scholarly journals Predicting Duration of Mechanical Ventilation in Acute Respiratory Distress Syndrome Using Supervised Machine Learning

2021 ◽  
Vol 10 (17) ◽  
pp. 3824
Author(s):  
Mohammed Sayed ◽  
David Riaño ◽  
Jesús Villar
Keyword(s):  
Machine Learning ◽  
Mechanical Ventilation ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Distress Syndrome ◽  
Supervised Machine Learning ◽  
Gradient Boosting ◽  
Mimic Iii ◽  
Over Time

Background: Acute respiratory distress syndrome (ARDS) is an intense inflammatory process of the lungs. Most ARDS patients require mechanical ventilation (MV). Few studies have investigated the prediction of MV duration over time. We aimed at characterizing the best early scenario during the first two days in the intensive care unit (ICU) to predict MV duration after ARDS onset using supervised machine learning (ML) approaches. Methods: For model description, we extracted data from the first 3 ICU days after ARDS diagnosis from patients included in the publicly available MIMIC-III database. Disease progression was tracked along those 3 ICU days to assess lung severity according to Berlin criteria. Three robust supervised ML techniques were implemented using Python 3.7 (Light Gradient Boosting Machine (LightGBM); Random Forest (RF); and eXtreme Gradient Boosting (XGBoost)) for predicting MV duration. For external validation, we used the publicly available multicenter database eICU. Results: A total of 2466 and 5153 patients in MIMIC-III and eICU databases, respectively, received MV for >48 h. Median MV duration of extracted patients was 6.5 days (IQR 4.4–9.8 days) in MIMIC-III and 5.0 days (IQR 3.0–9.0 days) in eICU. LightGBM was the best model in predicting MV duration after ARDS onset in MIMIC-III with a root mean square error (RMSE) of 6.10–6.41 days, and it was externally validated in eICU with RMSE of 5.87–6.08 days. The best early prediction model was obtained with data captured in the 2nd day. Conclusions: Supervised ML can make early and accurate predictions of MV duration in ARDS after onset over time across ICUs. Supervised ML models might have important implications for optimizing ICU resource utilization and high acute cost reduction of MV.

Dynamic prediction of ICU duration of mechanical ventilation in Acute Respiratory Distress Syndrome: A machine learning approach (Preprint)

2020 ◽  
Author(s):  
Mohammed Sayed ◽  
David RIAÑO
Keyword(s):  
Machine Learning ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Distress Syndrome ◽  
Total Duration ◽  
Gradient Boosting ◽  
Dynamic Prediction ◽  
Extreme Gradient Boosting ◽  
Over Time

BACKGROUND Acute respiratory distress syndrome (ARDS) is a pulmonary failure, commonly observed and associated with significant mortality and cost in intensive care units (ICUs). However, only few studies investigated prediction over time, in particular for more than two ICU days. OBJECTIVE Our study aimed at characterization of the best dynamic scenario in terms of the first two ICU days to predict the total duration of MV for the whole treatment in ARDS using three Machine learning (ML) techniques. METHODS We used data of 1,341 ARDS patients of MIMIC III database (MetaVision, 2008- 2012) according to the Berlin definition for ARDS severity. Patients’ evolution was tracked along the first 3-ICU days after ARDS onset to assess "true" severity. Three ML classification techniques were evaluated: Light Gradient Boosting Machine (LightGBM), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost). RESULTS LightGBM models outperformed RF and XGBoost in predicting the total duration of MV for the whole treatment in ARDS over time (RMSE 146.59 -152.08 hours). CONCLUSIONS Dynamic prediction of the total duration of MV for the whole treatment in ARDS can be provided using ML techniques when "true" severity of ARDS patient is considered. Consequently, this leads to better resource allocation in ICU.

scholarly journals Supervised machine learning for the early prediction of acute respiratory distress syndrome (ARDS)

2020 ◽  
Vol 60 ◽  
pp. 96-102 ◽  
Author(s):  
Sidney Le ◽  
Emily Pellegrini ◽  
Abigail Green-Saxena ◽  
Charlotte Summers ◽  
Jana Hoffman ◽  
...  
Keyword(s):  
Machine Learning ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Distress Syndrome ◽  
Supervised Machine Learning ◽  
Early Prediction

scholarly journals Supervised Machine Learning for the Early Prediction of Acute Respiratory Distress Syndrome (ARDS)

