The Effect of Parent-Mediated Joint Attention Intervention on Joint Attention and Language Skills in Children with Autism Spectrum Disorder - A Systematic Review

Abstract

This review aimed to systematically estimate the effects of parent-mediated joint attention intervention on joint attention and language skills in children with ASD. A thorough literature search was conducted in six databases via systematic electronic literature search between January 2000 and July 2023. A total of 11 studies were included after following a stepwise systematic screening of the searched literature. Our findings showed improvements in joint attention skills in young children with ASD following parent-mediated joint attention intervention. Further, it showed insufficient evidence for the effect on language skills of children with ASD following parent-mediated joint attention intervention.

Keywords:

• Autism spectrum disorder
• children
• joint attention
• language skills
• parent-mediated intervention

Introduction

Young children communicate using a non-verbal mode of communication (i.e., eye-contact, pointing, etc.) during the early developmental years, even before they start expressing verbally (Mundy & Newell, 2007). Joint attention is a pivotal skill set of nonverbal behaviours, including eye gaze, pointing, and showing, generally used to share a common goal during a communicative exchange that emerges during early developmental years (Bruinsma et al., 2004). Joint attention is also defined as a complex set of behaviours whereby two individuals communicate about, and focus attention on, a third object or event (Bakeman & Adamson, 1984). Joint attention has been defined differently by many scholars in the literature. For this study, we operationally define joint attention as ‘any aspect of sharing attention with another individual regarding an object or event of mutual interest’ (Mundy & Newell, 2007). Joint attention behaviours could be of two types, i.e., Initiation of Joint Attention (IJA) and Responding to Joint Attention (RJA). IJA refers to infants’ use of gestures and eye contact to direct others’ attention to objects, events, and to themselves (Mundy et al., 2007). The RJA, on the other hand, refers to the child’s ability to follow the direction of gaze and gestures of others to share a common reference. Usually, RJA and IJA are involved as two domains in the assessment of joint attention skills.

Before acquiring more complex cognitive and verbal abilities, most young infants establish efficient social communication by learning joint attention skills together with other crucial social skills including eye contact, vocalizations, and social referencing (Neimy et al., 2017). The first stage of joint attention, responding to joint attention, appears between 3 to 4 months of age, while the second stage, initiating joint attention, appears between 8 and 13 months of age (Thorup et al., 2018; Van Hecke et al., 2016). Delays in initiating joint attention typically appear between 12 and 24 months of age, which frequently restricts crucial interactions with others required for later language development (Adamson et al., 2019).

Joint attention is associated with later language development in typical as well as atypical development (Bottema‐Beutel, 2016). One of the early developmental disorders where the joint attention is affected is autism spectrum disorder (ASD). A systematic review and meta-regression by Bottema‐Beutel (2016) reported strong associations between joint attention and language in children with ASD and typically development, that varied according to the type of joint attention and between groups of children with ASD and typically development. Research indicates that the role of joint attention is pivotal such that, when trained, it produces collateral improvements in areas, including language, affective expression, and social initiations in children with ASD (Kasari et al., 2008; Whalen et al., 2006; Yoder & Stone, 2006). Increase in RJA correlates with communication gains in children with ASD, such as the increased length of utterances and scores on receptive language skills (Murray et al., 2008). Kasari and associates (Kasari et al., 2014) designed the Joint Attention Symbolic Play Emotion Regulation (JASPER) intervention which targets joint attention episodes and demonstrated language gains even for minimally verbal children with autism.

Parent-mediated interventions for children with ASD

In the field of ASD, the caregiver-focused approach to early intervention is also known as parent-implemented or caregiver-implemented intervention (Boyd et al., 2014; Woods & Brown, 2011). Caregiver-implemented intervention is an approach where professionals use coaching or collaborative consultation to support children’s learning using natural, everyday activities (Woods & Brown, 2011). During Parent-Mediated Intervention (PMI), parents or caregivers receive training and coaching from the qualified professionals to facilitate children’s learning by building upon parents’ skills across various contexts (Shalev et al., 2020). Parent-implemented interventions designed for young children with ASD can be effectively taught and used by parents in their homes (Schertz et al., 2013). Thus, it would seem most appropriate that parents be involved in the early intervention sessions to alleviate the core social skill difficulties such as joint attention in children with ASD (Green et al., 2010). Parents play a pivotal role in intervention delivery, specifically when it targets infants and toddlers. The theory and body of research on parent training have resulted in several empirically supported intervention methods to address the symptoms exhibited by children with ASD. As a result, there is increasing literature on parent-mediated interventions targeting autism symptoms during early intervention. A few approaches focus on training parents in facilitating or delivering the intervention in children with ASD. By training parents to deliver the intervention to their children, who are at increase likelihood of developing ASD, young children can be guided to a more typical developmental trajectory (Landa, 2018). Parents can incorporate the naturalistic techniques into a range of daily routines in children with autism, thereby increasing the intervention dosage and facilitating the generalization and maintenance of treatment gains (Schreibman et al., 2015). Recently, Shire et al. (2018) reported that caregivers’ strategy implementation, when investigated as a time-varying covariate, was associated with children’s spontaneous communication. As per the study conducted by Jones & Feeley (2009), parents could be taught effective treatment strategies that target core difficulties in children with autism, such as joint attention. The authors also highlighted the importance of joint attention in early intervention and parent training programs involving new parents. The JASPER model of parent-mediated intervention showed increased parents’ responsive behaviour during the intervention and at a 6-month follow-up (Shire et al., 2016). Another home-based autism intervention program targeting joint attention for 18 to 40-month-old children demonstrated the efficacy and feasibility of training parents in the role of a full-time therapist to carry out the intervention at home (Perera et al., 2016). For the present review, we considered studies that followed parent-mediated joint attention interventions that targeted joint attention skills as primary or secondary outcomes in children with ASD.

Study purpose

Systematic reviews on parent-mediated intervention and joint attention interventions have been undertaken in children with ASD (Althoff et al., 2019; Murza et al., 2016; Oono et al., 2013) in the past. However, the effect of parent-mediated joint attention intervention on joint attention and language skills in children with autism is seldom reported or known. We, therefore, systematically reviewed the effects of parent-mediated joint attention intervention on joint attention and language skills in children with ASD. The present study attempts to address the following research questions: (a) can parent-mediated joint attention intervention facilitate improvement of joint attention skills in children with ASD, and (b) can parent-mediated joint attention intervention bring improvement in the language skills of children with ASD?

Method

Registration

The current systematic review followed the methodology and reporting by the Preferred Reporting Items for Systematic Review and Meta-Analyses guidelines (PRISMA-P) (Moher et al., 2015). Additionally, this review was specified and registered in advance with the International prospective register of systematic reviews PROSPERO (registration number: CRD42020206439).

Information sources and search strategy

The first author conducted an initial electronic database search in October 2020, and rechecked in August 2023. A thorough investigation was conducted on six databases, namely PubMed, Psych INFO, CINAHL complete, SCOPUS, ProQuest, and Web of Science, based on the complete list of searches according to the PICOS framework (Patient/Population, Intervention, Comparison, Outcome, and Study; Amir-Behghadami & Janati, 2020) criteria. In our study, only three components of the framework were used, i.e. patient or population (P), intervention (I), and outcome (O), as there was no comparison group. The references of the included studies were also searched to identify additional relevant studies. Table 1 presents the search keywords and the hits obtained after the search in each of the six databases mentioned above. The search was restricted to original articles published in peer-reviewed journals between January 2000 and July 2023. This period was chosen to identify all parent-mediated joint attention intervention studies in the last two decades and encompass the current state of literature. All search results were imported into Rayyan- a web application for systematic reviews (Ouzzani et al., 2016), and deduplicated.

Table 1. Search keywords and total hits in various databases.

Keywords	Databases searched (Total hits)
Population Parent, Coach, Mediator, Caregiver Child, toddlers, pre-schoolers, young children Autism, Autism Spectrum Disorder, Pervasive Developmental Disorder, Autistic Disorder, ASD, Asperger	Web of Science (n = 3812) PsycINFO (n = 1299) PubMed (n = 3565) CINAHL Complete (n = 167) Scopus (n = 252) ProQuest (n = 4989)
Intervention Intervention, treatment, therapy, training Parent training, parent-mediated training, parent-mediated joint attention, parent education, parent psychoeducation, parent delivered, caregiver-mediated
Outcome Joint attention, responding joint attention, initiating joint attention, receptive language age, expressive language age

Eligibility criteria

Articles retrieved from the initial search were included in this review based on the following inclusion criteria: (a) The included studies should have children participants (age 6 or younger) with a confirmed diagnosis of ASD; (b) The study must include a parent-mediated joint attention intervention only, or parent-mediated intervention with joint attention as a significant part of the intervention program (i.e., parents were trained to actively deliver joint attention intervention strategies); (c) The parent-mediated joint attention intervention must have the physical presence of a parent-child-interventionist triad. Other types of parent-mediated joint attention interventions (such as virtual or teleintervention) were not considered, given the methodological variations; (d) The studies using parent-mediated joint attention training should have included and reported primary outcome measures related to either joint attention, such as RJA, IJA outcomes, eye contact, pointing, showing, either using a standardized tool or task, and/or secondary outcome measure such as language skills (e.g. receptive language and expressive language); (e) Articles that were written in English; (f) Original (peer-reviewed) studies that incorporated any of the intervention designs such as Randomized Controlled Trial (RCT), quasi-experimental design, pretest-post-test design or Single Subject Experimental Design (SSED). The authors did not include grey literature in this review, given the scientific rigor involved in the method and results and the possibility of a high risk of bias. Studies with the following criteria were excluded from the present review: (a) Case reports, conference proceedings, abstracts, letters, reviews, or studies that had no relevant outcome measures, (b) Children with ASD having co-morbid conditions such as ADHD, epilepsy; (c) Intervention services provided through tele-rehabilitation mode.

Review process

A two-stage screening process was carried out to identify if the relevant articles met the inclusion criteria. Two authors (CP and HM) completed the title and abstract screening independently to reach a consensus. The first and third authors (CP and RS) independently reviewed the full-text articles to determine whether they met the inclusion criteria. Inter-rater agreement was calculated by dividing the total number of agreed eligible studies by the total number of studies and multiplying by 100 (Martinho et al., 2022; Zhang & Chen, 2024). The inter-rater agreement between the two authors was 96% for title screening, 98% for abstract screening, and 92% for full-text screening. The fifth author (ST) resolved the disputes for full-text articles to finalize the eligible articles until the inter-rater agreement reached 100%. The inter-rater reliability of data extraction was 91%; disagreements were resolved by the fifth author (ST) through discussion until the inter-rater agreement reached 100%.

Risk of bias/quality assessment

We used appropriate tools to assess the risk of bias according to the research design of the included studies. Cochrane’s risk of bias tool for single-case research design studies (Reichow, Barton, & Maggin, 2018) was used for studies with a single-subject experimental design. This tool has eight domains to assess the risk of bias (sequence generation, participant selection, blinding of participants and personnel, procedural fidelity, blinding of outcome assessors, selective outcome reporting, dependent variable reliability, and data sampling). A three-point rating scale of high, unclear, and low is given for each domain-specific item. For studies that employed non-randomized intervention, the Risk of Bias in Non-Randomized Studies - of Interventions (ROBINS-I) tool was utilized (Sterne et al., 2016). ROBINS-I assesses the risk of bias on seven domains (i.e., bias due to confounding, bias in selection of participants into the study, bias in classification of intervention, deviations from intended interventions, missing data, measurement of outcomes, and selection of reported results). Each domain level judgment in ROBINS-I uses a five-point rating scale of low, moderate, serious, critical, or no information of bias. The internal validity and risk of bias for studies using RCTs were determined by using the Cochrane "Risk of Bias" (RoB 2) tool. The RoB 2 (Sterne et al., 2019) assesses the risk of bias on five domains (i.e., risk of bias arising from randomization process, risk of bias due to deviation from intended interventions, risk of bias due to missing outcome data, risk of bias in the measurement of outcome, and risk of bias in selection of the reported result), by answering each statement that describes as high (H), some concerns (SC), or low (L) risk of bias. For all the three tools used for risk of bias assessment in this review, a high risk of bias indicates low quality of studies, whereas a low risk of bias indicates high quality of studies.

Data extraction

A data extraction template was developed to extract data from the included studies. The response fields in the data extraction form consisted of specific parent and child characteristics. The parent characteristics reported were the number of subjects (parents) in each group, i.e., the number of fathers and/or mothers recruited in the study. The child characteristics included were mean age and number of children recruited in each study. Details of child’s and family’s race, ethnicity and cultural background were also collected. In addition, the intervention characteristics of the included studies were also retrieved, such as type of intervention used, frequency and duration of each session, duration of parent training, location of parent training, outcome measures used as RJA, IJA, Receptive Language skills (RLS), Expressive language skills (ELS), baseline/pre-intervention scores for joint attention, and the results obtained. The extracted data was entered in the pre-designed template according to the characteristics of the included studies. The first author (CP) extracted data from all the included studies using the data extraction template. This was followed by verifying and confirming the key variables in the data by the fifth author (ST).

Outcome measures

The measures of interest for this review were four variables related to children with ASD. The primary outcome measures included RJA and IJA skills, and secondary outcome measures were receptive language and expressive language skills in children with ASD. The use of terminology for these outcome measures varied across the literature, while our criteria for inclusion in terms of joint attention measures remained the same (as described in introduction section, page 4, para 1). Table 2 presents the various terms related to each outcome measure as reported in the included studies.

Table 2. Terms related to outcome measures from the included studies.

Citation(s)	Outcome measures	Terms used in literature
Hansen et al. (2018)	RJA	Child responses to joint attention bids
Jones et al. (2006)		Respond to joint attention
Jones & Feeley (2009)		Responding to others joint attention directives
Kasari et al. (2010)		Frequency of joint attention responses
Rocha et al. (2007)		Child joint attention responding
Schertz & Odom (2007)		Responding to joint attention
Chiang et al. (2016)	IJA	Pointing, Alternate gaze, and showing
Jones et al. (2006)		Initiate to joint attention
Jones & Feeley (2009)		Initiating joint attention
Kasari et al. (2015)		Initiating joint attention
Perera et al. (2016)		Pointing
Rocha et al. (2007)		Child initiated joint attention bids
Schertz & Odom (2007)		Initiating joint attention
Drew et al. (2002)	Receptive language	Words understood
Kasari et al. (2015)		Reynell Receptive Language
Oosterling et al. (2010)		Words understood
Drew et al. (2002)	Expressive Language	Words said
Jones et al. (2006)		Vocalizations, Phonemes, and words
Kasari et al. (2015)		Reynell Expressive language
Oosterling et al. (2010)		Words said

RJA: Responding joint attention; IJA: Initiating joint attention.

Results

A total of 14,362 records were identified through the initial search of six databases. After deduplication, 14,084 articles remained. Following title and abstract screening, 78 articles remained. As there were insufficient studies per outcome measure(s) of interest in each research design, a meta-analysis could not be performed for this review.

A total of 11 articles met the inclusion criteria for the narrative review (as shown in Figure 1). All selected studies incorporated parent-mediated joint attention intervention using different intervention designs, such as randomized controlled trials (n = 4), non-randomized controlled trials (n = 2), and single subject experimental design (n = 5). Results of the narrative synthesis is presented under the following sections: parent and child characteristics, intervention characteristics, and treatment outcomes for measures of interest.

Figure 1. Prisma flow diagram showing search and screening for the review.

Parent and child characteristics

Participants included were either parents or caregivers of children with ASD. Only one of the studies (Hansen et al., 2018) included caregivers as grandparents of children with ASD, while the remaining ten studies had parents of children with ASD as participants. The number of parents in the included study varied depending on the intervention design. The total number of parents in the studies having an RCT design ranged between 24 to 104 parents (Drew et al., 2002; Oosterling et al., 2010), while each group (experimental and comparison) did not exceed 43 participants. There was heterogeneity observed in the parental characteristics (i.e., race, ethnicity, culture, educational qualification) as these studies were conducted in different countries and regions. Mothers implemented the intervention in most of the studies in comparison to fathers. Besides, a few of the studies did not mention the mother-to-father ratio. Only 50% of the studies reported the educational qualifications or the years of education of the parents involved in the study (Chiang et al., 2016; Kasari et al., 2010; Kasari et al., 2015; Oosterling et al., 2010; Perera et al., 2016; Schertz & Odom, 2007). Only one study investigated family characteristics such as socioeconomic status and family income (Chiang et al., 2016). Table 3 presents the parent and child characteristics of the included studies.

Table 3. Parent and child characteristics of the included studies.

Citation	Study Design	Parent characteristics			Child characteristics
		Number of parents in the treatment Group	Number of mothers included in the study	Number of fathers included in the study	Number of children with ASD	Mean age (in months)	Race, ethnicity or cultural background
Drew et al. (2002)	RCT	24 parents	No information	No information	Treatment: 12 Comparison: 12	Treatment: 21.4 Comparison: 23.6	–
Kasari et al. (2010)	RCT	Intervention: 19 Waitlist: 19	No information	No information	Treatment: 19 Comparison: 19	Treatment: 33.3 Comparison: 34.3	Ethnicity Caucasian: Treatment: 10(53%) Comparison: 12 (63%) Minority Treatment: 9(47%) Comparison: 7 (37%)
Kasari et al. (2015)	RCT	Treatment (JASPER): 43 Comparator (PEI): 43	Treatment (JASPER):43 Comparator (PEI): 43	–	Treatment: 43 Comparison: 43	Treatment: 30.7 Comparator: 32.3	Race/Ethnicity African American Caucasian Hispanic Asian Other
Oosterling et al. (2010)	RCT	104	57	47	Treatment: 36 Comparison: 31	Treatment: 35.2 Comparison: 33.3	–
Chiang et al. (2015)	Quasi-experimental Research design	Intervention: 18 Comparison:16	Intervention: 18 Comparison:16	Intervention: 18 Comparison:16	Treatment: 18 Comparison: 26	Treatment: 35.9 Comparison: 39.4	–
Perera et al. (2016	Prospective intervention design	No information	No information	No information	Experimental group: 62 Comparison group: 42	Experimental group: 32 Comparison group: 54.2	–
Hansen et al. (2018)	Concurrent multiple baseline across caregiver-child dyads design	3	2 and 1 grandmother	–	3 children with ASD	53	–
Jones et al. (2006)	Multiple baseline probe design across two joint attention skills	3	2	1	2 children with ASD	24.6	–
Jones & Feeley (2009)	Multiple probe design across two joint attention skills	5	3	2	3 children with ASD	36.7	–
Rocha et al. (2007)	Multiple baseline design across participant pairs	3	2	1	3 children with ASD	31.6	–
Schertz & Odom (2007)	Mixed methods research design	3	3	–	3 children with ASD	23.6	–

JASPER: Joint attention, Symbolic Play, Engagement and Regulation; PEI: parent only psycho-educational intervention; RCT: randomized controlled trial, RJA: Responding joint attention; IJA: Initiating joint attention.

The number of children included in each study varied depending on the study design. A total of 215 children were recruited in studies with RCT design, 148 children in Non-randomised Controlled Trials (NRCT) studies, and 14 children were included in SSED studies. The included studies had children aged between 21.4 to 53 months. A few studies exclusively recruited children below 36 months of age (Jones et al., 2006; Kasari et al., 2010; Schertz & Odom, 2007). Of the 11 included studies, ten studies reported a total of 259 boys and 72 girls. While one study did not mention the gender of the children recruited (Chiang et al., 2016).

Intervention characteristics

The duration of parent implementation of joint attention intervention varied between 8 weeks to 2 years across studies. The duration and frequency of sessions also varied between studies. In this review, one study presented a parent-mediated joint attention intervention that notably combined training preschool teachers to facilitate joint attention skills as the initial part of their study (Jones et al., 2006). All included studies provided joint attention training using various models of interventions (e.g., JASPER, pivotal response training, discrete trial teaching, etc.). The method of delivering training to parents included group-based interventions, one-to-one coaching, and training using video modules. The intervention was carried out in various settings such as home (Drew et al., 2002; Jones & Feeley, 2009; Oosterling et al., 2010; Perera et al., 2016; Schertz & Odom, 2007; Kasari et al., 2015; Rocha, Schreibman, & Stahmer, 2007); community (Jones et al., 2006), and clinics (Kasari et al., 2010). Table 4 below presents the intervention characteristics of the included studies.

Table 4. Intervention characteristics of the included studies by research design.

Citation	Type of Parent-mediated treatment approach	Duration of intervention (total sessions)	Parent-training location	Frequency and duration of each session (in minutes)	Outcome measures of interest for the review				Measures reported in the studies	Results for joint attention outcome/language outcome	Joint attention scores during baseline/Pre-intervention
					RJA	IJA	RLS	ELS
Randomized Controlled Trial
Drew et al. (2002)	Social pragmatic joint attention parent training program v/s local services	1 year	Home	1 per 6 weeks Each session − 180 minutes	–	–	✓	✓	Words understood Words said	Positive effect on receptive and expressive language skills	Intervention: 12.8 (1.6) Comparison: 12.0 (2.4)
Kasari et al. (2010)	Joint attention intervention	8 weeks (24 sessions)	Clinics or home	24 sessions Each session- 30 minutes	✓	✓	–	–	- Frequency of joint attention initiations - Frequency of joint attention responses - Primary outcome measures - Unengaged/other Object engagement -Joint engagement	Positive for RJA and null effect was observed for IJA	Intervention: 0.42 (0.69) Comparison: 0.63 (0.23)
Oosterling et al. (2010)	Focus parent training v/s care as usual	2 year (approximately 12 sessions)	Home	First year 180 minutes home visits every 6 weeks Second year 1 visit every 3 months	–	–	✓	✓	-Words understood -Words said -Joint attention factor	Null effect on receptive and expressive language skills	Intervention: 4.84 (4.62) Comparison: 4.61 (4.40)
Kasari et al. (2015)	JASPER parent mediated approach v/s psychoeducational intervention	10 weeks (20 sessions)	Clinic	60 minutes/week 2 sessions of 30 minutes/week	–	✓	✓	✓	-Duration of joint engagement -IJA -Child play skills -Play level -Joint attention -Cognitive & language skills	Null effect on IJA. Positive effect on receptive and expressive language skills	Intervention: −0.8 (2.3) Comparison: −0.9 (0.2)
Non-randomized controlled trial
Chiang et al. (2016)	Caregiver-mediated joint engagement approach	8 weeks (20 sessions)	–	20 sessions 60 minutes/session Twice/week	–	✓	–	–	-Alternate gaze -Showing -Pointing -Giving	Null effect on IJA	Intervention: 0.47 (0.87) Comparison: 2.44 (3.03)
Perera et al. (2016)	Home-based intervention	–	Home	420 minutes 1 session/month	–	✓	–	–	-Sustained eye contact -Response to name -Social reciprocity -Imitative behavior -Pointing	Positive effect on IJA	Intervention: 0.32 (1.78) Comparison: 11.79 (17.34)
Single Subject Experimental Design
Hansen et al. (2018)	Caregiver training and coaching	Until the caregiver scored 80% or higher over 3 consecutive sessions	Institution lab	Average was 5 to 6 sessions/participant	✓	–	–	–	-Parent fidelity -Child responses to joint attention bids	Positive effect on RJA	Below 30% for all the participants
Jones et al. (2006)	Joint attention instruction (RJA followed by IJA)	Till the child mastered RJA with 36 original toys IJA also had to be mastered	Home or community settings	Participant 1 RJA-19 sessions IJA- 36 sessions Participant 2 RJA-22 sessions IJA-35 sessions	✓	✓	–	✓	-RJA -IJA	Positive effect on IJA, RJA, and expressive language	Below 20% for all the participants
Jones & Feeley (2009)	Discrete trial format along with pivotal response training strategies	90% independent responses across two consecutive sessions	Home	Participant 1 RJA-7 sessions IJA- 24 sessions Participant 2 RJA-54 IJA-117 Participant 3 RJA-107 IJA-107 sessions	✓	✓	–	–	-RJA -IJA	Positive effect on RJA and IJA	All participants scored 0 to 40% for RJA and only 1 participant scored 0-20% for IJA
Rocha et al. (2007)	Joint attention parent training	17 hours of intervention	Clinic	3 days/week Each session-1 hour and 15 min	✓	–	–	–	-Parent joint attention initiations -Child RJA	Positive effect on RJA	Below 40% for all participants
Schertz & Odom (2007)	Parent mediated, developmentally oriented, and interaction based intervention model	Dyad A- 16 weeks Dyad B- 26 weeks Dyad C- 9 weeks	Home	One session/week	✓	✓	–	–	-RJA -IJA -Parent fidelity and adherence	Positive effect on RJA and IJA in 2/3 child participants	Below 15% for all participants

JASPER: Joint attention, Symbolic Play, Engagement and Regulation; PEI: parent only psycho-educational intervention; RCT: randomized controlled trial; RJA: responding joint attention; IJA: initiating joint attention; RLS: receptive language skills; ELS: expressive language skills.

