Abstract
BACKGROUND:
The rate of employment for adults diagnosed with Autism Spectrum Disorder (ASD) is low. This may be due in part to repetitive or disruptive behaviors associated with the ASD diagnosis and challenges to delivering services in the workplace.
OBJECTIVE:
This paper outlines a behavior analytic approach to providing job coaching supports in collaboration with employers in an inclusive setting.
METHODS:
A case example for a 26-year-old male (22 at the time of this project) diagnosed with ASD, employed as a data entry specialist, is highlighted. An A-B design was used to evaluate the impact of behavior analytic interventions on the rate of disruptive vocalizations. Behavior analysts and job coaches from a university-based team worked with employers to identify feasible and effective strategies.
RESULTS:
Rates of vocalizations decreased and maintained for 5 months following intervention fading. Social acceptability ratings from key office personnel reflected improved coworker relations and a high degree of acceptability for procedures used by the team.
CONCLUSIONS:
Employer collaboration was necessary for intervention design and decision making. Service providers should consider response effort for employers, the degree of normalization of strategies, how to limit disruption to productivity and the work setting, and methods for fading interventions or components of interventions.
Introduction
As of 2020, only 17.9%of individuals with disabilities were employed compared to 61.8%of those without a disability (U.S. Bureau of Labor Statistics, 2021). In particular, adults diagnosed with Autism Spectrum Disorder (ASD) are underemployed and have a number of unserved job-related needs (Chen et al., 2015). Perhaps even more concerning is the typical rate of pay for these adults. Most employed individuals with ASD are paid below the federal poverty line (Roux et al., 2016). This is problematic given the many benefits to being employed, including greater independence and self-determination (Lee & Carter, 2012).
Several factors may contribute to low employment rates, including challenging behaviors (Schall, 2010). For example, repetitive and rigid behaviors can hinder productivity, limit precision with job tasks, and are disruptive at work (Wills et al., 2019). Interventions for these behaviors may not be feasible when individuals are completing job-related tasks (Lanovaz & Sladexzek, 2012). Additionally, job sites themselves pose barriers to implementation due to their public nature, a lower tolerance for behaviors that compete with organization missions, and limited resources and personnel with the needed expertise (Schall, 2010). For example, intensive and individualized interventions may require expertise to design, implement, and monitor.
Case studies provide initial evidence in support of applied practices and a better understanding of vocational needs for adults with ASD (Hendricks, 2010). The purpose of this brief report is to demonstrate how behavior analytic supports were delivered for an adult diagnosed with ASD employed in an inclusive setting by the PROMOTES (Providing Realistic Opportunities to Mentor On-Site Training for Employment Skills) Employment Project. PROMOTES was established to extend services provided by a community training center in Southwest Michigan to help students obtain and maintain paid employment.
Case example
The participant and referral
John is a 26-year-old male (22 at the time of this project) who was initially referred to PROMOTES due to deficits in job-related skills. After his referral, he provided written consent to participate in the program. John qualified for special education services through an autism eligibility (diagnosis through medical evaluation and educational impact) since age four. His last evaluation was conducted in 2012, where he met diagnostic criteria in all areas. Services provided by the community training center alone were not enough to secure John a paid job. John’s school obtained an unpaid work experience for him at a local company in need of data entry services. This allowed PROMOTES to provide coaching directly in a workplace setting and build rapport with John’s supervisor until a paid data entry position opened and he was hired as a part-time, paid employee.
Problem identification
After about two months as a paid employee and approximately a week prior to the current project/data set, John’s supervisor reported that he was engaging in a variety of disruptive vocalizations. These were defined as vocalizations that were audible to others but were not contextually appropriate. A variety of topographies were reported, including self-talk, imitating his coworkers and sounds made by the printer or copy machine, clucking, and singing. Given the typically quiet nature of the office, John’s employers were concerned with the disruption the vocalizations caused and expressed uncertainty on how to help John. Given the need to intervene quickly, PROMOTES did not have time to conduct a formal assessment on the function(s) of John’s vocalizations. Interviews were conducted with John’s classroom teacher and primary caregiver to identify if and under what conditions these vocalizations had occurred before. They reported only infrequent and small durations of vocalizations.
Collaborative decision making
Two Board Certified Behavior Analysts (BCBAs) from a Michigan university supervised undergraduate and graduate student job coaches during the implementation of procedures in this case study. These team members worked collaboratively with the employment site. John’s primary job coach (an advanced undergraduate student) and the PROMOTES program coordinator (a supervising BCBA) had frequent and brief touchpoints with John’s primary supervisor for data entry services (hereafter referred to as “supervisor”) throughout intervention implementation and decision making. An A-B design was used to evaluate the effects of a variety of behavior-based interventions on of the rate of vocalizations (Cooper et al., 2007). Ongoing decisions about whether to move or stay within a current intervention were driven by visual analysis of data throughout the project. A “goal line” for vocalizations was established at 0.57 vocalizations per minute as this was the lowest rate recorded during baseline. When John’s vocalizations exceeded this level, intervention changes were made.