2020 ◽  
Author(s):  
Sidney Le ◽  
Emily Pellegrini ◽  
Abigail Green-Saxena ◽  
Charlotte Summers ◽  
Jana Hoffman ◽  
...  
Keyword(s):  
Machine Learning ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Distress Syndrome ◽  
Supervised Machine Learning ◽  
Early Prediction ◽  
Radiology Reports ◽  
Tree Models ◽  
Boosted Tree

ABSTRACTPurposeAcute respiratory distress syndrome (ARDS) is a serious respiratory condition with high mortality and associated morbidity. The objective of this study is to develop and evaluate a novel application of gradient boosted tree models trained on patient health record data for the early prediction of ARDS.Materials and Methods9919 patient encounters were retrospectively analyzed from the Medical Information Mart for Intensive Care III (MIMIC-III) data base. XGBoost gradient boosted tree models for early ARDS prediction were created using routinely collected clinical variables and numerical representations of radiology reports as inputs. XGBoost models were iteratively trained and validated using 10-fold cross validation.ResultsOn a hold-out test set, algorithm classifiers attained area under the receiver operating characteristic curve (AUROC) values of 0.905, 0.827, 0.810, and 0.790 when tested for the prediction of ARDS at 0-, 12-, 24-, and 48-hour windows prior to onset, respectively.ConclusionSupervised machine learning predictions may help predict patients with ARDS up to 48 hours prior to onset.

A Ventilator-associated Pneumonia Prediction Model in Patients With Acute Respiratory Distress Syndrome

2020 ◽  
Vol 71 (Supplement_4) ◽  
pp. S400-S408
Author(s):  
Zongsheng Wu ◽  
Yao Liu ◽  
Jingyuan Xu ◽  
Jianfeng Xie ◽  
Shi Zhang ◽  
...  
Keyword(s):  
Mechanical Ventilation ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Distress Syndrome ◽  
Neuromuscular Blocking ◽  
Calibration Plot ◽  
Binary Logistic Regression Analysis ◽  
Ventilator Associated Pneumonia ◽  
Mechanical Ventilation Duration

Abstract Background Mechanical ventilation is crucial for acute respiratory distress syndrome (ARDS) patients and diagnosis of ventilator-associated pneumonia (VAP) in ARDS patients is challenging. Hence, an effective model to predict VAP in ARDS is urgently needed. Methods We performed a secondary analysis of patient-level data from the Early versus Delayed Enteral Nutrition (EDEN) of ARDSNet randomized controlled trials. Multivariate binary logistic regression analysis established a predictive model, incorporating characteristics selected by systematic review and univariate analyses. The model’s discrimination, calibration, and clinical usefulness were assessed using the C-index, calibration plot, and decision curve analysis (DCA). Results Of the 1000 unique patients enrolled in the EDEN trials, 70 (7%) had ARDS complicated with VAP. Mechanical ventilation duration and intensive care unit (ICU) stay were significantly longer in the VAP group than non-VAP group (P < .001 for both) but the 60-day mortality was comparable. Use of neuromuscular blocking agents, severe ARDS, admission for unscheduled surgery, and trauma as primary ARDS causes were independent risk factors for VAP. The area under the curve of the model was .744, and model fit was acceptable (Hosmer-Lemeshow P = .185). The calibration curve indicated that the model had proper discrimination and good calibration. DCA showed that the VAP prediction nomogram was clinically useful when an intervention was decided at a VAP probability threshold between 1% and 61%. Conclusions The prediction nomogram for VAP development in ARDS patients can be applied after ICU admission, using available variables. Potential clinical benefits of using this model deserve further assessment.

