Impact of a Contextual Intervention on Child Participation and Parent Competence Among Children With Autism Spectrum Disorders: A Pretest–Posttest Repeated-Measures Design

Research Design

We used a one-group repeated-measures pretest–posttest research design to investigate whether contextually relevant interventions led to (1) increased child participation and (2) increased parental competence and decreased perceived stress. We obtained approval from the Human Subjects Committee at the University of Kansas Medical Center. Parents provided informed consent; children ≥ age 7 provided assent.

Participants

We recruited families who have children with ASD (based on parent reports) ages 3–10 through schools and support groups in a Midwestern suburban area. We selected families on a first-come, first-served basis when parents indicated that they had unmet needs in their family life and when children had at least one sensory pattern outside typical (Sensory Profile [SP]; Dunn, 1999). We recorded other family services but did not control for them. We excluded children with typical sensory patterns.

Measures

Intervention

Two occupational therapists provided 10 intervention sessions per family; sessions lasted approximately 1 hr (total = 10 hr) over 12–15 wk. The intervention reflected principles of context therapy (Darrah et al., 2011) and contained three intervention characteristics: activity settings, daily life routines, and sensory processing patterns. Figure 1 illustrates the relationships among intervention characteristics and outcomes (successful child participation, increased parental competence).

OPEN IN VIEWER

Figure 1.

Illustration of contextual intervention and outcomes.

Parents identified goals and selected activity settings in which they needed support on the basis of their priorities and interests (e.g., home, park, grocery store). Activity settings distinguish this intervention from traditional therapy, which might occur in clinic interactions with therapists. Evidence has indicated that children get more practice and develop and generalize skills when interventions are embedded within routines (e.g., getting dressed, playing, sitting in grocery cart; Dempsey & Dunst, 2004; Law, Garrett, & Nye, 2004).

We used SP data, expertise about sensory processing, and task analysis to consider the impact of each child’s sensory patterns on routines and settings. In conversations with parents, we linked sensory processing principles to their family’s routines and settings, so that parents learned how their children’s sensory processing patterns might affect participation. One parent who had successfully used visual schedules with an older child made one for her younger child’s morning routine but found that it did not help. The therapist coached, “Let’s look at your daughter’s sensory patterns and see if we can see why the visual schedule didn’t work.” Through coaching, the mother realized her daughter’s auditory strengths and decided to try a kitchen timer instead, because that approach took advantage of the daughter’s strength in auditory processing. The timer prompted her daughter to get out of the shower and move through her routine.

During the first intervention session, therapists met with parents to review their priorities from the COPM, GAS, and activity configurations (i.e., outline of family’s daily schedule) to provide the settings and routines components of the contextual interventions. Using SP findings, therapists coached parents to design intervention plans for Week 1. At each subsequent visit, they discussed what had happened since the last visit. When closing each session, parent and therapist made a joint plan detailing the family’s strategies; the plan reflected how sensory knowledge informed the strategies in authentic settings and life routines for that week.

We followed coaching principles outlined by Rush and Shelden (2011) and consistent with context therapy (Darrah et al., 2011) to guide intervention planning. Therapists used reflective statements and questions to invite parents to discuss possible ways to achieve those goals. Therapists avoided yes–no questions; rather, they fostered insights. For example, they would ask, “How does he wash his face now?” (awareness); “What other times does he get things on his face; how does he act during those other times?” (analysis); “I wonder how we could use successful past strategies to change his routine?” (alternatives); and “What supports do you need to try that?” (action). Therapists did not offer expert advice or directives. Instead, they continued to ask questions to uncover what the child’s participation would look like (e.g., “If his performance was a little better, what would he be doing instead?”) and make comments until parents identified specific characteristics that would represent progress or goal achievement for them.

Maintaining Fidelity of Contextual Intervention

Two occupational therapists provided interventions to 10 families each. Therapists attended statewide training in coaching, which involved 4 days of training with 6 mo of follow-along coaching (Rush & Shelden, 2011). As director of a statewide evidence-based practice project, the senior researcher, Winnie Dunn, collaborated to develop materials used in the statewide training. After intensive training with simulated practice, therapists implemented coaching strategies with actual families and obtained feedback from experienced coaches. They kept coaching logs and analyzed them for reflective questioning strategies. This process continued until logs contained reflective questions and comments and did not contain directives. The intervention was manualized for future publication.

The research team met weekly to review research procedures and provide coaching to therapists; recordings from visits served as references for these meetings. The therapists also met weekly to ensure they were providing consistent intervention.

