A randomized controlled trial protocol for evaluating the feasibility, acceptability, and work outcomes of individualized placement and support adapted for autistic adults in the community

Abstract

Relatively few autistic adults, including those with average intellectual abilities, are competitively employed, meaning that they hold jobs together with non-disabled workers and receive comparable wages and benefits. In California, for example, most autistic individuals served by the state are placed in programs where they participate in skill-building and socialization but not in actual competitive jobs. Failure to participate in the labor force can diminish autistic workers’ sense of purpose, well-being, and ability to earn a living wage.

Available research suggests that supported employment that assists autistic adults in finding and keeping jobs, produces the highest sustained competitive employment rates. Thus, our team has been investigating the Individualized Placement and Support (IPS) model, which has an extensive evidence base for increasing competitive employment rates in individuals with chronic mental illnesses. In a California Department of Developmental Services Employment Grant investigating adults with autism and intellectual disabilities, we demonstrated a competitive employment placement rate of 52 % using IPS. Components of IPS were appropriate for this population, however there were implementation challenges related to IPS model fit with the vocational support agencies.

Based on focus groups and stakeholder input, we have adapted IPS to provide intensive agency training, leadership education, and record keeping support. Herein, we detail a protocol for a randomized controlled trial of the adapted model (IPS-AUT) to evaluate feasibility, acceptability, and work outcomes. We also investigate potential moderators and mediators of treatment effectiveness to provide a foundation for a larger more adequately powered randomized clinical trial.

This protocol is registered at ClinicalTrials.gov: NCT 06829264.

1. Introduction

As autistic individuals enter early adulthood (ages 21–40 years), they become increasingly disengaged from the community [1] and fail to enter the workforce at rates comparable to their non-autistic peers [2]. Available data suggests that only 14 % of autistic adults ever achieve competitive integrated employment (CIE a.k.a. CE), which is defined as work in an integrated setting with wages and benefits that are comparable to non-disabled workers [3]. In fact, in California, over 80 % of developmentally disabled consumers served by the state are placed in day programs (structured daytime services for skill-building and socialization) and sheltered settings (segregated work environments with supervised employment for sub-minimum wages) where they do not have the opportunity to engage in inclusive work [4]. This is a clinically significant problem given that engaging in meaningful work is one of the most important determinants of sense of purpose, general well-being, life satisfaction [5,6], self-esteem, and mental health functioning [7]. Furthermore, in California, legislative mandates eliminated the subminimum wages, which have been used in sheltered workshops and day programs, in January 2025. This has created an imperative for new vocational service options which promote CIE [8].

There have been multiple studies of brief interventions, services, and adjuncts to existing employment supports to improve the employment and retention of autistic adults in hospital [9], clinical [10], and educational settings (e.g. Ref. [11]). However, employment rates produced by these programs are typically low, involve only adolescents or young adults or CIE is not the endpoint. After considering overall societal costs and benefits, supported employment programs that help autistic workers to find and keep jobs that are consistent with their strengths, resources, priorities, concerns, abilities, and informed choices have been shown to be the most highly efficacious [[12], [13], [14], [15], [16]].

Consumers in the mental health system can receive an evidence-based vocational service, known as Individual Placement and Support (IPS; [17]), which is the most extensively studied, widely disseminated, and highly validated person centered model of supported employment for those with severe mental illness in the world [18]. Twenty-six of 28 randomized controlled trials (RCTs) have established IPS as the best vocational model for improving CIE to over 55 % in those receiving IPS, compared to 25 % in those not receiving IPS [19]. IPS services embrace 8 core principles including: 1) a focus on CIE in regular community jobs; 2) no exclusion for clients wanting to work; 3) attention to clients’ skills and preferences; 4) rapid job search; 5) systematic community-based job development by employment specialists (ESes); 6) personalized benefits counseling; 7) time-unlimited follow-along supports; and 8) integration with other client services. The core un-adapted IPS model has been used successfully in 1 small trial (n = 5) of autistic adults in a hospital setting with 100 % receiving competitive jobs [20]. The only community-based trial of IPS in intellectual disabilities and autism had a mean 36 % employment rate at 1 year [21] largely due to implementation challenges in the developmental disability services system versus the mental health service system that typically supports IPS.

