Abstract
Play in children with autism spectrum disorder (ASD) is often atypical, yet consensus regarding effective occupational therapy strategies for improving play is not established. To examine the efficacy of strategies used in occupational therapy to improve play in ASD, authors completed a systematic review of papers from January 1980 through January 2019. Search terms included autism, Asperger’s, ASD, autistic in combination with play, playfulness, pretend, imagination, praxis, creativity, and generativity. Twenty papers met inclusion criteria and were reviewed. Reviewed interventions included parent education, modified play materials or environments, imitation of the child, and modeling by an adult, a peer, or video. Moderate to strong support exists for the specific strategies of imitation of the child and modeling for the child, with lesser or mixed support for other strategies. Certain strategies commonly used in occupational therapy may be effective in improving the occupation of play in ASD.
Children and youth with autism spectrum disorder (ASD) demonstrate variable deficits in social communication, repetitive behaviors, and restricted interests that are often reflected in their play. Children with ASD demonstrate the ability to play, yet their play may be different in quality, duration, and/or type. The atypical play of children with ASD is commonly reported (Bodison, 2015; Kasari & Chang, 2014; Lee et al., 2016; Tomaino, Miltenberger, & Charlop, 2014). They play less frequently, with limited flexibility and creativity, and use play objects atypically or repetitively. They are more likely to engage in play that provides pleasurable sensation past the age when this is common. The most universally documented deficit in children with ASD is a lack of symbolic play. These difficulties in play may result in rejection by others, limited friendships, bullying, or victimization, in turn leading to greater anxiety during peer interaction over time (Mazurek & Kanne, 2010; Rowley et al., 2012; Schupp, Simon, & Corbett, 2013).
The occupation of play is important because it is meaningful to the child, provides satisfaction and joy, promotes quality of life, and makes them happy (Moore & Lynch, 2018 ; United Nations Committee on the Rights of the Child 2013). Although some children with ASD may struggle with the social aspects of play and may appear unmotivated to engage in play with peers (Carré et al., 2015; Chevallier, Kohls, Troiani, Brodkin, & Schultz, 2012), play may engender social inclusion (Lindsey & Colwell, 2013; Stagnitti, O’Connor, & Sheppard, 2012), foster better social outcomes and friendships at later ages (Freeman, Gulsrud, & Kasari, 2015), and may be important for reducing stress associated with social play with peers as children with ASD grow (Schupp et al., 2013). Play should be an important goal for intervention for all of those reasons and more (Warreyn, van der Paelt, & Roeyers, 2014). Play also may provide “environmental enrichment” of the sort that impacts neurologic function in areas of the brain implicated in responses to novelty, praxis, imitation, and social behaviors (Burgdorf, Kroes, & Moskal, 2017; Pellis, Pellis, & Himmler, 2014; Siviy, 2016). Therefore, play could be essential for children with ASD who demonstrate difficulties in those specific areas of function.
There are many studies of play as an intervention, but there are fewer studies with play as the dependent variable and no consistently agreed-upon intervention to improve play outcomes. Despite the interdisciplinary belief that play is important, for occupational therapists, there has often been a disconnect between embracing play as an outcome of intervention given, the available studies regarding play, the outcomes measured in those studies, and considerations of play as occupation. Limited evidence exists within the field of occupational therapy to guide practice and occupational therapists require better information to support the selection of appropriate strategies to improve play for children with ASD. Consequently, the authors examined the literature to answer the following focused question: What strategies, used in occupational therapy, are effective for improving the occupation of play in children and youth with ASD?
Method
This systematic review was conducted using PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (see Supplemental Figure 1). The first author searched the literature from January 1980 through January 2019 within PubMed, PsycINFO, CINAHL, OTDBase, and the Cochrane Database. The search terms are provided in Table 1. With a research assistant, the first author completed initial review of titles and removed duplicates. All authors then worked in pairs to determine the appropriateness of abstracts for further review. A master citation table was created with decisions reported for each abstract and codes for exclusion recorded. Any discrepancies between author pairs were discussed, and the final decision required agreement of all three authors for any disputed paper. Full text review was also completed in pairs following the same process.