Data collection
Data were collected in the natural work environment by undergraduate and graduate student job coaches. John’s desk was adjacent to the staff breakroom and near several employee cubicles. Job coaches were positioned in a vacant cubicle approximately 3.05 m from John’s desk and hidden from his view. This was done to reduce disruption to other office personnel and to minimize the likelihood that the rate of vocalizations was influenced by data collector presence. Initially, ABC (antecedent, behavior, consequence) data were collected by recording what happened immediately before (antecedent) and after (consequence) the vocalizations to identify their function(s) (outcome of the behavior; Cooper et al., 2019). Data were also collected on the frequency of vocalizations that were audible within a 3.05 m radius during 30–60 min observations and converted to rate per minute. The goal was not to eliminate the vocalizations but to achieve a more socially acceptable decibel level. No patterns were discerned throughout analysis based on different topographies (the behavior’s form) or function (Cooper et al., 2019). Thus, all data are presented in aggregate form in Fig. 1.
Rate of Disruptive Vocalizations for John in an Inclusive Workplace Setting.
Scheduled breaks
Scheduled breaks have been used to decrease problem behaviors maintained by escape from instructional or non-preferred activities. Brief breaks occur at consistent intervals (fixed-time) and are not dependent on the occurrence of any specific target behavior (non-contingent; Cooper et al., 2019). For example, Kodak et al. (2003) evaluated the impact of breaks from instructional requests scheduled every 10 sec to increase compliance and to reduce problem behaviors for children diagnosed with ASD. In the current case study, scheduled breaks were embedded in John’s work schedule as an initial intervention step given that scheduled breaks are highly normalized to work settings.
Despite a descending trend during baseline, the rate of vocalizations was still above the goal line, so John’s employers agreed a new intervention should be introduced. An initial schedule with 5 min breaks every 15 min was posted on John’s cubicle to give him direct access to the materials. Based on supervisor input that breaks every 15 min were disruptive and concerns that this would hinder productivity, the schedule was adjusted to 5 min breaks every 25 min. John’s primary job coach explained the schedule, modeled how to read it, and asked open-ended questions to check for John’s comprehension. John was also shown designated spaces where he could engage in the vocalizations without disrupting coworkers during his breaks. After introducing scheduled breaks, there was no change in the rate of vocalizations.
Modified habit reversal training
PROMOTES turned to the literature on repetitive behaviors maintained by automatic reinforcement, including habit reversals. Traditionally, this involves four steps. First, awareness training communicates what the behavior is and when it happens. Next, instruction is provided for an incompatible response during competing response practice. Third, habit control motivation includes identifying the consequences of the behavior. Finally, practice and additional instruction are provided during generalization training in natural settings (Azrin & Nunn, 1973; Spieler & Miltenberger, 2017). A modified habit reversal training package was devised and delivered across approximately 45 min. Since participating in the training prevented John from completing his work tasks, his job coach met with his supervisor to establish a plan for execution.
The first phases of the modified habit reversal training package occurred in a private space away from his coworkers. Phase one was a modified awareness training that lasted approximately 15 min. Training consisted of describing the vocalizations, when they occurred, and how they could impact his relationship with coworkers, productivity, and future at the company. John watched segments of video that captured the vocalizations and were recorded prior to this phase via a small camera attached to a cart behind his desk. Simultaneously, his job coach pointed out instances when the vocalizations did and did not occur. John then watched novel videos and tallied instances of disruptive vocalizations. After John identified all disruptive vocalizations for two consecutive clips, decibel level training was introduced.
During the 15 min decibel level training, the team taught a lower magnitude (decibel level) of vocalizations. A rule was introduced: “I can make noises, but they have to be quiet enough that no one can hear me”. John’s job coach used the phone application “Noise Light 5x” (available on Android devices and created by hdop) to model examples and non-examples of appropriate decibel levels. The application was programmed to display a green smiley face when vocalizations were below a quiet whisper, a yellow face when above a quiet whisper but less than normal conversation, and red frowny face when above normal conversation volume. These levels were set through repeated testing at John’s job site.
