Use of Work Systems to Increase the Independence of Adolescents With Autism Spectrum Disorder

Participant Recruitment

Once institutional review board approval was obtained for the study, participants were recruited through a parent listserv through a local school district and posted on the Autism Society website for the state the study was conducted in. Interested caregivers contacted the principal investigator via e-mail or phone. Participants had to meet the following inclusion criteria: (a) have an autism diagnosis as indicated by the Autism Diagnostic Observation Schedule (ADOS; Lord et al., 1999), (b) be between the ages of 10 and 19 years old, (c) live at home with a caregiver, (d) identify a household task they would like to learn in collaboration with their caregiver, and (e) provide assent and parental consent. Furthermore, at least one caregiver had to be willing to be present and support the adolescent during the identified task. Six caregivers contacted the principal investigator expressing interest in participating in the study. Two were not able to participate due to schedule constraints (e.g., upcoming travel that would not allow for frequent data collection as required by single case design), and one family identified a skill (showering independently) that would not allow for video-taping/observation, and thus three adolescents with ASD and three of their parents participated in this study.

Participants

Setting

The study was completed in each participant’s home. For Josh and William, the study primarily took place in their kitchen. In some instances, the participant would leave the kitchen to recycle materials. Both kitchens were mid-sized with large countertops for preparing food. For Carrie, the study primarily took place in the laundry room, which was a small room with a washer, dryer, and cupboard with laundry supplies. She folded laundry primarily on a counter in the same room. For all three participants, the only people present were the participant, the participant’s mother, and the researcher. During all observations, external noises were reduced (i.e., the television was turned off and no music was playing).

Materials

Each participant was provided with a work system developed by the researchers (see further description under procedures). Additional materials included all necessary items needed to complete the household task provided by the family. For Josh, materials included food, colander, knife, bags/containers, lunch bag, and cleaning supplies. For Carrie, materials included laundry detergent, fabric softener, dryer sheets, washer, dryer, and folding board. The folding board was a board to assist in folding shirts. For William, materials included food, stove, oven, utensils (e.g., spatula, colander), pots, pans, pie dish, serving dishes, and cleaning supplies. Other materials used in this study included a handheld camera, which was used to record each observation. A research assistant moved the camera into various, unobtrusive locations throughout the activity so all aspects were viewable for coding.

Definition and Measurement of the Dependent Variables

Data were collected on two dependent variables: steps initiated independently and steps completed independently. A step was coded as being initiated independently if the adolescent started the task in the absence of a parent prompt. A step was coded as completed independently if the adolescent completed the step without any parent prompts. Parent prompting was observed and recorded at two points for each step, the initiation of the step and during the completion of the step.

All observations were video recorded. A member of the research team observed each video recorded observation and recorded whether each step was initiated independently, completed independently, and if the parent prompted the adolescent to initiate or complete the step. More specifically, if the adolescent began a task without any parent prompts, the researchers scored a + on the coding sheet. If the parent prompted the adolescent, then a − was scored on the coding sheet. The same procedures were used for steps completed independently. See Figure 1 for an example of a task analysis coding sheet for doing laundry. A percentage of total steps initiated independently was derived by dividing the number of steps initiated independently by the number of total required steps and multiplying by 100. Likewise, a percentage of total steps completed independently was derived by dividing the number of steps completed independently by the number of total required steps and multiplying by 100.

OPEN IN VIEWER

Figure 1.

Task analysis coding sheet for laundry.

Design and Procedures

A single case, multiple baseline across participants design (Kazdin, 2011) was conducted to evaluate the effects of work systems on the independent initiation and completion of tasks in the home setting. To be consistent with the standards set forth by What Works Clearinghouse (Kratochwill et al., 2010), the following were implemented: (a) the independent variable (i.e., work systems) was systematically manipulated by the researchers across the three participants, (b) the outcome variables (i.e., steps initiated independently and steps completed independently) were measured over time and a secondary observer collected data on a minimum of 20% of sessions across all phases and across all participants, (c) the study included three baseline conditions, and (d) a minimum of five data points were collected in all phases. To systematically manipulate the implementation of the work system the following procedures were used. First, we examined the baseline data of all participants to determine which participant had the most/least variable data. The participant with the least variable data started intervention first. Once the first participant entered intervention, we waited until their data were stable in intervention before the next participant started intervention. Then, we examined which participant had the least variable data to determine who should enter the intervention phase next. The last participant entered the intervention phase once data for Participant 2 were stable in intervention.