Based on the success of this model and its seeming compatibility with the autistic and developmentally delayed populations, we implemented an 18-month Employment Grant from the California Department of Developmental Services (DDS) (Grant A22-334) to further investigate the provision of IPS to autistic adults. We enrolled participants from 4 supported employment agencies that were contracted with the Alta California Regional Center—one of the Regional Centers (RCs), that are the California state agencies charged with offering services to those with developmental disabilities. These agencies were trained to deliver IPS to 21 autistic consumers ages 21–53 years (Section 2.5). We were guided by a team of stakeholders referred to as the Community/Academic Policy Partnership (C/APP) in the implementation of this trial (see 2.2). Results were promising [22]. By the end of the trial, 11 of 21 (52 %) of consumers had been placed in a job and 80 % of these individuals maintained their jobs for 6 months; 3 of 4 agencies (75 %) continued their participation in the project until the end; and 100 % of consumer participants remained enrolled. At the 6-month fidelity review -- a structured 2-day review of agency interactions with the consumers, their family members, and other service agencies (See 2.5.1) -- 3 agencies achieved fair fidelity, which is considered acceptable, as specified by Bond et al. [23]. By 12 months, all had improved and were in the fair to good fidelity range. See Section 2.5.1 about fidelity reviews.

In October 2024, we conducted a series of focus groups with a subset of consumer participants, (n = 13), and their permitted parents/carers (n = 4), the employment specialists who worked directly with clients on job development and ongoing support (ESes; n = 4), and the ES supervisors (n = 6) to identify facilitators and barriers to IPS adaptation for autistic adults in California. Consumers were generally very satisfied with the supported employment model. Families, ESes, and ES supervisors made helpful suggestions about a need for more agency training, leadership involvement, family inclusion, and assistance with documentation. Using an iterative process informed by implementation science methods, results were integrated with input provided by the C/APP and the literature. Community-informed guidelines shaped discussions [24]. Based on this process, we developed a consumer support toolkit (CST) consisting of model adaptations. See Fig. 1 for the components of the IPS model and the adaptations made to form the CST. We refer to the IPS model plus the CST as IPS-AUT.

Fig. 1.

Comparison Chart of the Individual Placement and Support for Autism (IPS-AUT). This figure illustrates the core components of the evidence-based IPS model, including rapid job search, community-based job development, and time-unlimited follow-along support. It also highlights the added elements informed by stakeholder feedback and implementation science, including the Consumer Support Toolkit (CST), which features agency training, leadership engagement, family involvement, and strategies for data documentation and reimbursement alignment.

2. Methods

2.1. Study design

2.1.1. The trial

We chose a randomized wait-list controlled trial design with an “active” control group (See 2.6) with the following Aims:

Specific Aim 1: To evaluate the work outcomes (achievement of CIE, job retention, earnings, job satisfaction) of IPS-AUT, as well as consumer, parents/carers (if permitted by consumers), supported employment providers, and willing employers (if permitted by consumers), perspectives on model feasibility, acceptability, and whether they are satisfied they are with IPS-AUT.

Specific Aim 2: To preliminarily investigate signals of effect for proposed mediator (target) variables (parent/carer engagement and work-related social cognition [WRSC; these processes involve cognitive skills such as perspective-taking, social cue interpretation, and understanding workplace norms, culture, and dynamics]).

2.2. Study procedures

As in the DDS pilot, the C/APP consisting of consumers, family members, employment professionals, academics, employers, state agency leaders, policy makers and lobbyists involved in employment, will meet quarterly to provide guidance about how to make IPS-AUT most appropriate for autistic adults, agencies serving them, and achievement of sustainable funding enabling scale-up of the model. Many of the original DDS stakeholders have agreed to participate in the new grant. All partnerships will be formed using evidence-based best practices to ensure knowledge exchange and collaboration [24].

2.3. Eligibility criteria

2.3.1. Partner agencies

During the pre-award period and the first 4 months of the trial, we have been thoroughly educating potential partner agencies about IPS, requirements of the trial, and factors that make the model most successful. Each partner agency will have signed a document saying that they understand the requirements. (See also section 2.3.5). We anticipate recruiting 6 agencies, each contributing 10 participants (n = 60). Each partner agency will be paid a training fee of $1900 per VSP trained.

