Improving parents’ ability to advocate for services for youth with autism: A randomized clinical trial

Abstract

Youth with autism face challenges accessing services as they transition to adulthood. Improving parents’ ability to advocate for services on behalf of their youth may be an effective way to improve service access and ultimately transition outcomes in this group. In this study, we tested whether participating in an advocacy intervention improved parents’ ability to advocate for services for their transition-aged youth with autism. 185 parents of youth with autism ages 16–26, recruited across three states in the U.S., were randomized to one of two experimental conditions. The treatment condition received the ASSIST program, a 12-week (24-hr) group-based intervention. The control condition received the same written materials as the treatment condition. Primary outcomes for this report – parent advocacy ability – were collected at baseline (prior to randomization) and post-test (immediately after the treatment group finished the 12-week program) by survey. After taking ASSIST, the treatment condition had greater gains than controls in knowledge of adult services (B = −1.62, CI = −2.33 to −0.90) and perceived advocacy skills (B = −0.19, CI = −0.33 to −0.04). Participants who had less knowledge, lower perceived advocacy skills, and less active coping styles at baseline had the greatest treatment gains. Findings suggest that ASSIST is effective in improving parent advocacy ability and may be most beneficial for parents who experience greater challenges advocating for their son/daughter with autism. Future research will examine whether gains in parent advocacy ability leads to improvements in service access and post-school outcomes for transition-age youth with autism.

Lay Summary

Disability services can be important tools to improve the lives of autistic youth, but they are often difficult to understand and access. This study tested whether a services advocacy program improved parents’ ability to advocate for services on behalf of their son or daughter with autism. The study findings suggested that parents who took the program were better prepared to pursue disability services for their transition-aged youth.

Introduction

Over the past three decades, the prevalence rate of autism has risen dramatically, with approximately 1 in 44 children now identified with this condition (Maenner et al., 2021). As a result, with each passing year increasing numbers of these youth exit high school. The transition from high school into adulthood is a particularly challenging time for youth with autism, with many experiencing difficulties in employment, post-secondary education, independent living, mental health, and social relationships (Hollocks et al., 2019; Orsmond et al., 2013; Shattuck et al., 2012; Taylor & Seltzer, 2012).

Though youth with autism often require assistance in employment and activities of daily living, at transition they encounter an adult service system that is inadequately funded. For example, the number of adults with intellectual and developmental disabilities on waitlists for home and community-based services has tripled from 2005 to 2017 (The Case for Inclusion 2020: Key Findings Report, 2020; The Case For Inclusion Report: 2019, 2019). However, lack of funding for services is only one challenge. Even when one is granted eligibility, the service system offers many unconnected locations for services related to employment, independent living, and post-secondary education. Understanding each agency’s eligibility scheme, funding mechanism, and programming can be a significant barrier to service access. Failure to access needed services can have serious ramifications (Roux et al., 2021; Taylor & Mailick, 2014; Schott et al., 2021), and few evidence-based interventions are available to improve transition outcomes for these youth (Interagency Autism Coordinating Committee (IACC), September 2017).

Effective parental advocacy may play a critical role in ensuring youth with autism have access to services. Advocacy can be defined by speaking or acting on behalf of another person or group of people (Wolfensberger, 1977), and can encompass a broad range of activities such as direct help to another person, pursuing education to learn about an area in which one would like to effect change, or working to change systems. Interventions aimed at improving parent’s ability to advocate on behalf of their son or daughter have been shown to increase the initiation of and amount of mental health services received by children with mental health problems (Koroloff et al., 1996; Kutash et al., 2011). Parents of children with mental health problems who reported greater empowerment in areas such as advocacy and knowledge were more satisfied with the services their families were receiving (Resendez et al., 2000). Among young adults with autism, (Lee et al., 2022) found that the number of advocacy activities that parents reported engaging in was highly associated with the number of services their sons and daughters were receiving. In fact, parent advocacy activities accounted for nearly as much variance in service access as youth functioning (intellectual disability status, autism symptom severity) and family socio-demographic characteristics (parent race/ethnicity, household income) combined. Thus, preparing parents to better advocate for services on behalf of their autistic transition-aged youth may be a promising mechanism to improve service access and ultimately youths’ transition outcomes.

