Analysis and Prediction of Length of Stay in the Postanesthetia Care Unit. |
Won Oak Kim, Hae Keum Kil, Bon Nyeo Koo, Jeong Il Kim |
Department of Anesthesiology, Yonsei University College of Medicine, Seoul, Korea. |
|
|
|
Abstract |
BACKGROUND Optimal control for the management of the length of stay in the postanesthesia care unit (PACU) following general anesthesia in adults is an important strategy for surgical patients' care. A model to predict the results of the PACU stays could be used to improve the utilization of the PACU and resources of the operating room through a more efficient arrangement. The purpose of this study was to evaluate the performance of the decision tree based analysis using clinical sets of data from adult patients undergoing general anesthesia. METHODS The decision tree was trained with 351 clinical sets (86% in 409 data sets) using a Chi-squared automatic interaction detection (CHAID) algorithm and validated through independent testing of 58 cases (14%). Twenty-two independent variables were used to find determinant variables and to predict categorical dependent values (lengths of stay in the PACU). RESULTS The decision tree based analysis correctly predicted in 68% of real situations and identified influencing variables as intubation state, complication in the PACU, and intraoperative transfusion. CONCLUSIONS We concluded that the decision tree based analysis could provide a useful predictive and classifying model for the optimization of limited resources of the PACU.
The decision tree based analysis is an alternative way of classifying, and a predicting method for developing a model for lengths of stay in the PACU with easy interpretation and clear graphical displays of the structure of variables. |
Key Words:
Recovery: length of stay; prediction variables; Statistics |
|