Additional instruction and practice occurred at John’s desk during a 15 min generalization training since this was where the vocalizations occurred most frequently. John’s job coach stated the rule (“you can make the noises, but they have to be quiet enough that no one can hear you”), and John restated the rule once. John practiced speaking below a quiet whisper with and without visual feedback from the Noise Light 5X application because, outside of training, John’s supervisor did not want John’s phone out during work. Following implementation of the training package, the rate of vocalizations immediately decreased to below the decibel goal line but only maintained for four sessions. The team reintroduced only the rule rehearsal given the time intensity of the entire package. John vocally rehearsed the rule introduced during training a single time as a low response effort tactic.
Self-monitoring form with an incentive system
PROMOTES identified self-monitoring as the next intervention step. Self-monitoring has been recently applied to problematic behaviors in the workplace (Wills et al., 2019). One promising self-management tool is the “Self and Match”, which involves a self-score by the individual on the targeted behavior(s) and a matching score from another individual (Croce, 2015). This system has been used to decrease vocal stereotypy for a child diagnosed with ASD (Bulla & Frieder, 2017). A similar self-monitoring system was developed by PROMOTES. A modified form was posted in John’s cubicle and included a five-point scale for rating the disruptiveness of his vocalizations from 1 = “none” to 4 = “very high”. The scale was color-coded according to the Noise Light 5x application (i.e., green = “none” and red = “very high”). The form contained a space for John and an employer to rate his vocalizations each workday, and for John to circle “yes” or “no” if he achieved his weekly goal.
An employee near John’s cubicle provided a “match” score for John each workday (Monday through Thursday) given their ability to consistently monitor his vocalizations at work. PROMOTES facilitated the use of the incentive system given the time and effort required to run it. John earned points based on the correspondence of his score compared to his coworker’s and the overall rate of vocalizations. John earned zero points for matching scores in the green zone, one point for matching scores in the yellow or red zone, or two points for non-matching scores. Goals were established by PROMOTES each week based on a total number of points below which John had to score. At the end of each week, John reviewed his total points and scored if he met his goal with his job coach. If John met his goal, points were calculated and could be turned in for prizes. Most often, selected prizes were gift cards based on John’s reported preference.
Following implementation of the self-monitoring form and incentive system, vocalizations immediately decreased but maintained for only one session. This may have been due to the delay between implementing the system and John’s first opportunity to earn a prize. John had to use the form and count his points for one week. It is possible that John needed to practice rating his performance, receive a “matching” score, and contact earning points or prizes multiple times before his performance would sustain below the decibel goal line. Therefore, a new intervention was not introduced until vocalizations exceeded the goal line a second time.
Scheduled rule rehearsals
Like the rule rehearsals used in the modified awareness training, a schedule for rule rehearsals was embedded into John’s work schedule at the beginning and end of his week (on Mondays and Thursdays) based on ongoing data trends. Rules are beneficial when a behavior (like John’s vocalizations) could result in significant, but delayed or improbable, outcomes (Cooper et al., 2019). Rules can also lead to faster changes in behavior than when a person directly contacts those consequences (Skinner, 1974). For example, John’s vocalizations had the potential to negatively influence his relationship with others in the workplace or lead to termination. Rules were one way to provide frequent verbal reminders about workplace expectations without John directly contacting the natural consequences of his vocalizations (e.g., termination) or others programmed by the PROMOTES team. Rule rehearsals were combined with the self-monitoring form and incentive system because of concerns about the amount of time it could take for John’s performance to stabilize. Rule rehearsals were also minimally invasive and required little extra time by John or resources beyond what was already in place. Before shifts, John stated the rule (“I can make the noises, but they have to be quiet enough that no one can hear me”) a single time with his job coach. The job coach asked John if he had any questions, provided praise for rehearsing the rule, and reminded him of his goal.
Intervention fading
When John’s vocalizations remained consistently below the goal line, John’s supervisor and primary job coach collaborated on how to fade the interventions through frequent in-person check-ins. Decisions were based on John’s ongoing performance. First, the rule rehearsal schedule was faded from two to one day per week. After, the weekly rule rehearsals and self-monitoring system were removed entirely. Rates maintained across five months during monthly probes collected by John’s job coach. John was told that these data would be collected but was not aware of when. The incentive system was maintained to support other vocational goals and skill-building outside the scope of this case study.
Summary of results
The results of this project are displayed in Fig. 1 with sessions on the x-axis and rate of vocalizations per minute depicted on the y-axis. Solid phase lines represent when new interventions were introduced and remained in-place across a phase or several phases. Dashed phase lines indicate when a training or treatment package component was introduced a single time. Except for scheduled breaks, all interventions resulted in an immediate decrease in disruptive vocalizations. However, these low rates did not maintain following implementation of the modified habit reversal training or with the self-management form alone. With the addition of a rule rehearsal schedule, vocalizations remained below the goal line and maintained for 5 months following intervention fading.
