The Efficacy of a Culturally-Adapted Group-based Parent Coaching Program for Autistic Children in China via Telehealth: A Randomized Controlled Trial

Abstract

The study aimed to examine the efficacy of a culturally-adapted, group-based parent coaching program for autistic children in China delivered via telehealth. A randomized controlled trial was conducted, with 18 parents allocated to the self-directed group that received the intervention through an online learning platform, and 19 parents allocated to the web + group therapy group, which included the same program along with weekly 1.5-hour group coaching sessions via videoconferencing. The primary outcomes were parents’ mental health and children’s adaptive functioning, while the secondary outcomes focused on the child behaviors, parenting stress and parenting style, and family quality of life. Linear Mixed Models were used to evaluate treatment effects across time and to model longitudinal trajectories of outcomes in both children and parents. Both intervention groups showed significant improvements in children’s communication skills (F (1, 60.27) = 29.86, p < 0.001) and social engagement (F (1, 60.07) = 11.73, p = 0.001), as well as reductions in parenting stress (F (1, 59.07) = 8.76, p = 0.004) and anxiety levels (F (1, 57.62) = 4.84, p = 0.032). Additionally, the group-based parent coaching via videoconferencing was associated with greater improvements in children’s quality of life (F (1, 59.95) = 5.90, p = 0.018) and parents’ anxiety outcomes (F (1, 57.62) = 4.84, p = 0.032). This study demonstrated the efficacy of a culturally adapted telehealth intervention for both autistic children and their parents. The preliminary findings suggest positive outcomes in children’s adaptive functioning and parents’ mental well-being. Group-based parent coaching through videoconferencing could be a promising and practical model for in-home services, particularly for families with limited access to in-person services.

Introduction

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition characterized by impairments in social interaction and communication and by the presence of restricted, repetitive behaviors and interests (American Psychiatric Association & Association., 2013). Individuals with autism often have co-occurring medical conditions, such as attention deficit hyperactivity disorder (ADHD), anxiety disorders, epilepsy, gastrointestinal (GI) issues, and sleep disorders (O’Nions et al., 2018; Petrou et al., 2018). These co-occurring conditions not only have a significant impact on autistic child, limiting their daily activity participation, but also interact with the core symptoms of autism, leading to complex behavioral presentations, which in turn exacerbate additional financial and time demands on families with autistic children (Dovgan & Mazurek, 2019).

Parents of autistic children experience higher levels of stress (Barroso et al., 2018) and other negative affects compared to parents of typically developing children, and even higher than parents of children with other disabilities (Estes et al., 2009; Hayes & Watson, 2013; Phetrasuwan & Shandor Miles, 2009). Parents of autistic children started to report high levels of stress after the diagnosis of their children (Gupta, 2007; Meirsschaut et al., 2010). From then, the severity of the child’s autism symptoms (Ingersoll & Hambrick, 2011; Kasari et al., 1997; Lyons et al., 2010) and the co-occurring behavior problems (Lecavalier et al., 2006; McStay et al., 2014a; McStay, Trembath et al., 2014) significantly influenced parenting stress in parents of autistic children (Lin et al., 2021). Studies indicate a significant concurrent bidirectional relationship between parenting stress and child behavior problems in autism, with parenting stress contributing to an increase in autism symptoms and behavior problems, and vice versa (Lin et al., 2021; Rezendes & Scarpa, 2011; Rodriguez et al., 2019; Stephenson et al., 2023). Highly stressed parents are more likely to experience other mental health problems, such as depression and anxiety, due to the impact chronic stressors can have on their cardiovascular, immune, and gastrointestinal systems (Miodrag & Hodapp, 2010). Elevated levels of depression and anxiety were also reported by parents of autistic children (Padden & James, 2017), leading to a lower quality of life (Vasilopoulou & Nisbet, 2016) and decreased parenting self-efficacy (Giallo et al., 2013).

While much research has focused on Western contexts, examining the experiences of parents in China provides valuable insights into how cultural and societal factors may reflect and influence the needs of families and their response to available autism interventions. Parents of autistic children in China reported similar elevated parenting stress as western parents (Lu et al., 2015). However, Chinese mothers’ parenting stress was associated not only with the child’s behavioral symptoms but also with the limited medical and educational resources and financial burden (Wu et al., 2015). Additionally, the scarcity of specialists affects timely diagnosis and leads to delayed treatments and limited evidence-based services in China (Su et al., 2013; Xiong et al., 2011). McCabe (2012) discovered that intervention services for autism in China often used a mix of methods, and evidence-based practices coexisted with self-designed interventions (Clark et al., 2019). Liu et al. (2020) conducted a systematic review examining evidence of parent-mediated interventions in China. Only four identified programs (4/21) met the criteria in using evidence-based intervention models, and 3 of 21 identified studies reported the use of cultural adaptation, and over half (14/21) of the included studies failed their quality evaluation. According to a qualitative study, the fidelity of online parent training during the COVID-19 pandemic in China was limited by inexperienced teachers, and tensions between inexperienced teachers and parents demanded more culturally-relevant instructional materials (McDevitt, 2021). Under such circumstances, a well-structured, culturally-adapted parent program via telehealth may offer a promising solution.

Research has shown that parent psychoeducational (PPE) training (DaWalt et al., 2018; Hemdi & Daley, 2017) and parent-mediated intervention (PMI) (Ingersoll & Wainer, 2013; Kasari et al., 2015; Oono et al., 2013; Parsons et al., 2017; Stadnick et al., 2015) are both effective in reducing stress, anxiety, and autism severity in children (Baumann et al., 2015; Deb et al., 2020; Hemdi & Daley, 2017; Nevill et al., 2018). PPE provides parents with updated knowledge and coping strategies, while PMI actively engages parents in therapy delivery to promote skill acquisition or behavior change in the child (Green et al., 2015, 2017). Those results have been found in person mode of delivery (Liu et al., 2020; Nevill et al., 2018), in group-based setting (O’Donovan et al., 2019) or in telehealth format (Hemdi & Daley, 2017; Lindgren et al., 2016). Parent coaching has demonstrated its capacity to support families while waiting for intensive early interventions (Abouzeid et al., 2020; Hernandez-Ruiz, 2019). Several studies have shown that parent coaching interventions are feasible and acceptable to both parents and providers (Abouzeid et al., 2020; Hernandez-Ruiz, 2019). Additionally, tele-delivered parent coaching interventions have been successfully tested in rural areas (Guðmundsdóttir et al., 2018; Salomone & Maurizio Arduino, 2016), and for underrepresented population (Yllades et al., 2021). These interventions have been shown to enhance parents’ teaching skills (Rogers et al., 2012b) and increase their ability to reflect effectively (Hernandez-Ruiz, 2019; Siller et al., 2018).

