The transition to adulthood is challenging for many autistic individuals [1-4], with the transitioning period often coinciding with a loss of services, sometimes referred to as “dropping off a cliff,” rendering autistic individuals and their families without services [5,6]. To achieve the milestones of early adulthood as a transition-aged autistic youth necessitates skills for independence, which are reliant on self-knowledge about their own strengths and difficulties [1,7,8]. Psychoeducation is a well-established, evidence-based intervention for many psychiatric disorders, typically used to increase knowledge about their health condition, self-knowledge, and strategies in everyday life and to improve quality of life (QoL) [9,10]. Furthermore, psychoeducation aims to empower patients and actively engage them in their treatments as well as further service usage [10-12].

Indeed, psychoeducation is the recommended first-line intervention after a confirmed autism diagnosis [13-15]. However, the research field of psychoeducation for autistic individuals is still new and growing. In clinical settings, stepped-care models are recommended [14,16], by first providing a low-intensity intervention to most patients and later moving on to more intensive interventions if needed. In clinical practice in the United Kingdom [17] and in Sweden [13], the first step often entails studying psychoeducative materials, for example, reading an informative website or a pamphlet. The effects of self-study of informative websites about autism have not been evaluated; therefore, it is unknown whether it is an acceptable and effective intervention for autistic individuals. In addition, only a limited number of previous studies have described structured psychoeducational interventions for autistic individuals. In 2015, Gordon et al [18] conducted a randomized controlled trial (RCT) for autistic adolescents and their parents, with low attrition and high satisfaction among participants. Hidalgo et al [19] conducted an open trial for autistic adults and their close relations using group-based psychoeducational intervention, and Crane et al [20] used autistic-led peer groups; both were able to show low attrition as well as high satisfaction with program contents. However, in an open trial of a psychoeducational intervention for older autistic adults, Lenders et al [21] found a high level of attrition. Although all the abovementioned trials reported preliminary positive results of improved autism knowledge [18,19,21], QoL [19], and self-awareness [18], some studies had limitations such as small sample sizes [20,21] or no control group [19,21]. To further corroborate the evidence base of psychoeducation for autistic individuals, there is still a need for studies with rigorous trial designs, especially targeting older adolescents and emerging adults.

Psychoeducation using internet delivery could possibly contribute to a broader selection of intervention modalities and may also facilitate cognitive and communicative accessibility according to individual needs and preferences [22,23]. For example, autistic individuals explicitly ask for digital communication to improve comprehension, feel in control of the communication, and be able to express themselves freely and with humor [24,25]. Furthermore, recent studies of internet-delivered interventions aiming to promote health or behavioral changes for autistic adults showed positive preliminary results [26,27], and a continued use of technology-based interventions has been recommended in a review of technology-based interventions for autistic individuals [28]. Regarding internet-delivered psychoeducational interventions specifically, only 1 uncontrolled feasibility study has been published, showing good feasibility and acceptability for the internet-delivered psychoeducation SCOPE (Spectrum Computerized Psychoeducation) for transition-aged autistic youth [29].

In conjunction with examining the effectiveness of internet-delivered psychoeducation, it is necessary to consider the possible adverse effects of psychological interventions [30,31]. The tentative adverse effects of psychoeducation include lowered self-confidence and increased self-stigma, which are associated with reduced mental well-being [18,32]. Therefore, it is necessary to examine potentially harmful effects in terms of adverse changes in symptoms of anxiety, depression, and reduced self-acceptance following psychoeducation.

In summary, combining the asked-for, internet-based delivery with the recommended first-line psychoeducational intervention could potentially provide autism knowledge and empower transition-aged autistic youth. Therefore, promising interventions showing high acceptance and feasibility [29] need further evaluation in an RCT.

The objectives were to evaluate the effectiveness of SCOPE—a therapist-supported, internet-delivered psychoeducative program—using a 3-armed RCT design. We tested the hypotheses in Textbox 1.

This study is a 3-armed pragmatic RCT [33] of the internet-delivered psychoeducational intervention SCOPE for transition-aged autistic youth, conducted at Habilitation & Health (Stockholm Region outpatient disability services) in Stockholm, Sweden. The trial was registered at ClinicalTrials.gov (NCT03665363). Participants were enrolled between 2016 and 2021. Therapists were existing Habilitation and Health clinical staff trained in the SCOPE method.