The potential promise of parent advocacy in improving outcomes for youth with autism is further supported by the success of other interventions aimed at improving advocacy ability among families of children with disabilities. One example is the Volunteer Advocacy Project, a 36-hour special education advocacy program teaching parents (primarily) how to advocate for their children with disabilities and for other families. The Volunteer Advocacy Project has been shown to increase self-reported advocacy ability in multiple studies (Burke, Goldman, et al., 2016; Burke, Goldman, et al., in press). Other advocacy trainings have been shown to increase self-reported advocacy ability in parents of Latino children with autism (Burke, Magaña, et al., 2016) and Black families of children with autism (Pearson & Meadan, 2021; Pearson et al., 2022). Of note, all these advocacy programs are focused on special education policy, which pertain to school-based services and supports. To our knowledge, there are no studies of programs (outside of our pilot work described below) that teach parents to navigate adult service systems on behalf of their youth with a disability.

In response, we developed an intervention to support parents in advocating for services on behalf of their autistic youth who are transitioning to adulthood (Taylor et al., 2017b; Taylor et al., 2022). Findings from a pilot randomized controlled trial (RCT) suggested that participating in the intervention increased parents’ ability to advocate on behalf of their son or daughter with autism (Taylor et al., 2017b), leading to improved service access and post-school outcomes (Taylor et al., 2017a). Though promising, this pilot intervention was only relevant to the adult service system in the state in which it was developed (as adult service systems differ from state to state). Thus, we adapted the program – called ASSIST or Advocating for Supports to Improve Service Transitions – to be relevant across states and localities in the United States. Taylor and colleagues (2022) found that ASSIST could be implemented by community providers with fidelity at study sites in three states in the U.S. and was highly feasible and acceptable to study participants.

The objective of the current study was to examine, using a multi-site RCT design, whether participating in ASSIST led to improvements in advocacy ability for parents of transition-aged autistic youth. Adopting an experimental therapeutics approach (Insel & Gogtay, 2014), parent advocacy ability was hypothesized as the mechanism of intervention (or treatment target) by which ASSIST would impact service access and post-school outcomes for youth with autism; thus, in this study we examined whether that treatment target was engaged as a result of the intervention. Based on our pilot work (Taylor et al., 2017b), we hypothesized improvements in three indicators of parent advocacy ability: knowledge about adult disability services, perceived advocacy skills, and empowerment.

We also investigated who was most likely to benefit from the program, by testing whether ASSIST was more or less effective in improving parent advocacy ability depending on the timing of when it was delivered (before or after high school exit) and characteristics of the autistic youth (intellectual disability, autism symptom severity). We hypothesized that ASSIST would be more effective in improving advocacy ability for parents of youth with more significant impairments, as they are often eligible for more services (Taylor & Henninger, 2015). We also tested whether characteristics of the parent (race/ethnicity, education, mental health, coping styles) impacted intervention effectiveness. Outside of race/ethnicity and socioeconomic status, parent characteristics are rarely considered as moderators of treatment effectiveness – even for parent-mediated treatments or programs aimed at equipping parents to better support their children. In their review on effects of early intervention for children with autism, Estes and colleagues (2019) discussed the importance of considering parent characteristics as moderators of intervention success and point specifically to the potential role of problem-focused coping in increasing parental efficacy. Studies also suggest that parental mental health may be an important moderator to consider; in a review of studies of attention deficit hyperactivity disorder (ADHD) treatments, children whose mothers were experiencing elevated depressive symptoms had attenuated responses to treatment (Hinshaw, 2007). Further, our research suggests that youth with autism whose mothers have more (versus fewer) depressive symptoms may be at higher risk for instability in work and educational activities after high school (Taylor & DaWalt, 2017). For the present study, we hypothesized that parents with fewer mental health symptoms and more problem-focused coping styles would be most likely to benefit from ASSIST.

Method

Participants

Eligibility criteria for the RCT were as follows: (a) parent or legal guardian of a youth with autism who was 16–26 years old; (b) parent provided documentation that the youth had received an autism diagnosis from a school-based or medical professional; (c) parent lived in a state in which the study was conducted (IL, TN, WI) and was able to attend the 12-week group program at the intervention site on the day and time it was delivered; (d) youth had lifetime scores on the Social Communication Questionnaire (Rutter et al., 2003) – a parent report screener for autism – of 12 or greater, indicating a likelihood of autism (a cut-off of 12 is recommended for samples that include adults with intellectual disabilities; Brooks & Benson, 2013). We required a diagnosis of autism because youth without an autism or developmental disability diagnosis would likely not have access to many of the services discussed in ASSIST. We included families of those who had and had not left high school to explore when a program about adult services is most effective.

Families were recruited through autism studies, research registries, disability agencies, school personnel, and autism support groups. To reach as wide a group as possible, multiple recruitment methods were used including targeted emails, dissemination of recruitment materials through various newsletters and email distribution lists, mailed letters, phone calls to families on research registries, and word of mouth.