The parent-mediated ESDM (P-ESDM) focused on the quality of social interaction between autistic children and their caregivers, as well as a framework to embed specific objectives across all developmental domains (Rogers et al., 2012b). Previous research indicated P-ESDM has the potential to be delivered successfully through telehealth methods (Vismara et al., 2012) and be adapted to different context (Colombi et al., 2018; Holzinger et al., 2019). Two studies in China examined the effectiveness of P-ESDM on autistic preschoolers in mainland China. Xu et al. (2017) conducted a pilot study examining the efficacy of an 8-week ESDM plus other eclectic intervention services delivered by local teachers. Zhou et al. (2018) conducted a non-randomized trial to compare a 26-week, high-intensity, P-ESDM intervention with a general community-based treatment program using the Chinese version of P-ESDM. Both studies were conducted in a 1-on-1 clinical in-person setting and showed positive results in participating children.

Group-based parent training may help to offset the shortage of professionals in China who specialize in autism-related services. O’Donovan et al. (2019) conducted a systematic review to examine group-based parent training interventions for families of autistic children. According to their findings, group-based parent training interventions improved parent behavior, children’s behavior, parent health, and social support. However, with 12 out of 13 included studies falling below the quality rating threshold, O’Donovan et al. (2019) stated that those findings should to be interpreted with caution and be further investigated using better study design and methodology. Banbury et al. (2018) found that health professional-led group videoconferencing was a feasible option for that to provide education and social support into the home setting. Group-based interventions via videoconferencing enhanced accessibility, and the effects of the videoconferencing groups were comparable to those of in-person groups (Banbury et al., 2018).

The lack of resource and professional shortages limited evidence-based interventions in low and middle-income countries (Durkin et al., 2015; Olusanya et al., 2018). One solution for autism care in low-resource settings could be parent-mediated interventions (PMIs) (Blake et al., 2017; Lee & Meadan, 2021; Liu et al., 2020; Parsons et al., 2017). Research has shown the effectiveness of PMIs in both in-person (Abouzeid et al., 2020; Liu et al., 2020; Nevill et al., 2018; Oono et al., 2013) and telehealth settings (Hao et al., 2021; Jurek et al., 2021; Parsons et al., 2017). A systematic review by Parsons et al. (2017) found that remotely delivered parent-mediated intervention showed promising results in improving caregiver expertise and intervention fidelity. However, only one program out of seven included studies conducted randomized controlled trials (Ingersoll & Berger, 2015; Ingersoll et al., 2016). Additionally, the majority of parent interventions have been designed to teach parents to apply specific intervention strategies focused on supporting the child (Ingersoll et al., 2016). Interventions designed mainly for the benefit of parents focused only on parental well-being and parenting competence (Ferraioli & Harris, 2013). A critical gap in current parent training studies is that the direct beneficiary of intervention often focused on the child’s improvements or the parent’s progress, rather than considering the progress of both parties concurrently. Weitlauf et al. (2022) emphasized the significance of supporting parents’ well-being in parent-mediated early intervention for autism. However, since the study lacked specific measures for parental outcomes, it is challenging to determine the effectiveness of MBSR on both child and parents. To better serve families in low-resource settings, more culturally appropriate PMI materials and interventions should be developed and tested using rigorous RCT design. This could foster a collaborative and supportive relationship between caregivers and professionals, leading to improved access to autism services for families in low-resource settings.

This group-based parent coaching program was culturally adapted based on P-ESDM. The program focused on teaching caregivers how to facilitate child development in a natural home environment, and it was also designed to improve parental competence and alleviate parenting stress and mental health outcomes. The process of cultural adaptation (Qu et al., in review) and the implementation evaluation (Qu et al., 2022) of the intervention were reported in other two articles. The cultural adaptation process followed Bernal et al. (1995)’s Ecological Validity Model(EVM) and the family-centered capacity building approach (Trivette & Keilty, 2017) to determine “what” to adapt and “how” to adapt. Eight dimensions in the EVM were assessed to ensure the fit in the intervention content and telehealth delivery. Five dimensions were determined to modify: language, persons, content, methods, and context. Metaphors, concepts, and goals stayed as original P-ESDM (Qu et al., in review). For example, we still used the P-ESDM concepts and goals which based on the behavioral and developmental principles. We emphasized to set families goals which not only focused on facilitating child development but also promoting the well-being of the parents. This culturally-adapted program differed from the original P-ESDM in following aspects. First, the telehealth intervention applied two service delivery models, a self-directed digital learning condition that connected people to the training, and an enhanced version that included synchronous group-based parent coaching therapy (web + group therapy), which connected people to people (professionals and peers). The program was delivered through the Canvas online learning management system and the Chinese videoconferencing software DingTalk. Second, we created four types of learning materials tailored to better meet the needs of our target population. (1) Pre-recorded Lectures, consisting of multiple short videos that presented parent-friendly and culturally appropriate content; (2) Demonstration Videos, featuring real parent-child interactions to model intervention strategies; (3) Commentary Videos, adding professional comments to help parents reflect on their own practices; and (4) Module-specific Practice Manuals, offering step-by-step instructions and guidance for parents. Additionally, the group-based coaching sessions were designed using mediated parent learning principles and the family-centered capacity-building approach. The telehealth delivery of this culturally adapted program was found to be acceptable, appropriate, and feasible for Chinese parents, parents in both groups reported high level of satisfaction and increased parent perceived competence and self-efficacy (Qu et al., 2022).

The current study aimed to examine the efficacy of this culturally adapted telehealth intervention on children’s adaptive functioning, child behaviors, and children’s quality of life, as well as the parents’ mental health, parenting stress and parenting styles, and parents’ quality of life. The primary outcomes of the study were parents’ mental health and child’s adaptive functioning over time. The secondary outcomes focused on the child behaviors, parenting stress and parenting style, and quality of life of the families.

Methods

An RCT was conducted to investigate the efficacy of two telehealth service conditions on parent and child outcomes. Outcome measures were assessed at baseline (T1), post-intervention (T2), and 1-month follow-up (T3). Ethical approval was granted by the Health Sciences and Behavioral Sciences Institutional Review Board at the University of Michigan (HUM00182526).

