Clinical Decision Making for Problematic Behavior

Part 1: Development of decision-making tool Part 1 will involve developing a decision-making tool for clinicians for selecting appropriate and effective interventions for problematic behaviour maintained by access to tangible items. We will do this by a) conducting a meta-analysis of the empirical literature and b) surveying behavioural practitioners. We will deliver the survey to behavioural practitioners electronically and it will take approximately 30 minutes to complete. We will provide eight client scenarios based on literature and the investigators' experience, asking them to decide which intervention best suits each example. We will also ask them to consider additional factors, e.g., client characteristics, the setting, and practical limitations such as time and resources. We will provide a list of interventions, ask them to rank the top three for each scenario. Based on survey responses and synthesis with the literature, we will create a step-by-step guide on client characteristics, goals, and the intervention that aligns with these. This will include a decision tree based on the expert-suggested and empirically-tested interventions. This tool will ask yes/no questions that lead clinicians to a new question or a potential treatment option for challenging behaviour maintained by access to tangibles. Part 1 will be complete when we have received at least 30 responses to the survey. Adherence to Part 1 will be measured by the number of surveys returned. Part 2: Testing the use of the decision-making tool. Part 2 will commence approximately 2 weeks after Part 1 is complete. Part 2, we will recruit 3-9 participants with cognitive impairment. Participants could be people with dementia, brain injury, learning disability, intellectual disability and may be a child or an adult. The factor they will all have in common is that they engage in behaviour that is thought to be maintained by access to a preferred tangible item. They represent a client group that behavioural practitioners are likely to encounter in their clinical work. Assessment: 1. Descriptive assessment: Interview the client (if appropriate) and caregivers/ support staff to understand the challenging behaviour and conditions under which it is likely to occur. We will use an array of assessment tools, including open-ended functional assessment interviews, and the Functional Assessment Screening Tool (FAST). These interviews will take between 30 min and 60 min per participant. 2. We will then conduct direct observations of the behaviour (at times informed by the descriptive assessment data). These are baseline sessions (please see outcome measures). We will record the behaviour, in addition to variables that may be related (e.g., provision of tangible item by staff, attention, time of day etc). These sessions will be approximately 30 minutes, and there will be one to four sessions conducted with each participant (no more than two sessions a day). The number of sessions will be determined through data-based decision making of the repeated measures data. We will analyse the data after every session, looking specifically at variability and richness of the data (e.g., stability of occurrence of problematic behaviour). The more variable the data, the more likely we are to conduct another session (up to the maximum number). These sessions will be conducted over approximately 4 weeks. These sessions will occur in where the person lives (i.e., the care home) or attends school. Procedural integrity will be measured in approximately 30% of sessions by a second observer. They will record integrity (adherence) based on a task analysis (a list of the investigator behaviours in which we will engage in each session). 3. Experimental functional analysis: This assessment is used to empirically-demonstrate the function of a problematic behaviour. (Iwata et al., 1982/1994) This will involve 3-10 sessions, each lasting 5 to 15 minutes. In each session, the way in which we react to the problematic behaviour will change. In the first condition, we will ignore any instances of the behaviour, then in the following condition we will provide attention (the type of attention (positive or negative) will depend on the individual's learning history and how the staff normally react- this will be individually tailored and defined for each participant). In the third condition, we will give frequent attention whether the individual engages in the behaviour or not, and in the fourth session, we will present a task or instruction (again, this will be individually defined based on what the individual is usually asked to do) and will remove the demand if the behaviour occurs. In the fifth and final condition, we will provide an item contingent on problematic behaviour. Each session will comprise each of the five conditions. The condition in which the most problematic behaviour occurs indicates the likely function. We expect that all of our participants will engage in the most behaviour in the fifth condition, confirming their behaviour to be maintained to access to a tangible item. The number of sessions will be determined through data-based decision making of the repeated measures data. We will analyse the data after every session, looking specifically at variability and richness of the data (e.g., stability of responding). The more variable the data, the more likely we are to conduct another session (up to the