Intervention

The intervention involved three parts: (a) developing a work system for a parent-selected task for each adolescent, (b) training parents on how to use the work system, and (c) introducing the work system to the adolescents. Parents served as the intervention agents. Each parent selected a task they felt was a priority for their adolescent to learn. Each of the tasks were then confirmed with the adolescent. More specifically, a member of the research team asked each adolescent if they felt it was important for them to know how to complete the identified task and if they were willing and interested in learning how to complete the task. All adolescents agreed to the identified tasks and had been exposed to the tasks by their parents prior to the start of the study. The difficulty level of the tasks varied by participant. For example, Josh’s parent selected making lunches as a priority and William’s parent selected cooking. Cooking involved several appliances and steps and was more difficult. However, all identified tasks were socially important for all parents and adolescents.

Work systems were developed by the second author and a graduate student in special education and introduced to each adolescent by the parent. The work systems were designed to provide prompts for the participants in an effort to reduce the reliance on parents to provide information about the task and expectations. The work systems were also designed to meet the individual’s needs related to reading comprehension, sequencing, and the complexity of the design. See Figures 2 to 4 for work system examples.

OPEN IN VIEWER

Figure 2.

Work system for Josh.

OPEN IN VIEWER

Figure 3.

Work system of Carrie.

OPEN IN VIEWER

Figure 4.

Work system for William.

Josh and Carrie used the same work system every day for packing lunches and doing laundry, respectively. Josh’s work system included three main steps: prepare for lunch packing, pack lunch, and clean up. For Josh’s lunch, he made a peanut butter and jelly sandwich, and cleaned grapes and tomatoes. No appliances were used. He was also expected to clean up his supplies once his lunch was made. If Josh was running low on supplies, the work system prompted him to tell his mother or father. A total of 43 steps were included in his task analysis coding sheet, although not all steps were required. For example, he only recycled materials when the container was empty.

Carrie’s task was doing laundry which involved washing, drying, and folding clothes, towels, and sheets. Carrie’s work system first asked her if the items in the laundry were clothes, sheets, or towels and then prompted her to select the correct dryer card that corresponded with the items in the laundry. If the washing machine was empty, then the work system prompted her to check the laundry basket and again select the appropriate card that corresponded with the items in the dirty laundry. Six different cards were available depending on what she was washing and/or drying. A total of 38 steps were included in her task analysis coding sheet, although not all steps were required. For example, if she was washing towels she did not need to “separate socks.” See Figure 1 for the task analysis coding sheet used for Carrie.

William used four different work systems depending on what recipe he was making. Throughout the intervention, he made apple crisp, hamburgers, pasta, and scrambled eggs. He made each recipe a minimum of 2 times. The recipes for the day were selected based on what ingredients were available and parent/adolescent preference. Each work system included four sections: wash hands, to get, to do, and to clean. He used the oven and stove depending on what he was making. The number of steps ranged from 35 to 63 depending on the recipe. Not all steps were required during each session. For example, when making scrambled eggs if an eggshell did not fall into the bowl, it was not necessary to get the eggshell out of bowl; therefore, that step was not required.

Parents received a 45-min training on the use of the work system prior to the intervention phase. This was a one-on-one training session with the second author. Each parent was taught how to use the work system and how to prompt their child to use the system. Parents were taught how to direct their child’s attention to the prompts provided by the work system with verbal or gestural prompts (e.g., “Look at your work system” or a point to the next step in the work system), rather than providing the directions about what to do and when to do it. This allowed parents to remove themselves from the activity and allowed the adolescent to use the visual cues in the work system to provide answers to the four questions (what to do, how much to do, how do I know when I am finished, and what do I do next). The second author modeled the work system and the parent practiced with the second author. The same training procedure was used across all three parent participants. No mastery criterion was used for parent training. The only material provided for the training was the work system each participant was expected to use during intervention (e.g., the laundry work system was used for Carrie’s parent’s training session). No further instruction was provided to the parents throughout the intervention phase.

The length and frequency of the intervention sessions were dependent on the nature of the task. Laundry and cooking occurred twice per week and lunch packing occurred once per week (Josh packed his lunches for the week on Sundays). Sessions lasted from approximately 15 min (laundry) to 90 min (cooking apple crisp). Laundry sessions varied by session depending on if items were already in the washer and/or dryer. If items were in the washer and dryer, the lesson was longer as more steps were required to transfer items and fold items.

Procedural Integrity and Interobserver Agreement

As a measure of treatment fidelity, throughout each observation the observer recorded (a) whether or not the work system was visually available, (b) whether or not the work system was in proximity to the participant, and (c) whether or not the parent prompted the use of the work system if the participant was not using it (e.g., pointing to the work system). A research assistant collected treatment fidelity data in person by recording information on those three steps.

