Young autistic children and their siblings often need support to play and engage with each other. The inclusion of older siblings in the provision of support may improve outcomes for autistic children and strengthen the sibling relationship. This study evaluated the inclusion of older siblings in the delivery of Early Start Denver Model (ESDM) support to autistic children aged between 3 and 4 years. It used a non-concurrent multiple baseline across participants design to evaluate the effects of sibling-mediated ESDM for four young autistic children and their older, non-autistic siblings. For most dyads, there were improvements in autistic child engagement and in sibling initiations during the sibling-mediated ESDM, which were generally maintained at follow-up. There was some improvement in sibling responses and minimal improvement in autistic child imitation and functional utterances. Sibling initiations and responses were generally positively correlated with autistic child engagement, functional utterances, and imitation. The parents of children in all four dyads found the sibling-mediated ESDM to be acceptable. These preliminary results suggest that sibling-mediated ESDM may be beneficial for improving the interaction between autistic children and their non-autistic siblings while benefits for teaching additional child skills might be more limited.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Autism is a form of neurodivergence which is characterized, in part, by differences in social communication and social interaction (American Psychiatric Association, 2013). This brain difference is associated with strengths and challenges that are unique to each individual (Kapp et al., 2013; Pellicano & den Houting, 2022). Many autistic people and members of the broader autism community in countries like the United Kingdom, Australia, and New Zealand have indicated a preference for identity-first language (e.g., “autistic person”) rather than person-first language (e.g., “person with autism”). This is because they view autism as an inherent part of their identity (Bury et al., 2023; Kenny et al., 2016; Monk et al., 2022). As such, identity-first language is used in this paper. Some autistic people may desire and benefit from support to develop and maintain relationships with others, including family members. In turn, their family members may also need support to better understand the autistic child and their unique way of interacting (Gillespie-Lynch et al., 2017; Pellicano & den Houting, 2022).
The provision of high-quality support (e.g., therapy, intervention, or treatment) is often associated with improvements in outcomes such as language, social communication, and play skills for young autistic children (Whitehouse et al., 2021; Sandbank et al., 2020). Support is increasingly provided within the child’s everyday environments, including home and early childhood settings, to encourage the maintenance and generalization of skills over time (Spector & Charlop, 2018; Walton & Ingersoll, 2012). Indeed, research suggests that individuals who frequently interact with the child, such as teachers and parents, are generally successful in learning to use strategies to support young autistic children (Nevill et al., 2018; Tupou et al., 2019).
Children learn many skills from playing with their peers. Peer-mediated support, that is support delivered by peers with or without the involvement of adults, often leads to improvements in social initiations and responses, communication, and imitation skills for pre-school and school-aged autistic children (Chang & Locke, 2016; Gunning et al., 2019). These peer-mediated supports generally focus on improving peer interactions and play in a range of contexts including classrooms, playgrounds, and camps. Autistic school-aged children may enjoy participating in peer-mediated support (Zanuttini & Little, 2021). Non-autistic pre-school-aged peers also generally appear able to implement the strategies and improve in their ability to interact and communicate with the autistic child (Gunning et al., 2019).
Siblings could serve as ideal play partners to autistic children as they spend a considerable amount of time together across everyday environments (Spector & Charlop, 2018). Due to this familiarity, autistic children may also be more comfortable around their siblings than other peers (Ferraioli et al., 2012). Many non-autistic siblings view their relationship, or aspects of their relationship, with their autistic sibling positively (Burnham Riosa et al., 2023; Douglas et al., 2023; Schmeer et al., 2021; Wright et al., 2024). Non-autistic siblings often report that they have a strong relationship with their autistic sibling and enjoy spending time with them. Many also embrace their autistics siblings’ nature and autism diagnosis, sharing that interactions with their sibling have helped them to improve their own empathy and understanding. However, research suggests that there may be more stress and conflict in sibling relationships when one child is autistic and that the quality of the sibling relationship may be “poorer” compared to relationships between non-autistic siblings (Douglas et al., 2023; Guidotti et al., 2021; Shivers et al., 2019). Non-autistic siblings appear to make few attempts to interact with their autistic sibling during free play, suggesting that they could benefit from some additional support to facilitate these interactions (Walton & Ingersoll, 2012).
A meta-analysis of single-case design studies suggests that sibling-mediated supports lead to improvements in play, social, functional, and academic skills for autistic children aged 0–15 years (Bene & Lapina, 2021). These studies focused on supporting sibling play and interactions at home and at school. Techniques used to teach the siblings included social stories, video modeling, and direct instruction. Autistic children have also been reported to show signs of enjoyment during sibling-mediated support (Spector & Charlop, 2018). Similarly, siblings were reported or observed to enjoy participating in this type of support and generally learned to use the techniques (Pacia et al., 2021; Shivers & Plavnick, 2015). Several studies have also found that the quality of interaction between the siblings improves during sibling-mediated support (Lu et al., 2021; Pacia et al., 2021). This suggests that sibling-mediated support may be of benefit to both the autistic child and their sibling.
The Early Start Denver Model (ESDM) is a promising approach to supporting young autistic children that has been successfully delivered by individuals within the child’s everyday environment, such as parents and teachers (Fuller et al., 2020). The ESDM is a comprehensive, naturalistic, developmental, behavioral approach that targets a wide range of skills for children under the age of 6 who are, or may be, autistic (Rogers & Dawson, 2009; Schreibman et al., 2015). This model combines developmental principles and behavior analytic techniques to teach a comprehensive range of skills through fun interactions embedded in daily routines and play. Several meta-analyses suggest that ESDM leads to improvements in cognition and language for autistic children, although this approach was found to have more limited effects on standardized measures of adaptive behavior and social communication (Fuller et al., 2020; Sinai-Gavrilov et al., 2020; Wang et al., 2020). There was also no difference in the effectiveness of the ESDM depending on whether it was implemented by parents or professionals (Wang et al., 2020).
The naturalistic and flexible nature of ESDM means that it may be particularly well-suited to implementation by similar age siblings or peers. van Noorden et al. (2020) have conducted the only study to date examining peer-mediated ESDM. This case study evaluated the effectiveness of 3 weeks of therapist-facilitated ESDM “playdates” for an autistic 5-year-old child and a non-autistic same-age peer. The peer did not receive any training, instead, the therapist used ESDM techniques to model and support the interaction between the two children. The researchers found some increase in parallel play and engagement between the autistic child and the peer, which was partially maintained 7-weeks later. However, there was no increase in social initiations for the peer or communication for the autistic child. The mothers of the autistic child and the peer both viewed the playdates positively. As this study only had one participating dyad, the design means that it was not possible to establish whether the implementation of these playdates caused any improvements seen for the autistic child or the peer.