Participants

Table 1 presents participants’ demographic information. The inclusion criteria for participating in the feasibility study were: (1) parents over the age of 18, (2) with a child aged 2–5 years old at entry, with a diagnosis or a strongly suspected diagnosis of ASD made by a psychiatrist within 6 months, (3) with access to the Internet and a digital device to receive the web-based intervention, and (4) the participating child met or exceeded cutoff scores of 10 (social interaction), 8 (communication and language), and 3 (restricted and repetitive patterns of behavior or interest) on the Autism Diagnostic Interview-Revised (ADI-R) (Lord et al., 1994). Exclusion criteria were: (a) parents attending another parenting program during the intervention phase; (b) children with genetic defects or inherited diseases (e.g., Rett Syndrome, Fragile X Syndrome); (c) children with physical or sensory disabilities; (d) family in crisis (e.g., the child at risk of residential placement); (e) refusal to give consent to take part in the research. There was no exclusion criteria based on co-occurring intellectual disability. Demographic characteristics were examined for group difference using independent samples t-tests for continuous variables and Chi-square tests for categorical variables. No significant group difference was found at baseline.

Table 1.

Participants demographics at baseline

Characteristic	Self-directed (N = 18)	Web + Group therapy (N = 19)
Child Age (years)
Mean ± SD	3.17 ± 0.70	3.25 ± 0.95
Range	1.68–4.62	1.70–5.16
Child Sex, N (%)
Males	16 (88.89)	16 (84.21)
Females	2 (11.11)	3 (15.79)
M-CHAT
Mean ± SD	9.39 ± 3.05	7.21 ± 3.82
ADI-R, Mean ± SD
Social impairment	20.61 ± 3.57	19.16 ± 4.25
Communication	10.94 ± 2.01	9.42 ± 2.81
Repetitive interest	3.78 ± 1.73	3.32 ± 1.97
Total	38.72 ± 6.33	35.26 ± 7.70
Participating Parent, N (%)
Mother	17 (94.44)	15 (78.95)
Adoptive Mother	0 (0.00)	1 (5.26)
Father	1 (5.56)	3 (15.79)
Family structure, N (%)
Nuclear Family	6 (33.33)	4 (21.05)
Three Generations Family	10 (55.56)	14 (73.68)
Single Parent Family	1 (5.56)	0 (0.00)
Others	1 (5.56)	1 (5.26)
Monthly Family Income, N (%)
< 460 Dollars	4 (22.22)	1 (5.26)
460–770 Dollars	0 (0.00)	3 (15.79)
770–1540 Dollars	4 (22.22)	4 (21.05)
1540–3080 Dollars	6 (33.33)	7 (36.84)
3080–7690 Dollars	4 (22.22)	4 (21.05)
Location, N (%)
Urban	15 (83.33)	16 (84.21)
Rural	3 (16.67)	3 (15.79)
Mother’s Education, N (%)
High school or below	2 (11.11)	2 (10.53)
Junior college	6 (33.33)	6 (31.58)
Undergraduate	7 (38.89)	8 (42.11)
Graduate or above	3 (16.67)	3 (15.79)
Mother’s Employment, N (%)
Full-time	6 (33.33)	10 (52.63)
Part-time	0 (0.00)	1 (5.26)
Unemployed	11 (61.11)	7 (36.84)
Self-employed	1 (5.56)	1 (5.26)

Technology Requirements and Preparation

The technology requirements for participating families included a digital device (A computer, iPad, a tablet, or smartphones with a screen), Webcam and microphone, and access to Email and Internet, etc. Group videoconferencing through DingTalk required a minimum bandwidth speed of 2Mbps.

Before the intervention started, a Program Workbook was distributed to participants in both conditions. The Program Workbook included: (1) participant username and code for the program website, (2) picture tutorials for using Canvas and DingTalk. We added (3) general guidelines of the group therapy, “Rules and Etiquette”, for parents in the web + group therapy condition. Part (3) listed ground rules for language use in group discussion and a checklist for the participating in virtual group sessions, including hardware preparation (i.e., device, electricity, earphones, mic, and internet connection), environment setup (i.e., camera background, light, noise, and privacy), and confidentiality.

Intervention Procedures

The study recruited 67 participants through a digital portal. 10 families were not enrolled because the study was part of a Ph.D. dissertation, and there was a lack of funding to recruit and retain additional therapists to provide such intervention service to all eligible families. In the end, 44 families randomly assigned to either of the two conditions using a computer-generated random group sequence. Prior to enrollment, parents completed a virtual consultation session lasting 1.5–2 h to complete eligibility screening for their children using Modified Checklist for Autism in Toddlers, Revised with Follow-up (Robins et al., 2009) and ADI-R (Lord et al., 1994). Children were included in the study if they scored above cutoff scores on M-CHAT-R/F and the ADI-R. The ADI-R interviews were conducted by a trained therapist who completed ADI-R clinical training. The screening of eligibility, including an ADI-R interview and the M-CHAT-R/F, was completed before allocation, and the assessor of the eligibility assessments was blinded to the treatment allocation. As a result, 37 families enrolled in the program, while 7 declined to participate due to reasons such as time conflicts (3/7), privacy concerns (1/7), work (1/7), and pregnancy (2/7).

37 families were randomly assigned to either a self-directed (comparison) group or a web + group therapy (treatment) group. Both groups self-navigated the same program website, delivered in Chinese through the Canvas online learning management system. The program consisted of 12 modules covering topics in P-ESDM: (1) Autism and what to expect, (2) Increasing the child’s attention, (3) Sensory social routines, (4) Dyadic engagement, (5) Nonverbal communication, (6) Imitation, (7) ABC’s of learning, (8) Joint attention, (9) Play, (10) Pretend play, (11) Speech development, and (12) Program summary. Each module took an average of 71 min to complete (Supplement Table 1, The Characteristics of Learning Materials on Program Website). New modules were released on Mondays and included four types of learning materials, pre-recorded lectures, demonstration videos, commentary videos, and a module-specific practice manual. Online Q&A sessions were provided for each module, and parents were given the option to submit a module-specific assignment.

Parents in the self-directed web group were encouraged to complete one module per week and practice the intervention strategies taught at home. Those who submitted homework received individualized feedback based on their performance in parent-child interaction video clips. Homework video-clips of their home practice also served as a valuable resource for conducting fidelity checks, ensuring consistent delivery of the intervention. In addition to the online program, the treatment group participated in weekly group-based parent coaching sessions via videoconferencing, with each subgroup consisting of 6–7 parents to enable full participation. The interventionist met with each subgroup once per week for 1.5 h through DingTalk, a videoconferencing software. The weekly group therapy session followed the module topics and used a structured model based on mediated learning principles (Schertz & Horn, 2017; Schertz et al., 2018). Five structured group activities were modified to guide parent learning (Supplement Table 2; Activities and Roles in Weekly Group Session). The treatment group attended 11 group therapy sessions with an average attendance rate of 92%.

Measures

Child and parent outcome measures were collected digitally through Qualtrics at baseline (T1), immediately post intervention (T2), and 1-month follow-up (T3). The interventionist remained blinded to all test results during the intervention.

Child Outcome Measures

Adaptive Functioning

Vineland Adaptive Behavior Scales -Third Edition.