2.3.2. Consumers

The study will be described to consumers at each partner agency. Those consumers interested in participating will then undergo informed consent procedures to indicate they want to participate and subsequently will be referred to our team. To qualify, they must have: 1) Received a community diagnosis of autism spectrum disorder as demonstrated by a letter from a health care provider, psychologist or other mental health professional, RC representative, or school psychologist, 2) A T-score on Social Responsiveness Scale (SRS-2; [25]) > 60 that is indicative of autism, 3) An age of 21–40 years, 4) At least a 4th grade reading level (approximately mild ID as demonstrated by the Wide Range Achievement Test, 4th edition, Reading-Recognition subtest (WRAT-4 [26]), and 5) shown that they want to work by their request to be referred to a supported employment agency and their verbal affirmation of this goal during the development of their IPS career profile. We will exclude individuals who have active psychosis or suicidal ideation or pose an elopement risk because trial budget constraints prevent us from providing adequate care for them. In the event any consumers evidence suicidality, the PI (a clinical psychologist) will make appropriate referrals.

2.3.3. Parents/carers

Parents/carers will be included in assessments if permitted by the young adult consumer and after undergoing informed consent procedures. Parents/carer non-inclusion will not disqualify consumer participants.

2.3.4. Employers

Employers will be included in assessments if permitted by the young adult consumer. They will be properly consented. Employer non-inclusion will not disqualify consumer participants.

2.3.5. Vocational Support Professionals (VSPs)

Vocational support professionals, consisting of ESes and their supervisors who work at the partner agencies will also be consented and are viewed as highly valued members of the team. They will: 1) Agree to participate in upfront and weekly training by certified IPS Trainers and our team as part of their regular jobs, 2) Permit the IPS team to attend weekly supervision meetings with their agency until good fidelity is achieved, 3) Participate in 1 monthly meeting of a learning community with other providers and DDS professionals where difficult cases, special topics of member choosing, and information exchange that will occur after the first fidelity review is completed, and 4) Permit fidelity reviews at 6–8 months and then 6 months later until good fidelity is achieved. To mitigate the possibility of coercion, we will affirm with their supervisors that their failure to consent to participate in the trial should not impact their performance evaluations and will not disqualify the agency or the client from participating. They also will be compensated for completing evaluations of their clients and the program.

2.4. Participant recruitment, screening, and randomization

Once partner agencies are selected, we will randomize at the level of the agency such that 4 of the 6 agencies will be trained in IPS first (treatment first group; n = 40), and 2 of the 6 will be trained in IPS after 1 year (active wait list-control group; n = 20). Once recruited partner agencies are randomized, VSPs recruited, and trained, we will initiate the 1-year RCT. For the treatment first group, fidelity reviews will occur after 6–8 months as recommended by IPS (this will give agencies 2 months to onboard participants before the first fidelity review), and then again, 6 months after the first fidelity review (approximately 12 months). The active wait-list control group will be the comparator for the treatment first group in Stage I. In Stage II, the wait-list control group agencies will receive IPS training and services. Consumers, their parents/carers, VSPs, and employers will complete various employment outcome and other measures throughout their participation to provide important information about feasibility, acceptability, outcomes, and satisfaction with IPS as well as potential moderators and mediators of trial success.

Fig. 2 is a schematic of the RCT.

Fig. 2.

Schematic of the Individual Placement and Support for Autism (IPS-AUT) Trial. This figure presents the schematic of the Hybrid Type I Waitlist-Controlled Randomized Controlled Trial (RCT) Design. It presents the two-stage cluster-randomized trial structure comparing the immediate IPS-AUT implementation (Treatment First group) with a waitlist control group. It outlines participant flow across agencies, fidelity review timepoints, and data collection periods for employment outcomes, feasibility, acceptability, and mediator/moderator assessments.

2.5. IPS-AUT training

We will conduct a two-day upfront training for VSPs. Day 1 will include education about autistic workers, the IPS model, the IPS fidelity review scale and assessment process, and the basics of obtaining state reimbursement. Day 2 will focus on job development and culminate in an activity where all trainees go out into the community to generate job leads and then discuss the experience with the group.

