scholarly journals Reinforcement learning assisted oxygen therapy for COVID-19 patients under intensive care

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Hua Zheng ◽  
Jiahao Zhu ◽  
Wei Xie ◽  
Judy Zhong
Keyword(s):  
Intensive Care ◽  
Reinforcement Learning ◽  
Mortality Rate ◽  
Critically Ill ◽  
Flow Rate ◽  
Critically Ill Patients ◽  
Cross Validation ◽  
Standard Of Care ◽  
Oxygen Flow Rate ◽  
Oxygen Flow

Abstract Background Patients with severe Coronavirus disease 19 (COVID-19) typically require supplemental oxygen as an essential treatment. We developed a machine learning algorithm, based on deep Reinforcement Learning (RL), for continuous management of oxygen flow rate for critically ill patients under intensive care, which can identify the optimal personalized oxygen flow rate with strong potentials to reduce mortality rate relative to the current clinical practice. Methods We modeled the oxygen flow trajectory of COVID-19 patients and their health outcomes as a Markov decision process. Based on individual patient characteristics and health status, an optimal oxygen control policy is learned by using deep deterministic policy gradient (DDPG) and real-time recommends the oxygen flow rate to reduce the mortality rate. We assessed the performance of proposed methods through cross validation by using a retrospective cohort of 1372 critically ill patients with COVID-19 from New York University Langone Health ambulatory care with electronic health records from April 2020 to January 2021. Results The mean mortality rate under the RL algorithm is lower than the standard of care by 2.57% (95% CI: 2.08–3.06) reduction (P < 0.001) from 7.94% under the standard of care to 5.37% under our proposed algorithm. The averaged recommended oxygen flow rate is 1.28 L/min (95% CI: 1.14–1.42) lower than the rate delivered to patients. Thus, the RL algorithm could potentially lead to better intensive care treatment that can reduce the mortality rate, while saving the oxygen scarce resources. It can reduce the oxygen shortage issue and improve public health during the COVID-19 pandemic. Conclusions A personalized reinforcement learning oxygen flow control algorithm for COVID-19 patients under intensive care showed a substantial reduction in 7-day mortality rate as compared to the standard of care. In the overall cross validation cohort independent of the training data, mortality was lowest in patients for whom intensivists’ actual flow rate matched the RL decisions.

Inhaled Epoprostenol Through Noninvasive Routes of Ventilator Support Systems

2018 ◽  
Vol 52 (12) ◽  
pp. 1173-1181 ◽  
Author(s):  
Mahmoud A. Ammar ◽  
Madhu Sasidhar ◽  
Simon W. Lam
Keyword(s):  
Critically Ill ◽  
Flow Rate ◽  
Critically Ill Patients ◽  
Support Systems ◽  
Oxygen Flow Rate ◽  
Oxygen Flow ◽  
Arterial Oxygen ◽  
Positive Pressure ◽  
Ventilator Support ◽  
Respiratory Status

Background: The administration of inhaled epoprostenol (iEPO) through noninvasive routes of ventilator support systems has never been previously evaluated. Objective: Describe the use of iEPO when administered through noninvasive routes of ventilator support systems. Methods: Critically ill patients admitted to the intensive care unit who received iEPO through noninvasive routes were analyzed. Improvements in respiratory status and hemodynamic parameters were evaluated. Safety end points assessed included hypotension, rebound hypoxemia, significant bleeding, and thrombocytopenia. Results: A total of 36 patients received iEPO through noninvasive routes: high-flow oxygen therapy through nasal cannula, n = 29 (81%) and noninvasive positive-pressure ventilation, n = 7 (19%). Sixteen patients had improvement in their respiratory status: mean decrease in fraction of inspired oxygen (FiO2), 20% ± 13%; mean increase in partial pressure of arterial oxygen to FiO2 (PaO2/FiO2) ratio, 60 ± 50 mm Hg; and mean decrease in HFNC oxygen flow rate, 6 ± 3 liters per minute (LPM). Eight patients had declines in their respiratory status (mean increase in FiO2, 30% ± 20%; mean decrease in PaO2/FiO2 ratio, 38 ± 20 mm Hg; and mean increase in HFNC oxygen flow rate, 15 ± 10 LPM), and 12 patients had no change in their respiratory status. Conclusion and Relevance: This represents the first evaluation of the administration of iEPO through noninvasive routes of ventilator support systems and demonstrates that in critically ill patients, iEPO could be administered through a noninvasive route. Further evaluation is needed to determine the extent of benefit with this route of administration.