Given the relatively promising preliminary findings of the van Noorden et al. (2020) study and the potential benefits of sibling-mediated supports more generally, this study aimed to evaluate the effects of sibling-mediated ESDM using a non-concurrent multiple baseline design. In this design, participants take part in a baseline phase in which they do not receive any support, followed by provision of the sibling-mediated ESDM, then a follow-up phase evaluating whether changes continue after sibling-mediated ESDM is completed. Such designs allow for a rich and nuanced evaluation of emerging approaches for supporting autistic children (Schreibman et al., 2015). This appears to be the first study to examine the effects of sibling-mediated ESDM. ESDM has been widely implemented by parents and teachers in naturalistic environments with relatively promising results (Fuller et al., 2020; Sinai-Gavrilov et al., 2020; Wang et al., 2020), as such implementation by siblings may be a logical next step. Specifically, the study examined the effect of an initial training followed by 2 h/week of therapist-supported sibling-mediated ESDM for 8 weeks on outcomes for four autistic children and their older, non-autistic siblings. The primary outcome was autistic child engagement with their older sibling, as the main aim was to support successful interactions between the sibling dyads. Secondary outcomes included functional utterances (i.e., spontaneous, meaningful spoken language) and imitation for the autistic child, and social initiations and responses for the older sibling. We also used correlations to determine whether there were any associations between the sibling and autistic child variables. Finally, we examined parent perceptions of the social validity of this approach. The specific research questions for the four autistic child-older sibling dyads were:
1.
Is there a replicable functional relation between implementation of sibling-mediated ESDM and autistic child engagement? (Primary). A replicable functional relation would be established if the introduction of the sibling-mediated ESDM led to improvements in autistic child engagement with their non-autistic sibling for each dyad (Ledford & Gast, 2018).
2.
Is there a replicable functional relation between implementation of sibling-mediated ESDM and autistic child functional utterances and imitation?
3.
Is there a replicable functional relation between implementation of sibling-mediated ESDM and older sibling social initiations and responses?
4.
What is the association between autistic child and sibling outcomes during sibling-mediated ESDM?
5.
Do parents find sibling-mediated ESDM to be socially valid?
Method
Participants
Participants were recruited by an invitation emailed to all families in the database of the Victoria University of Wellington Autism Clinic in New Zealand. A researcher met with interested families at their home to go through further study information and to obtain informed consent. Autistic children were eligible for inclusion if they were aged between 3 and 6 years old and had a clinical diagnosis of autism. Children with other medical, genetic, neurological, or sensory conditions (e.g., Down’s syndrome or fragile X syndrome) were excluded. Siblings of the autistic children were eligible for inclusion if they were aged between 5- and 10-year-old and had no known learning disability or psychological/neurodevelopmental diagnoses. Siblings who had scores one standard deviation or more below what was expected for their age on the communication and socialization subscales of Vineland Adaptive Behavior Scale-Third Edition (VABS-3; Sparrow et al., 2016) were excluded. Parents and children were required to live in the same home.
Table 1 describes the demographic characteristics of the four participating sibling dyads. Two dyads were comprised of a female autistic child and their older, non-autistic, sister while two included a male autistic child and their older, non-autistic, brother. Three autistic children had minimal spoken language, while one child (Dyad 1) had fluent speech. The siblings had average (100) or above average standard scores on most domains of the VABS-3 (Sparrow et al., 2016), while the autistic children had below average scores in all domains. Parents of three of the four dyads were married with college-level education and at least one parent in full-time employment. The fourth dyad was supported full-time by a single father with some high school education. At least one parent in all four dyads had participated in one or more parent coaching programs. Three of the autistic children attended a multidisciplinary support service once per week, while the remaining child (Dyad 4) received intermittent support from a speech language therapist. Three of the four autistic children had also had at least one term of therapist delivered ESDM therapy. Three of the autistic children also attended kindergarten 3 days/week, while one child (Dyad 1) attended Montessori kindergarten five mornings per week.
Table 1
Demographic characteristics and VABS-3 domain standard scores for each participating dyad
Dyad 1
Dyad 2
Dyad 3
Dyad 4
Ethnicity
NZ European
Rotuman
NZ European + Indian
Afrikaans + NZ European
Sibling
Age
6 years
7 years
7 years
5 years
Gender
Female
Male
Female
Female
VABS-3 domain standard scores
Communication
100
106
110
102
Daily living skills
110
104
114
98
Socialization
110
114
120
120
Motor skills
105
87
121
93
Autistic child
Age
4 years
3 years
4 years
4 years
Gender
Female
Male
Male
Female
Age of diagnosis (years)
3
3
2
1
Additional diagnoses or medical conditions
None
None
GDD; verbal dyspraxia
Cleft palette
Spoken language level
Fluent
Minimal
Minimal
Minimal
Language(s) spoken at home
English
English
English
English
VABS-3 domain standard scores
Communication
62
42
34
41
Daily living skills
72
67
65
61
Socialization
61
60
48
50
Motor skills
77
79
71
68
Parent
Relationship status
Married
Married
Married
Single father
Full-time employment (≥1 parent)
Yes
Yes
Yes
No
Education
College
College
College
Some high school
Previous parent coaching (≥1 parent)
Yes
Yes
Yes
Yes
NZ New Zealand, GDD global developmental delay
Settings and Personnel
Sessions were conducted in each family’s home. The location within the home varied based on the activity that the children chose. Generally, sessions took place in the living room, children’s bedrooms, or on play equipment (e.g., trampolines) in their garden. All baseline, therapist support, and follow-up sessions were conducted by researchers (EW, LvN, and HW) who were also certified ESDM therapists.
Materials
An Apple iPhone® was used to videotape sessions. Immediately after each session, videos were transferred to a password-protected OneDrive folder and deleted from the iPhone®. A paper handout of relevant ESDM techniques was used in the sibling coaching. A laptop was used to show example videos of a consenting family with similar aged children who did not participate in the current research. Children used their own play materials during sessions. Researchers provided extra toys such as board games, bubbles, and play-doh as a procedural modification for Dyad 1. The activity or activities chosen in each session varied. Across phases of the study, siblings were free to choose their own activities. If they did not know what they wanted to play during baseline or follow-up, their parent would often suggest some possibilities, and data were only collected when the children had chosen joint or individual activities. The choice of shared activities is a key component of the sibling-mediated support (see Supplementary Table 1), as such, the therapist supported the sibling to choose activities when necessary, during the sibling-mediated ESDM phase.