Our team will attend all weekly agency supervision meetings where we will teach the principles of IPS and fidelity in greater depth. We will produce 8–10 video segments of important concepts in IPS including working with new clients, the role of the family in achieving good employment outcomes for consumers, helpful resources in the community, how and when to reduce supports to consumers, elements of model fidelity, and agency reimbursement strategies (20 min each) and play them followed by discussion of potential trial participants; management of difficult cases; and community visits to improve job development skills. Video segments can be used to help train new providers if there is staff turnover. Additionally, we will help agencies to align their data collection with IPS requirements.

Once agencies have completed the first fidelity assessment, we will initiate a monthly learning community meeting for all VSPs where discussion of difficult cases, mental health challenges, and issues affecting employment can occur.

2.5.1. IPS fidelity reviews

These reviews are completed to assess how well an agency is implementing IPS core principles, with a focus on continuous quality improvement. In a fidelity review, 1–2 evaluators visit with an agency for 1–2 days and rate agencies on a well-validated 25-point scale [23], with a score of 100 being considered “good.” During the evaluation, reviewers meet with consumers, parents, agency staff, and other community resource providers (e.g. DOR representatives) to assess how well the agency is adhering to IPS with respect to staffing, organization, and services provided. A written report of action items for the agency is produced.

2.6. Active control condition

To minimize attrition and provide community services to those in the active control group, which we believe is most ethical, we will implement a bi-weekly support group during the first year. We elected to include a job club group given that such support groups are very common and very different than the more intensive service we are providing in the treatment condition. This group will be designed to follow the principles of the Job Club model [27]. The facilitator will be a trained mental health or vocational service professional from the community. Participants in the Job Club can include consumers and their family members. Topics typically discussed include job interviewing, resume preparation, dealing with rejection, and difficult employment situations with colleagues and bosses.

2.7. Study measures

The study has been approved as an exempt study by the UC Davis IRB (2024578-4). All measures used in the battery can either be completed though phone or text or be completed online through a REDCap secure link. Assessors will be present on Zoom to assist participants in completing assessments and are specially trained to ensure that assessments optimize what participants can do and that the battery is well tolerated through frequent breaks and support. The database will note whether an assessment is believed to be an accurate reflection of participant abilities and whether a non-standard assessment was used.

2.7.1. Qualification and general assessments

The proposed schedule of consumer qualification assessments can be found in Table 1. As mentioned above, these will be completed by both the consumers in the intervention first group and the active control group once these individuals have consented to participate in the trial at their agencies and then are passed to our team for 1–2 h testing sessions.

Table 1.

Consumer Qualification Measures. This table summarizes the instruments used to determine eligibility for study participation, including assessments of autism symptom severity and basic reading level. Measures were selected to ensure that participants met the inclusion criteria related to autism diagnosis and cognitive functioning.

Qualification Measures	Source	Description	Frequency & Administration Time
Social Responsiveness Scale-2 [25]	Participant Questionnaire	65-item questionnaire used to assess autism symptom severity. T-score of 60 suggests an ASD. (CDE).	Baseline; 15 min
Wide Range Achievement Test, 4th Ed., Reading-subtest (WRAT-4 RR; [26]).	Direct	Brief test of word reading. Sixth grade reading level or better required. (CDE).	Baseline; 10 min

CDE = Common Data Element.

2.7.2. Primary outcome measures of employment success, engagement, retention, fidelity

Monthly calls or texts to consumers, parents, and VSPs will be used to collect information about job attainment, retention, and satisfaction in both the intervention first and active control groups using measures commonly implemented in IPS research [28,29]. Five-item Likert scale-based items also will be administered to them quarterly to rate the extent to which they see IPS-AUT as feasible, acceptable, and appropriate in the intervention first group only. This methodology has been adapted from Weiner et al., [28]. Two Trained Raters will complete a 2-day site visit to assess level of fidelity at which IPS is being used. Consumers, parents, VSPs, and Department of Rehabilitation (DOR) providers are interviewed. Scores >100 are considered “good” (See Table 2).

Table 2.

Trial Outcome and Fidelity Variables. This table outlines the primary outcome measures used to assess employment success, consumer engagement, service satisfaction, and IPS-AUT model fidelity. Data sources included consumers, parents/carers, vocational support professionals (VSPs), and trained fidelity raters. The frequency and administration times are indicated for each measure.