scholarly journals N-terminal pro-brain natriuretic peptide and high-sensitivity troponin T exhibit additive prognostic value for the outcome of critically ill patients

2018 ◽  
Vol 9 (5) ◽  
pp. 496-503 ◽  
Author(s):  
Max Lenz ◽  
Konstantin A Krychtiuk ◽  
Georg Goliasch ◽  
Klaus Distelmaier ◽  
Johann Wojta ◽  
...  
Keyword(s):  
Intensive Care ◽  
Mortality Rate ◽  
Brain Natriuretic Peptide ◽  
Critically Ill ◽  
Natriuretic Peptide ◽  
Critically Ill Patients ◽  
Plasma Levels ◽  
Troponin T ◽  
High Sensitivity ◽  
The Individual

Background: Patients treated at medical intensive care units suffer from various pathologies and often present with elevated troponin T (TnT) and N-terminal pro-brain natriuretic peptide (NT-proBNP) levels. Both markers may reflect different forms of cardiac involvement in critical illness. Therefore, the aim of our study was to examine the synergistic prognostic potential of NT-proBNP and high-sensitivity TnT (hs)TnT in unselected critically ill patients. Methods: We included all consecutive patients admitted to our intensive care unit within one year, excluding those suffering from acute myocardial infarction or undergoing cardiac surgery and measured NT-proBNP and TnT plasma levels on the day of admission and 72 hours thereafter. Results: Of the included 148 patients, 52% were male, mean age was of 64.2 ± 16.8 years and 30-day mortality was 33.2%. Non-survivors showed significantly higher NT-proBNP and TnT plasma levels as compared with survivors ( p<0.01). An elevation of both markers exhibited an additive effect on mortality, as those with both NT-proBNP and TnT levels above the median had a 30-day mortality rate of 51.0%, while those with both markers below the median had a 16.7% mortality rate (hazard ratio 3.7). These findings were independent of demographic and clinical parameters ( p<0.05). Conclusions: Our findings regarding the individual predictive properties of NT-proBNP and TnT are in line with literature. However, we were able to highlight that they exhibit additive prognostic potential which exceeds their individual value. This might be attributed to a difference in underlying pathomechanisms and an assessment of synergistic risk factors.

Outcomes of Critically Ill Gynecological Cancer Patients Admitted to Intensive Care Unit

2012 ◽  
Vol 30 (1) ◽  
pp. 7-11 ◽  
Author(s):  
Silvio A. Ñamendys-Silva ◽  
María O. González-Herrera ◽  
Julia Texcocano-Becerra ◽  
Angel Herrera-Gómez
Keyword(s):  
Intensive Care Unit ◽  
Intensive Care ◽  
Prognostic Factors ◽  
Mortality Rate ◽  
Hospital Mortality ◽  
Critically Ill ◽  
Critically Ill Patients ◽  
Gynecological Cancer ◽  
Apache Ii ◽  
Apache Ii Score

Purpose: To assess the characteristics of critically ill patients with gynecological cancer, and to evaluate their prognosis. Methods: Fifty-two critically ill patients with gynecological cancer admitted to intensive care unit (ICU) were included. Univariate and multivariate logistic regressions were used to identify factors associated with hospital mortality. Results: Thirty-five patients (67.3%) had carcinoma of the cervix uteri and 11 (21.2%) had ovarian cancer. The mortality rate in the ICU was 17.3% (9 of 52) and hospital mortality rate were 23%(12 of 52). In the multivariate analysis, independent prognostic factors for hospital mortality were vasopressor use (odds ratio [OR] = 8.60, 95% confidence interval [CI] 2.05-36; P = .03) and the Acute Physiology and Chronic Health Evaluation (APACHE) II score (OR = 1.43, 95% CI 1.01-2.09; P = .048). Conclusions: The independent prognostic factors for hospital mortality were the need for vasopressors and the APACHE II score.

scholarly journals Insulin therapy for hyperglycemia in critically ill patients

2016 ◽  
Vol 53 (5) ◽  
pp. 250
Author(s):  
Julianti Julianti ◽  
Silvia Triratna ◽  
Aditiawati Aditiawati ◽  
Irfanuddin Irfanuddin
Keyword(s):  
Intensive Care Unit ◽  
Intensive Care ◽  
Blood Glucose ◽  
Mortality Rate ◽  
Insulin Therapy ◽  
Critically Ill ◽  
Organ Dysfunction ◽  
Critically Ill Patients ◽  
Therapy Group ◽  
Control Group