Interobserver agreement (IOA) data were collected on each participant across all baseline and intervention phases. To obtain IOA, both observers (primary and secondary) watched the recorded session and used a task analysis to record if the participant initiated a step independently and/or completed a step independently. Percentage of IOA was calculated by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100. When IOA fell below 85% the primary coder and secondary coder met and revisited the task analyses and data collection procedures to ensure clarity.

Social Validity

Social validity data were collected from the parents, assessing their perceptions on the work system intervention and its impact on their child’s independent performance on the selected skill. Parents independently completed a pre- and post–social validity questionnaire. The pre–social validity questionnaire was completed in person with the demographic form and the post–social validity questionnaire was completed without the research staff present and the parent mailed the form back to the research staff in a self-addressed stamped envelope. Answer choices were analyzed on a 5-point Likert-type type scale (1 = disagree/never; 5 = agree/often).

Steps Initiated Independently

For all three participants, baseline data indicated initiating steps independently was a challenge. The range of steps initiated independently across participants was 13% to 66%.

The mean percentage of steps initiated independently was substantially higher during intervention when the work system was in place. Across all three participants there was an immediate increase in level from baseline to intervention phases and no overlapping data between baseline and intervention phases.

For Josh, the mean percentage of steps initiated independently increased from 63% in baseline (range = 56%–66%) to 91% in intervention (range = 85%–94%). Data were stable during both baseline and intervention phases and an immediate increase in level was observed from baseline to intervention. During the follow-up session, Josh initiated 90% of the steps independently.

Similarly, Carrie demonstrated increases in steps initiated independently from baseline to intervention. For Carrie, the mean percentage of steps initiated independently increased from 42% in baseline (range = 21%–61%) to 92% in intervention (72%–100%). Although data during baseline did indicate an upward trend prior to intervention, a rapid level change was observed in steps initiated independently after the introduction of the work system. Carrie maintained a high percentage of steps initiated independently during the follow-up session (100%).

For William, the mean percentage of steps initiated independently increased from 26% in baseline (range = 12%–52%) to 91% when the work system was implemented (range = 80%–100%). Data were variable during baseline; however, an immediate increase in level was observed when the work system was implemented. Data were stable during intervention. Furthermore, William maintained a high percentage of steps initiated independently during the follow-up session (98%).

Steps Completed Independently

For all three participants, the mean percentage of steps completed independently was noticeably higher when the work system was being used. Once the work system was introduced, an immediate increase in level was observed for William and Josh. Baseline data indicated completing steps independently was a challenge for Josh and William.

For Josh, the mean percentage of steps completed independently increased from 65% during baseline (range = 62%–70%) to 91% during intervention (range = 83%–97%). An immediate increase in level was demonstrated after the work system was introduced. Furthermore, the percentage of steps completed independently during the follow-up session was consistent with the intervention phase (93%).

For Carrie, the mean percentage of steps completed independently increased from 89% in baseline to 98% during intervention. Although data during baseline did demonstrate a slight upward trend, the data were more variable during the baseline phase ranging from 79% to 97% and were more stable during the intervention phase ranging from 91% to 100%. Furthermore, Carrie maintained a high percentage of steps completed independently during the follow-up session (100%).

For William, the mean percentage of steps completed independently was substantially higher during intervention. The mean percentage of steps completed independently increased from 52% during baseline (range = 28%–83%) to 93% when the work system was being used (range = 83%–100%). During baseline, data were variable; however, a downward trend was demonstrated from the fifth data point onward. An immediate increase in level was observed once the work system was being used. During the follow-up session, William completed 100% of the steps independently.

Procedural Integrity and IOA

Procedural integrity data were collected on 100% of intervention observations. Data indicated during every observation in the intervention phase all criteria were met in entirety (i.e., the work system was visually available, the work system was in proximity to the participant, the parent prompted the use of the work system if the participant was not using it). Therefore, the intervention was implemented with 100% accuracy for each participant.

IOA data were collected on each participant across all baseline and intervention phases. For Josh, IOA was collected on 42% of sessions (40% baseline; 43% intervention). The mean IOA during baseline was 87% (range = 87%–87%) for steps initiated independently and 84% (range = 80%–95%) for steps completed independently. The mean IOA during intervention was 92% (range = 91%–95%) for steps initiated independently and 90% (range = 84%–98%) for steps completed independently. For Carrie, IOA was collected on 40% of sessions (43% baseline; 38% intervention). The mean IOA during baseline was 82% (range = 80%–86%) for steps initiated independently and 86% (range = 84%–88%) for steps completed independently. The mean IOA during intervention was 90% (range = 88%–92%) for steps initiated independently and 92% (range = 88%–95%) for steps completed independently. For William, IOA was collected on 42% of sessions (44% baseline; 40% intervention). The mean IOA during baseline was 83% (range = 78%–90%) for steps initiated independently and 80% (range = 65%–89%) for steps completed independently. The mean IOA during intervention was 88% (range = 83%–93%) for steps initiated independently and 89% (range = 85%–92%) for steps completed independently.