Measure	Source	Description	Frequency & Administration Time
Trial Outcomes (Aim 2)
Vocational Survey to Assess Competitive Employment	Participant, Parent, VSP Reports	Calls/texts to consumers will assess attainment of CIE, wages, and job retention. Parents and VSPs will also be queried to triangulate veridical reports.	Monthly (12x) for duration of 1 year trial; 5 min
Survey of rates of engagement, retention, and satisfaction [28]	Participant, Parent, VSP Reports	Calls/texts to consumer will collect engagement, retention, and satisfaction ratings. Parents/carers and VSPs also queried to triangulate veridical reports.	Monthly (12x) for duration of 1 year trial; 5 min
Participant, Parent/Support Provider, Employment Specialist Experience Ratings [28]	Participant, Parent, VSP, employer Reports	These 5-item Likert scale-based items will be used to rate the extent to which these individuals view IPS-AUT as implemented in the trial as feasible, acceptable, and appropriate.	Quarterly (4x) for duration of 1 year trial; 5 min
IPS Fidelity Scale (IPS Center) [17]	Trained Raters and all Stakeholders	Using the IPS Fidelity Scale, 2 Trained Raters assess level of fidelity to the IPS Model. Consumers, parents, VSPs, RC, and DOR are interviewed. Scores >100 are considered “good” fidelity.	8 and 12 months; 1.5-day site visit

VSP = Vocational Support Professionals; CIE = Competitive Integrative Employment; IPS-AUT = Individualized Placement and Support Autism; RC = Regional Center; DOR = Department of Rehabilitation.

2.7.3. Secondary outcome, moderator, and other measures of interest

To assess parent involvement, we will add questions about parent engagement in the job development and the ongoing employment process to the monthly calls or texts used for the primary outcomes mentioned above using the same 5-item Likert scale format. To assess the consumer WRSC skills, we will add a Questionnaire about consumer thinking skills at work from the Thinking Skills at Work Program (need a reference) to be administered at pre, mid-point, and post testing. We also will collect the Behavior Rating Inventory of Executive Function-2 Adult version (BRIEF-A) [30] and the Adaptive Behavior Assessment System (ABAS-3) [31] (See Table 3).

Table 3.

Measurements of Potential Moderator and Mediator Variables. This table presents the instruments used to assess the potential moderators and mediators of IPS-AUT outcomes, including parent engagement, work-related social cognition (WRSC), executive functioning, and adaptive behavior. Measures will be administered at multiple time points to support the exploratory mediation analyses.

Measure	Source	Description	Frequency & Administration Time
Survey of parent engagement	Participant, Parent, VSP Reports	This brief 5-item measure will be added to the calls/tests to participants, parents/carers, and VSPs that query engagement, retention, and satisfaction as shown in Table 2. It will ask about engagement of parents/carers in the job development and ongoing employment of their adult children. All respondents are queried to triangulate veridical reports.	Monthly (12x) for duration of 1 year trial; 5 min
Client Interview Part III from the Self-Management Portion of the Thinking Skills at Work Curriculum [29]	Participant, Parent, VSP Reports	This brief questionnaire queries clients about their planning, helpful routines, and other strategies used to aide their problem solving, attention and concentration, and memory. All respondents are queried to triangulate veridical reports.	Baseline, 6, and 12 months; 15 min
Behavior Rating Inventory of Executive Function-2 Adult version (BRIEF-A) [30]	Participant Questionnaire	75 items rating on a 3-point Likert scale. Assesses 9 dimensions of executive functioning and has several composite indices. Also has built-in validity scales. Commonly used in autism. (CDE).	Baseline, 6, and 12 months; 15 min
Adaptive Behavior Assessment System (ABAS-3) [31]	Participant Questionnaire	Measure of adaptive functioning that has been used extensively in autism. A conceptual composite and coping self-efficacy scales show changes in our adult transition study. (CDE).	Baseline, 6, and 12 months; 15 min

VSP = Vocational Support Professionals; IPS-AUT = Individualized Placement and Support Autism; RC = Regional Center; DOR = Department of Rehabilitation. CDE = Common Data Element.

