A 65-year-old, Caucasian male diagnosed with severe intellectual disability and bilateral deaf-blindness was included. Due to the level of cognitive impact, he was legally conserved by a family member who provided written consent. The participant provided assent prior to and throughout each session by remaining seated, standing in the immediate area, and/or remaining engaged with the researcher in the absence of maladaptive behavior. If the participant engaged in maladaptive or precursor behavior (e.g., clapping hands together firmly), sessions were terminated by the researcher prompting the “all done” sign by tapping on the participant’s elbows. The participant and his conservator did not receive compensation for participation, and the researcher informed the participant’s conservator that participation could be terminated at any time. The researcher conducted the study within the adult residential facility (ARF) where the participant resided with 24-h care from a licensed vocational nurse (LVN) and direct support professionals (DSPs). All aspects of the study were approved by the university’s institutional review board.

The participant was reported to work as an electronic parts handler from 1989–2005 in a vocational group in which he communicated using unspecified tactile prompts and adapted American Sign Language (ASL; signing into the palm of his hand). The participant resided in a developmental center from 2005–2017 prior to moving into his current residence. No additional documentation or skills assessments were made available to the researcher.

At the time of the study, the participant independently used the restroom, manipulated the bidet, dressed himself, fed himself, and navigated his home with tactile cues (i.e., rectangular/circular wooden plaques) mounted to the walls. The participant was reported to meet mastery criterion (i.e., 80% accuracy on average for 3 months) for the adapted ASL sign “more” but all other ASL targets (i.e., adaptive signs for “choices,” “all done,” and “eat”) required prompting from others. Per staff report, the participant engaged in self-injurious behavior (SIB; biting his hand) on average three times per month when he required assistance or when basic needs were not met (e.g., hunger, thirst, etc.).

The researcher selected a SadoTech Elderly Monitoring Pager (referred to as “device” hereafter) supplied with a monitoring pager worn on a lanyard and a luminating receiver for the participant to request help (see Appendix A). When the monitoring pager’s button was pressed, the receiver illuminated and emitted an auditory signal. Additionally, the researcher carried over a Sensory Slap Fidget Bracelet (see Appendix A) from the participant’s regular programming to indicate that the researcher was present. Prior to beginning each session, the researcher placed the bracelet on her right wrist, approached the participant, and placed her hand near the participant’s hand so that they could touch the bracelet.

The dependent variable was the percentage of independent correct responding for device usage. Correct responding was defined as independently reaching for and manipulating the device with one or both hands and pressing the button using one or more fingers when help was needed. The researcher used touch cues (i.e., tapping the participant’s elbow twice and waiting 1 s), partial-physical (i.e., wrapping palm around elbow and providing slight pressure while pushing forward toward the device), and full-physical (guiding with hand-over-hand assistance to use the device) prompts to teach the response. Incorrect responses on EO trials (i.e., when help was needed) were defined as holding the device without pressing the button or engaging in unrelated or maladaptive behavior; an incorrect response on AO trials (i.e., when help was not needed) was using the device.

Prior to beginning each session, the researcher placed the tactile bracelet on her right wrist, approached the participant, and placed her hand near the participant’s hand. If the participant did not move away or engage in maladaptive behavior, the researcher began the session. Sessions included both EO (i.e., help was needed to complete the routine) and AO (i.e., help was not needed to complete the routine) trials programmed during previously mastered behavior chains (i.e., toileting, meals, and dressing; see Table 1 and Online Supporting Information for details regarding manipulated chains) at times when they were normally completed (e.g., when the participant needed to use the bidet, after dinner was prepared, and before bed). The researcher remained the same distance from the participant on all trials across phases.

On EO trials, the researcher arranged for materials for one step in the chain to be broken, inoperable, or missing. For example, the bidet was made inoperable by switching it to the off position, the utensils were missing when a meal was served, and shirts were missing from the dresser drawer.

On AO trials, the researcher did not arrange for any materials to be broken, inoperable, or missing.

On EO trials in baseline, the researcher waited 15 s once the participant reached the contrived step. If the participant used the device independently, reinforcement in the form of help (i.e., providing the missing item or fixing the broken object) was provided. Following 15 s without the participant using the device, the researcher ended the trial and restored the environment (e.g., switched bidet to on position, provided utensils, placed shirts in dresser). During AO trials, the researcher used the least intrusive prompt for the participant to continue to the next step of the routine if they used the device. If the participant did not use the device during AO trials, the participant could complete the routine without interruption.

In the intervention phase, the participant completed trials for a maximum of 10 min up to three times per day (e.g., missing shirts during dressing routine was one trial which could have lasted one full 10 min period due to the duration of the dressing routine itself). The researcher began by providing the participant with errorless prompting in which the researcher contrived the same three situations (EO trials) from baseline and used most-to-least prompting. Errorless prompting was discontinued after completing trials for a total of 20 min. After 20 min, we advanced to a progressive-prompt delay (i.e., 1 s, 2 s, and 3 s delay) wherein the delay was lengthened following 10 trials without prompt resistance. Across intervention, the participant was exposed to each scenario simultaneously while rotating AO/EO trials. The researcher alternated AO trials in which identical scenarios were presented without researcher manipulation (with the researcher the same distance away). Once the participant reached 80% correct responding across all scenarios, we began the treatment extension.

After reaching the mastery criterion, the researcher manipulated a different step in each of the three routines (see Table 2 and Online Supporting Information for detailed scenario information) and alternated between EO and AO trials. The researcher used least-to-most prompting if the participant did not independently use the device.