From bartending interruptions to medication delivery interruptions: Managing the risks of a high-fidelity simulation study with pilot research.

Abstract

In this paper we describe the risks of complex applied research, especially in work domains where professional practitioners are scarce. For such research, careful preparation and piloting is needed, especially when estimating sample size is required for a full study. However, such pilot work may reduce the potential sample size for the full study. We describe how the these issues have been addressed in applied psychology contexts. We then present a case study illustrating how we determined sample size for a study investigating the impact of workplace interruptions on errors that intensive care unit nurses might make during medication preparation and administration. The pilot work was performed in a functionally related domain to nursing-bartending-and bartender participants filled cocktail orders. Pilot 1 investigated performance with 0 interruptions and applied a model from a field observation to estimate probable effect sizes and sample sizes with 1 or 4 interruptions per medication scenario. Pilot 2 collected empirical data on the effect of 1 or 4 interruptions per cocktail scenario on cocktail errors and estimated sample size for the medication study, which was subsequently successfully run. The applied community could benefit from further discussions about these issues and the means for addressing them. Public Significance StatementApplied experimental psychologists increasingly perform studies in complex work environments that can be hard to access. Careful piloting to determine feasibility and target sample size is valuable, but piloting might require only approximating the target work environment. This paper describes how a pilot study testing the impact of interruptions on bartending work was used to prepare for a major simulation study of the impact of interruptions to intensive care unit nurses preparing medications.