad set of skills, including toileting, putting on shoes, communication, joint attention, play, social skills, motor development, and sensory integration that could be learned through ABA. Conversely, some parents were critical of behavioural goals that promoted neurotypical behaviours, such as making the child appear ‘normal’ or encouraging them to fit into neurotypical society, as described above under potential harms and negative experiences.

Overall, parent respondents identified that Ethicality as a key aspect of acceptability of autism services. Respondents perceived that therapies could cause harms, and many discussed the importance of neurodiversity-affirming approaches. Respondents presented mixed views on the use of behavioural principles to reach goals.

Perceived Effectiveness

This construct is defined as “the extent to which the intervention is perceived as likely to achieve its purpose” [18]. S3 Fig presents a visual integration of qualitative and quantitative responses. For OT, SLP, and ABA, most parent respondents selected ‘Agree’ or ‘Strongly Agree’ for the statement “I think that [therapy type] can teach new skills to autistic children”. The other statement was “The skills that are emphasized in [therapy type] are not meaningful or impactful in the daily lives of autistic children”. Over 75% of parent respondents selected ‘Disagree’ or ‘Strongly Disagree’ for this statement for OT and SLP services, while comparatively fewer (30%) selected ‘Disagree’ or ‘Strongly Disagree’ for ABA.

Two topics within the construct of Perceived Effectiveness were identified in the parent free text responses: importance of seeing visible impact/progress and therapist fit. Some parents described seeing notable successes of therapies through goal achievement (e.g., learning practical skills) or improved child well-being (e.g., growing independence, reduced anxiety). For example, one non-autistic parent wrote about the perceived effectiveness of therapies for their child:

As controversial as ABA therapy is, its rapid effectiveness with parent involvement is apparent when compared to reaching similar goals in OT and ST [speech therapy] settings. We did OT and ST 2x/week before ABA with very slow or no results. Of course, this wouldn’t be the same for every child.

This quote suggested that the respondent placed value on the relative speed of goal attainment. Other parents did not notice progress or changes for their child. For example, one parent wrote:

Therapies are expensive and I often hear parents get discouraged and say therapies are ineffective.

Thus, some parent respondents described being discouraged from continuing a therapy or stopping it altogether because the therapy was expensive and lacked a perceptible impact.

The second category was the importance of therapist fit. Across all therapy types, some parents expressed that the acceptability of a service was dependent on the therapist delivering it. There were few responses that spoke to what constituted a strong ‘fit’ with families, which expressed the role of positive, respectful relationships. Responses in this category highlighted that parents’ views of acceptability were contingent on how they viewed the therapist and their approach, exemplified in the quote below about SLP services:

If they (SLP and client) do not mesh well nothing gets accomplished.

Another parent who had also engaged with SLP services for their child provided a more detailed experience with their therapist:

It is also tricky to know for all…therapies if my personal experiences were related to the therapy itself or if my experience was more related to the therapist. For example, I had SLPs that worked with my son that didn’t seem to pay any attention to our family’s goals. They were naive at best and mean at worst.

In summary, the perceived effectiveness, demonstrated both by progress attained through a therapy and the fit with the therapist was identified as an important feature of acceptable therapies by parent respondents.

Self-Efficacy

This construct is defined by the TFA as “the participant’s confidence that they can perform the behaviour(s) required to participate in the intervention” [18]. Respondents were asked two Likert questions related to this construct. The first question asked respondents to indicate their agreement that the expectations for children’s participation in the therapy are reasonable. For OT and SLP, most respondents selected ‘Agree’ or ‘Strongly Agree’; for this statement (78.1% n = 68, and 76.6%, n = 60, respectively). For ABA, there was more variation in the level of agreement; almost 30% of respondents selected ‘Agree’ or ‘Strongly Agree’, and almost half of respondents selected ‘Disagree’ or ‘Strongly Disagree’. The second self-efficacy question asked respondents to indicate their level of agreement as to whether parents could complete tasks as directed by a therapist. More than half of respondents selected ‘Agree’ or ‘Strongly Agree’ for OT and SLP. For ABA, nearly half of respondents selected neutral for this statement, and nearly 40% selected ‘Agree’ or ‘Strongly Agree’ (S4 Fig).

The qualitative category under this construct was the importance of therapists/practitioners engaging with parents. Responses emphasized the importance of teaching parents or helping them understand how to support their child, particularly in areas such as their communication strategies. As one parent described about the purpose of OT:

I was able to implement these techniques to help my child make improvements in the global gross motor skills delay.

This category spanned all therapies. One parent specifically identified their desire for more parent engagement in the SLP services they received. Other parents also described the role of therapists in helping them understand their child’s needs and equipping them with the tools to support their children. As one parent shared about the goal of SLP approaches:

To help parents get different tools to communicate with their kid.

Both examples demonstrated parents’ beliefs about the importance of a good fit between the child, family, and therapist for a therapy to effectively and acceptably achieve its intended purpose. In summary, parents’ perspectives on their self-efficacy with respect to autism services was influenced by working with therapists who effectively engaged them, helping them to understand their child and their needs, and enabling their adoption of techniques beyond designated clinical visits.