The CONSORT flow diagram is shown in Figure 1. One hundred eighty-five participants were recruited across study sites (IL = 61, TN = 63, WI = 61); 91 were assigned to the treatment condition and 94 to the control condition. Of the 185, 176 (95%) were retained at post-test. Demographic characteristics by experimental condition are presented in Table 1. Parent participants averaged around 50 years of age and were primarily female. There was limited diversity in parent race or education (see Table 1). Youth in both conditions averaged 19.5 years of age and were primarily male. About 60% were still in high school (versus having already exited) and 40% had a co-occurring intellectual disability. There were no statistically significant differences between treatment and control in demographic variables or baseline measures of parent advocacy ability.

Figure 1.

Consort flow diagram

Table 1.

Baseline characteristics by experimental condition

	Treatment N=91	Control N=94
Cohort
Cohort 1	48% 44⁄91	46% 43⁄94
Cohort 2	52% 47⁄91	54% 51⁄94
Site
TN	36% 33⁄91	32% 30⁄94
IL	32% 29⁄91	34% 32⁄94
WI	32% 29⁄91	34% 32⁄94
Parent characteristics
Age in years	51.85 (6.63)	50.24 (5.91)
Gender
Male	12% 11⁄ 91	9% 8⁄ 94
Female	88%80⁄ 91	92% 86⁄ 94
Race
White	82% 75⁄91	85% 80⁄94
Black/African American	10% 9⁄91	7% 7⁄94
Asian	3% 3⁄91	3% 3⁄94
More than one race	2% 2⁄91	2% 2⁄94
Other	2% 2⁄91	2% 2⁄94
Hispanic/Latino ethnicity	8% 7⁄91	7% 7⁄94
Highest level of education
Less than high school	0% 0⁄ 91	0% 0⁄ 94
High school degree (or equivalent)	4% 4⁄ 91	5% 5⁄ 94
Some college	17% 15⁄ 91	14% 13⁄ 94
Associates degree	12% 11⁄ 91	9% 8⁄ 94
Bachelor’s degree	37% 34⁄ 91	40% 38⁄ 94
Master’s degree	26% 24⁄ 91	25% 23⁄ 94
Ph.D. or Professional degree (JD, MD, etc.)	3% 3⁄ 91	7% 7⁄ 94
Household income	5.29 (2.62)	5.51 (2.51)
Youth characteristics
Age in years	19.47 (2.85)	19.47 (2.59)
Gender
Male	76% 69⁄91	73% 69⁄94
Female	24% 22⁄91	25% 23⁄94
Non-binary gender or other	0% 0⁄91	1% 1⁄94
I am not sure / I prefer not to respond	0% 0⁄91	1% 1⁄94
In high school	59% 54⁄91	61% 57⁄94
Has intellectual disability	40% 36⁄91	40% 38⁄94
Autism symptom severity	69.40 (8.33)	67.88 (8.14)
Retention at follow-up	97% 88⁄ 91	94% 88⁄ 94

Note. For age, household income, and autism symptom severity, means are presented with standard deviations in parentheses. A score of “5” on household income corresponds to a range of $80,001 to $100,000 per year, and a “6” corresponds to $100,001 to $125,000 per year. Autism symptom severity is measured by the total t-score from the Social Responsiveness Scale-2.

Procedures

The trial is registered in ClinicalTrials.gov, # NCT04173663. This was a multi-site, single-blind parallel-group study conducted in the U.S. (3 sites). All study procedures were approved by the Institutional Review Boards at the participating study sites (Vanderbilt University Medical Center, University of Illinois at Urbana-Champaign, University of Wisconsin at Madison). Participants were recruited from December 2019 to November 2020. After recruitment, screening, and informed consent, baseline data were collected from parents and youth with autism via interview and questionnaire. Interviews were conducted in-person at the respective universities/clinics at the start of the study, and then via web-conferencing once social distancing restrictions were implemented due to COVID-19. After completing baseline data collection, participants were randomized to either the treatment or control condition (1:1 ratio) using a computer-generated list of random numbers created by a biostatistician (who was not involved in data collection or allocation to experimental condition). To encourage equivalent distribution of factors known to impact service access (Shattuck et al., 2011; Taylor & Henninger, 2015), block randomization was used within each site stratified by whether the autistic youth was in high school and had an intellectual disability. Study personnel collecting baseline data were not aware of future group assignments.

Participants were treated in six cohorts that started in February 2020 (2 cohorts), August 2020 (3 cohorts), or November 2020 (1 cohort). After each cohort finished the ASSIST intervention, all participants in that cohort were emailed a link to an online post-test survey that included measures of parent advocacy ability. Participants had six weeks from the end of the ASSIST series to complete the post-test. Though participants were not blinded, research staff facilitating completion of the surveys were blinded to experimental condition.