The Vineland Adaptive Behavior Scales - Third Edition, Comprehensive Parent/Caregiver Form (Sparrow et al., 2016) is a 502-item questionnaire designed to be completed by an adult respondent who is knowledgeable about the examinee’s everyday adaptive functioning. The Comprehensive Parent/Caregiver Form generates standard scores for Communication, Daily Living Skills, Socialization, Motor Skills, and Adaptive Behavior Composite. Higher scores indicate better adaptive skills and the cut-off used to indicate low levels of adaptive functioning is < 70. All domains showed excellent internal consistency (α = 0.94–0.99), good test–retest reliability (r = 0.64–0.94) and concurrent validity (r = 0.67–0.81) (Pepperdine & McCrimmon, 2017). The Chinese version of VABS demonstrated adequate psychometric properties, with acceptable levels of test–retest reliability, interrater reliability, and split-half reliability for each subscale (ICC = 0.62–0.89) (Wu et al., 2004). VABS-2 were reported good internal consistency for all subscales in preschool children with autism (Balboni et al., 2016). The Chinese version of the Comprehensive Parent/Caregiver Form is purchased from the publisher with a permission to use digitally during the study period.

Child Behavior

Child Behavior Checklist (CBCL).

The CBCL/1.5–5 (Achenbach & Rescorla (2004) is a clinical measurement used to assess children’s problematic behavior. It involves 99 specific behaviors rated by caregivers as 0 (Not True of the child), 1 (somewhat or sometimes true), or 2 (very true or often true) and produces scores in Internalizing Problems, Externalizing Problems, and Total Problems. The CBCL/1.5–5 has shown good internal consistency for the syndrome scales in preschool population (Achenbach & Rescorla, 2000). Rescorla et al. (2019) reported the CBCL/1.5–5 had high internal consistency in differentiating children diagnosed with and without autism spectrum disorder. The Chinese version of the CBCL/1.5–5 has been shown to be reliable and valid (Liu et al., 2011).

Child Quality of Life

Pediatric Quality of Life Inventory–Fourth Version (Peds-QoL).

The Peds-QoL measures the health-related quality of life for children and adolescents aged 2 to 18 years (Varni et al., 2001). It includes subscales such as Physical Functioning, Emotional Functioning, Social Functioning, and School Functioning, and computes standard scores. The Peds-QoL is also reliable in the Chinese population. The PedsQoL-4.0 Generic Core Scales reported good internal consistency (0.82–0.92) in children with pediatric conditions, including autism (Limbers et al., 2011). The Peds-QoL has good internal consistency (0.86–0.90) in Chinese population (Hao et al., 2010).

Parent Outcome Measures

Parent Mental Health

Parent mental health was assessed using the Generalized Anxiety Disorder 7-item Scale (GAD-7) (Spitzer et al., 2006) and Patient Health Questionnaire 9-item Scale (PHQ-9) (Kroenke et al., 2001).

Generalized Anxiety Disorder 7-item Scale (GAD-7).

GAD-7 is a self-report questionnaire that identifies generalized anxiety disorder and assesses symptom severity (Spitzer et al., 2006). Scores range from 0 to 21, with > 5, >10, and > 15 indicating mild, moderate, and severe anxiety, respectively (Spitzer et al., 2006). The Chinese version of GAD-7 showed excellent reliability (α = 0.91) and good validity in Chinese population (Zeng et al., 2013).

Patient Health Questionnaire 9-item Scale (PHQ-9).

The PHQ-9 is a 9-item depression screening tool. The items rely on DSM-IV criteria for the diagnosis of depression. The PHQ-9 score ranges from 0 to 27, PHQ-9 scores of > 5, >10, > 15, and > 20 represent mild, moderate, moderately severe, and severe depression, respectively. The PHQ-9 is a depression screening tool consisting of 9 items based on DSM-IV criteria for depression diagnosis (Kroenke et al., 2001). The PHQ-9 score ranges from 0 to 27, with a score of > 5, >10, > 15, and > 20 representing mild, moderate, moderately severe, and severe depression, respectively. The Chinese version of PHQ-9 showed good reliability (α = 0.86) and validity (r = 0.29, p < 0.001) in Chinese population (Wang et al., 2014).

Parent Quality of Life

WHO Quality of Life Scale (WHOQOL-BREF).

The WHOQOL-BREF is a 26-item instrument based on the WHOQOL-100, which evaluates health-related functions across four domains: physical health, psychological health, social relationships, and environment (Group, 1998). The Chinese version of the WHOQOL-BREF has demonstrated good reliability (α = 0.73) and validity (CFI = 0.88) among the Chinese population(Zhang et al., 2012).

Parenting Stress and Parenting Styles

Parent Stress Index-4, Short Form (PSI-4, SF).

The PSI-SF is a parent-report questionnaire that assesses parenting stress in parents of children under 12 years old (Abidin, 2012). The PSI-SF includes 36 items, divided into three subscales: Parental Distress (PD), Parent-Child Dysfunctional Interaction (PCDI), and Difficult Child (DC). The PSI-SF has been widely used in both clinical and typical populations, with a percentile score falling at or above 90 indicates a clinically significant high level of stress (Abidin, 1995). The authors of this study obtained permission to use the Chinese version of the PSI-SF and collect data digitally.

The Parenting Styles and Dimensions Questionnaire (PSDQ).

The PSDQ is a parent-report questionnaire consisting of 32 items that evaluates parenting styles (Robinson et al., 1995). Each item is rated using a 5-point Likert Rating Scale (1 = Never, 5 =Always), with higher scores indicating more frequent use of the described behavior. The questionnaire calculates scores for three types of parenting styles: Authoritative, Authoritarian, and Permissive. The Chinese version of PSDQ has demonstrated good reliability (α > 0.62) and validity (TLI: 0.808–0.920, RMSEA: 0.052–0.07), and excellent internal consistency across all three subscales in Chinese parents(Fu et al., 2013). Additionally, it was widely used in research involving parents of children with special needs (Phillips et al., 2017; van Steijn et al., 2013).

Statistical Analysis

A priori power analysis was conducted using G*Power3 (Faul et al., 2007) to estimate the minimal sample size required to test the study hypotheses. In order to test the group difference and within-between interaction (Linear multiple regression: Fixed model) using a medium effect size (primary outcomes) = 0.40 (Da Paz & Wallander, 2017; Fuller et al., 2020) at a significance criterion of α = 0.05, the result showed that a total sample of 28 participants with 2 groups of N = 14 was required to achieve a power of 0.80. With an expected attrition rate of 20% (Karlson & Rapoff, 2008), a total sample of 36 families, with at least 18 participants per group were randomized in our study.