Data Collection

During testing, parents and therapists completed the OMB. The same therapist provided testing and interventions to families. Intervention therapists were blind to parents’ responses on the PSI–SF and PSOC; parents completed them, and other team members entered and scored their forms. They were not blind to parents’ results on the GAS and COPM, which they needed to guide interventions.

Figure 2 details data collection timelines. During TEST 1 (the initial visit after consent), therapists reviewed the study and completed the OMB and an activity configuration about daily routines of interest to the family. During baseline (4 wk), families continued usual routines and services. They completed the OMB at TEST 2, and therapists scheduled intervention sessions.

OPEN IN VIEWER

Figure 2.

Timeline for data collection.

During the intervention period, therapists met with parents (face to face or by phone) according to their needs and at their convenience (10 sessions). During these sessions, therapists coached families to support their children using three intervention characteristics. TEST 3 occurred after 10 intervention sessions. During sustainability (4 wk), therapists and parents did not meet; families continued routines and services from other sources. Finally, during TEST 4 therapists and parents completed the OMB again.

Data Analysis

We used a within-subject (repeated-measures) analysis of variance with post hoc testing to determine effects for time (Figure 1). We made four comparisons:

With four pairwise comparisons, Holm’s sequential Bonferroni procedure controlled for familywise error rates (Green & Salkind, 2011). With this correction, we set p < .0125 as our significance level. We used polynomial contrast calculations to identify the source of significant differences when there was overall significance.

Sample Characteristics

Participants were parents of 17 boys and 3 girls, ages 3–10 yr (mean = 6.5 yr). Twelve children had autism, 1 had Asperger syndrome, and 7 had ASD and another diagnosis (e.g., developmental delay, attention deficit hyperactivity disorder, depression). Physicians diagnosed 11 participants; a team diagnosed the other 9 participants. Ten children took supplements (e.g., melatonin, homeopathics) or medication (e.g., citalopram).

Nineteen mothers and 1 father served as participants. Two families had one child; others had 2–6 children. Eighteen families were suburban; 2 families were urban. One family was Latino, 17 were White, and 2 reported more than one ethnicity. Families reported household income levels (1, <$30,000; 5, $30,000–$75,000; 14, >$75,000). All parents had some college education. All children received other services.

Sensory Profile Patterns

For 10 children, all four quadrant scores were 1–2 standard deviations greater than those of their peers without autism. Nine children had three or four quadrants in the more-than-others categories. One child had only Registration in the more-than-others category.

Intervention Characteristics

The contextual intervention characteristics (active ingredients) were activity settings and routines and the child’s sensory patterns. Parents selected activity settings—home (74%; e.g., bedroom, bathroom, transitions) and community (22%; e.g., church, parking lots, stores)—and routines—self-care (49%; e.g., dressing, eating, sleeping, hygiene) and leisure (37%; e.g., playing, watching TV).

The third characteristic was the child’s sensory patterns. When creating joint plans, therapists referenced sensory patterns: seeking (21%), avoiding (1%), sensitivity (10%), registration (13%), and two or more patterns (55%). They used sensory system information: auditory (13%), visual (26%), touch (20%), movement (17%), body position (13%), and other (e.g., oral; 11%).

Children’s Participation

For the COPM, results indicated a significant time effect for Performance (Wilks’s Λ = .137, F[4, 16] = 27.408, p < .001, η² = .863) and Satisfaction (Wilks’s Λ = .181, F[4, 16] = 19.546, p < .001, η² = .819). Polynomial contrasts for COPM indicated a significant linear effect for Performance scores (F[1, 19] = 90.907, p < .001, η² = .858). Comparisons 2 and 4 were significant (both ps < .001). Ratings changed from 3.6 to 7.0 (10-point scale). Polynomial contrasts for COPM Satisfaction indicated a significant linear effect (F[1, 19] = 66.502, p < .001, η² = .816). Comparisons 2 and 4 (both ps < .001) and Comparison 1 (p = .011) were significant. Ratings changed from 3.2 to 7.0 (10-point scale). Figure 3 illustrates changes across time periods.

OPEN IN VIEWER

Figure 3.

Mean scores for child outcomes across testing periods (n = 20 children, 44 goals).

For GAS, results indicated a significant time effect as well (Wilks’s Λ = .070, F[4, 16] = 66.328, p < .001, η² = .930). Polynomial contrasts indicated a significant linear effect (F[1, 19] = 215.963, p < .001, η² = .927). Comparisons 2 and 4 (both ps < .001) were significant. The average change for goals was 2 points (4-point scale; Figure 3).

