Real Patient Intensive Care Data On A Patient Simulator
Jochen Vollmer, Stefan Mönk, Wolfgang Heinrichs, Thomas Uthmann · Simulation in Healthcare The Journal of the Society for Simulation in Healthcare · 2006
Introduction: Simulation of the course of disease of critically ill patients is a difficult but desirable task for the training of medical professionals. The usage of existing model-driven patient simulators in this field can be questioned as their educational models are based on “healthy” patient’s physiology and typical problems of intensive care patients (e.g. development of multi organ failures) cannot be simulated very well. To a big part this is due to a lack of mathematical models that describe the pathophysiological processes in critically ill patients and obtaining mathematical sound models for such patients will probably not be possible in the near future. How can we approach simulation of intensive care patients using the available tools? The first step could be educational reproduction of a real patient’s course of disease. Methods: Based on data from a patient data management system of about 3000 intensive care patients with 15000 total days of treatment we developed a method to automatically transfer the course of a real patient’s diseases into scenarios on a METI Human Patient Simulator. The baseline data codes bedside information about the patient and his disease, a record of the status of the patient during his stay in the intensive care unit dissected into the performance of individual organ systems on a 24h basis. The performance of each of these organ systems was qualified in one of 5 discrete classes from “fully functional” to “highly critical”. Additionally a tendency of the development of the status in each of the organ systems was recorded on a daily basis. Using METI HPS version 6.3 as a physiology simulator, we identified the parameters inside the HPS software that describe function and performance of each of the recorded organ systems. We calibrated the values for each of these parameter sets to the levels of the discrete classes that described the organ status. The tendency of the developments in each of the organ systems was then translated into an onset-speed of the change of each of the parameters for the organ system. The data of each of the recorded days of a patient’s stay in the ICU was coded into a state in a METI HPS scenario. Results: The translation of real ICU patient data into METI HPS patients and scenarios generates patient models which - within the limits of educational models - represent the underlying real patient. It enables us to look at the course of a disease in real time as well as in just a few minutes. This allows to discover general tendencies and correlations of different states in the disease. With this method we obtain a possibility to reproduce critically ill patients for demonstration and training purposes.Table