Statistical analysis were conducted using SPSS 28.0 (Spss, 2016), with a significance level of p < 0.05. Non-normality of outcome variables was assessed visually using boxplots and statistically using the Shapiro–Wilk Test. Standard scores, or Percentile scores were used for outcome measures such as VABS-3 and PSI-SF-4. The intention-to-treat analysis followed a statistical analysis plan using Linear Mixed Models (LMM) (Chakraborty & Gu, 2009; McCoy, 2017). We estimated treatment effect based on the initial randomized assignment while accounting for the longitudinal repeated-measured data. This approach allows participants with missing data to be included in the model using maximum likelihood estimation. Linear Mixed Models were used to model longitudinal trajectories of outcomes while accounting for subject-level variance. The effects of treatment (self-directed or web + group therapy), time (baseline (0), post-intervention (3), and 1-month follow-up (4), and treatment by time interactions were explored. Random intercepts were included to account for repeated-measure correlation. Missing values were considered valid data in the mixed linear models, and statistics were based on all subject cases with valid data for all variables in the model.

Results

Figure 1 outlines the design of the randomized trial and shows participants flow through each phase of the study. A total of 34/37 (86.49%) participants completed the post assessment, and 8/37 (21.62%) were lost from follow-up.

Fig. 1.

Consort flow diagram of the progress of two groups

Baseline Differences

The randomization procedure created baseline equivalency between treatment conditions. Baseline demographic characteristics are shown in Table 1, no significant group difference was identified in child’s age, gender, verbal status, M-CHAT, ADI-R, family structure, family income, location, mother’s education, and mother’s employment.

Shaprio–Wilkes tests indicated data were normally distributed, independent t-tests were conducted to compare baseline clinical measures (Tables 2 and 3). There were no significant group differences in baseline clinical measures in both child and parent at baseline, as p values for baseline measures ranged from 0.06 to 1 (Tables 2 and 3).

Table 2.

Children baseline outcomes

	Self-directed Web (n = 18)	Web + group therapy (n = 19)	p
	M (SD)	M (SD)
VABS
Communication	69.44 (15.02)	64.32 (17.97)	0.35
Daily Living	77.89 (9.68)	81.37 (14.00)	0.39
Socialization	64.78 (9.93)	61.95 (13.12)	0.47
Motor	88.00 (13.48)	92.47 (12.31)	0.30
Adaptive Behavior	70.44 (7.19)	69.53 (9.69)	0.75
Composite
CBCL
Internalizing	10.83 (3.79)	9.42 (3.20)	0.23
Externalizing	9.83 (3.62)	7.74 (2.77)	0.06
Total problems	34.83 (8.46)	30.26 (7.61)	0.09
Peds-QoL
Total Score	61.04 (16.13)	59.38 (15.70)	0.75
Physical Health Summary Score	69.97 (16.51)	69.74 (14.88)	0.96
Psychosocial Health Summary Score	54.11 (18.45)	52.01 (18.06)	0.73

Note M, mean; SD, standard deviation; VABS, Vineland Adaptive Behavior Scales (standard scores); CBCL, Child Behavior Checklist (raw scores); Peds-QoL, Pediatric Quality of Life Inventory

Table 3.

Parent baseline outcomes

	Self-directed Web (n = 18)	Web + group therapy (n = 19)	p
	M (SD)	M (SD)
GAD-7	6.17 (5.46)	6.16 (5.38)	1
PHQ-9	7.39 (6.50)	6.58 (5.34)	0.68
QoL
Physical health	0.59 (0.14)	0.66 (0.13)	0.13
Psychological	0.51 (0.19)	0.50 (0.17)	0.85
Social relationships	0.51 (0.13)	0.57 (0.15)	0.20
Environment	0.54 (0.18)	0.53 (0.15)	0.90
Total Score	81.83 (15.18)	84.53 (13.41)	0.57
PSI
Parental Distress	36.89 (7.95)	37.42 (11.57)	0.87
Parent-Child Dysfunction Interaction	32.44 (8.99)	30.00 (7.89)	0.38
Difficult Child	35.50 (8.10)	30.95 (7.67)	0.09
Total Stress Score	104.83 (22.13)	98.37 (22.67)	0.39
PSDQ
Authoritative	3.62 (0.40)	3.44 (0.50)	0.23
Authoritarian	2.27 (0.43)	2.29 (0.49)	0.91
Permissive	2.56 (0.43)	2.60 (0.68)	0.81

Note. M, mean; SD, standard deviation; GAD-7, Generalized Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire-9; QoL, WHO Quality of Life Scale (percentiles); PSI, Parenting Stress Index-4 (percentiles), Short Form; PSDQ, Parenting Styles and Dimensions Questionnaire

Linear Mixed Models Outcomes

Linear Mixed Models were conducted to estimate the treatment effect across time and to model the longitudinal outcomes trajectories while accounting for subject-level variance. Random effects were assumed to vary across participants, while fixed effects were assumed to be constant across groups or time. Descriptive characteristics of child and parent outcome measures by groups across three timepoints and Linear Mixed Models fixed effects results were presented in Tables 4 and 5, respectively.

Table 4.