Experimental Conditions

The treatment condition participated in the ASSIST program (see below for description). Though ASSIST was intended to be delivered in-person in the community (and began this way), social distancing restrictions due to COVID-19 were implemented mid-way through the intervention series for the first two cohorts. In response, ASSIST was moved to synchronous web-conferencing, and this method was the mode of delivery for all remaining sessions and series. Each week, handouts and resources that would be covered in the next week’s ASSIST session were mailed to all participants, which they assembled in a binder provided by the research team.

The control condition was an information control; participants received the same ASSIST binder and written materials as the treatment condition on the same schedule. Thus, both conditions had access to identical written information, but the treatment condition also had access to explanation of the materials and group-based discussion with local experts and other parents. After a 12-month waiting period, control condition participants had the opportunity to take the ASSIST program.

Intervention: ASSIST

Full information on the development, content, and implementation of ASSIST can be found in a prior publication (Taylor et al., 2022). ASSIST is a group-based 12-week program that teaches parents about the adult disability service system, how the services might benefit their autistic youth, and the best ways to access services and supports. ASSIST is led by a program facilitator with experience in adult services and in leading groups. In this study, program facilitators were members of local disability organizations with support provided as needed by the university team. Each weekly session focuses on a different aspect of adult services and supports and lasts two hours (24 hours total). Topics include person-centered planning, models of decision making, supplemental security income, social security disability insurance, health insurance, Medicaid waiver, employment, post-secondary education, housing, special needs trusts, and advocacy. An optional 13^th session on secondary transition planning can be delivered to participants if relevant (i.e., if autistic youth are still in high school). A full list of session topics is included in Table 2.

Table 2.

ASSIST topics and purpose of each session

Week	Topic	Session Purpose
1	Introduction to ASSIST & person-centered thinking and planning	An introduction to the course is covered in the video, with time for questions. Next, discuss the importance of person-centered thinking and planning. Definition, rationale, and implications are covered in the video. A local expert will discuss the types/methods of person-centered planning.
2	Models of decision-making	Discuss different models of decision-making including guardianship/conservatorship, power of attorney, and supported decision-making. Definition of options are covered in the video. A local expert will discuss the specifics of legal options, the process for establishing formal arrangements, and pros and cons of each option.
3	Supplemental Security Income (SSI)	Discuss the potential benefit of SSI. Definition, eligibility, and implications are covered in the video. A local expert will discuss the application process, details about benefits and work, and local resources.
4	Social Security Disability Income (SSDI) & Supplemental Nutrition Assistance Program (SNAP)	Discuss the potential benefits of SSDI and SNAP. Definition, qualifications and implications are covered in the video. A local expert will discuss the application process, eligibility, details about benefits, and local resources.
5	Health insurance	Discuss different health insurance options. Definition, eligibility, and implications of having Medicaid, Medicare, and private insurance are covered in the video. A local expert will discuss the state-specific Medicaid program, Medicare Parts for different types of coverage, access to private insurance, pros and cons of options, and application processes.
6	Medicaid waiver	Discuss Medicaid waiver programs. Definition, eligibility, and implications of Medicaid waiver programs are covered in the video. A local expert will discuss the state-specific waiver services, eligibility criteria, and application process.
7	Employment Part 1	Discuss employment options. Types of employment experiences and how to take a person-centered approach to seeking employment are covered in the video. A local expert will discuss employment training programs and ways to find meaningful employment.
8	Employment Part 2	Discuss Vocational Rehabilitation (VR) services. Definition, basic eligibility, and how to be prepared to apply are covered in the video. A local expert will discuss the VR application process, eligibility, benefits, dispute resolution, and local resources including Workforce Innovation & Opportunity Act.
9	Post-secondary education	Discuss post-secondary education options. Definitions, eligibility, supports, and implications are covered in the video. A local expert will discuss specific college programs, the Americans with Disabilities Act and Section 504 of the Rehabilitation Act, and local and federal funding resources.
10	Housing, independent living, & enabling technology	Discuss different housing options and technology. Options, eligibility, and implications are covered in the video. A local expert will discuss the application process, creative housing, other housing options in state as well as smart home options.
11	Special Needs Trusts & Achieving a Better Life Experience (ABLE) accounts	Discuss financial planning. Definition, eligibility, and differences between special needs trusts and ABLE accounts are covered in the video. A local expert will discuss the implications of different types of trusts, limits, and application process.
12	Advocacy	Discuss the importance of advocacy. Definition, steps, and empowering the young adult as a self-advocate are covered in the video. A local expert will discuss local advocacy organizations, and ways to practice individual and systemic advocacy.
13	Secondary transition	This is an optional session that may be relevant for groups where many of the youth with autism are still receiving school services. Discuss secondary transition. Assessments, goals, services and roles in Individualized Education Program (IEP) meetings are covered in the video. A local expert will discuss transition plans, diploma options, and local services.

