论文部分内容阅读
Aim This study was designed with the aim of comparing the performances of the Pediatric Risk of Mortality (PRISM), the Pediatric Index of Mortality (PIM), and the revised version PIM2 models in a pediatric intensive care unit in China.Methods A total of 885 critically ill pediatric patients were enrolled in this study between January 1, 2014 and June 31, 2014.Information on age, sex, diagnosis, length of stay in PICU, outcomes and the variables required to calculate PRISM, PIM and PIM2scores were collected.The logistic regression model developed in the learning sample was evaluated in the test sample by calculating the area under the receiver operating characteristic (ROC) plot (AUC) to assess discrimination between death and survival.Calibration across deciles of risk was evaluated using the Hosmer-Lemeshow goodness-of-fit x2 test.Results Of the 852 critically ill pediatric patients enrolled in this study, 745 patients survived until the end of the PICU stay (107 patients died, 12.56%).The median age of all patients was 6.5 (IQR:2-21) months.The AUCs (95% confidence intervals, CI) were 0.729 (0.670-0.788) for PRISM, 0.721(0.667-0.776) for PIM and 0.726 (0.671-0.781) for PIM2.The Hosmer-Lemeshow test gave a chi-square of 44.22 (P<0.000, v=10) for PRISM, 14.30 (P>0.05, v=10) for PIM and 12.47 (P>0.05, v=10) for PIM2.The standardized mortality rate was 1.00 for PRISM, PIM, and PIM2.Conclusions In this study, the PRISM, PIM and PIM2 scores displayed a good discriminatory performance.With the exception of PRISM, the PIM and PIM2 models display good calibration.A positive correlation was found among PRISM, PIM, and PIM2 scores.