From: Using machine learning methods to predict in-hospital mortality of sepsis patients in the ICU
Authors | Title | Dataset | Methodology | Predictors | Outcome | Sepsis definition |
---|---|---|---|---|---|---|
Masson, S. et al. [19] | Presepsin (soluble CD14 subtype) and procalcitonin levels for mortality prediction in sepsis: data from the Albumin Italian Outcome Sepsis trial | A multicentre, randomised Albumin Italian Outcome Sepsis trial, 100 patients | Cox regression model | Presepsin level, procalcitonin level and some covariates | 28-day/ICU/90-day mortality | Sepsis-2 |
Adrie C. et al. [20] | Model for predicting short-term mortality of severe sepsis | A multicentre database including data from 12 ICUs, 2268 patients | Generalised linear model | SAPS II and LOD scores at ICU admission, septic shock, multiple organ failure, comorbidities, procedures, agents, bacteraemia and sources of infection | 14-day mortality within ICU stay | Sepsis-2 |
Ripoll, V.J.R. et al. [21] | Sepsis mortality prediction with the Quotient Basis Kernel | MIMIC II | Support vector machines (SVMs), LR, SAPS | SOFA and SAPS scores at ICU admission | ICU mortality | Sepsis-2 |
Fang W-F et al. [22] | Development and validation of immune dysfunction score to predict 28-day mortality of sepsis patients | Sepsis patients admitted to ICU at a hospital in Taiwan, 151 patients | LR | Monocyte HLA-DR* expression, plasma G-CSF* level, plasma IL*-10 level, and serum SeMo* ratio | 28-day mortality | Sepsis-3 |
Xie, Y. et al. [23] | Using clinical features and biomarkers to predict 60-day mortality of sepsis patients | Protocol-based care in early septic shock trial, around 530 patients | LR | Clinical features and biomarkers obtained during the first 24 h of hospital admission | 60-day mortality | Not mentioned |
Poucke, S.V. et al. [31] | Scalable predictive analysis in critically ill patients using a visual open data analysis platform | MIMIC II | Naïve Bayes, LR, RF, AdaBoost, Bagging, Stacking, SVM | Demographics, comorbidities, types of care unit, platelet count | ICU mortality | NA |
Zhang, Z. & Hong, Y [32]. | Development of a novel score for the prediction of hospital mortality in patients with severe sepsis: the use of electronic healthcare records with LASSO regression | MIMIC III | LASSO, LR | Demographics, clinical and laboratory variables recorded during the first 24 h in ICU | Hospital mortality | Sepsis-2 |
Taylor, R.A. et al. [34] | Prediction of in-hospital mortality in emergency department patients with sepsis: a local big data-driven, machine learning approach | Adult ED* visits over 12 months, 4676 patients | RF, CART, LR | Demographics, previous health status, ED health status, ED services rendered and operational details | Hospital mortality | Sepsis-2 |
Pregernig, A. et al. [35] | Prediction of mortality in adult patients with sepsis using six biomarkers: a systematic review and meta-analysis | 44 articles in English | Qualitative analysis, meta-analysis | Angiopoietin 1 and 2, HMGB1*, sRAGE*, sTREM*-1, suPAR* | 28-day/30-day/ICU/hospital/90-day mortality | Sepsis-1/Sepsis-2/Sepsis-3 |