Social Validity

Main Findings

This study addressed the following research question: Does an independent work system implemented by parents increase independent task initiation and completion in home settings for adolescents with ASD? A functional relationship was demonstrated between the implementation of the work system and an increase in steps initiated independently, but not steps completed independently. All three participants demonstrated a low level of initiating tasks independently during baseline. Although five of the data points during Carrie’s baseline were increasing, the data were variable and the last data point began to indicate a downward trend. Josh and William’s data indicated a clear demonstration of effect for steps completed independently, however, Carrie’s data did not. Carrie’s baseline data demonstrated she could independently complete steps prior to intervention. Although the percentage of steps completed independently increased for all three participants from baseline to intervention, three demonstrations of effect were not found; therefore, a functional relationship was not established.

Social validity data collected from parents indicated that the goal of the study was important. More specifically, all parents felt independence was an area their child needed to improve on and specifically noted increasing quality and accuracy when completing the identified task was an important goal. All parents reported the procedures were appropriate and felt the work system could be used to support their child in completing other tasks at home and in school. Finally, parents clearly articulated improved outcomes as a result of the intervention. Taken together, social validity data highlight meaningful outcomes for the parents and adolescents who participated in the study. They demonstrate that independent work systems are an effective intervention for family members to implement in the home and that independent performance can increase quickly in the performance of everyday, household tasks.

Demonstrating these skills independently is key for improving postsecondary outcomes for adolescents on the autism spectrum. Secondary students with ASD who demonstrate independence and/or behavioral autonomy are more likely to be employed and live independently than those who are more dependent on caregivers (Wehmeyer & Palmer, 2003). More than 50% of youth with ASD who graduate high school are not engaged in postsecondary education or employment (Shattuck et al., 2012), thus implementing interventions, such as work systems, to support the independent functioning of individuals with ASD is imperative.

Work Systems

The findings from this study support the notion that works systems are a cost-effective, easy-to-implement visual strategy, which may assist in alleviating some of the challenges around independent living. More specifically, work systems address the characteristics of ASD that make independent functioning challenging, including lack of initiation, prompt dependency, and poor generalization of skills (Hume et al., 2009). There are several reasons why parent-implemented work systems may serve as an effective intervention to support the independent functioning of individuals with ASD. First, work systems are a visual support that utilizes the visual strengths of many individuals with ASD. Second, work systems support the initiation of tasks as the tasks are listed in a sequential order and the structure of the work system helps the individual predict what will come next. Therefore, the work system is a consistent strategy the individual knows they can refer to for guidance (Hume et al., 2009). Furthermore, work systems may encourage the completion of activities as the individuals with ASD may be motivated to complete an activity to have access to the next activity (Bennett et al., 2011). In this study, the participants may have been motivated to complete the task to receive the end product (e.g., eat the prepared meal). Finally, implementing the intervention in the adolescent’s natural environment with natural intervention agents may have promoted maintenance of skills (Bearss et al., 2015), as all participants continued to initiate almost all steps independently during the follow-up phase.

This study adds to the current literature base regarding use of work systems by investigating the impact on task initiation and completion. Previous studies have focused on other dependent variables such as percent of time on-task (Hume & Odom, 2007), percent of intervals a prompt was delivered/graduated guidance was provided (e.g., Bennett et al., 2011; Hume et al., 2012; Hume & Odom, 2007), and percent of steps completed accurately (e.g., Bennett et al., 2011; Hume et al., 2012). Focusing on initiation is timely, because research highlights that initiating becomes more challenging during adolescence for individuals with ASD (Rosenthal et al., 2013). Furthermore, this study investigated the use of work systems with adolescents—an age group in which few studies have been implemented. Although work systems have a solid evidence base, most studies have been conducted with early childhood and elementary aged children. Finally, in this study, the work system was implemented in the home setting, providing support that work systems can be utilized across settings. Previous research has focused on using work systems primarily in school settings.

