Different criteria affect prevalence of relapse of behavior targeted for treatment

Abstract

Several studies have examined the prevalence of behavioral relapse among individuals with intellectual and developmental disabilities following common treatment challenges (context changes, schedule thinning). Most applied studies compare behavior during the treatment challenges with the maximum level of behavior from five preceding treatment sessions. This max‐of‐5 criterion could inadvertently capture behavior in transition during the preceding treatment phase, thereby underestimating the prevalence of relapse. In the current study, we reanalyzed existing clinical data with the max‐of‐5 criterion and an alternative criterion less likely to capture target behavior in transition—the mean of the last two sessions (mean‐of‐2 criterion) of the treatment phase. As hypothesized, the max‐of‐5 criterion produced lower prevalence estimates relative to the mean‐of‐2 criterion. We encourage researchers conducting these analyses to weigh different approaches to reporting prevalence data and discuss considerations for future areas of research and practice related to measurement of relapse.