Examining the relationship between reinforcement ratios and seclusion rates in residential treatment

The use of highly restrictive interventions, such as seclusion, to manage disruptive behaviors within mental health treatment facilities is gaining increasing focus in clinical and research discussion, as is the identification of strategies for reducing the frequency of its use. Evidence-based interventions such as behavior modification have long been utilized as a means of altering maladaptive environmental response patterns and developing prosocial skills within the confines of residential treatment. It follows that the effective implementation of behavior modification in the teaching of adaptive strategies as replacements for aggressive and assaultive behaviors should decrease usage rates of restrictive interventions such as seclusion. A unique opportunity to examine both issues simultaneously exists at a long-term, residential treatment center for children and adolescents. This facility utilizes an intricate behavior modification system in combination with cognitively based therapeutic interventions and medication management while employing an extensive system for tracking reinforcement and seclusion rates. It was hypothesized that reinforcement ratios on days in close proximity to a seclusion event are more likely to differ from days unrelated to seclusion. The study utilized archival data consisting of rates of reinforcement over the 8-day period prior to a seclusion event. Rates of reinforcement were initially analyzed using a 4 x 2 x 2 mixed between/within-subjects analysis of variance to examine positive to negative ratio frequencies at four points in time prior to a seclusion event; those points were specified as 1, 2, 3, and 8 days prior to the event itself, with the 8th day considered an unrelated reference day. Ratio of reinforcement served as the dependent variable, day of reinforcement as the within-subjects independent variable, and gender (female and male) and race (non-Caucasian and Caucasian) served as between-subjects variables. After initial analysis produced significant main effects for day of reinforcement, with a strong difference noted between day 1 and day 8, a second mixed between/within analysis of variance was completed using those days as the independent variable. Significant main effects were again found for day of reinforcement, with additional significant effects found on the between-subjects independent variables of gender and race.