Design
An adapted non-concurrent multiple baseline design was used. In this design each participant serves as their own baseline. Data were collected at multiple timepoints prior to and following the implementation of the sibling-mediated ESDM. This design is used to determine whether any improvements seen upon the implementation of the sibling-mediated intervention are replicated across dyads and outcomes. The study included the following sequential phases: (a) pre-assessment, (b) baseline, (c) sibling coaching, (d) therapist support, and (e) follow-up. The minimum length of baseline phase for each participant was assigned before baseline data collection began. The first dyad was assigned a five-session baseline, and then baseline lengths were assigned in multiples of three for successive participants. Dyad 2 was assigned 6 baseline sessions, Dyad 3 was assigned 9 baseline sessions, and Dyad 4 was assigned 12 baseline sessions. After the initial five assigned baseline sessions for Dyad 1, several variables had an increasing trend, so an extra two baseline sessions were added, resulting in seven baseline sessions in total. Issues such as prolonged sickness and conflicting appointments meant that the first nine baseline sessions for dyad four took more than twice as long as expected to collect (over 3 months instead of 5 weeks). As such, the researchers decided, for ethical and logistical reasons, to commence sibling coaching after these 9 sessions rather than the initially assigned 12 sessions.
Procedures
Throughout the study, sessions were scheduled twice per week for each family. If a session was missed due to sickness or conflicting appointments, every effort was made to reschedule this session within the next week.
Pre-assessment
Prior to baseline, parents completed a demographic questionnaire and participated in a VABS-3 (Sparrow et al., 2016) interview with a researcher regarding their autistic child and older sibling. A researcher also used a play-based ESDM curriculum assessment to determine a set of individualized ESDM teaching goals for each autistic child in the therapist support phase (Rogers & Dawson, 2009). These goals were guided by the skills listed in the first five domains of the ESDM curriculum checklist: (a) receptive communication, (b) expressive communication, (c) social skills, (d) imitation, and (e) joint attention. Up to 10 goals, just beyond the child’s current developmental level, were set for each autistic child in collaboration with their parents. These goals were more specific than the dependent variables used in this study and were tailored to each participant. The overarching dependent variables were designed to compare broader changes across participants during the sibling-mediated ESDM.
Baseline
Baseline sessions involved a 10-min video of the children playing together. During these sessions, the researcher asked the children to play as they normally would, and recording began when the children were settled into play (either together or in parallel). The researcher did not interact with the children, and no feedback or coaching was provided.
Sibling Coaching
This phase, implemented by EW, involved teaching the sibling strategies for interacting with the autistic child and comprised of two 30-min sessions. The sibling was present for the duration of the training while the autistic child only participated during the practice at the end. The procedures used to teach the sibling strategies included: (a) discussion of a short, child-friendly booklet of techniques with pictures and diagrams, (b) video examples of the techniques being implemented, (c) role play of the techniques with the therapist, and (d) practice of the techniques with the autistic child with live coaching from the therapist. The siblings were taught techniques for capturing attention and increasing motivation including sitting face-to-face, putting away toys that were not being used, and holding interesting objects near the face. Siblings also learned how to join the autistic child’s play, how to be a helpful play partner, and how to elaborate on the play and take the lead. Finally, siblings were taught to provide opportunities for the autistic child to learn new skills through a simplified explanation of an antecedent-behavior-consequence teaching structure. This explanation included examples of clear cues, waiting for the autistic child to perform a behavior (with prompting from the therapist if needed), and providing an immediate naturalistic positive consequence. For instance, the sibling was told that they could offer the autistic child two toys that they might like and then wait (antecedent). If the autistic child showed any sign of interest in one of the toys (behavior), the sibling should hand it over straight away (consequence). They were also told that if they were finding this difficult, the therapist would be able to help them. No data were collected in this phase.
Therapist Support
This phase involved 15, 45–60-min-long support sessions. These support sessions were conducted by LvN for Dyad 1 and EW for the remaining dyads. These therapists were Masters- (EW) and PhD-level (LvN) educational psychology students. Both were trained and certified in ESDM direct therapy and had been delivering this approach with autistic and potentially autistic children for over a year. Neither had previously worked directly with any of the dyads to which they were assigned. The first 10-min of each session was video recorded for data collection. During these sessions, the therapist prompted and supported the sibling to use the techniques covered in their training. An example of a prompt could include the therapist verbally pointing out to the sibling that the autistic child appeared interested in a specific toy or action. The therapist also supported the autistic child to respond to the sibling’s cues, and to be successful in performing new behaviors or approximations of new skills. For example, if the sibling was offering the autistic child a toy, and the autistic child did not appear to notice, then the therapist would also point at the toy and label it. Detailed procedures used by the therapist to support both the autistic child and the older sibling are included in Supplementary Table 1. Due to on-going sickness and scheduling issues, Dyad 4 did not complete this phase or participate in the follow-up phase.
Procedural Modification Dyad 1
This modification occurred after the 8th therapist support session for Dyad 1 and involved the therapist bringing a small number of extra, novel toys. This was because both children reported being uninterested in the toys available at home. This modification only applied to Dyad 1.
Follow-up
Follow-up was identical to baseline and commenced 1 week after the final therapist support session for each Dyad. These sessions were conducted by the same therapist who implemented the therapist support sessions in all instances, except that HW rather than LvN conducted follow-up for Dyad 1.
Dependent Variables
Definitions of the primary and secondary dependent variables (DV) are presented in Table 2. Data were coded from 10-min. videos collected in the baseline, therapist support and follow-up phases. Each video was divided into 60 × 10-s intervals, and interval recording was used to measure all DVs. Intervals where either child was out of frame for longer than 5 s were excluded from coding and analysis. A baseline session for Dyad 2 and a therapist support session for Dyad 1 and Dyad 4 were only 8 min because of technical issues or a perceived lack of autistic child assent to continue participating in the activity. This lack of assent involved the child leaving the area in which the activity was taking place and appearing unwilling to return. One therapist support session video for Dyad 4 could not be used because of technical difficulties. For each DV, an overall percentage of intervals with the DV present was calculated using the following formula: (number of intervals with DV present)/(total number of included intervals) × 100.