Background Hyperglycemia in critically ill patients is associated with higher mortality. Insulin therapy may improve outcomes, not only by preventing deleterious effects of hyperglycemia, but by improving the molecular dynamics in organ dysfunction.Objectives To assess the effects of insulin therapy on critically ill patients in an intensive care unit (ICU) setting and the risk of hypoglycemia.Methods An open-label, clinical trial was conducted in the Pediatric Intensive Care Unit (PICU) of Dr. Moh. Hoesin Hospital, Palembang, from November 2011 to March 2012. Subjects were consecutively assigned to receive either regular insulin at a dose of 0.05 U/kg/h if the blood glucose level reached >200 mg%, or standard therapy (control group). Blood glucose levels were measured hourly until they reached 80-110 mg%. Dose adjustments were made when the blood glucose level reached 145 mg%, by reducing the insulin dose to 0.025 U/kg/h. Outcomes of therapy were measured by Pediatric Logistic Organ Dysfunction (PELOD) score improvement, mortality rate and the occurrence of hypoglycemia.Results Forty subjects were enrolled in this study, with 20 subjects assigned to the insulin therapy group and 20 subjects to the standard therapy group. Two subjects, one from each group, were not included in the final analysis due to their deaths within 24 hours. There was no significant difference in distribution of PELOD scores before intervention between the groups (OR=0.5; 95%CI 0.1 to 1.9, P=0.32). However, after intervention, the PELOD scores was significantly lower in insulin therapy group compared to control group (OR 0.2; 95% CI 0.05 to 0.8, P=0.02). In the insulin group after intervention, fewer subjects had scores >20.5 and more subjects had scores ≤20.5, indicated a lower risk of organ dysfunction. There was also a significantly lower mortality rate in the insulin group compared to the control group (OR 0.2; 95% CI 0.05 to 0.8, P=0.02). None of the subjects suffered hypoglycemia.Conclusion Insulin is beneficial in improving organ dysfunction and decreasing mortality for critically ill patients.

scholarly journals Invasive Blood Pressure Measurement and In-hospital Mortality in Critically Ill Patients With Hypertension

2021 ◽  
Vol 8 ◽  
Author(s):  
Bin Zhou ◽  
Liang-Ying Lin ◽  
Xiao-Ai Liu ◽  
Ye-Sheng Ling ◽  
Yuan-Yuan Zhang ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Intensive Care Unit ◽  
Intensive Care ◽  
Mortality Rate ◽  
Confidence Interval ◽  
Hospital Mortality ◽  
Critically Ill ◽  
Critically Ill Patients ◽  
Odds Ratio ◽  
Invasive Blood Pressure

Background: Invasive blood pressure (IBP) measurement is common in the intensive care unit, although its association with in-hospital mortality in critically ill patients with hypertension is poorly understood.Methods and Results: A total of 11,732 critically ill patients with hypertension from the eICU-Collaborative Research Database (eICU-CRD) were enrolled. Patients were divided into 2 groups according to whether they received IBP. The primary outcome in this study was in-hospital mortality. Propensity score matching (PSM) and inverse probability of treatment weighing (IPTW) models were used to balance the confounding covariates. Multivariable logistic regression was used to evaluate the association between IBP measurement and hospital mortality. The IBP group had a higher in-hospital mortality rate than the no IBP group in the primary cohort [238 (8.7%) vs. 581 (6.5%), p &lt; 0.001]. In the PSM cohort, the IBP group had a lower in-hospital mortality rate than the no IBP group [187 (8.0%) vs. 241 (10.3%), p = 0.006]. IBP measurement was associated with lower in-hospital mortality in the PSM cohort (odds ratio, 0.73, 95% confidence interval, 0.59–0.92) and in the IPTW cohort (odds ratio, 0.81, 95% confidence interval, 0.67–0.99). Sensitivity analyses showed similar results in the subgroups with high body mass index and no sepsis.Conclusions: In conclusion, IBP measurement was associated with lower in-hospital mortality in critically ill patients with hypertension, highlighting the importance of IBP measurement in the intensive care unit.

scholarly journals Atrial fibrillation increased in-hospital and 4-year all-cause mortality in Critically ill patients with liver cirrhosis