Student outcomes
Initially, John’s rate of disruptive vocalizations reached almost 2.75 per minute, or 165 vocalizations per hour. Though neither PROMOTES nor John’s employer wanted him to lose his job, this was clearly an issue that had to be addressed. Through collaboration with John’s employers and the implementation of several behavior analytic strategies, vocalizations dropped to near-zero rates and maintained for five months. John has successfully maintained employment as a data entry specialist for over three years. He has earned $10 per hour while working 12 hr per week during the peak season for the organizations (i.e., October –May) and four during summer (i.e., June –September).
One major outcome was the successful fading of interventions with continued low rates of disruptive vocalizations. Despite the initial intensity of on-site supports, many long-term benefits were achieved. By reducing disruptive vocalizations, John has built a variety of other skills that make him an essential member of the team. Before this case example, several other employers had to assist John with data entry tasks. Data entry is critical to this organization for sustaining smooth day-to-day operations. John is now responsible for all data entry tasks for the entire organization and does so on a completely independent basis. His supervisor and coworkers report that he is faster and more accurate than they would be in the same position.
Social acceptability
Surveys were distributed to three personnel from John’s office: John’s primary supervisor and two coworkers who sat close to John’s desk. The surveys were administered by the job coach during intervention fading using a five-point Likert scale (i.e., 1 = highly disagree to 5 = highly agree) and completed anonymously within two days after distribution. Completed surveys were left for John’s job coach to pick up in a manila envelope positioned at the front entry reception desk. The results are summarized in Table 1. Overall, results indicated that vocalizations decreased following intervention and caused less disruption to employees.
A Summary of Social Acceptability Ratings by Work Personnel Before and After Intervention
A Summary of Social Acceptability Ratings by Work Personnel Before and After Intervention
Achieving a better understanding of how to address challenging behaviors in inclusive employment settings is important given the numerous negative consequences that can ensue. The outcomes of this project point to several recommendations for navigating this difficult topic. One recommendation is to make small but meaningful changes to the work environment using strategies that have a high degree of normalization to employment settings. For example, scheduled breaks are common in work settings and involve small adjustments that are relatively easy to make. However, frequently leaving one’s desk can impact the flow of the organization or productivity. Weighing the cost and benefits of procedures is therefore necessary.
More intensive training techniques, such as the modified habit reversal training, may be necessary when small adjustments to the work environment do not address performance concerns. Several factors that could diminish the impact of these strategies should be planned for, like the amount of time they take the individual away from job tasks. Too much time away from John’s desk could have impacted the quality of his data entry performance or the amount of work completed during a shift. This is certainly not ideal from an employer’s perspective.
Self-monitoring systems with a match component have the potential to be both effective and acceptable in workplace settings. Having a “match” creates a programmed opportunity for employers to regularly deliver feedback. However, employer buy-in is critical and special consideration should be made when determining how the system is used. For example, employer “matches” every minute or even every hour require a high response effort and could compete with other job requirements. This could do more harm than good by leading to low treatment integrity or diminished rapport between the client, job coaching team, and employers. Additionally, regular feedback may not have a high degree of normalization to certain organizations, which could make using the system burdensome.
Dividing responsibilities for implementing the self-monitoring form allowed PROMOTES and John’s supervisor to capitalize on each team’s strengths and their available resources. This was particularly important given that any employer involvement in the system added more work for them. Systematically setting goals and incentive system implementation were led by PROMOTES due to limited employer expertise in this area. John’s job coach was well-positioned to deliver feedback on whether he did or did not meet his goal and ensured everyone was comfortable with the system.
Conclusion
Practitioners should employ strategies that are likely to be effective while prioritizing the goals and preferences of clients (Wilczynski et al., 2013) and incorporating employer input. For example, modifications were made to the self-monitoring form when John reported that recording data on vocalizations throughout his shift would be distracting. Additionally, though leaving the Noise Light 5X application at John’s desk during work could have been helpful for delivering continuous visual feedback, his supervisor was concerned about the disruption this could have caused to others and John’s productivity. When possible, modifications should be made to ensure interventions are both impactful and other factors related to an individual’s work experience are not negatively impacted. Additionally, programming for long-term maintenance of intervention outcomes and data collection can allow service providers to reallocate resources to other target behaviors or clients when interventions can be faded.
Footnotes
Acknowledgments
Special thanks to the Michigan Department of Health and Human Services (MDHHS) who provided grant funding in 2015 to the Department of Psychology at Western Michigan University (WMU) to help adults with ASD obtain and maintain paid employment in the state of Michigan. Additional thanks to Rachel Popp and other members of the PROMOTES team including John, the employment site, and the school district for their support of this project. This would not have been possible without their time, insight, and dedication.
Conflict of interest
None to report.