Shahani A K, Ridley S A, Nielsen M S
School of Mathematics, GeoData Institute, University of Southampton, Southampton, UK.
Anaesthesia. 2008 Oct;63(10):1074-80. doi: 10.1111/j.1365-2044.2008.05577.x. Epub 2008 Jul 10.
Using real data from a number of hospitals, we predicted the patient flows following a capacity or organisational change. Clinically recognisable patient groups obtained through classification and regression tree analysis were used to tune a simulation model for the flow of patients in critical care units. A tuned model which accurately reflected the base case of the flow of patients was used to predict alterations in service provision in a number of scenarios which included increases in bed numbers, alterations in patients' lengths of stay, fewer delayed discharges, caring for long stay patients outside the formal intensive care unit and amalgamating small units. Where available the predictions' accuracy was checked by comparison with real hospital data collected after an actual capacity change. The model takes variability and uncertainty properly into account and it provides the necessary information for making better decisions about critical care capacity and organisation.
利用多家医院的真实数据，我们预测了容量或组织变革后的患者流量。通过分类和回归树分析获得的临床上可识别的患者群体，被用于调整重症监护病房患者流量的模拟模型。一个准确反映患者流量基础情况的调整模型，被用于预测多种情景下服务提供的变化，这些情景包括病床数量增加、患者住院时间变化、延迟出院减少、在正式重症监护病房之外护理长期住院患者以及合并小单元。在有实际数据的情况下，通过与实际容量变化后收集的真实医院数据进行比较，检验了预测的准确性。该模型充分考虑了变异性和不确定性，并为做出关于重症监护容量和组织的更好决策提供了必要信息。
