The Use of Simulation to Assess Dynamic Decision Making.

Abstract

INTRODUCTION: In dynamic situations such as in the intensive care unit, decisions are often made by applying precompiled rules for a particular event relying on past experience to predict future outcomes. Klein has described this type of thinking as recognition primed decision-making 1. Two common clinical conditions in critically ill patients include hypotension from sepsis and chest pain from myocardial ischemia. If presented with a patient with either one of these conditions, a precompiled response could be used to start treatment. Treatment challenges arise if clinicians are presented with a patient with both problems. We retrospectively reviewed the treatment of a simulated septic patient who subsequently developed myocardial ischemia. MATERIALS AND METHODS: Following IRB approval, we reviewed ten audiovisual recordings of fully immersive high fidelity ICU course simulations taught at the Palo Alto VA from June 2003-June 2004. Medical management was first provided by an intern with a senior resident arriving, on average 12 minutes, into the scenario. The METI-HPS was used for all of the simulations. The scenario involved an elderly male admitted to the ICU for low urinary output, increasing heart rate and hypotension, with the presumptive diagnosis of urosepsis. The patient developed chest pain and ischemic EKG changes 8 minutes into the scenario. The teams’ overall performance was scored as either management of primary sepsis, primary ischemia, or both. RESULTS: All ten teams began the appropriate treatment of sepsis (additional intravenous access and fluid). Upon the development of chest pain (HR 120’s BP 80’s/50’s) 5 teams deviated from the sepsis treatment plan and converted to a primary cardiac treatment plan, including giving aspirin, nitrates, heparin and morphine. Four teams continued with the sepsis pathway, including either starting or escalating vasopressors, and one team treated the patient as having both a cardiac and a septic event. DISCUSSION: Medical decision-making often relies on pattern recognition. Using precompiled responses, for example treatment pathways, cognitive aids, and mnemonics; a physician can make decisions more efficiently. If faced with novel situations or if conflicting treatment goals arise, physicians are required to use abstract reasoning to determine the most appropriate next step. In this scenario, the patient developed myocardial ischemia secondary to increased demand from the hypotension and tachycardia caused by sepsis. Half the teams converted to a primary cardiac pathway treating the patient’s ischemia as a primary cardiac event. This conversion not only delayed appropriate treatment but also resulted in morbidity. Cognitive tunneling may have occurred as the teams relied on familiar medical management even when it was inappropriate. We could not question the participants on why they chose a primary cardiac treatment pathway, as this was a retrospective study. Further investigation is needed to determine if these errors were secondary to lack of experience or represent a more fundamental flaw in the way dynamic decision-making is taught in medical education. Simulation is a useful method to probe such issues.