Search Terms.
Note. ASD = autism spectrum disorder.
Decisions about inclusion or exclusion were made based on predetermined criteria (Table 2). Papers were included if the subjects were 3- to 18-year-old children with ASD and whether the diagnosis of ASD was independently confirmed by the researchers was noted. Studies were excluded if they did not have some specific outcome measure of engagement in play with some statistical measure of change post intervention (Level III or above; Level IV papers were only included as part of a systematic review that also included higher level papers). To maintain a holistic focus on play as viewed by occupational therapists, articles were excluded if the outcomes were merely a component of play (i.e., proximity seeking) or a specific skill, ability, or trait (such as motor skills, imitation, joint attention, language, problem solving). Specific intervention strategies that were included were those that are considered hallmarks of occupational therapy such as modifying or altering activities, materials, or contexts or facilitation of engagement with prompting, cueing, directing, or modeling (American Occupational Therapy Association [AOTA], 2014; National Board for Certification in Occupational Therapy [NBCOT], 2018). Papers were also included if the strategies were related to therapeutic use of self, such as modifying adult interaction style. Papers were excluded if the strategies or interventions would not be typically used by an occupational therapist clinically, for example, provision of a robot.
Inclusion and Exclusion Criteria.
Note. ASD = autism spectrum disorder; PDD-NOS = Pervasive Developmental Disorder–Not Otherwise Specified; AOTA = American Occupational Therapy Association.
Author pairs analyzed each of the included papers for the quality of the paper, the type/level of the study, sample characteristics, specifics about the intervention, and outcome measures, and the results and the evidence table was completed with input from all three authors (Table 3). A risk of bias table was completed (Table 4) for the selected studies with specific notes about quality also provided in Table 3. Levels of evidence for individual papers and strength of the evidence for specific strategies were rated based on the AOTA guidelines for systematic reviews (AOTA, 2017) (see Table 5).
Evidence Table of Included Studies.
Note. ASD = autism spectrum disorder; dx = diagnosis; ADOS = Autism Diagnostic Observation Schedule (Lord et al., 2000); R = randomization; tx = treatment; IOA = interobserver agreement; IRR = inter-rater reliability; PDD-NOS = Pervasive Developmental Disorder–Not Otherwise Specified; SES = socioeconomic status; ID = intellectual disability; ADI-R = Autism Diagnostic Interview–Revised (Lord, Rutter, & Le Couteur, 1994); PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses; CARS = Child Autism Rating Scale (Schopler et al., 1980); EI= early intervention; PND = Percentage of Nonoverlapping Data; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses; GAS = Goal Attainment Scaling; SPA = Structured Play Assessment (Kasari et al., 2006).
Risk of Bias Table.
Source. Risk of bias table format adapted from Higgins, Altman, and Sterne (2011).
Note. Categories for risk of bias are as follows: low risk of bias (+), unclear risk of bias (?), and high risk of bias (−). NA = not applicable due to study design.
AOTA Guidelines for Systematic Reviews.
Source. AOTA (2017).
Note. RCT = randomized controlled trial; AOTA = American Occupational Therapy Association.
In particular, the authors were interested in the strategies used to improve play and these interventions were categorized based on the descriptions provided by the studies reviewed, using the categories of parent education, modified play materials, modified play environments, imitation of the child, modeling of the child, and modified interaction other than imitation or modeling. Parent education meant the parent was trained to perform some specific task or activity with his or her child. Modified play materials included alterations to the toys used or the specific selection of toys meant to elicit a specific type of play. Modifications also occurred during naturalistic play and were meant to allow for greater engagement or independence with the materials. Modifications to materials also included the provision of materials involving preferred themes or interests or those that were reinforcing. Modified play environments included the specific use of visuals to enhance understanding and performance, alterations to the sensory features of the play environment, or the specific choice of context for the data collection such as home versus school environments, or added peers. Imitating the child typically meant that an adult followed the child’s lead and imitated what the child was doing, using the same materials, verbalizations, and intonation. Modeling for the child occurred via an adult, a peer, or video and generally provided some form of demonstration or specific teaching. Author pairs made these categorization decisions with the third author providing input when there were any discrepancies. This categorization allowed for consideration of the strength of evidence for specific strategies used to promote play. Decisions about strength of evidence for specific interventions were made using guidelines provided by the AOTA (2017), which were based on designations provided by the U.S. Preventive Services Task Force (See Table 5).