Child outcome measures and linear mixed model results

	Self-directed Web						Web + Group therapy						Fixed Effect
	Baseline (T1)		Post intervention (T2)		Follow-up (T3)		Baseline (T1)		Post intervention (T2)		Follow-up (T3)		Group		Time		Group × Time
	n	M (SD)	n	M (SD)	n	M (SD)	n	M (SD)	n	M (SD)	n	M (SD)	F	p	F	p	F	p
VABS
Communication	18	69.44 (15.02)	14	77.21 (13.25)	12	75.75 (12.74)	19	64.32 (17.97)	18	73.11 (18.06)	17	72.06 (16.60)	0.87	0.36	29.86	0.00	1.30	0.26
Daily Living	18	77.89 (9.68)	14	78.29 (8.06)	12	78.08 (9.46)	19	81.37 (14.00)	18	80.72 (12.75)	17	78.29 (11.36)	0.93	0.34	0.89	0.35	0.42	0.52
Socialization	18	64.78 (9.93)	14	69.79 (13.62)	12	65.33 (10.06)	19	61.95 (13.12)	17	64.24 (11.91)	17	66.71 (12.71)	0.68	0.42	11.73	0.00	1.06	0.31
Motor	18	88.00 (13.48)	14	90.79 (13.20)	12	89.25 (13.90)	19	92.47 (12.31)	17	92.82 (12.09)	17	92.47 (15.58)	0.88	0.36	0.59	0.45	0.36	0.55
Adaptive Behavior Composite	18	70.44 (7.19)	14	73.93 (7.99)	12	72.25 (6.97)	19	69.53 (9.69)	17	72.18 (10.61)	17	71.76 (9.74)	0.09	0.76	12.91	0.00	0.44	0.51
CBCL
Internalizing	18	10.83 (3.79)	14	11.79 (4.32)	12	10.17 (3.93)	19	9.42 (3.20)	18	9.44 (4.12)	16	8.75 (3.44)	1.41	0.24	0.55	0.46	0.51	0.48
Externalizing	18	9.83 (3.62)	14	10.21 (4.00)	12	10.75 (3.41)	19	7.74 (2.77)	18	8.17 (3.17)	16	8.19 (3.10)	3.47	0.07	0.61	0.44	0.03	0.86
Total problems	18	34.83 (8.46)	14	35.43 (12.96)	12	35.67 (9.78)	19	30.26 (7.61)	18	29.83 (9.38)	16	28.31 (9.29)	1.93	0.17	0.43	0.52	0.92	0.34
Peds-QoL
Total Score	18	61.04 (16.13)	14	64.27 (14.51)	12	64.43 (15.73)	19	59.38 (15.70)	18	67.03 (14.99)	16	68.24 (15.78)	0.11	0.75	17.99	0.00	5.90	0.02
Physical Health Summary Score	18	69.97 (16.51)	14	74.78 (13.28)	12	75.78 (16.60)	19	69.74 (14.88)	18	76.74 (14.67)	16	76.95 (15.47)	0.00	0.97	9.77	0.00	1.18	0.28
Psychosocial Health Summary Score	18	54.11 (18.45)	14	56.40 (16.59)	12	55.88 (17.71)	19	52.01 (18.06)	18	59.99 (17.68)	16	61.97 (17.79)	0.13	0.72	14.73	0.00	6.12	0.02

Note M, mean; SD, standard deviation; VABS, Vineland Adaptive Behavior Scales (standard scores); CBCL, Child Behavior Checklist (raw scores); Peds-QoL, Pediatric Quality of Life Inventory

Table 5.

Parent outcome measures and linear mixed model results

Self-directed Web				Web + group therapy			Fixed Effect
	T1 (n = 18)	T2 (n = 13)	T3 (n = 12)	T1 (n = 19)	T2 (n = 18)	T3 (n = 16)	Group		Time		Group × Time
	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	F	p	F	p	F	p
GAD-7	6.17 (5.46)	4.69 (4.64)	4.67 (2.74)	6.16 (5.38)	3.50 (4.19)	3.81 (3.92)	0.00	0.96	11.36	0.00	4.84	0.03
PHQ-9	7.39 (6.50)	6.46 (6.21)	6.25 (4.75)	6.58 (5.34)	4.50 (5.44)	5.63 (4.81)	0.24	0.62	1.12	0.30	1.81	0.18
QoL
Physical health	0.59 (0.14)	0.68 (0.13)	0.69 (0.17)	0.66 (0.13)	0.67 (0.14)	0.64 (0.18)	2.16	0.15	5.45	0.02	3.84	0.06
Psychological	0.51 (0.19)	0.55 (0.16)	0.58 (0.20)	0.50 (0.17)	0.59 (0.22)	0.51 (0.19)	0.00	1.00	2.94	0.09	0.20	0.66
Social relationships	0.51 (0.13)	0.58 (0.16)	0.59 (0.18)	0.57 (0.15)	0.64 (0.18)	0.58 (0.17)	1.62	0.21	4.64	0.04	0.24	0.63
Environment	0.54 (0.18)	0.57 (0.17)	0.59 (0.18)	0.53 (0.15)	0.59 (0.23)	0.57 (0.16)	0.01	0.95	3.16	0.08	1.60	0.21
Total	81.83(15.18)	87.92(15.01)	89.83(16.99)	84.53(13.41)	90.11(17.87)	85.75(15.75)	0.38	0.54	6.12	0.02	0.00	0.95
PSI
PD	36.89 (7.95)	35.85 (9.49)	33.17 (9.30)	37.42(11.57)	33.17(10.87)	35.50 (9.35)	0.00	0.96	6.04	0.02	0.69	0.41
PCDI	32.44 (8.99)	29.31 (7.09)	29.25 (8.09)	30.00 (7.89)	27.56 (7.49)	31.13 (7.66)	1.05	0.31	5.20	0.03	0.72	0.40
DC	35.50 (8.10)	30.62 (8.46)	32.42 (9.29)	30.95 (7.67)	29.50 (5.55)	31.81 (6.01)	3.15	0.08	5.27	0.03	1.25	0.27
TSS	104.83(22.13)	95.77(23.71)	94.83(24.79)	98.37(22.67)	90.22(19.15)	98.44(19.04)	0.91	0.35	8.76	0.00	0.10	0.75
PSDQ
Authoritative	3.62 (0.40)	3.74 (0.57)	3.80 (0.46)	3.44 (0.50)	3.78 (0.33)	3.66 (0.33)	1.32	0.26	12.90	0.00	1.25	0.27
Authoritarian	2.27 (0.43)	2.14 (0.39)	2.10 (0.33)	2.29 (0.49)	2.27 (0.60)	2.38 (0.55)	0.01	0.94	0.57	0.45	1.07	0.31
Permissive	2.56 (0.43)	2.22 (0.73)	2.40 (0.56)	2.60 (0.68)	2.51 (0.75)	2.68 (0.57)	0.07	0.79	1.55	0.22	1.04	0.31

Note M, mean; SD, standard deviation; GAD-7, Generalized Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire-9; QoL, WHO Quality of Life Scale (percentiles); PSI, Parenting Stress Index-4 (percentiles), Short Form; PSDQ, PD, Parental Distress; PCDI, Parent-Child Dysfunctional Interaction; DC, Difficult Child; TSS, Total Stress Score; Parenting Styles and Dimensions Questionnaire

Difference Between Groups

Although both self-directed and web + group therapy conditions showed improvements or decreases in children’s and parents’ outcomes during the intervention period, there was no significant difference between the two groups in the outcome variables. Nonetheless, marginally significant group differences were observed in children’s Externalizing Behaviors, F (1, 61.74) = 3.47, p = 0.067 (Table 4), and in parents’ Dificult Child in Parenting Stress Index, F (1, 45.40) = 3.15, p = 0.083 (Table 5).

Change Across Time

Significant fixed effects of time were detected on children’s outcomes (Table 4), indicating the significant improvement across time: (1) VABS Communication, F (1, 60.27) = 29.86, p < 0.001, (2) VABS Socialization, F (1, 60.07) = 11.73, p = 0.001, (3) VABS Adaptive Behavior Composite, F (1, 59.54) = 12.91, p = 0.001, (4) Peds-QoL Total scores, F (1, 59.95) = 17.99, p < 0.001, (5) Peds-QoL Physical Health Summary Score, F (1, 60.93) = 9.77, p = 0.003, (6) Peds-QoL Psychosocial Health Summary Score, F (1, 60.24) = 14.73, p < 0.001.