2.8. Data management, monitoring, sharing

We will conform to all NIH guidelines in submitting required elements to the National Data Archive (NDA) starting at 6 months post award and every 6 months thereafter. We have submitted a Data Submission Agreement signed by the principal investigators and an institutional business official.

2.9. Data analysis

2.9.1. Data analyses to test the results of the clinical trial (Aim 1)

Data from the trial are multilevel, with the units of analysis being timepoint-specific measurements from clients who are nested within the supported employment agencies that are the units of randomization for our cluster-randomized design. For each of the trial outcomes, study arm specific means and proportions will be estimated with 95 % confidence intervals for multilevel data. Similarly, treatment effects (e.g. between-arm comparisons of intervention vs wait-list) and 95 % confidence intervals will be estimated using all available data and with group membership coded according to assigned arm, as in intention-to-treat analysis, using a generalized linear mixed model framework to accommodate all outcome types. This will allow us to rigorously test Aim 1 hypotheses about feasibility, acceptability, and treatment effects with robust 95 % confidence intervals while exploring heterogeneity of treatment effects (moderation). [[32], [33], [34], [35], [36], [37]]. These analyses will be restricted to Stage I data, where the wait-list control participants provide a concurrent control group for IPS. To account for cluster effects (unmeasured sources of between-unit heterogeneity), models will include random effects for agency and, for longitudinal data, client. Independent variables will include a binary indicator for whether the patient was from an agency assigned to IPS-AUT and a parsimonious and pre-specified set of demographic covariates and, when available, baseline measures (from time of enrollment), to improve the precision of estimated effects [33,38,39]. Moderation will be explored for each candidate effect modifier, including sex and reading ability, by fitting an additional model that includes an interaction term for the candidate effect modifier and the IPS indicator. Our primary outcome will be CIE attained over the 12-months following enrollment, which we will analyze using a Poisson mixed-effects regression model, to permit characterizing treatment effects as an adjusted relative rate, while accounting for the possibility of within-client transitions into and out of CIE, as well as possible variations in follow-up time among clients.

2.9.2. Specific plans to address missing data

We anticipate a relatively small amount of missing data arising from failure to complete follow-up surveys from attrition, from a client intermittently missing an assessment, or from a client completing an assessment with some missing items. Nonresponse and missing data patterns will be characterized and compared between the two arms of the study (intervention and wait list). Given that the missing at random assumption is untestable, we will follow a principled approach to addressing missing data that includes conducting primary analyses using the generalized linear modeling framework under the standard missing at random assumption and that supplements this with sensitivity analyses that make alternative assumptions about the ignorability of the missing data [40].

2.9.3. Power and sample size considerations

Consistent with recommended practices for early-phase studies, we aim to provide estimates with suitable precision to inform decision-making about whether and how to pursue further intervention development. We opted for an unbalanced 2:1 allocation with 60 total participants because such a study provides equivalent precision/power as for a 1:1 allocation with 53 total participants while providing 40 participants in the intervention group, to better inform feasibility and acceptability assessments. To account for the cluster-randomized design and the regression strategy [33,41], we simulated data in SAS [[42], [43], [44]], using an intra-cluster correlation coefficient of 5 % at the agency level and up to 55 % at the client level, reflecting moderately high stability for repeatedly measured outcomes like CIE status and client satisfaction. Combining an assumed missing data rate of 10 % and design effects that reduced the effective size of the actual number of outcome measurements from 40 % (for single-timepoint measures) to 86 % (for the 6 monthly reports of CIE status in the Stage-1 analysis), we anticipate being able to estimate intervention group means and between-group differences with sufficient precision that the margins of errors will be less than 0.67 and 0.58 standard deviations, respectively, providing similar performance as a two-group individually randomized clinical trial design of 38, slightly higher than the 30 patients frequently used for pilot studies. Effect modification analysis will be considered exploratory, due to low power for potentially meaningful heterogeneity in treatment effects.

2.9.4. Data analysis (Aim 2 mediation analyses)

The same generalized linear mixed modeling framework used for Aim 1 will be adapted for use in Aim 2, with regression models for functional outcomes revised to include candidate mediators as independent variables. To correlate changes in functional outcomes with changes in mediators, we can substitute these change-score variables into the regression model specification for ease-of-interpretation but will also specify and assess cross-lagged panel data models to test the direction of the association between mediators and outcomes measured at multiple occasions. In addition, we will fit statistical mediation models that include the IPS-AUT treatment indicator (as upstream cause) and assess direct and mediated effects of IPS-AUT, although for this early-stage study we consider these models exploratory.