maximum number of 10). Procedural integrity will be measured in approximately 30% of sessions by a second observer. They will record integrity (adherence) based on a task analysis (a list of the investigator behaviours in which we will engage in each session). Intervention: We will use an alternating treatments design under which we will alternate applying two interventions to the problematic behaviour to allow for a comparison of effectiveness. In some sessions, we will implement the intervention suggested by the decision-making tool and in the other sessions, we will implement the ‘universal intervention' suggested by the literature. The 'universal intervention' will be noncontingent reinforcement (NCR) and was chosen because it is the most common reported intervention in the empirical literature. NCR involves giving the participant access to the desired item at set time intervals (e.g., 5 minutes), regardless of their behaviour, therefore removing the 'need' for the behaviour. Which intervention (i.e., that selected by the tool or NCR) will be delivered in each session will be determined using a random number generator. 10 minutes prior to beginning the intervention phase, the clinician will use the tool to decide which intervention will be compared to NCR for the rest of the intervention comparison. Interventions suggested by the tool will fall under two categories: (1) Access-Related Interventions and (2) Functional communication training. Examples of access-related interventions are free access to the desired items at all times and differential reinforcement. Differential reinforcement involves providing reinforcement (reward) for desired behaviour while reducing reward for challenging behaviours. For example, the person will have more time to engage with their preferred tangible item for desired behaviour compared to less time when engaging in challenging behaviour. Functional communication training interventions involve teaching new skills to our participants to receive the item they desire, e.g., verbally asking, exchanging a picture or visual, or sign language. Each participant will receive the intervention the tool suggests and therefore may all receive a different intervention to be compared with NCR. We will implement the Intervention in the client's natural environment, such as school, home, or rehabilitation centre. The number of sessions will be individualised to the participant, depending on the behaviour of concern, skills learned, progress, participant availability etc. There will be a maximum of 30 sessions. The number of sessions will be determined through data-based decision making of the repeated measures data. We will analyse the data after every session, looking specifically at variability and richness of the data (e.g., stability of responding). The more variable the data, the more likely we are to conduct another session (up to the maximum number of 10). Procedural integrity will be measured in approximately 30% of sessions by a second observer. They will record integrity (adherence) based on a task analysis (a list of the investigator behaviours in which we will engage in each session). Sessions will be 30 min to 60 min in duration. This phase will finish when the behaviour has reduced to an acceptable level under one or both interventions (as determined with each participant and their support people / whanau; for example, an acceptable behaviour change might be a reduction in rate of behaviour to less than 1 occurrence per hour) or when the maximum number of sessions (30) has been implemented (this mirrors clinical practice). We will ask the support person who cares for the participant to complete social valid survey. This participant will accompany the researcher and participant for at least one session of each of the two compared intervention so that they can observe and therefore comment on each. In the survey, we will ask which intervention they think was the most appropriate, meaningful, and effective. Support people will provide feedback but will not implement the intervention. We will implement follow-up sessions one month and three months following the conclusion of the intervention comparison. Follow-up sessions will involve observing the participant without any intervention to determine whether the behaviour change achieved during the intervention remains. These sessions will be between 30 min and 60 min. Overall, Part 2 will be approximately 9 months for each participant. Part 3. Feedback and usefulness survey We will ask behavioural practitioners to complete a follow-up survey (this part will commence when Part 2 is completed). This survey will take approximately 30 minutes to complete. Some may have completed Part 1 and some may not have; we will not be attempting to match responses within a respondent. We will provide four case studies (different but equivalent to those in the survey in Part 1) and ask the respondent to determine the best intervention by using the decision-making tool. We will ask about its utility, comprehensiveness. We will also ask for feedback regarding the tool, i.e., do they like it and find it useful? This will allow the final version of the tool to be developed. Part 3 will be complete when we have received at least 30 responses to the survey. Adherence to Part 3 will be measured by the number of surveys returned. Participants will proceed to each subsequent phase immediately following completion of the previous part.