There were significant fixed effects of time on the parents’ outcomes (Table 5), indicating the significant changes across time: (1) GAD-7, F (1, 57.62) = 11.36, p = 0.001, (2) QoL Total scores, F (1, 59.46) = 6.12, p = 0.016, (3) QoL Physical health, F (1, 59.65) = 5.448, p = 0.023, (4) QoL Social relationships, F (1, 63.43) = 4.64, p = 0.035, (5) PSI Total Stress score, F (1, 59.07) = 8.76, p = 0.004, (6) PSI Parental Distress, F (1, 59.96) = 6.04, p = 0.017, (7) PSI Parent-Child Dysfunctional Interaction, F (1, 59.64) = 5.2, p = 0.026, (8) PSI Dificult Child, F (1, 59.19) = 5.27, p = 0.025, (9) PSDQ Authoritative style, F (1, 63.20) = 12.9, p = 0.001.

Interaction Effect

There were significant group by time interaction effect on children’s outcomes in terms of Peds QoL Total Scores, F (1, 59.95) = 5.90, p = 0.018, and Peds QoL Psychosocial Health Summary Score, F (1, 60.24) = 6.12, p = 0.016, and on parents’ mental health outcome in GAD-7, F (1, 57.62) = 4.84, p = 0.032 (Tables 4 and 5).

Outcome Estimate Statistics

Adaptive Functioning: VABS

Significant fixed time effects were found on Communication, Socialization, and Adaptive Behavior Composite without significant interaction effects. Supplement Table 3 showed estimate statistics for Communication, Socialization, and Adaptive Behavior Composite. Specifically, Communication standard scores showed significantly positive changes across time in self-directed, b = 1.40, t (60.83) = 2.96, p = 0.006, and the web + group therapy condition, b = 2.14, t (59.48) = 5.04, p < 0.001, which indicated a child spent 1 month in the telehealth intervention (time), the communication increased by 1.4 of a unit in the self-directed condition and by 2.14 of a unit in the web + group therapy condition, suggesting the time spent in the intervention positively predicts communication. Similarly, Socialization standard scores only demonstrated a significant positive change in treatment group across time, b = 1.34, t (58.93) = 3.39, p = 0.001, indicating that one unit change in Time (1 month), the socialization increased by 1.34 of a unit and suggesting the time spent in the treatment condition positively predicts socialization. Moreover, the web + group therapy condition had a significant positive impact on the adaptive behavior composite over time, b = 0.75, t (58.80) = 3.24, p = 0.002. Participating in the web + group therapy for one month had a 0.75 unit increase in the adaptive behavior composite score, suggesting time spent in the group therapy condition positively predicts overall adaptive skills.

Child Quality of Life: Peds-QoL

Parents’ ratings of their children’s QoL did not differ significantly between the two groups. However, there were significant changes in all three Peds-QoL summary scores over time. A significant interaction effect was found in parents’ report of their children’s QoL in the Total Score and Psychosocial Health Summary Score (Supplement Table 4). Parents in the treatment group reported a significantly positive change in their ratings of children’s QoL over the intervention period, including Total Scores, b = 2.87, t (58.66) = 5.04, p < 0.001, Physical Health Summary Score, b = 2.36, t (58.79) = 3.16, p = 0.002, and Psychosocial Health Summary Score, b = 3.14, t (58.92) = 4.76, p < 0.001. The increase in time positively predicted parents’ report of their children’s QoL in the treatment group. Conversely, no significant increase was reported from parents in the comparison condition over time. The group × time effect was significant in Total Score, b = −2.09, t (59.95) = − 2.43, p = 0.018, and Psychosocial Health Summary Score, b = −2.46, t (60.24) = − 2.47, p = 0.016. The negative coefficients indicate that the change in comparison group was significantly lower than the treatment group by 2.09 of a unit in Total Score, and by 2.46 of a unit in Psychosocial Health Summary Score. Figure 2 displays the mean scores of Peds QoL Total Score and Psychosocial Health Summary Score for the self-directed and web + group therapy, indicating that the web + group therapy condition had a greater time effect than the self-directed condition.

Fig. 2.

Mean of peds-QoL by groups over the intervention period

Parents’ Mental Health: GAD-7

A significant time effect and a significant group by time interaction effect were found on GAD-7 (Supplement Table 5). Parents in the treatment group reported a significant decrease in anxiety over time, b = − 0.79, t (55.73) = − 4.24, p < 0.001, while the comparison group did not, b = −0.17, t (59.07) = − 0.78, p = 0.44. The coefficients in the treatment group suggested that 1 unit change in Time was associated with 0.79 unit decrease in anxiety, suggesting that more time spent in treatment predicts a decrease in anxiety outcome in parents. The group × time effect was significant, b = 0.63, t (57.62) = −2.43, p = 0.032, with the treatment group showing a larger reduction in GAD-7 over time than the comparison group, as shown in Fig. 3.

Fig. 3.

Mean of GAD-7 by groups over the intervention period

Parents’ Quality of Life: WHOQOL-BREF

Analyses revealed significant fixed effect of time without significant interaction effects for Physical Health, Social Relationships, and Total Score in WHOQOL-BREF. Parents in both conditions experienced improvement in their overall quality of life (Total QoL score), with gains noted in the Physical Health and Social Relationships domains (Supplement Table 6).

Parenting Stress: Parenting Stress Index - Short Form

Parent stress level was measured by the PSI-SF. A significant fixed effect of time without a significant interaction effect was found on all subscales in PSI, including Parental Distress, Difficult Child, Parent-Child Dysfunctional Interaction, and Total Stress Score. Supplement Table 7 displayed the estimate statistics for the four subscales. Parents in both conditions demonstrated a decrease in all parenting stress variables over time. Specifically, total stress score demonstrated significantly decrease across time in comparison group, b = −2.13, t (60.45) = − 2.17, p = 0.034, and in treatment group, b = −1.72, t (57.28) = −2.01, p = 0.049. 1 unit change in Time was associated with 2.13 unit decrease in parenting stress for parents in the self-directed condition, and with 2.01 unit decrease for parents in the web + group therapy condition. The results suggested the Time negatively predicts parenting stress total score.

Parenting Style: Parenting Styles and Dimensions Questionnaire

Parenting style was measured by PSDQ, and the z-score of three parenting styles were yielded. A significant time effect was found on the authoritative parenting style, as parents in both conditions demonstrated increased scores across the intervention period. In the treatment group, there was a significate positive relationship between the authoritative parenting style and time, b = 0.08, t (60.42) = 3.56, p = 0.001, with one-unit change in Time corresponded to a 0.08-unit increase in the use of authoritative parenting behaviors (Supplement Table 8).