Because service systems differ in each state, ASSIST includes standardized information that is relevant across the United States, while also allowing for localization to be sure that group members are getting the relevant information for their geographic area. Specifically, each session begins with an icebreaker relevant to the topic of that session as well as group introductions (Week 1), or a recap and discussion of the previous week (Weeks 2–13). Next, the program facilitator introduces and plays a 10–15-minute video for that session, which includes an introduction to the topic and background information. After taking time to answer questions about the video, the program facilitator introduces the “local expert” for that session, who is a guest speaker with expertise in the topic at hand. Local experts are typically from disability organizations, state agencies, adult service providers, or attorneys with expertise in disability-related law. The local expert discusses geographically specific aspects of the topic, following a set of learning objectives that remain constant across all sites. A full set of learning objectives for the introductory videos and the local expert presentations can be found in Taylor and colleagues (2022). The session ends with time for discussion and questions.

At each session, participants are given “tips sheets” that present key points for each session with national and local resources. As they continue through the ASSIST series, participants complete a letter of intent regarding their son/daughter. The letter is broken up by sections that correspond to ASSIST sessions. Families are asked to complete the relevant section each week, thereby exiting ASSIST with a planning tool to facilitate continuity of care for their youth.

Analysis of implementation outcomes suggested that ASSIST is feasible, acceptable, and can be delivered with fidelity. As detailed in Taylor and colleagues (2022), treatment fidelity was good, with over 85% of fidelity items and 95% of local expert learning objectives met for nearly all sessions. Participant attendance and acceptability were high; 85% of participants attended at least three-quarters of the sessions, and group members reported being highly satisfied with the intervention. Importantly, implementation outcomes were similar across study sites (Taylor et al., 2022).

Measures

Indicators of parent advocacy ability at post-treatment – the focus of this report – were pre-specified as primary outcome measures (NCT044173663) and are hypothesized to be the mechanism by which ASSIST impacts services and post-school outcomes for autistic youth.

We collected three indicators of parent advocacy ability at baseline and post-treatment. We measured parents’ knowledge about the adult service system through 22 multiple-choice questions asking factual information about adult disability services. This scale was based on instruments used in our related research (Burke, Goldman, et al., 2016; Taylor et al., 2017b). Each response was coded as “correct” (1) or “incorrect” (0), with potential scores ranging from 0 through 22. Example questions include “Which of the following automatically qualifies your child for Medicare?” and “What does substantial gainful activity mean?” Floor and ceiling effects were avoided in this sample, with ranges from 3–20 (Mdn = 11, IQR = 9–13) at baseline and 5–21 (Mdn = 13, IQR = 11–15) at post-test.

We assessed parents’ perceived skill in advocating for their son/daughter with autism with a 10-item measure that has been used in our previous work (Burke, Goldman, et al., 2016; Taylor et al., 2017b). Participants responded to each item on a 5-point Likert scale and items were averaged to get an overall score of parents’ perceived advocacy skills. Example items include, “How able are you to effectively communicate with providers/agencies/professionals?” and “How able are you to apply your rights and knowledge of the laws in navigating the adult service system?” Internal reliability was high at baseline and post-test (Cronbach alphas = .88 and .91).

Parental empowerment was measured using the Family Empowerment Scale (Koren et al., 1992; FES). The FES measures the extent to which parents feel empowered across three dimensions: family; the service system; and the larger community and political environment. Item scores are summed to form an overall score. Reliability and validity of this measure have been established (Koren et al., 1992). We slightly modified the wording of the items to be relevant to parents of young adults (versus parents of children, for whom the measure was originally developed). For example, instead of “I believe I can solve problems with my child when they happen,” the modified version says “I believe I can solve problems with my son/daughter when they happen.” Good internal consistencies were found in this sample for the slightly modified version (Cronbach’s alphas of .93 and .94 at baseline and post-test), which has also been successfully used in our previous research (Taylor et al., 2017b).

Demographic and behavioral measures

were collected at baseline from parents and youth with autism to characterize the sample and serve as effect modifiers. Youth information included gender (coded as male vs. other for analysis, given the preponderance of males) and whether they had exited the school system. Autism symptom severity was measured using the total t-score from the Social Responsiveness Scale-2 (Constantino & Gruber, 2012).

