Construction of a Risk Prediction Model for Hypothermia During Surgery in Pediatric Under General Anesthesia Based on Machine Learning Algorithm
Construction of a Risk Prediction Model for Hypothermia During Surgery in Pediatric Under General Anesthesia Based on Machine Learning Algorithm
Intraoperative hypothermia refers to a core body temperature below 36.0°C during surgery, which is common in surgical patients. Due to the fact that children's body temperature regulation function is not yet fully developed, they are light in weight, and their blood vessels are superficial, children are very susceptible to the influence of environmental temperature. With the effects of anesthetic drugs, exposure of the surgical field, and disinfection of the surgical area, children face a higher risk of intraoperative hypothermia than adult patients. Studies have shown that the incidence of intraoperative hypothermia can be as high as 80%. Intraoperative hypothermia can lead to increased adverse cardiovascular events, poor coagulation, slower healing of surgical incisions, or wound infection, threatening the health of children, resulting in prolonged postoperative hospitalization and increased hospitalization costs. Therefore, there is an urgent need to develop a tool for predicting intraoperative hypothermia suitable for children, identify high-risk groups early, and take preventive measures as soon as possible, thereby reducing a series of complications caused by hypothermia. The purpose of this study is to clarify the current status and risk factors affecting intraoperative hypothermia in children, to provide a theoretical basis for clinical medical staff to provide intraoperative thermal insulation care for children, and on this basis to construct an intraoperative hypothermia risk prediction model to identify the probability of hypothermia in children at an early stage, so as to take targeted thermal insulation measures.
The study first clarified the overall incidence and influencing factors of IOH in children through meta-analysis, combined with expert consultation to determine potential predictive variables, and developed a data questionnaire for this study. The second part of the study will collect data prospectively, use four algorithms in machine learning, including logistic regression, decision tree, random forest, and support vector machine, to build an excellent intraoperative hypothermia risk prediction model, and conduct internal verification of the model. The third part of the study prospectively collects sample data to complete external verification of time period and space. The constructed model still has good predictive ability in the new data set. Finally, some models are visualized to form intuitive and easy-to-understand charts or interfaces.
Inclusion Criteria:
Exclusion Criteria: