Dynamic Prediction of ICU Mortality Risk Using Domain Adaptation

Abstract Background The identification of factors associated with Intensive Care Unit (ICU) mortality and derived clinical phenotypes in COVID-19 patients could help for a more tailored approach to clinical decision-making that improves prognostic outcomes. Methods Prospective, multicenter, observational study of critically ill patients with confirmed COVID-19 disease and acute respiratory failure admitted from 63 ICUs in Spain. The objective was to utilize an unsupervised clustering analysis to derive clinical COVID-19 phenotypes and to analyze patient’s factors associated with mortality risk. Patient features including demographics and clinical data at ICU admission were analyzed. Generalized linear models were used to determine ICU morality risk factors. The prognostic models were validated and their performance was measured using accuracy test, sensitivity, specificity and ROC curves. Results The database included a total of 2022 patients (mean age 64 [IQR 5–71] years, 1423 (70.4%) male, median APACHE II score (13 [IQR 10–17]) and SOFA score (5 [IQR 3–7]) points. The ICU mortality rate was 32.6%. Of the 3 derived phenotypes, the A (mild) phenotype (537; 26.7%) included older age (< 65 years), fewer abnormal laboratory values and less development of complications, B (moderate) phenotype (623, 30.8%) had similar characteristics of A phenotype but were more likely to present shock. The C (severe) phenotype was the most common (857; 42.5%) and was characterized by the interplay of older age (> 65 years), high severity of illness and a higher likelihood of development shock. Crude ICU mortality was 20.3%, 25% and 45.4% for A, B and C phenotype respectively. The ICU mortality risk factors and model performance differed between whole population and phenotype classifications. Conclusion The presented machine learning model identified three clinical phenotypes that significantly correlated with host-response patterns and ICU mortality. Different risk factors across the whole population and clinical phenotypes were observed which may limit the application of a “one-size-fits-all” model in practice.

Download Full-text

Learning to Predict in-hospital mortality risk in intensive care unit with attention-based Temporal convolution network

10.21203/rs.3.rs-44310/v2 ◽

2020 ◽

Author(s):

Yu-wen Chen ◽

Yu-jie Li ◽

Zhi-yong Yang ◽

Kun-hua Zhong ◽

Li-ge Zhang ◽

...

Keyword(s):

Time Series ◽

Mortality Risk ◽

Time Series Data ◽

Risk Model ◽

Series Data ◽

Dynamic Prediction ◽

Physiological Variables ◽

Prediction Of Mortality ◽

Low Sensitivity ◽

Mimic Iii

Abstract Background: Dynamic prediction of patients’ mortality risk in ICU with time series data is limited due to the high dimensionality, uncertainty with sampling intervals, and other issues. New deep learning method, temporal convolution network (TCN), makes it possible to deal with complex clinical time series data in ICU. We aimed to develop and validate it to predict mortality risk using time series data from MIMIC III dataset. Methods: 21139 records of ICU stays were analyzed and in total 17 physiological variables from the MIMIC III dataset were used to predict mortality risk. Then we compared the model performances of attention-based TCN with traditional artificial intelligence (AI) method. Results: Area Under Receiver Operating Characteristic (AUCROC) and Area Under Precision-Recall curve (AUC-PR) of attention-based TCN for predicting the mortality risk 48h after ICU admission were 0.837(0.824 -0.850) and 0.454. The sensitivity and specificity of attention-based TCN were 67.1% and 82.6%, compared to the traditional AI method yield low sensitivity (<50%). Conclusions: Attention-based TCN model achieved better performance in prediction of mortality risk with time series data than traditional AI methods and conventional score-based models. Attention-based TCN mortality risk model has the potential for helping decision-making in critical patients.

Download Full-text

Comparison of mortality risk evaluation tools efficacy in critically ill COVID-19 patients

BMC Infectious Diseases ◽

10.1186/s12879-021-06866-2 ◽

2021 ◽

Vol 21 (1) ◽

Author(s):

Vaidas Vicka ◽

Elija Januskeviciute ◽

Sigute Miskinyte ◽

Donata Ringaitiene ◽

Mindaugas Serpytis ◽

...

