Predicting 30-days mortality for MIMIC-III patients with sepsis-3: a machine learning approach using XGboost

• DOI: 10.1186/s12967-020-02620-5
• Publication Date: 2020-12-07T10:05:56Z
• Authors: Nianzong Hou; Mingzhe Li; Lu He; Bing Xie; Lin Wang; Rumin Zhang; Yong Yu; Xiaodong Sun; Zhengsheng Pan; Kai Wang
• Keywords: Nomogram; Stepwise regression

Abstract Background Sepsis is a significant cause of mortality in-hospital, especially in ICU patients. Early prediction sepsis essential, as prompt and appropriate treatment can improve survival outcomes. Machine learning methods are flexible algorithms with potential advantages over conventional regression scoring system. The aims this study were to develop machine approach using XGboost predict the 30-days for MIMIC-III Patients sepsis-3 determine whether such model performs better than traditional models. Methods Using v1.4, we identified patients sepsis-3. data was split into two groups based on death or within 30 days variables, selected clinical significance availability by stepwise analysis, displayed compared between groups. Three predictive models including logistic model, SAPS-II score XGBoost algorithm constructed R software. Then, performances three tested AUCs receiver operating characteristic curves decision curve analysis. At last, nomogram impact used validate model. Results A total 4559 included study, which, 889 3670 days, respectively. According results (0.819 [95% CI 0.800–0.838], 0.797 0.781–0.813] 0.857 0.839–0.876]) analysis models, best. risk verify that possesses value. Conclusions technique XGboost, more be built. This may prove clinically useful assist clinicians tailoring precise management therapy