Discussion

The study implemented a culturally-adapted, group-based parent coaching intervention delivered via telehealth and tailored to the Chinese context. The primary purpose of this study was to explore the efficacy of two formats of telehealth service on both children’s and parents’ outcomes. The results showed that both conditions effectively improved children’s communication skills and social engagement, as well as alleviated parenting stress and anxiety in parents of autistic children. Additionally, this telehealth intervention improved the quality of life for both children and parents. The group-based parent coaching via videoconferencing was found to be more effective in improving children’s quality of life and reducing parent’s anxiety.

The findings of the current study showed similar effects on autistic children and their parents compared with two studies in China (Xu et al., 2017); Zhou et al., 2018). However, both studies were conducted in a clinical in-person setting. Our findings added to the literature that a culturally-adapted, group-based telehealth intervention could bring about change in core developmental domains for autistic children, alleviate parents’ anxiety level, and improve the quality of life for both autistic children and parents. The findings also add to the literature on group-based parent training for young autistic children, which has reported mixed or minimal results on children’s outcomes (O’Donovan et al., 2019). Most child outcomes did not show significant group differences in our study. The possible reason could be that parents in two conditions received the identical modules, and the efficacy of the two telehealth service formats could not be differentiated given the dosage of the group therapy was only 1.5 h per week. Furthermore, over 50% of parents in the self-directed condition submitted homework and received tailored feedback (Qu, in review). This factor may have contributed to the lack of differences observed between two telehealth services. Future study should follow up with different levels of assistance in telehealth studies to differentiate the efficacy of various telehealth service formats, such as self-directed only verse self-directed with feedback.

Parents in both groups showed significant improvements over time, including reduced parenting stress and anxiety levels, and increased quality of life. The GAD-7 scale was used to measure anxiety, with scores above 5 indicating mild anxiety. The mean scores of GAD-7 in both groups exceeded the mild threshold at baseline and was decreased significantly below the cutoff after the intervention. The informal social support from other families during the group therapy sessions may have contributed to greater reduction in parental anxiety symptoms observed in the web + group therapy condition. However, the treatment group experienced a slight increase in anxiety levels at follow-up, which may result from the end of group therapy. A similar pattern was observed with regard to the Total Stress Percentile Score on the PSI-SF scale, which defined clinically significant parenting stress as a Total Stress Score of 90 or higher (Abidin, 2012; Abidin & Abidin, 1990). Both groups had high levels of parenting stress at baseline, but the mean Total Stress Score in treatment group dropped to around 90 after the intervention. Similarly, there followed a climb in the treatment group during the follow-up phase, while the self-directed condition kept dropping. This pattern may likely be due to the end of the group-based coaching session, when parents in the treatment group may not know how to proceed once the assistance of the therapist stopped. The web + group therapy condition may be more appropriate than the self-directed condition for parents who were experiencing severe depression or anxiety. Future studies should explore strategies for maintaining family engagement post-intervention. Additionally, further investigation is needed on better understanding the relationships between the parenting stress, parental anxiety, and children’s behavior problems, and its impact on the quality of life in both parents and their autistic children.

Group-based parent training has received less attention due to a lack of theories to guide intervention development, evaluation and implementation (Farmer & Reupert, 2013; Kowalkowski, 2012; Lodder et al., 2020; Lunsky et al., 2021; O’Donovan et al., 2019; Todd et al., 2010). Meanwhile, virtual parent training has the potential to help with service access in areas where autism services are scarce (Lindgren et al., 2016; Parsons et al., 2017). This study was among a few studies using videoconferencing to deliver a small-group parent coaching intervention and was tested in a RCT design. The intervention was culturally adapted (Qu, in review) and acceptable to Chinese parents (Qu et al., 2022). The preliminary findings of this RCT indicated that the successful adaptation and strategies used to actively engage parents in the group-based coaching. This study adds to the evidence supporting group-based parent training through videoconferencing.

Limitations

One limitation is that the retention rate at post intervention was 94.74% for the treatment group and 77.78% for self-directed group, and the self-directed condition had a lower intervention completion rate compared to the treatment group (50% vs. 90%). Reasons are unclear but suggesting the low dosage of professional supervision and assistance was critical and could be a factor that effectively promotes the parental engagement. In addition, the study relied on parent-reported measures which may introduce bias. Future telehealth study should consider using objective tools or measures to provide a more comprehensive and multiple perspectives. Another limitation is that it was difficult to separate maturation from intervention effects since both conditions received active treatment. However, when these results are comparing to studies that report longitudinal trajectories for toddlers with autism, it appears that the expected developmental trajectory is a worsening of standard scores (Landa & Garrett-Mayer, 2006; Lord et al., 2012; Ozonoff et al., 2008, 2010; Varcin & Jeste, 2017). Our findings contribute to current literature by showing the preliminary effectiveness of a 12-week culturally-adapted, group-based parent coaching intervention, which led to a positive developmental trajectory in autistic children and positive promotion to mental well-being in parents.

Implications for Clinical Practice

Several lessons that we learned from this telehealth intervention would sheds light on current clinical practice. First, video-based learning materials including lectures, demonstrations, and commentary videos could work as strategies in digital or telehealth intervention. We used two telehealth delivery formats in the current study which connected patients to knowledge or training, as well as patients to professionals or peers. In addition, different levels of assistance from the therapist were provided to these two conditions, tailored written feedback and group coaching with written feedback. Clinicians should consider providing different levels of assistance according to families’ needs and priority. Telehealth strategies that were accepted by parents of children with autism should be integrated in the telehealth practice, such as tailored feedback, peer commenting, live coaching, and guided reflection (Qu et al., 2022). Lastly, the use of videoconferencing in delivering group-based coaching sessions for parents of autistic children has emerging evidence. The structured group reflection using the family-centered capacity-building approach may lead to increased quality of life and positive outcomes for both children and parents.

Conclusion

This pilot RCT assessed the effectiveness of a culturally-adapted telehealth intervention for autistic children in China. Both intervention groups showed significant improvements in children’s adaptive functioning and quality of life, along with reductions in parenting stress and anxiety, and an increased use of authoritative parenting behaviors. Additionally, group-based coaching via videoconferencing resulted in even greater improvements in children’s quality of life and parents’ anxiety outcomes. These findings highlight the use of videoconferencing for group-based parent coaching as a promising and effective model for delivering in-home services while families wait for in-person care. Moreover, group-based parent coaching shows potential as a clinical approach to promote family functioning and quality of life, as well as to alleviate parental anxiety and parenting stress for parents of autistic children.