Co-occurring intellectual disability was determined by consensus using all available information at baseline. This information included: the adaptive behavior composite score from the Vineland Scales of Adaptive Behavior-III (Sparrow et al., 2016) collected from all participants; information and documentation relevant to intellectual functioning provided by parents for all participants (e.g., Individualized Education Program information, clinical or educational records with results of previous IQ testing, intellectual disability diagnosis documentation, whether the youth was pursuing/had received a regular or special education diploma); and research-administered IQ scores from the Weschler Abbreviated Scales of Intelligence (Wechsler, 1999) for those participants whose baseline data were collected in-person prior to COVID-19.

Parent information included gender, education (1 = less than a high school degree to 7 = Ph.D. or professional degree), household income (1 = $20,000 or less to 9 = greater than $200,000), race, and ethnicity. There was limited racial or ethnic diversity in this sample, so race/ethnicity was coded as white non-Hispanic versus others for analysis. Parental mental health was measured using the anxiety, depression, and stress subscales of the Depression Anxiety Stress Scales (Lovibond & Lovibond, 1995). Parental coping styles were measured by the Coping Orientation to Problems Experiences (COPE) scale (Carver et al., 1989). We chose three of the 14 COPE subscales (active coping, planning, use of information support) that reflect problem-focused coping (Dias et al., 2012), hypothesizing that parents who were more likely to use problem-focused coping would better attend to and use the information delivered through ASSIST.

Analysis Plan

Our targeted sample size (n = 180) was determined based on treatment effects observed in our pilot research (Taylor et al., 2017b) and was sufficient to detect a .42 standard deviation modification of treatment effect. Initial descriptive analysis of baseline demographic variables was done by experimental condition, site, and cohort to identify if important imbalances in covariates existed. We prespecified a common regression framework to estimate and test for a significant effect of treatment on outcomes at follow-up. Separate linear regression models were fit for each outcome using an indicator variable for experimental condition (intent to treat; predictor of interest) while controlling for site (2 d.f.), cohort, and the baseline measure of the outcome to improve precision of the estimated treatment effect. We present plots of the result for simpler models that did not control for site or cohort when neither of these effects were significant.

For outcomes that had significant treatment effects, we tested for effect modification. Separate models were fit in which we included an additional covariate defined as the product of the treatment indicator variable and the potential modifier. Modifiers considered were site, cohort, high school exit, characteristics of the autistic youth (intellectual disability, autism symptom severity) and characteristics of the parent (race/ethnicity, education, mental health, coping). Results for treatment effects are presented as mean differences between experimental conditions with corresponding 95% confidence intervals, and tests for effect modification were done using likelihood ratio tests with a two-sided significance level of 0.05. Analyses were conducted using the R statistical package.

Results

Table 3 presents estimates from the multivariable models examining treatment differences in post-test scores of parental advocacy ability, controlling for baseline scores, site, and cohort. After taking ASSIST, the treatment condition averaged higher scores on knowledge and perceived advocacy skills (with a marginally significant effect for empowerment) relative to controls. Scatterplots and model estimates are plotted in Figure 2. Because we controlled for baseline scores, higher post-test scores are synonymous with greater gains. Thus, relative to those who received all of the information presented in ASSIST in written format, participating in the group-based session was associated with greater gains in knowledge about adult services and perception of greater advocacy skills. Visual inspection of the plots in Figure 2 suggested that, for knowledge and perceived advocacy skills, treatment effects were most pronounced for those at the lower end of advocacy abilities at baseline.

Table 3.

Estimates from multivariable models predicting follow-up scores of parent advocacy ability

		Knowledge		Perceived Advocacy skills		Family Empowerment
Covariate	Level	Unadjusted Mean (SD)	Multivariable model estimate (CI)	Unadjusted Mean (SD)	Multivariable model estimate (CI)	Unadjusted Mean (SD)	Multivariable model estimate (CI)
Baseline		12.8 (3.2)	0.62 (0.52 to 0.73)***	3.8 (0.6)	0.62 (0.51 to 0.74)***	127.6 (17.5)	0.71 (0.60 to 0.82)***
Group	Treatment	13.3 (3.0)	Reference	3.9 (0.6)	Reference	128.4 (18.2)	Reference
	Control	12.3 (3.4)	−1.62 (−2.33 to −0.90)***	3.7 (0.7)	−0.19 (−0.33 to −0.04)*	126.9 (16.9)	−3.27 (−7.09 to 0.54)†
Site	TN	12.3 (3.5)	Reference	3.9 (0.5)	Reference	129.9 (15.1)	Reference
	IL	12.6 (3.0)	−0.00 (−0.88 to 0.88)	3.7 (0.6)	−0.15 (−0.33 to 0.04)	126.0 (17.9)	−3.88 (−8.57 to 0.8)
	WI	13.5 (3.1)	0.63 (−0.23 to 1.50)	3.8 (0.7)	−0.10 (−0.28 to 0.08)	127.0 (19.4)	−1.72 (−6.34 to 2.91)
Cohort	Cohort 1	13.1 (3.3)	Reference	3.7 (0.6)	Reference	126.6 (17.8)	Reference
	Cohort 2	12.5 (3.1)	−0.51 (−1.22 to 0.20)	3.8 (0.6)	−0.15 (−0.30 to 0.01)†	128.6 (17.4)	−3.32 (−7.21 to 0.57)†

^†p = .092

^*p < .05

^**p < .01

^***p < .001

Figure 2.