Only three out of the 11 studies (Kasari et al., 2010; Rocha et al., 2007; Schertz & Odom, 2007) reported data on parents’ implementation of strategies at home. Overall, studies examining parent-mediated joint attention interventions revealed that parents could promote joint attention skills within given timeframes, i.e., a minimum of 8 weeks and each session with a duration of 30 minutes (Kasari et al., 2010). Parent or child improvement was observed within various intervention approaches such as JASPER (Kasari et al., 2010; Kasari et al., 2015), social-pragmatic approach (Drew et al., 2002), and home-based autism intervention (Perera et al., 2016).

Treatment outcomes for measures of interest (i.e., joint attention and language skills in children with ASD)

Most of the studies used a video-taped session of parent-child interaction to code the joint attention behaviors (7 out of 11 studies), and 2 out of the 11 studies have used the Early Social Communication Scale (ESCS; Mundy et al., 2007), which is one among the frequently used standardized measure to assess joint attention skills in children with ASD.

All the five studies that used SSED performed visual analysis and descriptively reported the results. All five included studies found the presence of a functional relation between parent-mediated intervention and child’s joint attention skills. Over half of the seven group design studies that met the RCT and NRCT group design standards, reported findings indicating small to strong effects of the intervention group on child’s joint attention skills. For this review, the treatment outcome (i.e., joint attention skills) of all included studies were categorized into two categories: positive and null. The positive result across studies refers to an increase in joint attention or language skills in children with ASD after receiving the intervention. Null results indicate no significant change or improvement in joint attention or language skills after receiving the intervention. Treatment outcomes for measures of interest from the included studies are provided in Table 4. Six of eleven studies used direct teaching methods to teach joint attention (RJA) strategies to parents of children with ASD (Hansen et al., 2018; Jones et al., 2006; Jones & Feeley, 2009; Kasari et al., 2010; Rocha et al., 2007; Schertz & Odom, 2007). Of the six studies that targeted RJA skills, five studies (with a single-subject intervention design) reported a significant increase in RJA skills in children with ASD after parent-mediated joint attention intervention (Hansen et al., 2018; Jones et al., 2006; Jones & Feeley, 2009; Rocha et al., 2007; Schertz & Odom, 2007). Another RCT study reported improvement in RJA post-intervention (Kasari et al., 2010). These studies evaluated joint attention skills using coding from the video recordings and sections from different standardized assessment tools.

Unlike RJA, mixed results were observed for IJA skills in children with ASD. Only four of the eight studies showed improvement in IJA (Jones et al., 2006; Jones & Feeley, 2009; Perera et al., 2016; Schertz & Odom, 2007). Few other studies showed a null effect, where a few of the children with ASD exhibited improvements in IJA skills, while the scores were not statistically significant (Chiang et al., 2016; Kasari et al., 2010; Kasari et al., 2015). In comparison, all three parent participants showed positive results by increasing the occurrence of IJA bids with their child with ASD (Rocha et al., 2007). Besides, in the study by Jones et al. (2006), 4 of the 5 participants acquired IJA skills but required more sessions than RJA. Interestingly, after receiving an intervention to improve initiating skills, two-child participants showed an increase in IJA skills, and one child participant did not exhibit improvement in IJA skills (Schertz & Odom, 2007).

Only three of the eleven included studies examined receptive language skills in children with ASD following parent-mediated joint attention training (Drew et al., 2002; Kasari et al., 2015; Oosterling et al., 2010). One among these three studies showed a positive effect of parent-mediated joint attention training, where children with ASD showed 17 months of gain in receptive language over a 9-month study (Kasari et al., 2015). The other two studies showed a null effect for receptive language in children with ASD (Drew et al., 2002; Oosterling et al., 2010), such that children with ASD in the experimental group showed marginally higher comprehension compared to the children from the comparison group. Among the four studies that assessed expressive language as an outcome measure, two studies (with RCT design) reported significant gains in children with ASD from the intervention group (Drew et al., 2002; Kasari et al., 2015). In the study by Drew et al. (Drew et al., 2002), children went from being non-verbal to a single word or phrase speech in the experimental group (parent training group) than in the comparison group (local services group). In another study, ten months of gain in expressive language was observed over the 9-months intervention period (Kasari et al., 2015). Another study that used a single-subject intervention design, reported improvements in expressive language and social-communicative characteristics post-intervention (compared to baseline) in both the child participants with ASD (Kjellmer et al., 2012). Yet another study reported no significant improvement or null effect on expressive language in the intervention group after receiving a parent-mediated joint attention intervention (Oosterling et al., 2010).