Parental Competence

For the PSI–SF, results indicated a significant time effect (Wilks’s Λ = .436, F[4, 16] = 6.037, p < .007, η² = .564). Polynomial contrasts for subtests indicated a significant linear effect for defensive responding (7 items; F[1, 19] = 17.049, p = .001, η² = .516) and parental distress (12 items; F[1, 19] = 13.031, p = .002, η² = .449). Comparison 4 was significant (ps = .001). For defensive responding, parents began the study at the 96th percentile and ended the study at the 70th percentile on the basis of PSI–SF scoring criteria. For parental distress, parents went from the 85th percentile to the 50th percentile.

PSOC results indicated that parents experienced a significant improvement in efficacy. The analysis of variance indicated a significant time effect (Wilks’s Λ = .335, F[4, 16] = 10.065, p = .001, η² = .665). Polynomial contrasts indicated a significant linear effect (F[1, 19] = 22.078, p < .000, η² = .580). Only Comparison 4 was significant (p < .001), with parents increasing their sense of efficacy from the first visit to the last. Compared with a normative Australian sample (Gilmore & Cuskelly, 2008), parental efficacy levels in this study were 2 standard deviations higher than average. Parental satisfaction did not change in our study; however, when compared with Gilmore and Cuskelly’s (2008) data, parental satisfaction in our study was about average.

Intervention Characteristics

Because our intervention focused on families in natural environments, it is not surprising that the bulk of activity settings were home (74%) and community (22%). This study illustrated that although children received special services through school, families continued to have significant needs for support in their daily lives. In general, therapy services for self-care are not the scope of practice for school personnel. Self-care represented nearly half of goals, and playing represented 24% of goals. Although all families focus on self-care and playing at home and in the community, the increased intensity of managing daily life with children who are vulnerable (e.g., those with ASD) may require more attention. Other researchers have reported that family life revolves around ASD and that families have only fleeting moments of feeling like a family (DeGrace, 2004). Occupational therapists might consider how to reform services for school-age children to include family needs.

Parents talked about intervention characteristics as they told us how the joint plans had worked the previous week. In early team meetings, we discussed what we thought parents could do to achieve their goals. Because our intervention was not an expert approach, therapists did not give parents these ideas. We asked ourselves how to support parents to identify strategies in their own lives and gain insight into their situation. We presumed that parents had the resources to achieve their goals (strengths approach) and that we were responsible for providing reflective feedback and questions to reveal their capacity and resources. Using this reflective support, parents identified and implemented strategies that we could not have imagined. For example, one parent wanted her son to behave better after school. In coaching sessions with each other, we thought he needed a snack. However, in subsequent weeks, this mother realized through reflective discussions and joint plans that giving her son attention for a few minutes after school fostered his playing alone until dinner. Had we given our expert advice, we would have misguided this family.

Sometimes it was helpful to understand why plans did not work. One mother used music for morning routines because her son loved music, but it did not motivate him to get out of bed. In joint planning, the mother wondered whether he was hungry; he was also motivated by sweet flavors, so she decided to provide sweet breakfast foods.

Children’s Participation

Study findings indicate significant improvements in children’s performance (COPM and GAS). Parents exhibited positive perceptions of their children’s participation through the sustainability period, suggesting that families discovered successful methods for managing their daily lives.

Families set 44 goals (mean = 2.4 goals, range = 1–5 goals). When they felt satisfied, parents identified new goals based on GAS. Sometimes parents discontinued goals because they realized during coaching what the basis of their child’s behavior was or determined that the behavior was appropriate. For example, one mother decided to address playing and realized that traditional playing routines were not enjoyable for her son. She discovered that he preferred helping to fold laundry. By redefining what was satisfying, she recognized his strengths and interests.

Parents asked for help in many areas, including dressing, eating, getting shots, riding in cars, and transitioning from bus to home. Therapists listened carefully to parents’ explanations about goal areas and guided discussions so that together they specified levels of behavior for GAS. Therapists asked questions such as, “What would it look like if she cared for her own hair properly?” Although this process took time, asking reflective questions to specify measurement levels on GAS invited parents to think about what aspects of routines were going well and which needed adjustments.

