Strategies to reduce low-value care – An applied behavior analysis using a single-case design

INTRODUCTION: Implementation science has traditionally focused on the implementation of evidence-based practices, but the field has increasingly recognized the importance of addressing de-implementation (i.e., the process of reducing low-value care). Most studies on de-implementation strategies have used a combination of strategies without addressing factors that sustain the use of LVC and there is a lack of information about which strategies are most effective and what mechanisms of change might underlie these strategies. Applied behavior analysis is an approach that could be a potential method to gain insights into the mechanisms of de-implementation strategies to reduce LVC. Three research questions are addressed in this study: What contingencies (three-term contingencies or rule-governing behavior) related to the use of LVC can be found in a local context and what strategies can be developed based on an analysis of these contingencies?; Do these strategies change targeted behaviors?; How do the participants describe the strategies' contingencies and the feasibility of the applied behavior analysis approach? MATERIALS AND METHODS: In this study, we used applied behavior analysis to analyze contingencies that maintain behaviors related to a chosen LVC, the unnecessary use of x-rays for knee arthrosis within a primary care center. Based on this analysis, strategies were developed and evaluated using a single-case design and a qualitative analysis of interview data. RESULTS: Two strategies were developed: a lecture and feedback meetings. The results from the single-case data were inconclusive but some of the findings may indicate a behavior change in the expected direction. Such a conclusion is supported by interview data showing that participants perceived an effect in response to both strategies. CONCLUSION: The findings illustrate how applied behavior analysis can be used to analyze contingencies related to the use of LVC and to design strategies for de-implementation. It also shows an effect of the targeted behaviors even though the quantitative results are inconclusive. The strategies used in this study could be further improved to target the contingencies better by structuring the feedback meetings better and including more precise feedback.