Scatter plot of baseline and follow-up scores on parental advocacy ability by experimental condition. Solid lines represent slope estimates for each group, and the dotted line represents 1:1 correspondence (i.e., same score at baseline and follow-up). Slope estimates correspond to models that include experimental condition and baseline scores of the corresponding follow-up measure. As can be seen from the figure, for knowledge and perceived advocacy skills, there is greater treatment effects on post-test scores at lower levels of baseline.

Modifiers were next explored to understand for whom and when ASSIST might be most beneficial. None of the tested interaction terms were statistically significant, including study site, cohort, timing (i.e., high school exit), youth functioning, parental education, or parental mental health (see Table 4). The exception was the interaction between parent active coping and experimental condition on knowledge, B = 0.48, SE B = 0.15, p < .01. Model estimates by group are presented in Figure 3; greater group differences were observed for those who were low (versus high) in active coping. For those high in activity coping, change in knowledge of adult services from baseline to post-test was similar whether parents received written material about adult services (control group) or participated in the group-based ASSIST program (treatment group). However, for parents who were low in active coping, greater gains in knowledge of adult services were made by those who took the group-based ASSIST program, versus those who received written materials.

Table 4.

P-values for modifiers of the impact of treatment group on parent knowledge and advocacy skills

Modifier	Knowledge	Perceived Advocacy Skills
Site	0.85	0.25
Cohort	0.73	1.00
Timing (before vs. after high school)	0.21	0.73
Youth intellectual disability	0.70	0.45
Youth autism symptom severity	0.74	0.88
Parent race/ethnicity	0.33	0.93
Parent education	0.67	0.70
Parent depressive symptoms	0.83	0.39
Parent anxiety symptoms	0.20	0.64
Parent stress	0.35	0.28
Parent active coping	0.00	0.73
Parent planning (coping)	0.09	0.28
Parent use of instrumental support (coping)	0.16	0.62

Figure 3.

Graphical depiction of interaction between experimental condition and active coping on knowledge. For ease of interpretation, change in knowledge from baseline to post-test is plotted on the x-axis, and coping scores are plotted on the y-axis. Greater differences between treatment and control conditions were observed at lower levels of active coping.

Discussion

Our results suggest that, relative to receiving written information about services and supports, taking the group-based ASSIST program led to increases in advocacy ability for parents of youth with autism. The gains in knowledge about adult service systems and perception of advocacy skills observed in the current large-scale RCT are consistent with findings from our pilot work (Taylor et al., 2017b), providing further evidence for the benefits of ASSIST in improving parents’ ability to advocate for services and supports which may, in turn, improve access to critical services for transition-aged youth with autism (Lee et al., 2022).

In addition to being statistically significant, the effects of ASSIST on parents’ knowledge of adult services and their perception of their advocacy skills are also likely clinically significant. As can be seen in the Supplemental Material, the baseline to post-test changes in parent knowledge and perceived advocacy skills for the treatment group fall in the range of a medium effect (Cohen, 1988) and thus are likely to be clinically significant. In contrast, the effect size for baseline to post-test change in advocacy ability for the control group was near zero for knowledge and fell in the range of a small effect for perceived advocacy skills. Further, in summative evaluations of their experience with ASSIST, treatment group families reported finding value in the intervention and nearly all families said they would recommend ASSIST to others (Taylor et al., 2022). Families’ satisfaction with the ASSIST program provides further evidence that the statistically significant gains observed in our analyses are likely clinically significant.

ASSIST may have similar benefits for families regardless of where and when it is delivered. Though non-significant findings should be interpreted with caution, the impact of treatment did not differ based on study site, which is consistent with our previous work showing similar implementation outcomes across sites (Taylor et al., 2022). The impact of ASSIST on parent advocacy ability also did not differ based on if youth were in high school or youth disability severity. Most interventions for autistic youth target only those with average or above intellectual functioning (Baker-Ericzén et al., 2022; Gilmore et al., 2022; Stedman et al., 2019); in contrast, ASSIST may be a useful program for families of autistic youth at a range of ability levels.