2021 ◽  
Author(s):  
Yanting Luo ◽  
Bingyuan Wu ◽  
Yuankai Wu ◽  
Long Peng ◽  
Zexiong Li ◽  
...  
Keyword(s):  
Atrial Fibrillation ◽  
Liver Cirrhosis ◽  
Intensive Care ◽  
Mortality Rate ◽  
Critically Ill ◽  
Clinical Data ◽  
Critically Ill Patients ◽  
Cox Regression ◽  
Kaplan Meier ◽  
All Cause Mortality

Abstract ObjectiveThe purpose of this study was to use a large database that contains information on patient intensive care unit (ICU) admissions to study critically ill patients with cirrhosis and the relation with atrial fibrillation and short-term and 4-year mortality. MethodsThe Monitoring in Intensive Care Database III database was used to identify patients with cirrhosis hospitalized in an ICU from 2001 to 2012. Demographic and clinical data were extracted from the database. Clinical data and demographic information were collected for each patient in our study. Kaplan-Meier analysis and multivariate Cox regression models were performed to examine the relation between atrial fibrillation and in-hospital and 4-year all-cause mortality. ResultsA total of 1,481 patients (mean age 58 years, 68% male) with liver cirrhosis treated in an ICU were included in the analysis, and the prevalence of atrial fibrillation was 14.2%. The in-hospital all-cause mortality rate was 26.60%, and patients who had a significantly higher rate of atrial fibrillation (21.57% vs. 11.50%, P < 0.001). Multivariate analysis indicated that atrial fibrillation was significantly associated with in-hospital all-cause mortality (hazard ratio [HR] = 1.52, 95% confidence interval [CI]: 1.19 to 1.95; P < 0.001), and 4-year all-cause mortality (HR = 1.55, 95% CI: 1.12 to 2.13; P = 0.008). Kaplan-Meier survival analysis showed that patients with atrial fibrillation had a significantly higher in-hospital and 4-year all-cause mortality rate than patients without atrial fibrillation. ConclusionsCritically ill patients with liver cirrhosis have a significantly increased rate of atrial fibrillation, and the presence of atrial fibrillation is an independent risk for in-hospital and 4-year all-cause mortality.

scholarly journals Five novel clinical phenotypes for critically ill patients with mechanical ventilation in intensive care units: a retrospective and multi database study

2020 ◽  
Vol 21 (1) ◽  
Author(s):  
Longxiang Su ◽  
Zhongheng Zhang ◽  
Fanglan Zheng ◽  
Pan Pan ◽  
Na Hong ◽  
...  
Keyword(s):  
Machine Learning ◽  
Mechanical Ventilation ◽  
Intensive Care ◽  
Mortality Rate ◽  
Clinical Outcomes ◽  
Critically Ill ◽  
Critically Ill Patients ◽  
The Other ◽  
Clinical Phenotypes ◽  
Mimic Iii

Abstract Background Although protective mechanical ventilation (MV) has been used in a variety of applications, lung injury may occur in both patients with and without acute respiratory distress syndrome (ARDS). The purpose of this study is to use machine learning to identify clinical phenotypes for critically ill patients with MV in intensive care units (ICUs). Methods A retrospective cohort study was conducted with 5013 patients who had undergone MV and treatment in the Department of Critical Care Medicine, Peking Union Medical College Hospital. Statistical and machine learning methods were used. All the data used in this study, including demographics, vital signs, circulation parameters and mechanical ventilator parameters, etc., were automatically extracted from the electronic health record (EHR) system. An external database, Medical Information Mart for Intensive Care III (MIMIC III), was used for validation. Results Phenotypes were derived from a total of 4009 patients who underwent MV using a latent profile analysis of 22 variables. The associations between the phenotypes and disease severity and clinical outcomes were assessed. Another 1004 patients in the database were enrolled for validation. Of the five derived phenotypes, phenotype I was the most common subgroup (n = 2174; 54.2%) and was mostly composed of the postoperative population. Phenotype II (n = 480; 12.0%) led to the most severe conditions. Phenotype III (n = 241; 6.01%) was associated with high positive end-expiratory pressure (PEEP) and low mean airway pressure. Phenotype IV (n = 368; 9.18%) was associated with high driving pressure, and younger patients comprised a large proportion of the phenotype V group (n = 746; 18.6%). In addition, we found that the mortality rate of Phenotype IV was significantly higher than that of the other phenotypes. In this subgroup, the number of patients in the sequential organ failure assessment (SOFA) score segment (9,22] was 198, the number of deaths was 88, and the mortality rate was higher than 44%. However, the cumulative 28-day mortality of Phenotypes IV and II, which were 101 of 368 (27.4%) and 87 of 480 (18.1%) unique patients, respectively, was significantly higher than those of the other phenotypes. There were consistent phenotype distributions and differences in biomarker patterns by phenotype in the validation cohort, and external verification with MIMIC III further generated supportive results. Conclusions Five clinical phenotypes were correlated with different disease severities and clinical outcomes, which suggested that these phenotypes may help in understanding heterogeneity in MV treatment effects.