Results
Of the over 2,000 records initially reviewed, 20 papers met the inclusion criteria and were reviewed for this systematic review. The majority of studies were excluded for level of evidence, as many were case reports or reports of single-subject research, or for either lack of an outcome measure of play or lack of a specific intervention strategy meant to improve play that would typically be used by an occupational therapist.
Level of Evidence and Quality
The 20 papers were relatively evenly dispersed between the three levels of evidence included in this review. There were six Level I papers (Chang, Shih, Landa, Kaiser, & Kasari, 2018; Jung & Sainato, 2013; Kasari, Freeman, & Paparella, 2006; Kasari, Paparella, Freeman, & Jahromi, 2008; Lang et al., 2009; Luckett, Bundy, & Roberts, 2007) and of these, three were randomized controlled trials (RCTs; Chang et al., 2018; Kasari et al., 2006; Kasari et al., 2008). Seven papers were rated Level II (Dawson & Adams, 1984; Field, Field, Sanders, & Nadel, 2001; Murdock & Hobbs, 2011; O’Connor & Stagnitti, 2011; Pinchover, Shulman, & Bundy, 2016; Strauss et al., 2014; Wong, 2013), and seven papers were rated Level III (Dawson & Galpert, 1990; Dewey, Lord, & Magill, 1988; Field et al., 2014; Rogers & Lewis, 1989; Stagnitti et al., 2012; Stahmer, 1995; Wolfberg, DeWitt, Young, & Nguyen, 2015).
The quality of the evidence varied considerably. Three of the twenty papers provided details of well-designed and executed RCTs. Three more papers were systematic reviews; however, all three of these reviews included Level IV studies. The remainder of the 14 papers demonstrated less rigorous designs and a variety of methodological issues such as the potential for selection bias, a lack of blind assessors, limited description regarding the interventionists, no examination of fidelity to intervention, and outcome measures with questionable reliability and validity (see Tables 3 and 4). In addition, the group of studies demonstrated significant variability in intervention type and duration. Interventions were frequently grouped together making it difficult to determine which aspect of the intervention promoted better play. In addition, these studies included minimal discussion of, or determination of, the child’s motivation to engage with the toy or activities chosen by the researcher. Children with ASD were also grouped together, often with limited information provided about their functional levels or language abilities. There was minimal reporting of the racial and socioeconomic background of participants.
Subjects
The subjects included over 700 children and adolescents with ASD and as is typical with the diagnosis, the majority were male. Although our target range was age 3 to 18, the actual subjects in our review ranged in age from birth to 18, as some systematic review papers included subjects who were under the age of 3. The subjects’ race or family socioeconomic background was reported in only five papers (Chang et al., 2018; Field et al., 2001; Pinchover et al., 2016; Rogers & Lewis, 1989; Wong, 2013).
Strategies and Interventions
Specific interventions were adult-structured (six papers) or child-directed (11 papers), and some studies provided both types. Strategies or interventions were provided to individual children in nine of the studies and to small groups of children or dyads in nine of the studies. Settings included schools or research locations, the home, and a summer camp. Intervention duration varied greatly from 3 min to 4.5 hr per day for 6 months. The papers reported the interventionists in 15 of the 20 studies. The most frequent professional providing play intervention was a psychologist or psychology student. Occupational therapists were interventionists or co-interventionists in only three studies (O’Connor & Stagnitti, 2011; Rogers & Lewis, 1989; Stagnitti et al., 2012). In one study, either the parent or the teacher played with the child (Pinchover et al., 2016).