Modifier analyses indicated that those who might have the greatest difficulty advocating on behalf of their son or daughter gained the most from the intervention. Though we had hypothesized that those with more active coping styles would be best prepared to benefit from ASSIST, instead greater gains were made for those who were low on active coping. Similarly, greatest gains were made by those who began ASSIST with less knowledge of adult services and supports and who felt less skilled advocating on behalf of their youth (see Figure 2). Thus, ASSIST may be a helpful tool for parents who experience more uncertainty and difficulty supporting their son or daughter in accessing services. Alternatively, it may be that families who are more “primed” to advocate – whether because they have more active coping styles or because they already possess greater knowledge of adult services and supports – will benefit similarly from a series of classes like ASSIST as they would from a less intensive intervention, such as receiving written materials about adult services. It will be important for future research to continue to probe which modality and intensity of intervention is most effective in supporting families to advocate on behalf of their son or daughter. Families’ preferences should also be considered in these investigations; anecdotally, many families in our study who would likely have benefitted similarly from less intensive intervention (receiving written materials about adult services) nonetheless expressed preference for the more intensive intervention (full ASSIST classes).

In our pilot work with ASSIST, the largest treatment effect was observed for parental empowerment (Taylor et al., 2017b). Though supplemental analyses suggested that the treatment group gained in empowerment over time (see Supplemental Materials), our primary analyses showed that the impact of treatment on empowerment was only marginally significant. It may be that the COVID-19 pandemic limited potential empowerment gains. Post-test data was collected in late 2020/early 2021 – during this time, many children were participating in virtual schooling, agencies administering disability-related services were difficult to reach, and social distancing reduced opportunities for community integration (Burke, Cheung, et al., in press; Jeste et al., 2020). Parents of children with disabilities particularly struggled during the COVID-19 pandemic (Lee et al., 2021; Werner et al., 2022), and thus is it not surprising that there were minimal gains in empowerment. Future research during non-pandemic times will be helpful to better gauge the impact of ASSIST on empowerment.

There are important limitations of this study. Though our sample was diverse in terms of functioning level of the youth and high school exit status, it was not diverse in race, ethnicity, family income, or parental education. Our future work will examine what adaptations will be needed for ASSIST to be maximally effective for youth and families with diverse racial/ethnic backgrounds or socioeconomic position. Second, indicators of parent advocacy ability were all assessed via self-report. This is a common practice, as self-report is the most frequent way that advocacy ability and skills are measured (e.g., Burke, Goldman, et al., 2016; Koroloff et al., 1996; Kutash et al., 2011; Pearson et al., 2022; Taylor et al., 2017b). Further, significant correlations between self-reported advocacy ability and service access suggest that measuring advocacy ability in this way may be valid (Lee et al., 2022; Taylor et al., 2017a). However, in single-blinded trials such as this, and particularly when participants are informed of whether they have been assigned to treatment or control conditions (such as in this trial), use of self-reported outcomes raises the possibility that treatment effects could be explained by expectancy bias. Future research should consider other perspectives on advocacy ability, such as asking the autistic youth their perceptions of how well their parent is advocating on their behalf. The development of novel paradigms such as advocacy simulations may provide a more objective way to determine whether a person’s ability to advocate has improved (for examples in the context of Individualized Education Program meetings see Holden, 2017; Toledo, 2023).

Given the COVID-19 pandemic, it was necessary to shift delivery of ASSIST from in-person to online during the study. ASSIST participants recognized advantages and disadvantages of both intervention delivery modalities, and there was variability in preference of in-person versus online (Taylor et al., 2022). Further research is needed to better understand the effectiveness of in-person versus online ASSIST delivery, including which modality might work best for whom. Finally, in this project we hypothesized that parent advocacy ability would be the intervention mechanism by which ASSIST impacts service access and postschool outcomes for youth with autism (in other words, parent advocacy ability was the treatment target). Results from this analysis suggest that this target was engaged as a result of ASSIST. Future research will examine whether engagement of the treatment target – namely increasing parent advocacy ability – leads to improvements in service access and youth outcomes.

Conclusion

Many youth with autism have difficulty accessing services and supports during the transition to adulthood. Teaching parents to better advocate on behalf of their youth may be one avenue to improve service access and transition outcomes. Results from our RCT suggested that parents of transition-aged autistic youth who participated in ASSIST had gains in advocacy ability relative to a control group who received written materials. Those least prepared to advocate for their youth appeared to benefit most from participating in ASSIST. Future research will examine whether these gains in parent advocacy ability lead to improvements in service access and post-school outcomes for youth with autism.

Data Availability

Data collected as part of this project will be shared to the National Database for Autism Research in accordance with NIH guidelines.