Simplified Useful Respiratory Failure Evaluation System2

[Object]To design and evaluate simplified useful respiratory failure evaluation system2(SURFES2) for respiratory failure patients admitted to RICU/ICU.[Methods]100 respiratory failure patients' data from RICU/ICU(estimation group) were prospectively collected to screen prognostic factors which affect the outcome of respiratory failure patients by logistic regression and Delphi.A new respiratory failure scoring system was made by giving screened prognostic factors a defined weight combining death risk estimates of each variable with pulmonary expert evaluations. The data of 100 cases(validation group) were collected to evaluate the scoring system's discrimination and calibration and compare with APACHEⅡ.[Results]Arterial pH,oxygen index,age,mean arterial pressure,hematocrit, conscious state are associated with prognosis of respiratory failure,the highest score of each factor were 4.Death risk equation:In[R/(1—R)]=-5.231+0.508×score of simplified useful respiratory failure evaluation system2.The area under the receiver operating characteristic curve of simplified useful respiratory failure evaluation system2 and APACHEⅡfor the validity group is 0.901 and 0.925 respectively (z=0.462),there is no statistical difference.P values of Hosmer-Lemeshow test of both systems are greater than 0.05.The calibration of the two systems is reasonable.[Conclusion]Simplified useful respiratory failure evaluation system2 has high accuracy,less variables.It is suitable for evaluating the patients' prognosis of respiratory failure in daily clinical practice.APACHEⅡperform better than the simplified useful respiratory failure evaluation system2,but it is unpractical to evaluate the prognosis of respiratory failure patients for too many variables.