Risk of bias

The two reviewers (CP & HM) independently coded the risk of bias for each study using the appropriate tools based on the research design. A total of three discrepancies among the two reviewers were resolved through consensus, including the fifth author (ST). Inter-rater agreement was 91%. All disagreements were resolved by the fifth author (ST) until the inter-rater agreement reached 100%. Table 5 presents the RoB ratings for the studies with RCT. Only two studies with RCTs obtained a low risk of bias in all domains (Kasari et al., 2010; Kasari et al., 2015). Another two studies with RCTs indicated a high risk of bias related to the randomization process and measurement outcome (Drew et al., 2002; Oosterling et al., 2010). Table 6 shows the RoB ratings for the included NRCT studies. Both the included NRCTs showed an overall serious or high risk of bias (Chiang et al., 2016; Perera et al., 2016), mainly related to confounding and blinding of the assessors. Nearly half of the included studies in this review used a single subject experimental design (n = 5; Jones & Feeley, 2009; Hansen et al., 2018; Schertz & Odom, 2007; Jones et al., 2006; Rocha et al., 2007). For the SSED studies, sequence generation indicated a high risk of bias (100%)*; participant selection indicated a low risk of bias (100%)*; blinding of participants and personnel showed unclear (60%)* and high risk of bias (40%)*. Procedural fidelity indicated a low (60%)* and high risk of bias (40%)*. Blinding of outcome assessors showed a high risk of bias (60%)*, and selective outcome reporting indicated a low risk of bias (60%)*. Dependent variable reliability indicated a low risk of bias (80%)*, and data sampling indicated having a low risk of bias (100%)*. Tables 5 and 6 present the risk of bias ratings for the included RCT and NRCT studies, respectively. Summary of risk of bias by domain for included studies with SSED in the current parent-mediated joint attention review is presented in Figure 2.

Figure 2. Summary of Risk of Bias by domain for studies with single subject experimental design.

Table 5. Risk of bias ratings for the RCT studies included in the review.

Studies	Risk of bias arising from randomization process	Risk of bias due to deviation from intended interventions	Risk of bias due to missing outcome data	Risk of bias in the measurement of outcome	Risk of bias in selection of the reported result	Overall
Drew et al. (2002)	High	Some concerns	Low	High	Some concerns	High
Kasari et al. (2010)	Low	Low	Low	Low	Low	Low
Kasari et al. (2015)	Low	Low	Low	Low	Low	Low
Oosterling et al. (2010)	High	Some concerns	Low	High	Some concerns	High

Table 6. Risk of bias ratings for the NRCT studies included in the review.

Studies	Bias due to confounding	Bias in selection of participants into the study	Bias in classification of intervention	Deviations from intended interventions	Missing data	Measurement of outcomes	Selection of reported results	Overall
Chiang et al. (2016)	No information	Moderate risk of bias	Low risk of bias	low risk of bias	low risk of bias	Serious risk of bias	Moderate risk of bias	Serious risk of bias
Perera et al. (2016)	No information	Moderate risk of bias	No information	low risk of bias	low risk of bias	Serious risk of bias	Moderate risk of bias	Serious risk of bias

*the percentages shown within parentheses indicate the proportion of studies within SSED (n = 5).

Social validity

Social validity was measured in only 3 of the 11 (27%) studies, all 3 articles of which specifically claimed only the intervention. Schertz & Odom (2007) used a researcher-developed questionnaire with a five-point Likert scale rating, which allowed for differentiated levels of responses. The social validity ratings of all 3 studies suggest an improvement in child joint attention behaviours as a result of the intervention along with an improvement in confidence of parent’s ability to support child interaction (Jones et al., 2006; Rocha et al., 2007; Schertz & Odom, 2007).

Discussion

A systematic review aimed to evaluate the benefits of parent-mediated joint attention intervention in children with ASD is an important step to support the utility of parent-mediated joint attention intervention for families with ASD. The present review focused on the effects of parent-mediated joint attention intervention with specific outcomes such as joint attention and language skills in children with ASD. Given the moderate to high risk of bias ratings in some of the included studies, results of the review should be interpreted with caution. The review found encouraging evidence to suggest that parent-mediated joint attention intervention can effectively improve joint attention skills in children with ASD. Furthermore, the review’s results showed positive, though limited, evidence for improvement in language skills in children with ASD following parent-mediated joint attention intervention.

Our review showed positive results for joint attention skills of children with ASD in eight of eleven studies. In other words, the parent-implemented joint attention training helped the children with ASD improve their joint attention abilities. Our finding is comparable with the findings reported from direct (or clinician-driven) joint attention intervention on child’s joint attention skills in children with ASD (Murza et al., 2016). Improvements in RJA skills were reported by almost all the studies (n = 6) that measured RJA in children with ASD. In contrast, only half of the studies (that measured IJA) showed significant improvements following parent-mediated joint attention intervention. An increase in the number of parent initiations (because of parent training and coaching) has been linked to an increase in the child’s RJA skills (Hansen et al., 2018; Rocha et al., 2007; Schertz & Odom, 2007). Besides, the natural environment (i.e., the child’s residence with familiar people), in the case of parent-mediated intervention, facilitates better results (including joint attention skills) for children with ASD (Perera et al., 2016). Better results for RJA over IJA could be explained because of the higher demands of the latter type of joint attention skills. Children with ASD have known difficulty with initiation and spontaneous communication (Kjellmer et al., 2012). Another reason for better gains in RJA over IJA could be explained due to differences in the duration of the intervention across studies, indicating that IJA requires more extended intervention (Schertz & Odom, 2007).