Keyword(s):

Regression Analysis ◽

Critically Ill ◽

Mortality Risk ◽

Multivariate Regression Analysis ◽

Apache Ii ◽

Risk Scores ◽

Icu Mortality ◽

Icu Patients ◽

Point Increase ◽

Number Of Patients

Abstract Background As the COVID-19 pandemic continues, the number of patients admitted to the intensive care unit (ICU) is still increasing. The aim of our article is to estimate which of the conventional ICU mortality risk scores is the most accurate at predicting mortality in COVID-19 patients and to determine how these scores can be used in combination with the 4C Mortality Score. Methods This was a retrospective study of critically ill COVID-19 patients treated in tertiary reference COVID-19 hospitals during the year 2020. The 4C Mortality Score was calculated upon admission to the hospital. The Simplified Acute Physiology Score (SAPS) II, Acute Physiology and Chronic Health Evaluation (APACHE) II, and Sequential Organ Failure Assessment (SOFA) scores were calculated upon admission to the ICU. Patients were divided into two groups: ICU survivors and ICU non-survivors. Results A total of 249 patients were included in the study, of which 63.1% were male. The average age of all patients was 61.32 ± 13.3 years. The all-cause ICU mortality ratio was 41.4% (n = 103). To determine the accuracy of the ICU mortality risk scores a ROC-AUC analysis was performed. The most accurate scale was the APACHE II, with an AUC value of 0.772 (95% CI 0.714–0.830; p < 0.001). All of the ICU risk scores and 4C Mortality Score were significant mortality predictors in the univariate regression analysis. The multivariate regression analysis was completed to elucidate which of the scores can be used in combination with the independent predictive value. In the final model, the APACHE II and 4C Mortality Score prevailed. For each point increase in the APACHE II, mortality risk increased by 1.155 (OR 1.155, 95% CI 1.085–1.229; p < 0.001), and for each point increase in the 4C Mortality Score, mortality risk increased by 1.191 (OR 1.191, 95% CI 1.086–1.306; p < 0.001), demonstrating the best overall calibration of the model. Conclusions The study demonstrated that the APACHE II had the best discrimination of mortality in ICU patients. Both the APACHE II and 4C Mortality Score independently predict mortality risk and can be used concomitantly.

Download Full-text

Differences in pediatric ICU mortality risk over time

Critical Care Medicine ◽

10.1097/00003246-199810000-00032 ◽

1998 ◽

Vol 26 (10) ◽

pp. 1737-1743 ◽

Cited By ~ 48

Author(s):

John M. Tilford ◽

Paula K. Roberson ◽

Shelly Lensing ◽

Debra H. Fiser

Keyword(s):

Mortality Risk ◽

Icu Mortality ◽

Pediatric Icu ◽

Over Time

Download Full-text

Prognostic Factors to Predict ICU Mortality in Patients with ARDS Who Received Early and Prolonged Prone Positioning Therapy

10.21203/rs.3.rs-247253/v1 ◽

2021 ◽

Author(s):

Po-Hsin Lee ◽

Pin-Kuei Fu

Keyword(s):

Intensive Care Unit ◽

Intensive Care ◽

Prognostic Factors ◽

Mortality Risk ◽

Risk Group ◽

Apache Ii ◽

Apache Ii Score ◽

Prone Positioning ◽

Icu Mortality ◽

Factors Associated

Abstract Background: Early and prolonged prone positioning (PP) could reduce the mortality in patients with moderate to severe ARDS, however, factors associated with mortality in the intensive care unit (ICU) remain unclear. The aim of this study is to identified factors associated with mortality and create the prognostic score in patients with ARDS who underwent early and prolonged PP. Methods: This retrospective study included patients with moderate to severe ARDS admitted to the intensive care unit (ICU) from January 2015 to June 2018 in a tertiary referral center in Taiwan and who received early and prolonged PP. Demographic data, disease severity score, comorbidities, and clinical outcomes were recorded. Univariate and multivariate regression models were used to estimate the odds ratio (OR) of ICU mortality. Receiver operating characteristic (ROC) curve analysis were performed to identify the cutoff value of parameters. Results: A total of 116 patients were enrolled. In the multivariate analysis, three factors were significantly associated with mortality: renal replacement therapy (RRT; OR: 3.38, 1.55–7.36), malignant comorbidity (OR: 7.42, 2.06–26.70), and noninfluenza-related ARDS (OR: 3.78, 1.07–13.29). Age, RRT, noninfluenza-related ARDS, malignant comorbidity, and APACHE II score were included in a composite prone score, which demonstrated an area under the curve of 0.816 for predicting mortality risk. The mortality risk in ICU was 27.1% in the low-risk group (prone score: 0–2) and 84.2% in the high-risk group (prone score: 3–5). Conclusions: For patients with moderate to severe ARDS even receiving early and prolonged PP in ICU, poor prognostic factors were age, RRT, malignant comorbidity, noninfluenza-related ARDS, and higher APACHE II score. High mortality should be informed to the family of patient if their prone score was more than 3 points.