3. Discussion

The unemployment and underemployment of autistic adults is a significant societal problem given employment's human benefits (e.g. sense of purpose, well-being, life satisfaction, and income) and unemployment's economic costs (e.g. lost wages, lost family member wages, entitlements, and services costs). Although past studies have demonstrated supported employment's superiority with respect to costs and benefits in autism [12,13,16], there are few studies of supported employment characterized by integrated job development and job support, aimed at CIE, for post-secondary education graduates. To address this gap in the literature, our team has been evaluating the IPS model, which has an extremely strong evidence base in the mental health literature, targets CIE as an endpoint, and whose strategies appear to be largely compatible with the strengths and challenges profiles of autistic adults.

In a first early-stage investigation funded by the California DDS, we were able to demonstrate a CIE rate >50 %. This trial indicated that, while the components of the services offered by the model were generally appropriate for serving individuals with autism, complexities due to the way we implemented the model coupled with challenges in the service delivery system (i.e. agency structure, reimbursement pathways), posed a larger challenge for the IPS model. The CST outlined in this protocol is designed to remedy some of these problems. More specifically, most adults with disabilities that are clients of California service programs are referred to day programs where they are not engaged in inclusive paid employment. Given the sunsetting of sheltered workshops in our state, there is a great need for improving evidence-based supported employment services as is done in IPS. In our RCT, we will cover a larger geographical region and work with new Regional Centers and supported employment providers. This will provide us with an opportunity to work with a different set of services policy makers in various regions in our state. While our last trial enrolled consumers requesting supported employment only, in the new trial, we will attempt to enroll individuals that meet qualification criteria but who previously were clients of day programs, given this is where most individuals in California services are placed despite the fact they could be working.

3.1. Limitations of the protocol

Although ours will be the largest community based RCT of IPS for autistic adults, it remains a relatively small clinical trial. Our current plan is to take what we learn, and if we are successful, to write a grant for a larger and better-powered trial that includes providers throughout the state. The current trial will be helpful for expanding the expertise, system's knowledge, and relationships that would enable us to do so. A second limitation of the current trial is that supported employment is not the best intervention for all autism individuals (and notably those who either do not require this level of employment services and those who would be excluded due to their cognitive abilities) so our results would not generalize to all autistic individuals. There may be some individuals with autism who would require supports but are not clients of either the Department of Rehabilitation or DDS. We have not yet found an agency that will accept private pay or insurance, thus limiting who can participate.

Finally, our protocol aims to contribute to a broader understanding of how to successfully adapt and implement evidence-based supported employment interventions like IPS within the developmental disability service system. Through our collaboration with service providers and RCs across diverse areas of California, we hope to generate practical insights into what facilitates or hinders IPS implementation in community settings. These findings will inform future adaptations of IPS and related interventions, ultimately supporting the goal of increasing access to meaningful, competitive employment opportunities for autistic adults across a wide range of functioning levels, backgrounds, and regions of our country and the world.

CRediT authorship contribution statement

Marjorie Solomon: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Jo A. Yon-Hernández: Writing – original draft, Supervision, Methodology, Investigation, Data curation. Steve Ruder: Writing – review & editing, Conceptualization. Susan R. McGurk: Writing – review & editing, Methodology, Funding acquisition, Conceptualization. Daniel Tancredi: Writing – review & editing, Methodology, Funding acquisition, Conceptualization. Yukari Takarae: Writing – review & editing. Aubyn C. Stahmer: Writing – review & editing, Methodology, Funding acquisition, Conceptualization.

Data statement

As required by our funder, NIH, we will upload all data to NDA (C5706).

Declaration of generative AI and AI-assisted technologies in the writing process

AI technologies were not used in the writing process for this manuscript.

Funding

This current work was supported by the National Institutes of Health R34 MH138725, and a California Department of Developmental Disability Employment Grant (A22-334) to Marjorie Solomon, PI. We also relied on contributions from the MIND Institute and Jeff and Ulrike Savage.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

No data was used for the research described in the article.