Expanding Interventions for Automatically Maintained SIB

Individuals with intellectual and developmental disabilities (IDD) such as autism spectrum disorder (ASD) often engage in restricted and repetitive behaviors (RRBs). In fact, RRB is a diagnostic hallmark of ASD. For some, these behaviors cause, or have the potential to cause, injury to the individual exhibiting the behavior. Research has demonstrated that operant contingencies, and particularly automatic reinforcement, play a role in the maintenance of such behavior. Automatic reinforcement means that the behavior directly produces the outcomes that maintain it, without the requirement of a social interaction. For example, scratching a mosquito bite is a behavior maintained by the automatic negative reinforcer of attenuating an aversive sensation, while eating a piece of candy is maintained by the automatic positive reinforcer of sweet taste. Epidemiologic studies indicate that between 15% and 30% of cases of self-injurious behavior (SIB) are maintained by automatic reinforcement. This type of SIB has been noted to be particularly difficult to treat. However, a relatively recent review of the literature reported that reinforcement-based interventions such as non-contingent reinforcement (NCR), which involves response-independent reinforcer delivery, and combinations of differential reinforcement (DR) with NCR can be highly effective interventions for this type of behavior. Further, others provided preliminary evidence that response patterns during pre-treatment assessments could be used to identify cases for whom NCR would be effective, versus cases for which DR would be effective.

While NCR and DR have demonstrated efficacy, the literature lacks critical evaluation of successful application of these interventions to reduce automatically maintained SIB along at least two dimensions. 1) No studies demonstrate longevity of treatment effects. In most studies, demonstrations of efficacy have occurred during brief (5-15 min) sessions conducted one to several times per day. Interventions may be successful during those brief time periods, but it remains unknown whether intervention effects would continue if implemented over longer periods (e.g., 60 min). 2) Across existing studies, dense reinforcement schedules are used and little is known about whether intervention effects would continue as reinforcement schedules are thinned to levels practical in natural settings. The investigators propose a project to address these gaps in the literature. Our proposed research plan builds off prior research investigating differential effects of matched (stimuli that produce effect similar to the presumed reinforcing effect of SIB) and unmatched (produce an effect dissimilar to that presumed to result from SIB) stimuli used as reinforcers in NCR and other DR-based interventions. The literature on intervention for automatically maintained SIB demonstrates that both matched and unmatched stimuli may reduce SIB in NCR and DR-based interventions. The investigators propose to evaluate relative effects of matched and unmatched stimuli (a) over time and (b) during schedule thinning. Because of the nature and importance of the topic, questions related to participant recruitment, extended session times planned for treatment evaluations, and measurement of procedural fidelity need to be addressed in a preliminary investigation. The current proposal will thus have the following aims.

Specific Aim 1: Demonstrate ability to meet enrollment requirements. The larger project (R01) proposed an enrollment of 50 participants over 5 years, with 40 participants anticipated to complete the studies. Enrollment target for this R56 will be seven to 10 participants evenly split across the two proposed studies. The investigators will randomize the first participant and then alternate assignment for subsequent participants. The investigators anticipate five to seven participants completing. Based on published data regarding reinforcement-based interventions for SIB maintained by automatic reinforcement, the investigators anticipate two-thirds of the participants will receive NCR and one-third will receive DRO. To assess feasibility of the defined recruitment targets, the investigators will collect data on the following: number of responses to recruitment efforts including, when possible, the proportion of possible participants who responded to recruitment efforts; proportion of potential participants who did not continue in the consent process following screen and the reason (e.g., declined further participation, screened out due to inclusion criteria); and proportion of participants and potential participants who exited the study at any point following initial screen as well as the reason (e.g., decided to withdraw).