Download Full-text

Deploying unsupervised clustering analysis to derive clinical phenotypes and risk factors associated with mortality risk in 2,022 critically ill patients with COVID-19 in Spain

10.21203/rs.3.rs-125422/v2 ◽

2021 ◽

Author(s):

Alejandro Rodríguez ◽

Manuel Ruiz Botella ◽

Ignacio Matín-Loeches ◽

María Jiménez Herrera ◽

Jordi Solé-Violan ◽

...

Keyword(s):

Risk Factors ◽

Intensive Care ◽

Critically Ill ◽

Mortality Risk ◽

Critically Ill Patients ◽

Clustering Analysis ◽

Older Age ◽

Icu Mortality ◽

Factors Associated ◽

Clinical Phenotypes

Abstract Background: The identification of factors associated with Intensive Care Unit (ICU) mortality and derived clinical phenotypes in COVID-19 patients could help for a more tailored approach to clinical decision-making that improves prognostic outcomes. Methods: Prospective, multicenter, observational study of critically ill patients with confirmed COVID-19 disease and acute respiratory failure admitted from 63 Intensive Care Units(ICU) in Spain. The objective was to utilize an unsupervised clustering analysis to derive clinical COVID-19 phenotypes and to analyze patient’s factors associated with mortality risk. Patient features including demographics and clinical data at ICU admission were analyzed. Generalized linear models were used to determine ICU morality risk factors. The prognostic models were validated and their performance was measured using accuracy test, sensitivity, specificity and ROC curves. Results: The database included a total of 2,022 patients (mean age 64[IQR5-71] years, 1423(70.4%) male, median APACHE II score (13[IQR10-17]) and SOFA score (5[IQR3-7]) points. The ICU mortality rate was 32.6%. Of the 3 derived phenotypes, the A(mild) phenotype (537;26.7%) included older age (<65 years), fewer abnormal laboratory values and less development of complications, B (moderate) phenotype (623,30.8%) had similar characteristics of A phenotype but were more likely to present shock. The C(severe) phenotype was the most common (857;42.5%) and was characterized by the interplay of older age (>65 years), high severity of illness and a higher likelihood of development shock. Crude ICU mortality was 20.3%, 25% and 45.4% for A, B and C phenotype respectively. The ICU mortality risk factors and model performance differed between whole population and phenotype classifications.Conclusion: The presented machine learning model identified three clinical phenotypes that significantly correlated with host-response patterns and ICU mortality. Different risk factors across the whole population and clinical phenotypes were observed which may limit the application of a “one-size-fits-all” model in practice.

Download Full-text

The protective association of endogenous immunoglobulins against sepsis mortality is restricted to patients with moderate organ failure

10.1101/116665 ◽

2017 ◽

Author(s):

Ignacio Martin-Loeches ◽

Arturo Muriel-Bombín ◽

Ricard Ferrer ◽

Antonio Artigas ◽

Jordi Sole-Violan ◽

...

