列线图
医学
急诊科
接收机工作特性
曲线下面积
内科学
重症监护室
红细胞分布宽度
队列
逻辑回归
败血症
队列研究
急诊医学
精神科
作者
Chenyan Zhao,Yao Wei,Dongyu Chen,Jun Jin,H. S. Chen
标识
DOI:10.1016/j.intimp.2019.106145
摘要
To develop an inflammatory biomarker-based, simple-to-use nomogram for the early identification of septic patients at high risk of mortality in the emergency department (ED). All patients diagnosed with sepsis admitted to the intensive care unit (ICU) from the ED were screened from the Medical Information Mart for Intensive Care III database and divided into two cohorts: the primary cohort and the validation cohort. We used bivariate logistic regression analysis to determine independent risk factors and to construct a predictive nomogram and subsequently evaluated the calibration, discrimination and clinical usefulness of the nomogram. The gradient boosting machine (GBM) model was used to more accurately evaluate these variables. A total of 5663 admissions were enrolled, of which 3964 comprised the primary cohort and 1699 comprised the validation group, with 28-day mortality rates of 21.2% and 20.4%, respectively. Age > 69, neutrophil-to-lymphocyte ratio (NLR) > 9.8, platelet-to-lymphocyte ratio (PLR) > 249.89, lymphocyte-to-monocyte ratio (LMR) ≤ 2.18, and red cell distribution width (RDW) were detected as important determinants of 28-day mortality and included in the nomogram. The calibration plot revealed an adequate fit of the nomogram for predicting the risk of 28-day mortality. Regarding discriminative ability, receiver operating characteristic curve analysis showed that the nomogram had an area under the curve (AUC) of 0.826 (95% CI: 0.811–0.841, P < 0.001) in the primary cohort, which was greater than that of all individual parameters and other scores. Decision curve analysis also indicated that our nomogram was feasible in clinical practice, as the threshold probabilities were 0–0.62 for the primary cohort. The GBM model yielded a significantly greater AUC of up to 0.867. This proposed simple-to-use nomogram based on age, NLR, PLR, LMR and RDW provides a relatively accurate mortality prediction for septic patients in the ED.
科研通智能强力驱动
Strongly Powered by AbleSci AI