Specific Aim 2: Evaluate acceptability and feasibility of participation relative to observation length and study duration. The larger study includes observation periods with intervention implemented for up to 180 min (3 hours) and an anticipated study duration of 24 weeks. To determine parent/caregiver acceptability of these longer session appointments, social validity will be measured over the course of the project using modified versions of the Parent Evaluation Inventory (PEI) and the Therapist Evaluation Inventory (TEI). Each measure includes 19 items and assesses the extent to which caregivers view interventions positively. The investigators will modify the instruments so that (a) they are appropriate for caregivers who are not parents (for May Institute participants whose caregivers will be teachers and direct care staff), and (b) assess the acceptability of procedural features (e.g., pre-intervention assessments, session duration) in addition to acceptability of the intervention. The PEI and TEI have been widely used in behavioral intervention studies. Investigators will administer the measures after completion of (1) pre-intervention assessment (preference assessment, functional analysis, COAB), (2) efficacy evaluation, and (3) intervention tests. As a further assessment of social validity, attendance and dropouts will be monitored and the investigators will continuously monitor participant assent to determine whether participants are more likely to withdrawal assent as observation periods increase. Data related to study duration for participants completing the study protocols will be used to more accurately propose individual participation timelines for the subsequent R01 application.

Specific Aim 3: Track participant progression. Participant progress through the study protocols will be tracked to determine if the individual timeline proposed is adequate, or requires adjustment. In particular, the investigators will evaluate whether the stated phase change criteria (currently based on IES SCD criteria) need to be adjusted to allow for either more efficient progress through the study phases or more convincing demonstration of experimental control.

Specific Aim 4: Demonstrate acceptable procedural fidelity with intervention implementation and Interobserver agreement (IOA).

Fidelity. The study will be conducted across two demonstration sites (May Institute and the University of Georgia). Thus, uniformity of implementation could vary. During the R56 period, the proposed training program will be implemented, and data will be collected on procedural fidelity and compared across sites.

Specifically, the investigators will directly measure procedural fidelity across all conditions to ensure the integrity of the procedures. Conforming with generally accepted practices in SCD, the investigators will use direct observation instruments reflective of the individual study protocols and evaluate for errors of omission and commission across at least 30% of all sessions, evenly distributed across study conditions and phases, for all participants and all implementers. These instruments typically consist of task analyses that specify what implementers should do (e.g., provide response-independent access to the reinforcer every 30 s), and what they should avoid (e.g., no vocal response from the therapist if SIB occurs). Investigators will then compute a measure of fidelity related (e.g., conditional probabilities of correct procedures and incorrect therapist behavior) to these two forms of errors relative to the planned steps and components for those conditions. Fidelity data will largely be collected in vivo, though recordings can be accessed if necessary (sometimes therapists and clients move out of frame of the camera so the investigators do not want to rely on video for fidelity data collection). To assist with procedural uniformity, May Institute will train implementers across sites and meet weekly with observers to review fidelity of implementation data and provide additional training. Adjustments to training procedures will be made or training boosters will be conducted if procedural fidelity drops below 90% for three consecutive implementations of any given procedure.

IOA: Mulitple data collectors will be responsible for in vivo data collection over the course of the study. To ensure that observation records reliably reflect what occurred during sessions, IOA will be tracked across all phases and conditions. Specifically, a second, naive, observer will collect data for IOA across all conditions using the same instrument as the primary data collector. Aligning with the generally accepted standards in SCD, these data will be gathered across at least 30% of sessions in each condition for each participant. The investigators will compare data between the observers to compute a measure of IOA. Data collectors will undergo a review of definitions and practice if IOA scores fall below 90% for three consecutive observations.

Specific Aim 5: Post hoc evaluation of subtype/intervention match. Following the conclusion of the study for each participant, the subtype of their SIB, the prescribed intervention, and the intervention outcome will be denoted. This process will allow for additional data sets related to subtype prevalence and subtype by intervention outcome to be collected. Individuals who exhibit SIB that fits Subtype 3 (i.e., SIB with self-restraint) will be reviewed on a case-by-case basis. If their self-restraint eliminates the occurrence of SIB, they will not be included in the study as SIB will not occur during the assessment phase.