This study utilized a low-tech modality; however, work systems can be implemented using high-tech modalities (e.g., iPads or mobile devices). Both modalities have utility, and the decisions related to the appropriate form (e.g., words or pictures) and modality (e.g., paper- or iPad-based) rest with the team—including researchers, practitioners, family members, and the individual with ASD. When making the decision to use a low-tech modality in this study, the team considered what was (a) easily available and an efficient use of resources (e.g., cost, time); (b) preferred, understood, and easily used by the individuals with ASD; and (c) could be supported by the parent. Although a high-tech intervention may be the norm in our decade, research indicates that both low-tech and high-tech options can be effective (Bouck et al., 2014) and that students respond differently to high-tech and low-tech interventions (Hill & Flores, 2014). This study further demonstrates the efficacy of low-tech interventions.

Parent-Implemented Interventions

The findings from this study indicate utilizing parents as intervention agents was effective at implementing the work system. This is consistent with a study conducted by Kucharczyk and colleagues (2019) which found parents were effective intervention agents in implementing work systems with a young child with ASD. This expands the literature by investigating parent-implemented interventions with adolescents with ASD. Although parent-implemented interventions have a solid evidence base, most studies have been conducted with early childhood and elementary-aged children (Hong et al., 2016). It further expands the literature by combining visual supports and parent-implemented interventions, two evidence-based practices.

Research has clearly demonstrated parents are capable of learning new strategies and implementing them accurately (e.g., Dunlap et al., 2006), which was further supported by this study. Parents as intervention agents in the home setting create opportunities for generalization to occur in the adolescent’s natural environment, further highlighting the benefits of parent-implemented interventions. This study supports the fact that utilizing parents as intervention agents is an effective means to support individuals with ASD.

Limitations and Future Research

Although the results of this study are promising, several limitations and future areas of research must be addressed. First, a protocol was not given to parents regarding how to provide feedback to their child if they completed a step incorrectly. We encourage future research teams to examine the type of feedback parents provide to their child during the intervention as different forms of feedback may impact the future performance of the participant.

Second, this study employed two interventions: a parent mediated intervention and a work system intervention. Therefore, it is unknown to what extent each intervention contributed to the increased independent functioning of the participants.

Third, there are limitations with the data for Participants 2 and 3. Participant 2, Carrie, demonstrated the ability to complete steps independently during baseline. It could be that Carrie found the tasks reinforcing and therefore was highly motivated to complete the steps. Once she found the tasks reinforcing, it could have promoted initiating the task, which could explain the upward trend of initiations during baseline. Although she was able to complete steps independently during baseline, she inconsistently initiated steps independently during baseline. Initiating steps independently is needed to successfully start and finish a task independently. Carrie’s mother acknowledged the social significance of initiating laundry independently, as she felt she needed to be nearby to get her started on many steps. Although a functional relationship was only established between the implementation of the work system and steps initiated independently, data from Josh and William show promising results for initiation and completion. We encourage future teams to replicate this study to examine if a functional relation can be established between work systems and steps completed independently. For Participant 3, William, IOA fell below 80% during baseline. The recipes William followed were complex, making it challenging for the researchers during baseline sessions to reliably observe each step; however, reliability improved during intervention sessions.

Fourth, there are limitations related to lack of diversity among the adolescents and parents who participated in the study and narrow demographic data. We encourage future research teams to replicate this study and seek participants with diverse characteristics and to gather additional details about parent participants. It would be beneficial to know if parents had previous experience implementing interventions with their child. Furthermore, it would be useful to have a description of independent living skills, including strengths and support needs, of adolescent participants. It would also be beneficial to replicate this study with adolescents with ASD and comorbid ID, as the findings from this study cannot generalize to all individuals with ASD due to small sample size and limited diversity among the participants.

Finally, although this study provides evidence for the support of work systems to aid in the independent functioning of adolescents with ASD, more research in this area is needed. More research is needed to determine the efficacy of work systems in supporting the independent functioning of individuals with ASD across age ranges and environmental settings. More specifically, we encourage future research teams to investigate the efficacy of work systems with adolescents in home, educational, and employment settings. Replication using parents as intervention agents is needed, as well. Furthermore, we encourage future research teams to collect follow-up data. This study included a maintenance phase, but the data were limited as only one data point was collected. It would be beneficial to collect multiple data points several weeks after the intervention phase to better understand the long-term impact of the parent-implemented work system. More specifically, it would be helpful to know if participants continued to use the work system and if so, how levels of independence were maintained.

To close the gap between what individuals with ASD and their families want for their future (Chen et al., 2019; Cribb et al., 2019) and what is actually happening (Shattuck et al., 2012), interventions that support functional independence are needed. The combination of two evidence-based practices—work systems and parent-implemented interventions—show promise to enhance independent initiation and completion of daily living tasks, and thus have potential to contribute to closing that gap.