For children’s outcomes, Comparison 2 revealed changes from the intervention period, and Comparison 4 indicated overall changes from the beginning to the end of the study; outcomes at both time points improved significantly. No changes occurred during waiting periods, confirming that children’s participation did not change during baseline (Comparison 1) and remained high after the study (Comparison 3). Parents said they benefited from dedicated time to talk about their family and being accountable to someone for trying things each week.

There was one exception: The COPM Satisfaction score improved during Comparison 1. Perhaps during baseline, parents became more aware of satisfaction factors regarding their child’s participation and so had a different perspective when intervention began because they had thought about it for 4 wk. During the second visit, parents and therapists may have been more comfortable with their new relationships. Parents expected to say what was wrong with their children; because our approach emphasized strengths, TEST 1 may have reflected dissatisfaction, thereby dampening initial scores.

Parental Competence

On the PSI–SF, parents reported significantly lower distress and less defensive responding (e.g., “I feel trapped by my responsibilities as a parent”). Additionally, parental distress was at the 50th percentile at study’s end. Considering that these parents are handling challenging daily lives with children who have ASD, this finding is notable. As parents continued to reflect on their lives and identify strategies to support their family, perhaps they began to understand their own capacity to find solutions for situations that were overwhelming before intervention. These findings are consistent with those of other studies of parent empowerment (e.g., Nachshen & Minnes, 2005).

Parents reported significant improvements in parental efficacy. According to Rogers and Matthews (2004), the PSOC Efficacy scale assesses capability and problem-solving ability (e.g., “If anyone can find the answer to what is troubling my child, I am the one”). Efficacy may be inversely related to distress, because distress reduced and efficacy increased significantly. Parents demonstrated efficacy by reporting how the joint plan worked, planning with the therapist, and offering their own suggestions about how to make adjustments to achieve goals. Sometimes, parents made adjustments to joint plans without direct therapist support. Parents also demonstrated efficacy when they had “aha” moments during coaching conversations; for example, 1 parent said she realized that going to his brother’s soccer game was not fun for her son and that they were all more satisfied when she went to the game and got him a babysitter. Before the study, she thought that she was neglecting him by leaving him at home.

Efficacy levels were also notably stronger than those of 98% of typical parents (Gilmore & Cuskelly, 2008). Perhaps parents who must pay vigilant attention develop a stronger sense of themselves as parents. This example may be one of an “uber” strength, a characteristic of the person that has required his or her attention and so becomes both a unique feature of the person’s identity and a strength because of the attention the person has given to its development (Dunn, 2010).

Efficacy levels were significantly improved only for Comparison 4 (i.e., from the beginning to the end of the study). Perhaps parents need time to feel less distressed and more confident. These data suggest that parents need time to process their child’s progress, contextual factors, and their own reactions to their child’s behavior. Sometimes professionals can feel rushed with families because of caseloads and other responsibilities; this intervention takes time to implement and achieve results. Perhaps parents’ experience of trying strategies with their children and achieving goals they set allows them to build their competence.

Although parents’ satisfaction ratings did not change (PSOC), they maintained high satisfaction as parents (70%). The Satisfaction scale reflects parental emotions (Rogers & Matthews, 2004); parents’ scores indicated that they have relatively low levels of frustration and anxiety about parenting. Even with the intense life they experience with their children, these parents indicated that they were satisfied with being a parent.

Parents made plans every week; not all strategies were effective. Therapists completed an intervention documentation form at each session that included a question about how effective parents felt with joint plans. Across all plans, parents said joint plans worked well 26% of the time, were “okay” 23% of the time, and were not effective 51% of the time. Because efficacy improved and distress diminished, perhaps the iterative problem-solving process led to changes for parents. An expert approach might bypass the growth that is possible when a parent has to consider how to meet a goal.

Study Limitations

Because families volunteered, we might have had families who were more motivated to follow through with plans. Therapists collected test data with families they served, which may have affected results because they had developed a relationship and were invested in positive outcomes. However, comments such as those reported here suggest that parental insights had a strong part to play in the outcomes. In future studies, we will need to specify the details of the intervention process using transcripts from sessions and from coaching meetings; until then, a provider may not have enough information to implement these interventions and achieve similar outcomes.

Future Research

We need to establish clear fidelity parameters so we can replicate interventions and findings with other providers. We need to conduct additional studies to build evidence for contextual interventions; even this small sample yielded good power and effect size estimates, suggesting that this approach has promise. Parents and therapists both felt meeting in person at first was helpful, so identifying the most salient features of the intervention process so we can train experienced occupational therapists and faculty will be critical to moving this evidence into practice.