Table 2
Definitions and measurement type for each dependent variable
Target child
Dependent variable
Outcome level
Definition
Interval recording
Autistic child
Coordinated engagement
Primary
The autistic child is engaged and actively involved with the same object, or with just their sibling in face-to-face or person play, without a direct prompt from the adult
Whole
Autistic child
Supported engagement
Primary
As for “coordinated engagement”, but following a direct prompt by the therapist, or with active therapist participation
Whole
Autistic child
Imitation
Secondary
Any instance of the autistic child performing an action with or without an object, or producing a vocalization, within 10 s of the sibling model, and without any prompting from the therapist or sibling
Partial
Autistic child
Functional utterances
Secondary
A spontaneous utterance from the autistic child that is clearly directed toward the sibling: (a) with no prompting or modeling of the utterance within 10 s of its occurrence, (b) contextually related to the interaction or task, and (c) containing a phonetically correct approximation of the correct word or word combination
Partial
Older sibling
Initiations
Secondary
Any clear instance of the sibling spontaneously initiating a social interaction with the autistic child. Must be directed towards the autistic child, and not an adult. Can be an utterance or vocalization (initiating turn, comment, or inquisitive sound, but not laughing/yelling, etc.), or gestures (such as offering an object, elaborating on current play, touching, pointing or showing)
Partial
Older sibling
Sibling responses
Secondary
Any instance of the sibling independently responding appropriately to the verbal and non-verbal cues of the autistic child. This could include commenting, restatement, affirmation, help with a collaborative task, responding to gestures, copying of actions, and packing away
Partial
Social Validity
After the follow-up phase, parents completed an adapted version of the Treatment Acceptability Rating Form-Revised (TARF-R) to assess their perceptions of the acceptability of the sibling-mediated support for the autistic child (Reimers et al., 1992). This measure had 19 questions assessing (a) reasonableness, (b) willingness to participate, (c) side effects, (d) effectiveness, and (e) disruptiveness, each scored on a 7-point Likert-type scale ranging from strongly disagree to strongly agree. Side effects and disruptiveness were reverse coded, meaning that higher scores indicated a more favorable rating. Parents and non-autistic siblings were invited to participate in an interview to further share their perceptions of the sibling-mediated support, but this is not reported in the current study.
Interobserver Agreement
Videos were randomly assigned to two research assistants who were blind to the study phase and had not met the families. These research assistants were primarily trained using existing videos of a sibling dyad that did not participate in the study, as well as videos of the participating dyads that were not included in the interobserver agreement calculations. They each needed to reach 80% agreement for independent coding of several such practice videos before beginning to code the videos that were included in the interobserver agreement calculations. Interobserver agreement (IOA) was assessed for at least 20% of the videos for each dyad, across phases. The intraclass correlations indicate high interobserver agreement and were as follows: imitation 0.954, functional utterances 0.934, sibling initiations 0.954, sibling responses 0.909. For engagement, coder drift was identified for Dyad 2 during the therapist support phase, with one coder rating more interactions as engagement than met the operational definition. Coding was halted until further training was completed. After training, two further videos of Dyad 2 were coded for IOA for engagement only. The intraclass correlations for engagement were as follows: total engagement (including drift) 0.829, total engagement (excluding drift) 0.881, supported engagement (including drift) 0.824, supported engagement (excluding drift) 0.868.
Implementation Fidelity
Implementation fidelity was coded on at least 20% of sessions across phases and dyads.
Baseline and Follow-Up
Implementation fidelity during baseline and follow-up was examined by the same two research assistants who coded IOA using a specifically developed checklist for the same videos selected for IOA (see Supplementary Table 2). This was coded on the entire duration of each randomly selected 10-min video-taped play sample. EW explained this checklist to the research assistants and they learned to use it on a subset of the practice IOA videos. During these phases procedural integrity was 100%.
Sibling Coaching
Implementation fidelity during the sibling coaching phase was assessed using a specifically developed checklist that described each essential element of the coaching procedures across the 30-min session. A second researcher (HW) attended at least one sibling coaching session for each dyad and administered the procedural integrity checklist (see Supplementary Table 3). HW helped to develop this checklist so knew how to administer it. The average procedural integrity for coaching sessions was 95%. While the therapists were encouraged to reflect on their own fidelity after each session, this was only formally recorded by the second researcher on the session(s) they attended.
Therapist Support
Implementation fidelity for the therapist support sessions was assessed using an adapted version of the G-ESDM Group Fidelity Rating Scale (Vivanti et al., 2017; see Supplementary Table 1). The scale was modified to make it more appropriate and operational for implementation with sibling dyads. The modified scale included adapted versions of the 13 categories of therapist behavior used in the original scale: (a) management of children’s attention, (b) the use of the A-B-C teaching format, (c) use of behavioral teaching techniques, (d) management of the children’s arousal levels, (e) management of unwanted behaviors, (f) promoting sibling interactions, (g) optimizing child motivation, (h) use of positive affect, (i) sensitivity and responsivity, (j) use of multiple and varied communication, (k) use of appropriate language, (l) use of joint activity structure with elaboration, and (m) managing transitions between activities. Each item was rated on a scale from 1 to 3, where 1 indicated no use of the technique, 2 indicated partial use of the technique, and 3 indicated correct use of the technique. The same two research assistants coded fidelity from randomly selected 10-min play samples collected during the therapist support phase. EW explained this checklist to the research assistants and they learned to use it on a subset of the practice IOA videos. The percentage of implementation fidelity for each session was calculated using the following formula: (raw score)/(total number of possible points) × 100. The average implementation fidelity was 93%.
Data Analysis
The baseline, therapist support, and follow-up data were graphed and visually analyzed. The mean percentage of intervals containing each dependent variable was also calculated for each dyad and phase. The Tau-U method was used to calculate the size of changes from baseline to therapist support and baseline to follow-up. This is a common effect size in single-case research. For Dyad 1, Tau-U was calculated separately for those sessions with and without extra toys. Tau-U values represent the ratio of nonoverlapping pairs across phases and indicate the size and the direction of change across phases (Parker et al., 2011). To demonstrate that the sibling-mediated ESDM led to improvements in the target dependent variables, the values for target variables in the therapist support and follow-up phases should be higher than those in the baseline phase. If many values in baseline overlap with values in the therapist support and follow-up phases, this indicates that the implementation of the sibling-mediated support has not resulted in meaningful change in these target variables. Vannest and Ninci’s (2015) guidelines were used to interpret Tau-U values as small, that is many overlapping points (≤0.20); moderate, that is some overlapping points (0.21–0.60); and large (0.61–0.80) or very large (≥0.81), that is few overlapping points. A correlation analysis was performed to examine the relations between sibling initiations and responses, and the autistic child variables (engagement, imitation, and functional utterances). Spearman’s correlations were used as many variables were not normally distributed. Regardless of the direction, values between 0.1 and 0.3 indicated a weak relationship, values between 0.3 and 0.5 indicated a moderate relationship, and values between 0.5 and 1 indicate a strong relationship (Xiao et al., 2016).
Results
Variables for the Autistic Child
Table 3 shows the mean percentage, range, and Tau-U values for the autistic child dependent variables across all study phases and dyads.
Table 3
Mean percentage and TAU-U values for coordinated engagement, imitation, and functional utterances by the autistic child across all dyads and phases
Dyad
Coordinated engagement mean %, (range), TAU-U
Imitation mean %, (range), TAU-U
Functional Utterances mean %, (range), TAU-U
BL
Support Home toys
Support Extra toysa
F.U.