Surfactant therapy in combination treatment of COVID-19 related acute respiratory distress syndrome

2021 ◽  
pp. 50-56
Author(s):  
Renat R. Gubaidullin ◽  
◽  
Aleksandr P. Kuzin ◽  
Vladimir V. Kulakov ◽  
◽  
...  
Keyword(s):  
Mechanical Ventilation ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Failure ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Standard Therapy ◽  
Distress Syndrome ◽  
Severe Pneumonia ◽  
Lung Ventilation ◽  
Surfactant Therapy

ntroduction. The COVID-19 pandemic caused an outbreak of viral lung infections with severe acute respiratory syndrome complicated with acute respiratory failure. Despite the fact that the pandemic has a lengthened run, none of the therapeutic approaches have proved to be sufficiently effective according to the evidence-based criteria. We consider the use of surfactant therapy in patients with severe viral pneumonia and acute respiratory distress syndrome (ARDS) as one of the possible methods for treating COVID-19 related pneumonia. Objective. To prove the clinical efficacy and safety of orally inhaled Surfactant-BL, an authorized drug, in the combination therapy of COVID-19 related ARDS. Materials and methods. A total of 38 patients with COVID-19 related severe pneumonia and ARDS were enrolled in the study. Of these, 20 patients received the standard therapy in accordance with the temporary guidelines for the prevention, diagnosis and treatment of the novel coronavirus infection (COVID-19) of the Ministry of Health of the Russian Federation, version 9. And 18 patients received the surfactant therapy in addition to the standard therapy. Surfactant-BL was used in accordance with the instructions on how to administer the drug for the indication – prevention of the development of acute respiratory distress syndrome. A step-by-step approach to the build-up of the respiratory therapy aggressiveness was used to manage hypoxia. We used oxygen inhalation via a face mask with an oxygen inflow of 5–15 l/min, highflow oxygen therapy via nasal cannulas using Airvo 2 devices, non-invasive lung ventilation, invasive lung ventilation in accordance with the principles of protective mechanical ventilation. Results and discussion. Significant differences in the frequency of transfers to mechanical ventilation, mortality, Intensive Care Unit (ICU) and hospitalization length of stay (p <0.05) were found between the groups. Patients receiving surfactant therapy who required a transfer to mechanical ventilation accounted for 22% of cases, and the mortality rate was 16%. In the group of patients receiving standard therapy without surfactant inhalation 45% were transferred to mechanical ventilation, and 35% died. For patients receiving surfactant therapy, the hospital stay was reduced by 20% on average, and ICU stay by 30%. Conclusion. The inclusion of surfactant therapy in the treatment of COVID-19 related severe pneumonia and ARDS can reduce the progression of respiratory failure, avoid the use of mechanical ventilation, shorten the ICU and hospitalization length of stay, and improve the survival rate of this patient cohort.

An Evidence-Based Protocol for Manual Prone Positioning of Patients With ARDS

2021 ◽  
Vol 41 (6) ◽  
pp. 55-60
Author(s):  
Patrick Ryan ◽  
Cynthia Fine ◽  
Christine DeForge
Keyword(s):  
Mechanical Ventilation ◽  
Acute Respiratory Distress Syndrome ◽  
Respiratory Distress Syndrome ◽  
Respiratory Distress ◽  
Distress Syndrome ◽  
Medical Center ◽  
Academic Medical Center ◽  
Prone Positioning ◽  
Hypoxemic Respiratory Failure ◽  
Unplanned Extubations

Background Manual prone positioning has been shown to reduce mortality among patients with moderate to severe acute respiratory distress syndrome, but it is associated with a high incidence of pressure injuries and unplanned extubations. This study investigated the feasibility of safely implementing a manual prone positioning protocol that uses a dedicated device. Review of Evidence A search of CINAHL and Medline identified multiple randomized controlled trials and meta-analyses that demonstrated both the reduction of mortality when prone positioning is used for more than 12 hours per day in patients with acute respiratory distress syndrome and the most common complications of this treatment. Implementation An existing safe patient-handling device was modified to enable staff to safely perform manual prone positioning with few complications for patients receiving mechanical ventilation. All staff received training on the protocol and use of the device before implementation. Evaluation This study included 36 consecutive patients who were admitted to the medical intensive care unit at a large academic medical center because of hypoxemic respiratory failure/acute respiratory distress syndrome and received mechanical ventilation and prone positioning. Data were collected on clinical presentation, interventions, and complications. Sustainability Using the robust protocol and the low-cost device, staff can safely perform a low-volume, high-risk maneuver. This method provides cost savings compared with other prone positioning methods. Conclusions Implementing a prone positioning protocol with a dedicated device is feasible, with fewer complications and lower costs than anticipated.

Sign in / Sign up

Export Citation Format

Share Document