scholarly journals Anemia and Blood Transfusions in Critically Ill Patients

2012 ◽  
Vol 2012 ◽  
pp. 1-7 ◽  
Author(s):  
M. Kamran Athar ◽  
Nitin Puri ◽  
David R. Gerber
Keyword(s):  
Intensive Care Unit ◽  
Intensive Care ◽  
Mortality Rate ◽  
Blood Cell ◽  
Critically Ill ◽  
Critically Ill Patients ◽  
Blood Transfusions ◽  
The Past ◽  
Growing Body ◽  
The Impact

Anemia is common in critically ill patients. As a consequence packed red blood cell (PRBC) transfusions are frequent in the critically ill. Over the past two decades a growing body of literature has emerged, linking PRBC transfusion to infections, immunosuppression, organ dysfunction, and a higher mortality rate. However, despite growing evidence that risk of PRBC transfusion outweighs its benefit, significant numbers of critically ill patients still receive PRBC transfusion during their intensive care unit (ICU) stay. In this paper, we summarize the current literature concerning the impact of anemia on outcomes in critically ill patients and the potential complications of PRBC transfusions.

scholarly journals Prediction of mortality with unmeasured anions in critically ill patients on mechanical ventilation

2014 ◽  
Vol 71 (10) ◽  
pp. 936-941 ◽  
Author(s):  
Milos Novovic ◽  
Jasna Jevdjic
Keyword(s):  
Mechanical Ventilation ◽  
Intensive Care ◽  
Metabolic Acidosis ◽  
Mortality Rate ◽  
Critically Ill ◽  
Critically Ill Patients ◽  
Apache Ii ◽  
Strong Ion Difference ◽  
Acid Base ◽  
Prediction Of Mortality

Background/Aim. Acid-base disorders are common within critically ill patients. Physicochemical approach described by Stewart and modified by Figge gives precise quantification method of metabolic acidosis and insight into its main mechanisms, as well as influence of unmeasured anion on metabolic acidosis. The aims of this study were to determine whether the conventional acid-base variables are connected with survival rate of critically ill patients at Intensive care unit; whether strong ion difference/strong ion gap (SID/SIG) is a better predictor of mortality rate comparing to conventional acid-base variables; to determine all significant predictable parameters for the 28-day mortality rate at intensive care units. Methods. This retrospective observational analytic study included 142 adult patients requiring mechanical ventilation, survivors (n = 68) and nonsurvivors (n = 74). Apparent strong ion difference (SIDapp), effective strong ion difference (SIDeff) and SIG values were calculated with the Stewart-Figge?s quantitative biophysical method. Descriptive and analytical statistical methods were used in the study [t-test, Mann-Whitney U test, ?2-test, binary logistic regression, Reciever operating characteristic (ROC) curves, calibration]. Results. Age, Na+, acute physiology and chronic health evaluation (APACHE II), Cl-, albumin, SIG, SID app, SIDeff, and aninon gap (AG) were statistically significant predictors. AG represented a model with imprecise calibration, i.e. a model with little predictive power. APACHE II had p-value more than 0.05 if it was near it, and therefore it could be considered potentially unreliable for outcome prediction. SIDeff and SIG represented models with well-defined calibration. ROC analysis results showed that APACHE II, Cl-, albumin, SIDeff, SIG i AG had the largest area bellow the curve. By creation of logistic models with calibration methods, we found that outcome depends on SIG and APACHE II score. Conclusion. Based on our data, unmeasured anions provide prediction of mortality of critically ill patients on mechanical ventilation, unlike the traditional acid-base variables which are not accurate predictors of the 28-day mortality rate.

Sign in / Sign up

Export Citation Format

Share Document