Outcome Measures
Studies included various outcome measures. Specific assessment tools included the Penn Interactive Peer Play Scale (Fantuzzo et al., 1995) used in two studies (O’Connor & Stagnitti, 2011; Stagnitti et al., 2012), the Child-Initiated Pretend Play Assessment (Stagnitti, 2007) used in two studies (O’Connor & Stagnitti, 2011; Stagnitti et al., 2012), the Play Observation Scale (Rogers, Herbison, Lewis, Pantone, & Reis, 1986) used in one study (Rogers & Lewis, 1989), the Structured Play Assessment (Kasari et al., 2006) used in three studies (Chang et al., 2018; Kasari et al., 2006; Kasari et al., 2008), the Test of Playfulness (Bundy, 2010) used in one study (Pinchover et al., 2016), and the Play History Interview (Rogers et al, 1986) reported in one study (Luckett et al., 2007). The majority of the papers included researcher-created play coding schemes to rate behaviors from videotaped or observed play sessions. The types of coding that occurred included ratings of play novelty, mother–child interaction during play, developmental play level, affect, play complexity, involvement, fun, frequency and duration of types of play, and symbolic action.
Key Findings
Interventions targeting play were effective in improving play. This is documented in the outcomes of well-designed RCTs (Chang et al., 2018; Kasari et al., 2006; Kasari et al., 2008) as well as multiple studies of lower levels of evidence. All of the 20 papers reported improvements in play in some manner such as the complexity or type of play, the frequency or the duration of play, or creativity in play.
The most consistent finding regarding benefits of a particular strategy was for imitating the child, as described specifically in four studies (Dawson & Adams, 1984; Dawson & Galpert, 1990; Field et al., 2001; Field et al., 2014). Imitating the child was also a part of the more comprehensive play intervention used in the studies by Kasari and colleagues (Kasari et al., 2006; Kasari et al., 2008) and Chang et al. (2018) that also led to improved play outcomes in the amount and level of play. Thus seven studies, three of which were RCTs, two of which were Level II, and two of which were Level III, reported success with the strategy of imitating the child, providing strong support for this strategy. Imitation of the child with ASD was found to be more effective than modeling and contingent play responses for improving play creativity and duration (Dawson & Adams, 1984; Field et al., 2001).
There is also moderate to strong evidence that modeling for a child improved specific aspects of play (Jung & Sainato, 2013; Lang et al., 2009; Luckett et al., 2007; Murdock & Hobbs, 2011; Stahmer, 1995) such as functional play skills and the ability to engage in play dialogue. Modeling is part of the play intervention used in the RCTs as well (Chang et al., 2018; Kasari et al., 2006; Kasari et al., 2008). However, some of the evidence for modeling comes from review papers that included Level IV studies (Jung & Sainato, 2013; Lang et al., 2009; Luckett et al., 2007).
There is moderate evidence that modification of the environment influences play performance. This strategy was provided in seven studies that found improvements in play, including two RCTs, two Level II studies, and three Level III (Kasari et al., 2006; Kasari et al., 2008; O’Connor & Stagnitti, 2011; Pinchover et al., 2016; Rogers & Lewis, 1989; Stagnitti et al., 2012; Wolfberg et al., 2015). However, the specific modifications were quite varied between the studies, making it difficult to determine which particular type of environmental modification was best and leading to the designation of moderate as opposed to strong. Some studies altered the physical location of play, others the play partner, and still others altered either the sensory aspects of the environment, provided visuals to improve performance, or selected materials to be especially motivating for the child.
There was either insufficient evidence or mixed evidence for the rest of the strategies. For example, modifying materials appeared effective in increasing play complexity, involvement, and “fun” but there were too few studies to be certain (Dewey et al., 1988). Modification of adult interaction with the child while playing, through altering affect, encouraging involvement, cuing attention, or responding contingently, improved certain aspects of play such as complexity and play level or type in two studies (Field et al., 2014; Rogers & Lewis, 1989) but not in another (O’Connor & Stagnitti, 2011).
Discussion
Findings from this review support the ability of a practitioner to effectively intervene to improve play for children with ASD. The authors sought to determine which specific strategies appropriate for use during occupational therapy intervention were effective in improving the occupation of play in children and adolescents with ASD. Similar to prior reports, our results suggest a variety of methods can be effective in promoting the play of children with ASD (Barton & Wolery, 2008; Godin, Freeman, & Rigby, 2017; Jung & Sainato, 2013; Luckett et al., 2007; Tanner, Hand, O’Toole, & Lane, 2015). Specifically, this review finds that children with ASD can improve their participation, complexity, creativity, and novelty in play with targeted intervention. The strongest interventions in this review used imitation of the child and modeling for the child and combined adult-structured and child-directed approaches. However, this review also provides preliminary support for interventions such as modifying materials, environment, and interaction style.
Implications for Clinical Practice and Research
The research reviewed here provides support for specific strategies that have the potential to improve play for children with ASD. For example, occupational therapists should feel most confident including imitation of the child and modeling for the child, as they currently have the strongest evidence. Other strategies have emerging support such as the modification of materials, environment, and interaction style. Practitioners may choose to utilize these play interventions with emerging support; however, they must proceed cautiously, applying specific strategies while intervening to improve play with a client with ASD, while collecting data to document evidence of change for each individual client. In addition, practitioners must attend to tailoring the specific strategy to the desired outcome. For example, different strategies may be chosen based on whether the desired outcome is greater playfulness or greater creativity during play. Occupational therapists may consider writing goals to measure the complexity, novelty, frequency, duration, or type of play for their clients as these aspects of play appear alterable via intervention.
The need for additional research is clear. As occupational therapists were included as interventionists in so few of the papers, it will be important to study these strategies as part of comprehensive occupational therapy intervention with occupational therapy practitioners providing the interventions. Modification of materials is a key aspect of occupational therapy intervention; therefore, the efficacy of this approach deserves more attention as well. It will also be important to study the differing impact of occupational therapy strategies with modification of materials or tasks versus the occupational therapy practitioner’s therapeutic use of self in modifying interaction with the child. Additional future research should also compare the varied strategies directly to determine which are most effective and for which specific play deficits. Given the play-based approach of Ayres Sensory Integration, further study of this intervention to improve play outcomes is warranted, using appropriate play assessments as outcome measures.
Limitations
As with any systematic review, the process of selecting specific papers and excluding others potentially creates bias in the results. The authors attempted to reduce this bias by working as a team. An additional limitation of any systematic review is the quality of the evidence provided by the papers reviewed. The risk of bias for many of the studies was high. The limitations of the included studies greatly limit our ability to make definitive conclusions.
Summary
Children with ASD present with documented play deficits, and play is a primary concern of occupational therapy practitioners. Play is important not only for development and learning but also for its power to create joy and meaning. To aid the occupational therapy practitioner to engage in evidence-based practice, this review provided information regarding the efficacy of specific intervention strategies for children with ASD that are in accord with the philosophy of play as occupation. The findings support play as an area capable of growth in this population and validate the inclusion of certain specific strategies and methods of intervention.
Supplemental Material
prism – Supplemental material for A Systematic Review of Interventions to Improve the Occupation of Play in Children With Autism
Supplemental material, prism for A Systematic Review of Interventions to Improve the Occupation of Play in Children With Autism by Heather Kuhaneck, Susan L. Spitzer and Stefanie C. Bodison in OTJR: Occupation, Participation and Health
Supplemental Material
PRISMA_Diagram_FINAL_2 – Supplemental material for A Systematic Review of Interventions to Improve the Occupation of Play in Children With Autism
Supplemental material, PRISMA_Diagram_FINAL_2 for A Systematic Review of Interventions to Improve the Occupation of Play in Children With Autism by Heather Kuhaneck, Susan L. Spitzer and Stefanie C. Bodison in OTJR: Occupation, Participation and Health
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