Keyword(s):

Hospital Mortality ◽

Organ Failure ◽

Mortality Risk ◽

Multivariate Logistic Regression Analysis ◽

Predictive Ability ◽

Simultaneous Presence ◽

Icu Mortality ◽

Icu Admission ◽

The Impact ◽

Immunoglobulin Levels

AbstractBackgroundpre-evaluation of endogenous immunoglobulin levels is a potential strategy to improve the results of intravenous immunoglobulins in sepsis, but more work has to be done to identify those patients who could benefit the most from this treatment. The objective of this study was to evaluate the impact of endogenous immunoglobulins on the mortality risk in sepsis depending on disease severity.Methodsthis was a retrospective observational study including 278 patients admitted to the ICU with sepsis fulfilling the SEPSIS-3 criteria, coming from the Spanish GRECIA and ABISS-EDUSEPSIS cohorts. Patients were distributed into two groups depending on their Sequential Organ Failure Assessment score al ICU admission (SOFA < 8, n = 122 and SOFA ≥ 8, n = 156) and the association between immunoglobulin levels at ICU admission with mortality was studied in each group by Kaplan Meier and multivariate logistic regression analysis.ResultsICU / hospital mortality in the SOFA < 8 group was 14.8% / 23.0%, compared to 30.1 % / 35.3% in the SOFA ≥ 8 group. In the group with SOFA < 8, the simultaneous presence of total IgG <407 mg/dl, IgM < 43 mg/dl and IgA < 219 mg/dl was associated to a reduction in the survival mean time of 6.6 days in the first 28 days, and was a robust predictor of mortality risk either during the acute and the post-acute phase of the disease (OR for ICU mortality: 13.79; OR for hospital mortality: 7.98). This predictive ability remained in the absence of prior immunosupression (OR for ICU mortality: 17.53; OR for hospital mortality: 5.63). Total IgG <407 mg/dl or IgG1 < 332 mg/dl was also an independent predictor of ICU mortality in this group. In contrast, in the SOFA ≥ 8 group, we found no immunoglobulin thresholds associated to neither ICU nor to hospital mortality.Conclusionsendogenous immunoglobulin levels may have a different impact on the mortality risk of sepsis patients based on their severity. In patients with moderate organ failure, the simultaneous presence of low levels of IgG, IgA and IgM was a consistent predictor of both acute and post-acute mortality.

Download Full-text

Endothelial, Immunothrombotic, and Inflammatory Biomarkers in the Risk of Mortality in Critically Ill COVID-19 Patients: The Role of Dexamethasone

Diagnostics ◽

10.3390/diagnostics11071249 ◽

2021 ◽

Vol 11 (7) ◽

pp. 1249

Author(s):

Chrysi Keskinidou ◽

Alice G. Vassiliou ◽

Alexandros Zacharis ◽

Edison Jahaj ◽

Parisis Gallos ◽

...

Keyword(s):

Critically Ill ◽

Mortality Risk ◽

Roc Curves ◽

Mortality Prediction ◽

Prognostic Biomarkers ◽

Inflammatory Biomarkers ◽

Rank Test ◽

Icu Mortality ◽

Icu Admission ◽

Kaplan Meier

Endothelial dysfunction, coagulation and inflammation biomarkers are increasingly emerging as prognostic markers of poor outcomes and mortality in severe and critical COVID-19. However, the effect of dexamethasone has not been investigated on these biomarkers. Hence, we studied potential prognostic biomarkers of mortality in critically ill COVID-19 patients who had either received or not dexamethasone. Biomarker serum levels were measured on intensive care unit (ICU) admission (within 24 h) in 37 dexamethasone-free and 29 COVID-19 patients who had received the first dose (6 mg) of dexamethasone. Receiver operating characteristic (ROC) curves were generated to assess their value in ICU mortality prediction, while Kaplan–Meier analysis was used to explore associations between biomarkers and survival. In the dexamethasone-free COVID-19 ICU patients, non-survivors had considerably higher levels of various endothelial, immunothrombotic and inflammatory biomarkers. In the cohort who had received one dexamethasone dose, non-survivors had higher ICU admission levels of only soluble (s) vascular cell adhesion molecule-1 (VCAM-1), soluble urokinase-type plasminogen activator receptor (suPAR) and presepsin. As determined from the generated ROC curves, sVCAM-1, suPAR and presepsin could still be reliable prognostic ICU mortality biomarkers, following dexamethasone administration (0.7 < AUC < 0.9). Moreover, the Kaplan–Meier survival analysis showed that patients with higher than the median values for sVCAM-1 or suPAR exhibited a greater mortality risk than patients with lower values (Log-Rank test, p < 0.01). In our single-center study, sVCAM-1, suPAR and presepsin appear to be valuable prognostic biomarkers in assessing ICU mortality risk in COVID-19 patients, even following dexamethasone administration.

Download Full-text