In accordance with the ethical approval by the Swedish Ethical Review Authority (reference 2014/1868-31/5), prior to data collection, informed consent was obtained from all participants, and they were informed of their right to withdraw at any time. Participants were also informed about the purpose of the study and the methods used, and potential risks and benefits were discussed. Brief explanations about primary and potential secondary analyses of collected data were provided. Additionally, as approved by the Swedish Ethical Review Authority, all participants received a gift certificate of SEK 200 (US $22) after completing follow-up outcome measures.

Participants (n=141) were older adolescents and young adults, aged 16-25 years, with an autism diagnosis. Participants were recruited nationally through outpatient habilitation services, publicly funded psychiatric services, and private health care service providers. Swedish health care legislation allows nationwide service access, that is, individuals can seek health care in regions other than their listed address. All interested individuals were directed to a study recruitment website, provided by Habilitation & Health, to sign up and submit written consent. Professional staff and parents could assist participants in registering and accessing the study but could not apply on behalf of participants. Criteria for inclusion were (1) diagnosed autism (in accordance with the Diagnostic Statistical Manual of Mental Disorders or International Classification of Diseases, Tenth Revision) through community diagnosis with a documented IQ in the broad average range (ie, full-scale IQ>70); (2) aged 16-25 years; (3) sufficient Swedish language proficiency; and (4) the ability to log on and use the internet-based platform. Exclusion criteria were (1) diagnosed with acquired brain injury; (2) current problematic substance use; (3) severe suicidal ideation; (4) other severe psychiatric disorders (eg, psychosis); or (5) adverse personal circumstances (eg, being homeless) that could hinder participation. Ongoing pharmacological or other treatment was not considered a reason for exclusion.

A sample size of 126 participants was calculated to provide 80% power (1 – β) at a 2-sided 5% α level, with expectations of a moderate effect size. The effect size was estimated based on the results from the feasibility trial of SCOPE [29] and on a similar trial about psychoeducation for autistic adolescents [18]. To retain power with the expected attrition of 15%, we aimed to include 150 participants.

All statistical analyses were planned a priori; α was set at P=.05. The 2:1:1 ratio of randomization was chosen to enable a sufficient number of participants within the project time.

All variables were screened for violations of normality and sphericity; outcome data were approximately normally distributed. Statistical outliers for all outcome measures were screened using boxplots in SPSS (IBM Corp), and the few (n=7) extreme outliers (defined as outside the first or third quartile [SD 1.5]) were corrected by moving them to the closest whisker (ie, on the first or third quartile). Missingness was deemed completely at random for all outcome variables and time points using the Little missing completely at random test; we used allocation, age, gender, SRS-2, baseline BBQ, baseline HADS anxiety, and baseline HADS depression as assessment variables.

To determine changes between the intervention conditions (SCOPE vs TAU vs self-study) from the preintervention to postintervention time points and from the preintervention to follow-up time points, linear mixed models [49] used all available data without excluding cases with missing data. The models included random intercept, random time (slope), and other covariance structures; the model with the best fit was established using the log-likelihood ratio test. For outcomes where the slopes between self-study and TAU did not differ significantly (all outcomes apart from the ASD Quiz), the 2 control conditions were collapsed into one to increase power. For the ASD Quiz where the slopes of self-study and TAU differed significantly, post hoc tests were conducted to analyze the slopes of SCOPE with self-study and TAU, respectively. Between-group effect sizes, presented as Cohen d, were calculated by dividing the above-described coefficients of time×group interaction effect by the observed preintervention pooled SD [50]. The benchmark for Cohen d was based on the guidelines proposed by Cohen [51], where d values of 0.2, 0.5, and 0.8 are considered small, medium, and large effects, respectively.

The SCOPE intervention has been codeveloped and pilot-tested with autistic young adults. Several pilot iterations of the program were carried out online with the explicit purpose of codeveloping the program delivery (ie therapist involvement and feedback) and the contents evaluating the internet delivery, different aspects of the content (ie, module themes, phrasing, and video-vignette content), and therapist involvement and feedback. This work was carried out before the research projects commenced.

A total of 166 participants were assessed for eligibility; see Figure 1 for the CONSORT (Consolidated Standards of Reporting Trials) flowchart. In total, 4 participants either did not meet inclusion or met exclusion criteria, and another 17 declined participation. The remaining (n=145) participants were enrolled in the study; they were asked to complete baseline measures and subsequently randomized. Baseline characteristics for SCOPE, self-study active controls, and TAU controls are presented in Table 1. Recruitment to the study spanned 5 years to achieve the final number of participants. The lengthy recruitment process was mainly due to slow rates of interested participants.

Preintervention, postintervention, and follow-up assessments for all conditions based on the observed means and SDs are presented in Table 2, and interaction effects of between-group effects for preintervention to postintervention and preintervention to follow-up differences for all outcome measures are presented in Table 3.

