Association for Behavior Analysis International

Functional analyses and treatments of elopement remain critical areas for investigation given the serious and sometimes fatal consequences associated with elopement (e.g., traffic injury, drowning). The first paper in this symposium (Nevill et al.) summarized data on outcomes of functional analyses and treatment evaluations from a consecutive controlled case series of 27 individuals who engaged in elopement. In the second paper, Scheithauer et al. conducted a small randomized clinical trial to assess the feasibility of a manualized elopement intervention that included a functional behavioral assessment and subsequent function-based treatment. The second and third papers took a slightly different approach in which they began the functional analysis process by tailoring the test and control conditions based on information gathered from interviews. The focus of Jessel et al. was on evaluating a comprehensive assessment and treatment package with socially valid outcomes with two individuals diagnosed with autism. The focus of Blowers et al. was to evaluate the potential reinforcement value of a common consequence of elopement that has yet to receive attention in the literature—chase. We are fortunate to have Dr. Terry Falcomata, known for his research on the assessment and treatment of challenging behavior, serve as the discussant.

BCBAs

Prior research in the treatment of elopement indicates that developing a function-based treatment results in a more effective intervention (Lang et al. 2009; Piazza et al.1997). We conducted a consecutive controlled case series on the Neurobehavioral Unit with 27 individuals who engaged in elopement. For 20 individuals (74%) a functional analysis (FA) was conducted specifically targeting the participant's elopement while the other participants had a functional analysis that targeted a broad range of problem behavior including elopement. Of the 20 individuals with an elopement FA, 8 were inconclusive, and the most common function identified was social positive only (7), multiple functions (7), and automatic reinforcement (5). There were no participants who engaged in elopement to escape. The most common treatment for those individuals with an elopement FA were reinforcement-based (using differential or noncontingent reinforcement) whereas for those individuals for whom a more generic FA was conducted, response reduction interventions (e.g., response cost) or response blocking was more likely to be used. Treatments based on an elopement FA were also more likely to be effective. Figure 1 includes an FA and treatment from one participant who was included in the analysis and published in Piazza et al. (1997).

Elopement is common in autism spectrum disorder (ASD) and can be incredibly dangerous. There are examples of successful treatments, but most use single-subject design or summaries of clinical data. Randomized clinical trials (RCTs) are an important next step to support classification of behavioral interventions targeting elopement as an empirically supported treatment and to disseminate this treatment to individuals outside the field of behavior analysis. This study conducted a small (N = 24) RCT to assess feasibility of a manualized elopement intervention. The manual included a functional behavioral assessment and subsequent function-based treatment. Appointments took place in the home, with a heavy focus on caregiver training and safety measures. Primary outcome measures of feasibility were adequate: > 80% completion of treatment, therapist integrity > 80%, and caregivers rated the intervention as acceptable. Preliminary efficacy results demonstrated a significant improvement from pre to post for the treatment group on the irritability subscale of the Aberrant Behavior Checklist, the child-domain of the Parenting Stress Index, and a checklist designed to measure safety measures for elopement. Results suggest that an RCT with this population is feasible and that the manualized intervention shows promise for future efficacy trials using a larger sample.

Using a latency measure during a functional analysis of problem behavior increases the brevity of the assessment period and decreases risks resulting from exposure to contexts intended to evoke problem behavior. In addition, latency-based functional analyses may be especially suitable for discrete behaviors such as elopement, which require resetting the environment after each instance of the response. We evaluated a comprehensive assessment and treatment package (Hanley, Jin, Vanselow, & Hanratty, 2014) for the elopement of two children diagnosed with autism spectrum disorder. We observed short latencies to elopement during the test condition of the analysis and no elopement during the control condition. We then taught the participants multiple forms of functional communication of increasing complexity while elopement was blocked. This was followed by the thinning of reinforcement to a terminal goal, creating a treatment package that nearly eliminated elopement and was socially validated by caregivers.

Elopement can have serious or fatal consequences associated with a lack of supervision or getting lost (e.g., traffic injury, drowning). Given the potential dangers associated with elopement, caregivers are likely to chase after their child when elopement occurs. Like any other event that follows problem behavior, the chase and physical attention that occurs when retrieving the child may serve as a reinforcer and therefore maintain elopement in the natural environment. The potential reinforcing effects of chase, retrieval, or both has led researchers to attempt to control for, or at least minimize the influence of, these variables when conducting non-attention conditions (e.g., escape, tangible) of a functional analysis of elopement. However, no study to date has evaluated whether elopement is indeed sensitive to positive reinforcement in the form of chase. In the present study, we evaluated the effects of chase on elopement with an X-year old boy diagnosed with autism. Elopement was found to be maintained, at least in part, by chase. We then conducted a treatment evaluation in which we assessed the effects of various antecedent- and consequence-based treatment components within a multiple schedule.