Establishing a prediction system for the critically ill obstetric patients

Objective To establish a prediction system for the critically ill obstetric patients. Methods A retrospective study was conducted in 311 critically ill obstetric patients, who were admitted to the ICU of the Third Affiliated Hospital of Guangzhou Medical University from January 2000 to July 2008. The patients were divided into dead group(72 patients) and survival group (239 patients). Assigned the clinical indexes related to the mortality and established a prediction system for the critically ill obstetric patients. The clinical indexes were analyzed by using t test and the mortality prediction equation by using logistic regression model. Results The prediction system involved 18 clinical indexes, including body temperature, mean artery pressure, respiratory rate, heart rate, PaO2/FiO2, arterial pH, serum sodium, serum potassium, blood glucose, serum creatinine, serum total bilirubin, serum albumin, ALT, blood platelet count, fibrinogen, white blood count, hematocrit and scores of nervous system. The prediction equation for the mortality risk: log eY/(1-Y)=-7.275+ 0.245×total scores. Conclusions Establishing a prediction system for the critically ill obstetric patients not only improves the diagnostic level but also has an scientific and objective basis in improving medical quality and utilizing the medical resources. Key words: Pregnancy, high-risk; Severity of illness index; Forecasting; Evaluation