The primary outcome was autism knowledge, estimated using the ASD Quiz, which significantly increased for SCOPE and self-study participants from the baseline to postintervention time point compared to TAU, with a small effect size of d=0.47 (P=.04) between SCOPE and TAU at the postintervention time point. In addition, self-study participants had significantly higher autism knowledge compared to TAU at the postintervention time point, yielding a medium effect size of d=0.60 (mean change score: 1.20, 95% CI 0.11-2.30; P=.03). However, self-study participants' scores returned to baseline levels and did not differ significantly from TAU at follow-up (mean change score: –0.13, 95% CI –1.20 to 0.94; P=.81). In contrast, the ASD Quiz scores remained stable for SCOPE participants compared to self-study participants (d=0.53, P=.02) and TAU participants (d=0.48, P=.02) at 3-month follow-up. Figure 2 shows a visual representation of these interaction effects.

The self-study active control and TAU conditions did not differ significantly for any secondary outcomes and were collapsed into one control group (see Table 3 for details).

The preintervention to postintervention increase in perceived QoL (the BBQ) was significantly larger in the SCOPE condition than in the combined controls. The between-group effect size was small at the postintervention time point (d=0.37, P=.02). However, QoL increased significantly at follow-up for SCOPE participants compared to combined controls, with a medium effect size of d=0.60 (P=.001).

Neither HADS Anxiety nor HADS Depression scores increased significantly between time points for SCOPE participants or the combined control condition. In addition, acceptance of diagnosis (modified AAQ-II) did not increase significantly between time points for SCOPE participants compared to the combined control condition.

Reported intervention credibility was similar for SCOPE participants (n=60; mean 7.69, SD (1.53) and self-study participants (n=30; mean 7.32, SD 1.90), measured after the first module.

At the postintervention time point, 2 (4%) of 62 SCOPE participants had not completed the intervention, whereas 6 (20%) out of 30 did not complete the self-study. In total, 5 (8%) out of 29 participants allocated to TAU dropped out of the study.

Although no SAEs were reported by SCOPE participants, there were 2 AEs spontaneously reported at the postintervention time point, where one participant dropped out of the study following module 6, referring to mental health problems, with no further contact after therapist attempts. However, it is unclear if the problems were related to SCOPE participation. The other participant completed SCOPE and all measurements but gave several spontaneous comments about feeling angry and commented that she or he felt some examples of autistic individuals were unfair and misrepresentative. The comments were given as part of the weekly written communication with the assigned therapist.

To the best of our knowledge, this is the first RCT to compare a structured, therapist-supported, internet-delivered psychoeducative intervention for transition-aged autistic youth against an active control condition (self-study) and TAU. In this trial, SCOPE was superior to self-study and TAU conditions at follow-up in providing autism knowledge and increasing QoL without increasing mental health problems. The findings of this RCT align with a previous open feasibility trial of SCOPE [29], in that the completion rate was high and the intervention was perceived as credible.

Our findings offer further support for the SCOPE intervention. In terms of our first hypothesis, we observed that SCOPE and self-study active control participants increased their autism knowledge, with small- to medium-sized effects, from the baseline to postintervention time points, whereas TAU did not. However, the increase in knowledge was only retained by SCOPE participants, showing a medium-sized effect at the 3-month follow-up, where self-study active controls participants’ knowledge returned to baseline rates. Furthermore, autism knowledge gains at follow-up were not related to preintervention treatment credibility, which was rated equally by the SCOPE and self-study groups. Instead, the maintained autism knowledge may be related to the carefully selected information, the codeveloper’s (ie, autistic emerging adults) influence on content and language, and the self-reflection questions possibly facilitating a more thorough cognitive processing of the learned information. Moreover, therapist guidance, through weekly written contact, is included in the SCOPE program but not in the self-study condition. However, this study cannot differentiate the active components in the interventions. Nevertheless, participants’ knowledge gains align with previous studies reporting similar results following group-based, structured psychoeducation for autistic adolescents and older adults [18,19]. In addition, a unique aspect of this study is the addition of an active comparator conducting self-studies, which has not previously been reported for similar interventions. Therefore, the results may have important clinical implications, as it is common for mental health care providers to give reading tips for self-study in the postdiagnostic phase. Based on the results of this study, self-study of internet-based reading materials without therapist support provided only temporary knowledge gains.

In addition to measuring autism knowledge, it is important to consider QoL as an important outcome following psychological interventions [31]. Improving QoL aligns well with the aims of psychoeducation—to empower and increase self-esteem. In salient theories on QoL, self-esteem is an integral part of QoL [52,53]. Moreover, Schalock [52] and the autism-focused review of QoL [54] see empowerment as a component of self-determination, which in turn is a component of QoL. As such, our second hypothesis about increased QoL following SCOPE was confirmed by the results, in that SCOPE was superior to controls in increasing QoL at the postintervention time point and even more so at follow-up. This was in line with the preliminary results of Hidalgo et al [19], showing a small increase in QoL in autistic adults following a structured psychoeducational group. Furthermore, enabling an increased sense of QoL is in line with up-to-date goals of treatment guidelines for autism [16] and calls to action in efficient postdiagnostic care [15], explicitly recommending psychoeducation as a first-line intervention after a confirmed autism diagnosis. Attempting to reframe and guide an individual’s perception of what autism is for them aligns with modern goals for psychoeducation [55]. A few studies on psychoeducation for autistic individuals have seen tentative effects on self-awareness and self-esteem [18,20], also described in the qualitative responses from the feasibility trial of SCOPE [29]. Increased knowledge about autism could act as a mediator of changes in self-perception and recognition of how autism impacts themselves and their environment. Although we were not able to conduct that kind of mediator analysis, this could be addressed in future studies.

Moreover, related to the third hypothesis, the increased knowledge among SCOPE and self-study participants was not coupled with increased symptoms of anxiety or depression. Although the primary goals of psychoeducation are not expected to have direct effects on specific symptoms such as anxiety and depression [9,55], and consequently, such effects have not previously been consistently found in the literature on psychoeducational interventions about autism [18,19,21,29,56]. Therefore, positive effects on symptoms of anxiety and depression were not expected to be improved in this study. Notably, the stability of anxiety and depression symptoms over time is an important finding, as psychoeducation may be associated with initial adverse psychological effects [17,32].

Furthermore, this study also partly supports the open feasibility trial of SCOPE [29], demonstrating high completion rates in the SCOPE. However, the recruitment of study participants was slow, with approximately 30 participants included every year (2016-2020). Given the apparent willingness of transition-aged autistic participants to receive interventions [6,57,58], we expected a higher flow of participants. The slow recruitment could suggest that we did not reach the intended target audience; perhaps, the intervention was perceived as too time-consuming or unappealing to transition-aged autistic youth, or the mode of study entry (application through the study website using e-identification) was cumbersome. Unfortunately, this study was unable to clearly identify the complex reasons for participant recruitment difficulties in this trial. A qualitative analysis of participants’ written responses and evaluations could deepen our understanding of the lived experiences of SCOPE participation and thus inform future implementation of similar programs in clinical practice.

This study has several limitations. First, the primary outcome measure (ASD Quiz) was based on previous similar measures [19,40,59], but it has not been psychometrically evaluated. However, no consensus regarding autism knowledge assessment exists, with considerable heterogeneity in the field [59], and the ASD Quiz used in this study has previously achieved acceptable reliability [29]. Although the knowledge gains were modest regarding absolute numbers, the effect size was medium at 3-month follow-up. However, a risk for ceiling effects in the primary outcome could have limited our ability to measure the treatment effects. Second, identifying secondary outcome measures available in Swedish for autistic individuals was challenging, especially at the inception of this study. For future studies, using outcome measures for self-esteem and empowerment would be valuable to evaluate the core aims of psychoeducation. Failing to measure these intervention outcomes misses the mark of understanding the potential effects of SCOPE within these domains. Third, detection bias may have been introduced due to assessors (the autistic individuals) being aware of intervention allocation at postintervention assessments [60]. Finally, the study’s generalizability may have been limited due to potential sampling issues. As noted in Table 1, the sample consists mainly of newly diagnosed female participants. The female majority may be reflective of findings maintaining that female individuals obtain their diagnosis later than male individuals [61,62], which may explain the oversampling of female participants in this type of intervention study. Moreover, autistic female individuals also report higher rates of mental health problems [63], perhaps contributing to more treatment-seeking behavior in general. Additionally, our participants reported many health care contacts (see Multimedia Appendix 1). Overall, our sample could represent a treatment-seeking and receiving subset of transition-age autistic youth that is not representative of those that have “fallen off a cliff” [6].

This study shows that structured internet-delivered psychoeducation is effective in an outpatient context. The SCOPE program was superior to self-study active controls and TAU in imparting autism knowledge and increasing QoL. These results were maintained or further improved at the 3-month follow-up. In addition, providing internet-delivered services enables wide geographical accessibility and provides digital communication sought by transition-aged autistic youth.