BL
Support Home toys
Support Extra toys*
F.U.
BL
Support Home toys
Support Extra toysa
F.U.
Dyad 1
28.5% (0–60.0%)
60.1% (36.2–78.9%) 0.70
78.8% (66.7–90.0%) 1.0
56.7% (1.7–86.7%) 0.43
2.9% (0–8.3%)
10.7% (0–23.3%) 0.59
5.3% (0–13.3%) 0.25
2.2% (0–5.0%) −0.10
6.7% (0–13.3%)
11.1% (3.3–20.0%) 0.50
29.1% (16.7–45.0%) 1.0
18.9% (8.3–28.3%) 0.81
Dyad 2
2.4% (0–14.6%)
34.9% (11.3–66.6%) 0.98
N/A
64.5% (53.4–76.7%) 1.0
0%
0.5% (0–3.8%) 0.20
N/A
11.4% (6.7–18.3%) 1.0
1.0% (0–2.2%)
7.8% (1.7–13.2%) 0.9
N/A
30.5% (23.3–35.0%) 1.0
Dyad 3
5.6% (0–38.9%)
43.1% (7.7–72.0%) 0.93
N/A
46.0% (15.8–71.2%) 0.93
0%
5.7% (0–35.2%) 0.67
N/A
0.6% (0–1.8%) 0.33
0.6% (0–3.6%)
4.2% (0–18.2) 0.46
N/A
0% –0.22
Dyad 4
3.3% (0–18.3%)
61.1% (9.6–92.7%) 0.98
N/A
N/A
0.4% (0–1.7%)
2.2% (0–5.5%) 0.65
N/A
N/A
0.2% (0–1.8%)
1.0% (0–5.5%) 0.23
N/A
N/A
Tau-U values are underlined and bold font indicates large or very large effects. Dyad 4 did not participate in follow-up
BL baseline, F.U. follow-up
aThis procedural modification applies to Dyad 1 only
Engagement—Autistic Child
Engagement was the primary outcome. Figure 1 shows the percentage of whole intervals in which the autistic child was engaged with their sibling across dyads for all applicable phases. Engagement for all dyads increased from baseline to the therapist support and follow-up phases, with high Tau-U values indicating large effect sizes in all instances except baseline to follow-up for Dyad 1. The level of supported engagement was also relatively high for all dyads during the therapist support phase and followed similar trends to total engagement.
Fig. 1
Percentage of intervals containing full engagement for each dyad across phases. Gray, dashed lines indicate that there was more than a week between sessions and that the session(s) could not be made-up in the following week
×
For Dyad 1, during the baseline phase, a variable descending trend was observed for the child’s total engagement across the data path. For the therapist support phase, an initial increase in both supported and total engagement was observed followed by a variable descending trend for the use of toys from home; with the addition of extra toys, an immediate increase in engagement was observed with a variable but flat trend. These levels remained stable during the first and third follow-up session, with a large decrease in the second session because the older sibling did not want to play.
For Dyad 2, during the baseline phase, a stable trend was observed across the data path, with minimal levels of engagement. For the therapist support phase, a highly variable but flat trend was observed for both supported and total engagement with improvement compared to baseline. During follow-up, increased engagement was observed with a slight decreasing trend from the first to third session.
For Dyad 3, during the baseline phase, the first session contained comparatively high engagement, followed by low engagement with a relatively stable trend across the data path. For the therapist support phase, a highly variable trend was observed for both supported and total engagement with improvement compared to baseline and a slight increasing trend. During follow-up, an initial high but decreasing trend was observed.
For Dyad 4, during the baseline phase, a relatively stable trend with minimal engagement was observed across the data path except for the fourth session for which engagement was higher. For the therapist support phase, an initial increase was observed, followed by a highly variable trend and a decrease during the last session.
Imitation—Autistic Child
Figure 2 shows the percentage of intervals in which the autistic child imitated their sibling across all dyads and applicable phases. The effects of therapist support on imitation were mixed across dyads and phases, with Tau-U values showing weak to large effect sizes.
Fig. 2
Percentage of intervals containing at least one unprompted instance of imitation for each dyad across phases. Gray, dashed lines indicate that there was more than a week between sessions and that the session(s) could not be made-up in the following week
×
For Dyad 1, during the baseline phase, a stable increasing trend was observed for the first five sessions, with a decrease to no imitation in the final two sessions. For the therapist support phase, an initial increase was observed followed by a variable descending trend through the remainer of the sessions using toys from home. A variable, flat trend was observed in the sessions with extra toys. During follow-up, there was an immediate decrease in imitation, followed by a slight increasing trend.
For Dyad 2, during the baseline phase, a stable trend was observed across the data path, with no imitation in any session. For the therapist support phase, there was a stable trend with minimal levels of imitation. During follow-up there was an immediate increase in imitation, with a relatively stable trend.
For Dyad 3, during the baseline phase, a stable trend was observed across the data path, with no imitation in any session. For the therapist support phase, there was no imitation for the first three sessions, followed by a large increase and then a decrease to a stable trend with relatively minimal levels of imitation. During follow-up, there was a stable trend with minimal levels of imitation.
For Dyad 4, relatively stable trends were observed across the data paths during the baseline and therapist support phases with minimal levels of imitation across sessions.
Functional Utterances – Autistic Child
Figure 3 shows the percentage of intervals in which the autistic child directed functional utterances towards their sibling across all dyads and phases. The effect of the coaching and support on functional utterances was mixed across dyads, with Tau-U values showing weak to very large effect sizes.
Fig. 3
Percentage of intervals containing at least one unprompted utterance for each dyad across phases. Note: Gray, dashed lines indicate that there was more than a week between sessions and that the session(s) could not be made-up in the following week
×
For Dyad 1, during the baseline phase a slightly variable and increasing trend was observed for functional utterances across the data path, with a slight decrease in the final session. During the therapist support phase, a flat, variable trend was observed when using toys from home. When using extra toys there was an immediate increase followed by a variable and relatively flat trend. In follow-up, functional utterances remained at similar levels, except in the second session, in which utterances decreased.
For Dyad 2, a relatively stable trend was observed in baseline across the data path, with minimal functional utterances. In the therapist support phase, there was an increase in functional utterances from the second session which remained relatively stable for five sessions, followed by a decrease and then a variable but relatively flat trend for the remainder of the phase. During follow-up, increased functional utterances were observed with a slight decreasing trend from the first to third session.
For Dyad 3, a relatively stable trend was observed in baseline across the data path, with minimal functional utterances. This was followed by an immediate increase in functional utterances in the first session of the therapist support phase, followed by a variable and relatively flat trend for the remainder of the phase. There were no instances of functional utterances at follow-up.