Contrary to the positive outcomes in the studies mentioned above, the results of two parent-mediated intervention studies did not show significant improvement in the joint attention of children with ASD after the intervention. A study Chiang et al. (Chiang et al., 2016) reported no significant change in IJA skills after the joint engagement intervention. However, a marginal improvement was found during the follow-up period and not immediately after the intervention. The authors of these studies attribute the following two reasons for the lack of positive effects on joint attention: first, the parents in the intervention group displayed moderate fidelity, which showed a lack of implementation of treatment strategies with their children with ASD. Second, the children with ASD in the comparison group had higher joint attention skills; thus, even though the children in the intervention group improved, it was still less compared to the comparison group. The authors also reported that parents had insufficient time to practice the treatment strategies with their children during the intervention period.

The present review found limited evidence regarding the effect of parent-mediated joint attention intervention on the language skills of children with ASD. Only four of eleven studies on parent-mediated joint attention intervention evaluated the impact of the intervention on language skills in children with ASD. Children with autism showed significant improvement in receptive language skills only in one study and expressive language skills in two studies following parent-mediated joint attention intervention, respectively (Drew et al., 2002; Kasari et al., 2015). The other two studies revealed a null or non-significant effect of parent-mediated joint attention intervention on the language skills of children with ASD. In the study conducted by Oosterling et al. (2010), the parent training program results were non-significant, revealing that the intervention did not influence receptive or expressive language outcome, while the experimental group exhibited a slight improvement as compared to the comparison group. Given the strong association between joint attention and language in children with ASD (Bottema‐Beutel, 2016), a positive effect of parent-mediated joint attention intervention is expected on language skills. The available data is, however, not sufficient to demonstrate the effect. A possible reason also could be that language improvement might be expected following gains in joint attention in children with autism, that might require more time (Bottema‐Beutel, 2016).

The findings from the present review need to be considered with caution, given that nearly half of the studies were of moderate to low quality with potential biases. There was heterogeneity in terms of parent and child characteristics in the studies included in this review. There was also variability in the intervention used in terms of duration, frequency, model of intervention used, and location. These factors could have possibly influenced the child’s outcomes for joint attention and language skills. Further, parent fidelity was also a concern in a few of the included studies, which could be another reason for null or lack of treatment effect. The social validity was reported in a handful of studies. The available findings, thus, cannot be generalized to all parent-mediated interventions targeting children with ASD, and therefore, recommend the need for future research.

Taken together, the findings of the review revealed that parent-mediated joint attention intervention has a positive effect on a child’s joint attention skills. In this review, however, the effect of parent-mediated joint attention intervention on language skills of children with ASD could not be commented on, due to insufficient data. Thus, more rigorous research on parent-mediated joint attention intervention is needed to assess and confirm the long-term treatment effects on children with ASD. Additionally, results from the present review support the importance of promoting parent-mediated joint intervention during the child’s early years to optimize joint attention skills in children with ASD. As a result, clinicians can teach parents to implement joint attention strategies for their children with ASD, which can be integrated into daily routines and home settings.

Limitations of the study

The present review is first of its kind to investigate the effects of parent mediated joint attention intervention on child measures. This review had certain limitations as well. First, we did not include unpublished or grey literature. It is, thus, possible that we would have missed some significant data. Although, all major databases were included in the literature search for this study. Second, since we limited our search to original studies published in English language only, it is likely that we would have missed some of the relevant literature in languages other than English. Third, studies using telehealth mode of intervention were not included in this review thereby limiting or possibly missing out relevant evidence. Fourth, as the review included studies involving different intervention designs, a possible bias is unavoidable. This approach of including studies from different intervention designs, is however, feasible and been used in autism literature (Davis et al., 2019).

Conclusion and recommendations for practice

The findings of the review showed that parent-mediated joint attention intervention demonstrate the use of parent-mediated approaches in facilitating joint attention skills, especially RJA skills in children with ASD. However, the efficacy of parent-mediated joint attention training for language skills in children with ASD could not be estimated due to insufficient data. Given the promising results of parent-mediated interventions and the association between joint attention and developmental skills, there is a need for more robust research to establish the effects of parent-mediated joint attention intervention in children with ASD. Findings of the review have implications regarding the scope of joint attention training component to be added in all parent-mediated intervention programs to facilitate the development of joint attention skills and subsequent developmental skills in children with ASD. Results of the review also highlights the need for further research examining effects of parent mediated joint attention intervention on various child measures, as language, social and emotional skills. In clinical settings, joint attention training component needs to be added in all parent-mediated early intervention programs to facilitate the development of joint attention skills in children with ASD. Clinicians are encouraged to involve parents during the intervention process to increase the likelihood of generalizing joint attention skills across various settings.

Ethics statement

An ethics statement is not applicable because this study is based exclusively on published literature.

Disclosure statement

The authors declare no competing interests.

Data availability statement

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.

Additional information

Funding

There was no funding support for this study.