For Dyad 4, stable trends were observed during the baseline and therapist support phases with minimal levels of functional utterances across the data path, except for a slight increase in the fourth therapist support session.
Variables for the Non-Autistic Sibling
Table 4 shows the mean percentage and Tau-U values for the sibling-dependent variables across all study phases and dyads.
Table 4
Mean percentage and TAU-U values for sibling initiations and responses across all dyads and phases
Dyad
Initiations mean %, (range), TAU-U
Responses mean %, (range), TAU-U
Baseline
Support Home toys
Support Extra toysa
Follow-up
Baseline
Support Home toys
Support Extra toysa
Follow-up
Dyad 1
17.6% (9.0–41.7%)
26.9% (6.7–40.7%) 0.42
41.1% (21.7–72.9%) 0.80
33.3% (1.7–65.0%) 0.24
13.6% (5.3–30.0%)
21.3% (5.0–31.0%) 0.54
25.7% (18.3–35.0%) 0.78
23.3% (1.7–41.6%) 0.24
Dyad 2
7.5% (0–31.3%)
24.9% (7.1–45.2%) 0.73
N/A
58.5% (55.0–65.0%) 1.0
10.0% (0–28.3%)
15.2% (3.8–28.3%) 0.34
N/A
58.6% (55.8–61.7%) 1.0
Dyad 3
1.2% (0–11.1%)
20.9% (1.9–36.4%) 0.96
N/A
14.1% (12.3–17.3) 1.0
10.2% (0–40.7%)
26.2% (3.4–44.6%) 0.64
N/A
23.3% (15.8–29.1%) 0.70
Dyad 4
1.3% (0–8.3%)
14.7% (7.4–25.0%) 0.94
N/A
N/A
4.8% (0–11.7%)
6.9% (0–11.5%) 0.31
N/A
N/A
Tau-U values are underlined and bold font indicates large or very large effects. Dyad 4 did not participate in follow-up
aThis procedural modification applies to Dyad 1 only
Sibling Initiations
Figure 4 shows the percentage of intervals containing sibling initiations towards the autistic child for all dyads across applicable phases. The percentage of intervals containing initiations increased in the therapist support and follow-up phases for all dyads with Tau-U values indicating large and very large effect sizes except for Dyad 1 for whom the only large effect size was during the therapist support phase using extra toys.
Fig. 4
Percentage of intervals containing at least one instance of the older sibling initiating interaction with the autistic child for each dyad across phases. Gray, dashed lines indicate that there was more than a week between sessions and that the session(s) could not be made-up in the following week
×
For Dyad 1, a variable trend was observed for sibling initiations across the data path in baseline, with a comparative increase in the fourth baseline session. In the therapist support phase, there was a variable and slightly decreasing trend in sibling initiations for the sessions using toys from home. This was followed by an immediate increase in the first session in which extra toys were introduced, followed by a decrease to similar levels of initiations as with the toys from home, with a variable trend. In follow-up there was a variable and decreasing trend.
For Dyad 2, there were comparatively high levels of sibling initiations in the first baseline session, followed by a decrease to a stable trend with minimal levels of initiations for the remainder of the phase. In the therapist support phase, there was an immediate increase in initiations in the first and second sessions followed by a decrease in the following two sessions then a variable but relatively stable trend. In follow-up, an increased and relative stable trend for initiations was observed.
For Dyad 3, there was a stable trend with no instances of initiations across the data path, except for the final session where initiations increased. There was an immediate increase in initiations in the first therapist support session, there was then a variable trend for the remainder of the phase. Sibling initiations decreased in follow-up, with a flat trend.
For Dyad 4, there was a relatively stable, flat trend for sibling initiations across the data path in the baseline phase. In the therapist support phase, there was an immediate small increase, followed by a slightly variable but flat trend for the remainder of the phase.
Table 5 shows the correlations between sibling initiations and the autistic child dependent variables. Sibling initiations were significant and strongly positively correlated with total autistic child engagement for all dyads. Sibling initiations were significant and moderately-to-strongly positively correlated with autistic child utterances for all dyads except Dyad 4. Sibling responses were significantly and moderately-to-strongly positively correlated with autistic child imitation for all dyads except Dyad 1.
Table 5
Correlations between sibling initiations and autistic child variables by dyad
Dyad
Sibling initiations and autistic child total engagement
Sibling initiations and autistic child imitation
Sibling initiations and autistic child utterances
rs
p value
rs
p value
rs
p value
Dyad 1
0.63
<0.01
0.09
0.67
0.56
<0.01
Dyad 2
0.62
<0.01
0.66
<0.01
0.71
<0.01
Dyad 3
0.75
<0.01
0.73
<0.01
0.64
<0.01
Dyad 4
0.82
<0.01
0.47
0.05
0.21
0.40
Bold values are significant at p ≤ 0.05
Sibling Responses
Figure 5 shows the percentage of intervals containing sibling responses to the autistic child for all dyads across applicable phases. The effects of support on sibling responses were mixed, with Tau-U values showing moderate to very large effect sizes across dyads.
Fig. 5
Percentage of intervals containing at least one instance of the older sibling responding to the autistic child for each dyad across phases. Note: Gray, dashed lines indicate that there was more than a week between sessions and that the session(s) could not be made-up in the following week
×
For Dyad 1, a variable increasing trend was observed for sibling responses in the first six baseline sessions, followed by a decrease in the final baseline session. There was a flat and slightly variable trend for responses in the therapist support phase when using toys from home, followed by a decrease in the final session. When extra toys were introduced, there was an immediate increase with a relatively stable trend. In follow-up there was an increase in the first session, which was somewhat maintained in the third session, but a decrease in the second session.
For Dyad 2, sibling responses were comparatively high in the first baseline session. This was followed by a decreasing trend for the following two sessions, then a flat trend with minimal sibling responses for the remainder of the phase. A variable and slightly increasing trend was observed across the therapist support phase. In the follow-up phase there was an increase in sibling responses, with a slightly decreasing trend.
For Dyad 3, sibling responses were comparatively high in the first baseline session, followed by a variable trend for the remainer for the phase. During the therapist support phase, there was an immediate increase in the first session, followed by a variable trend for the remainder of the phase, with an increase from the seventh to the final session. There was a slight decrease in sibling responses in the first follow-up session, with a slight decreasing trend.
For Dyad 4, sibling responses were variable with a flat trend across the data paths for both the baseline and therapist support phases.
Table 6 shows the correlations between sibling responses and the autistic child-dependent variables. Sibling initiations were significantly and moderately-to-strongly positively correlated with autistic child engagement, utterances, and imitation in all cases except for autistic child engagement for Dyad 4 and sibling responses for Dyad 1 for which there was no significant correlation.
Table 6
Correlations between sibling responses and autistic child variables by dyad
Dyad
Sibling responses and autistic child total engagement
Sibling responses and autistic child utterances
Sibling responses and autistic child imitation
rs
p value
rs
p value
rs
p value
Dyad 1
0.59
<0.01
0.59
<0.01
0.06
0.782
Dyad 2
0.50
0.01
0.42
0.04
0.67
<0.01
Dyad 3
0.77
<0.01
0.57
<0.01
0.53
<0.01
Dyad 4
0.26
0.29
0.55
0.02
0.71
<0.01
Bold values are significant at p ≤ 0.05
Social Validity
Table 7 shows the TARF-R scores across participants. All total acceptability scores were well above the midpoint, indicated that the parents of all four dyads found the sibling-mediated support to be socially valid for the autistic child. The mother from Dyad 2 gave the highest total score, while parents from the remaining three dyads gave similar, and slightly lower, total scores. Parents also gave scores above the midpoint for all subscales, except the mother from Dyad 1 on the Disruption/Time subscale.
Table 7
TARF-R scores for all participants in relation to the acceptability of the sibling-mediated ESDM support for the autistic child
Scale/subscale
Dyad 1
Dyad 2
Dyad 3
Dyad 4
Midpoint
Maximum score
Mean (SD)
Total acceptability
80
101
84
88
49
98
88.3 (9.1)
Reasonableness
21
21
18
21
10.5
21
20.3 (1.5)
Willingness
15
19
15
12
10.5
21
15.3 (2.9)
Side effectsa
21
21
20
21
10.5
21
20.8 (0.5)
Effectiveness
15
21
19
18
10.5
21
18.3 (2.5)
Disruption/timea
8
19
12
16
10.5
21
13.8 (4.8)
Understandingb
7
7
7
6
3.5
7
6.8 (0.5)
aHigher scores indicate a more favorable rating
bNot included in total acceptability score
Discussion
This preliminary study evaluated the effects of therapist-supported, sibling-mediated ESDM on outcomes for four young autistic children and their older non-autistic siblings. We found some evidence of a replicable functional relation between implementation of sibling-mediated ESDM and both autistic child engagement with their non-autistic sibling (primary outcome) and sibling initiations towards the autistic child. That is, there was some evidence to suggest the introduction of sibling-mediated ESDM resulted in improvements in engagement for the autistic child and sibling initiations towards the autistic child for most dyads. Results for the remaining secondary outcomes were mixed, with some improvements in sibling responses for some dyads and minimal improvements in imitation or functional utterances for the autistic child in line with the introduction of sibling-mediated ESDM. As such, it was not possible to establish a strong, replicable functional relation between implementation of sibling-mediated support and changes in these variables. The variables for the autistic child were generally positively correlated with the variables for the older sibling, which suggests that there was an association between autistic child and sibling behaviors. Parents indicated on the TARF-R that the sibling-mediated support was acceptable for their autistic child (Reimers et al., 1992). They gave particularly high ratings on the TARF-R subscales related to the reasonableness and effectiveness of the support, as well as the absence of negative side effects.
While this appears to be the first study evaluating sibling-mediated ESDM, the findings for engagement for the autistic child and sibling initiations align with those of other sibling-mediated approaches. That is, several studies have reported improvements in engagement or joint attention for autistic children during the provision of sibling-mediated support (Ferraioli & Harris, 2011; Walton & Ingersoll, 2012; Spector & Charlop, 2018). Previous studies have also found improvements in non-autistic siblings’ abilities to interact with their autistic siblings (Oppenheim-Leaf et al., 2012; Walton & Ingersoll, 2012). The results of the current study provide further evidence that sibling-mediated support can be effective in supporting positive interactions between siblings. As the procedures used in each of these studies varied considerably, it is likely that there are numerous effective ways to support engagement between an autistic child and their non-autistic sibling.
In contrast to previous research, we did not find a functional relation between the implementation of sibling-mediated ESDM and changes in imitation or functional utterances for the autistic child. For example, Wainer and Ingersoll (2012) and Ferraioli and Harris (2011) both reported improvements in child imitation during sibling-mediated support, while Spector and Charlop (2018) found improvements in vocalizations for all three participants. The ESDM is a comprehensive support, meaning that it is designed to target a variety of child skills across domains (Rogers & Dawson, 2009), while the previous studies were all focused on one or two specific outcomes. These studies also generally involved more hours of support over a longer period. Thus, it is possible that a more focused, longer, and/or more intensive period of support could have led to greater improvements in imitation and functional utterances for the autistic child compared to the current study.
The primary aim of this sibling-mediated approach was to support the autistic children’s engagement with their non-autistic sibling. This is because research suggests that autistic children and their non-autistic siblings may need support to successfully interact and build relationships (Guidotti et al., 2021; Shivers et al., 2019; Walton & Ingersoll, 2012). While there were improvements in autistic children’s engagement with their older sibling for all dyads, there was also significant variability from session-to-session. This could be due in part to the activity that was chosen. If the activity was not motivating, the autistic child may have been less interested in engaging with their non-autistic sibling. Indeed, engagement improved significantly for the autistic child in Dyad 1 with the introduction of novel, and presumably more motivating, toys. There was also a strong correlation between sibling initiations and child engagement for all dyads. It is possible that the autistic child may have been more engaged when the sibling was actively making bids for their attention. Conversely, older siblings may also have been more likely to initiate interactions when this was rewarded by engagement from the autistic child. There also may have been times when the siblings simply preferred to play alone. For example, in the second follow-up session for Dyad 1, the older sibling repeatedly made statements about wanting to “draw by herself”.
Across dyads, siblings typically had greater improvements in initiations than responses during implementation of sibling-mediated ESDM. This could be because all siblings showed some degree of responsiveness to the autistic child during the baseline phase, whereas initiations were limited for three of the four siblings. Without training, it is possible that siblings found it easier or more natural to respond to the autistic child, rather than coming up with their own play ideas. Thus, there was more scope for improvement in initiations rather than responses. It is notable that, during the therapist support phase, the number of sibling responses remained stable or increased in line with their increasing use of initiations. This indicates that the total number of sibling interactions with the autistic child (initiations + responses) also increased during the sibling-mediated support. Sibling initiations were also generally more strongly positively correlated with autistic child outcomes than sibling responses. Initiations may be more likely to directly create an opportunity for the autistic child to imitate the sibling or use a functional utterance than a response. However, to determine whether this is the case, a more detailed examination of the nature of both sibling and autistic child initiations and responses is warranted.
Due to the age of the children, a therapist was actively involved in facilitating the sibling-mediated support. The therapist became involved when either of the siblings appeared to need help but attempted to fade this assistance wherever possible. However, during the therapist support phase, the percentage of intervals in which the therapist facilitated the interaction between the siblings (supported engagement) remained relatively high for all four dyads. This could indicate that the children were somewhat reliant on the support of the therapist and/or that the therapist was more involved than necessary. This therapist support was removed during follow-up. While autistic child engagement and sibling initiations and responses remained relatively high for the three dyads who participated in this phase, these outcomes did fall to baseline levels during at least one session for Dyads 1 and 3. It is possible that these children could have benefited from a longer period of therapist support and/or a more structured fading of this support (Cengher et al., 2018). In contrast, autistic child engagement, and sibling initiations and responses were consistently higher for Dyad 2 in follow-up compared to the therapist support phase. This suggests that on-going therapist involvement was perhaps hindering independent interactions between the siblings, and that this support could have been removed sooner.
Implications
Although preliminary, there are several potential implications of the current study. First, the results suggest that sibling-mediated ESDM may be beneficial for supporting interactions between young autistic children and their older, non-autistic siblings. The relatively low-intensity, home-based nature of this support suggests that it could be delivered relatively easily alongside existing child supports. However, while parents found the support to be acceptable overall, several reported that it was somewhat disruptive. Therefore, it is important to provide this support at a time and place that it convenient for the whole family. The therapists in this study were originally trained to provide one-on-one ESDM therapy but quickly learned to provide the sibling-mediated support. They implemented the support with high fidelity (above 90% correct implementation across therapists in both the sibling coaching and therapist support phases) which indicates that other ESDM therapists could also be trained to provide this sibling-mediated approach. This also emphasizes the importance of measuring implementation fidelity, for example, using the adapted version of the G-ESDM Group Fidelity Rating Scale created for this research (Vivanti et al., 2017; see Supplementary Table 1) and of putting supports in place to ensure that fidelity remains high. The lack of improvement in autistic child imitation and utterances suggests that it may be unreasonable to expect siblings to directly teach new child skills during sibling-mediated ESDM, at least in the context of a short-duration, low-intensity support program. It is possible that this support should focus on promoting positive interactions between siblings and perhaps generalization of skills learned in other contexts, rather than on explicitly targeting improvements in other autistic child outcomes.
Limitations and Future Directions
The current research is not without its limitations. First, only four dyads participated, which only allowed for four replications of the experimental effect, that is, four opportunities to examine whether implementation of sibling-mediated ESDM resulted in improvement in the target variables. A study with a greater number of dyads would have allowed more opportunities to experimentally examine the effects of this approach. Each dyad was allocated a different number of baseline sessions (5, 6, 9 and 12 sessions, respectively) however, increasing trends meant that Dyad 1 participated in seven baseline sessions and scheduling difficulties for Dyad 4 meant that two dyads participated in nine baseline sessions (Dyads 3 and 4). Due to the non-concurrent nature of this design, each participant began the sibling coaching at a different point in time which controls for general factors that may affect all participants. However, these similarities in baseline length for two dyads may possibly reduce the internal validity of the study. Due to diminished motivation, extra toys were introduced for Dyad 1. While this did result in improved engagement, it also meant that results for this dyad were not directly comparable to the other dyads. The use of introduced toys also reduced the naturalistic nature of the support. Third, the TARF-R (Reimers et al., 1992) was also only completed for the autistic child, so it is not possible to determine parents’ perceptions of the acceptability of the support for the non-autistic sibling. Parents reported on the TARF-R that there were few negative side effects related to the support, however, the researchers did not pro-actively record adverse events and effects related to the provision of sibling-mediated ESDM. In future, researchers should ensure that they actively record the occurrence of all events that result, or could have resulted, in harm to a child, therapist, or family member, as well as the actions taken to ensure everyone’s safety. While we included a measure of supported engagement, we only examined unsupported/unprompted instances of the other autistic child and sibling variables. Finally, Covid-19 restrictions meant that sessions could not take place if either child or the therapist were sick, which resulted in numerous missed sessions. While every attempt was made to make up missed sessions in the following week, there was often a week or more between sessions, and the therapist support phase lasted for longer than the specified 8 weeks.
The current study highlights several important avenues for future research. The provision of sibling-mediated ESDM should be evaluated with a larger sample size in a randomized controlled trial. This would increase the generalizability and rigor of the findings. Researchers could also examine the effects of longer-term or higher intensity sibling-mediated ESDM to determine whether this has a greater effect on autistic child outcomes. Given the high prevalence of autism across siblings (Ozonoff et al., 2011), researchers should also evaluate the effects of sibling-mediated ESDM when both siblings are autistic. It is also essential to use age-appropriate methods to understand the perspectives of the autistic child and their older sibling regarding the acceptability of this approach. In future, researchers should examine parent-led sibling-mediated ESDM. By empowering parents to facilitate play between their children, the support could possibly be more frequent and intensive, and could continue for a longer period (Jurek et al., 2022).
Conclusion
ESDM is a widely used and relatively promising naturalistic approach for supporting young autistic children (Fuller et al., 2020; Sinai-Gavrilov et al., 2020; Wang et al., 2020). This single-case study appears to be the first to examine the provision of sibling-mediated ESDM to young autistic children. The findings provide some evidence that the implementation of sibling-mediated ESDM led to improvements in engagement for the autistic child and sibling initiations across dyads. This suggests that sibling-mediated ESDM warrants further research, including replication and extension with a larger and more diverse samples of participants.
We would like to thank the children and families for generously giving their time to participate in this research. We would also like to thank Georgia Davies and Lauren McNeil for their work as research assistants on this project.
Compliance with Ethical Standards
Conflict of Interest
E.M., H.W., and L.v.N. are certified ESDM therapists. G.V. is a certified ESDM therapist and ESDM trainer and receives royalties from the book “Implementing the Group-Based Early Start Denver Model for Young Children with Autism.”
Ethical Approval
This research received ethical clearance from the New Zealand National Health and Disability Ethics Committee (reference no. 18/NTA/35).
Informed Consent
Parents of participating children read an information sheet and gave informed consent. The non-autistic older siblings were read an age-appropriate information sheet and gave assent by coloring in a thumbs-up icon. Assent from the autistic children was inferred by their willingness to engage in the play sessions with their sibling.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Met BSL Psychologie Totaal blijf je als professional steeds op de hoogte van de nieuwste ontwikkelingen binnen jouw vak. Met het online abonnement heb je toegang tot een groot aantal boeken, protocollen, vaktijdschriften en e-learnings op het gebied van psychologie en psychiatrie. Zo kun je op je gemak en wanneer het jou het beste uitkomt verdiepen in jouw vakgebied.
