scholarly journals Predicting High-Flow Nasal Cannula Failure in an ICU Using a Recurrent Neural Network with Transfer Learning and Input Data Perseveration: A Retrospective Analysis (Preprint)

10.2196/31760 ◽  
2021 ◽  
Author(s):  
George Pappy ◽  
Melissa Aczon ◽  
Randall Wetzel ◽  
David Ledbetter
Keyword(s):  
Neural Network ◽  
Retrospective Analysis ◽  
Transfer Learning ◽  
Recurrent Neural Network ◽  
Input Data ◽  
High Flow ◽  
Nasal Cannula ◽  
High Flow Nasal Cannula

HIGH FLOW NASAL CANNULA OXYGEN USE IN OCTOGENARIAN AND NONAGENARIAN PATIENTS: A RETROSPECTIVE ANALYSIS AMONG MICU PATIENTS

CHEST Journal ◽  
2018 ◽  
Vol 154 (4) ◽  
pp. 301A
Author(s):  
ARJUN SARADNA ◽  
BEN SHAMIAN ◽  
SHYAM SHANKAR ◽  
PRARTHNA CHANDAR ◽  
ANAND RAI ◽  
...  
Keyword(s):  
Retrospective Analysis ◽  
High Flow ◽  
Nasal Cannula ◽  
High Flow Nasal Cannula

Predicting High-Flow Nasal Cannula Failure in an ICU Using a Recurrent Neural Network with Transfer Learning and Input Data Perseveration: A Retrospective Analysis (Preprint)

2021 ◽  
Author(s):  
George Pappy ◽  
Melissa Aczon ◽  
Randall Wetzel ◽  
David Ledbetter
Keyword(s):  
Machine Learning ◽  
Transfer Learning ◽  
Airway Pressure ◽  
Input Data ◽  
Positive Airway Pressure ◽  
High Flow ◽  
Learning Models ◽  
High Flow Nasal Cannula ◽  
Non Invasive ◽  
Machine Learning Models

BACKGROUND High Flow Nasal Cannula (HFNC) provides non-invasive respiratory support for critically ill children who may tolerate it more readily than other Non-Invasive (NIV) techniques such as Bilevel Positive Airway Pressure (BiPAP) and Continuous Positive Airway Pressure (CPAP). Moreover, HFNC may preclude the need for mechanical ventilation (intubation). Nevertheless, NIV or intubation may ultimately be necessary for certain patients. Timely prediction of HFNC failure can provide an indication for increasing respiratory support. OBJECTIVE This work developed and compared machine learning models to predict HFNC failure. METHODS A retrospective study was conducted using the Virtual Pediatric Intensive Care Unit database of Electronic Medical Records (EMR) of patients admitted to a tertiary pediatric ICU from January 2010 to February 2020. Patients <19 years old, without apnea, and receiving HFNC treatment were included. A Long Short-Term Memory (LSTM) model using 517 variables (vital signs, laboratory data and other clinical parameters) was trained to generate a continuous prediction of HFNC failure, defined as escalation to NIV or intubation within 24 hours of HFNC initiation. For comparison, seven other models were trained: a Logistic Regression (LR) using the same 517 variables, another LR using only 14 variables, and five additional LSTM-based models using the same 517 variables as the first LSTM and incorporating additional ML techniques (transfer learning, input perseveration, and ensembling). Performance was assessed using the area under the receiver operating curve (AUROC) at various times following HFNC initiation. The sensitivity, specificity, positive and negative predictive values (PPV, NPV) of predictions at two hours after HFNC initiation were also evaluated. These metrics were also computed in a cohort with primarily respiratory diagnoses. RESULTS 834 HFNC trials [455 training, 173 validation, 206 test] met the inclusion criteria, of which 175 [103, 30, 42] (21.0%) escalated to NIV or intubation. The LSTM models trained with transfer learning generally performed better than the LR models, with the best LSTM model achieving an AUROC of 0.78, vs 0.66 for the 14-variable LR and 0.71 for the 517-variable LR, two hours after initiation. All models except for the 14-variable LR achieved higher AUROCs in the respiratory cohort than in the general ICU population. CONCLUSIONS Machine learning models trained using EMR data were able to identify children at risk for failing HFNC within 24 hours of initiation. LSTM models that incorporated transfer learning, input data perseveration and ensembling showed improved performance than the LR and standard LSTM models.

scholarly journals A Retrospective Analysis of COVID-19 ARDS Patients Successfully Extubated to Extubated to High-Flow Nasal Cannula Oxygen Therapy

2021 ◽  
Vol 0 (0) ◽  
pp. 0-0
Author(s):  
Gülçin Hilal Alay ◽  
Perihan Ergin Özcan ◽  
Özlem Polat ◽  
İlkay Anaklı ◽  
Günseli Orhun ◽  
...  
Keyword(s):  
Retrospective Analysis ◽  
Oxygen Therapy ◽  
High Flow ◽  
Nasal Cannula ◽  
High Flow Nasal Cannula

Failure of High Flow Nasal Cannula and Subsequent Intubation Is Associated With Increased Mortality as Compared to Failure of Non-Invasive Ventilation and Mechanical Ventilation Alone: A Real-World Retrospective Analysis

2020 ◽  
pp. 088506662096804
Author(s):  
David C. Miller ◽  
Jie Pu ◽  
David Kukafka ◽  
Christian Bime
Keyword(s):  
Mechanical Ventilation ◽  
Respiratory Failure ◽  
Retrospective Analysis ◽  
Real World ◽  
High Flow ◽  
Nasal Cannula ◽  
Invasive Ventilation ◽  
High Flow Nasal Cannula ◽  
Non Invasive ◽  
Non Invasive Ventilation

Background: Despite the increasing use of high flow nasal cannula oxygenation systems (HFNC) in clinical practice, little is known about its role in all cause respiratory failure as compared to traditional non-invasive ventilation (BiPAP). Furthermore, the effect of HFNC on mortality is unknown. Methods: We conducted a retrospective analysis of 49,853 patients with respiratory failure treated with non-invasive respiratory support (HFNC or BiPAP) and/or invasive mechanical ventilation (IMV) between 2017 and 2018. Results: Patients initially treated with HFNC who underwent subsequent intubation and IMV had a higher mortality rate as compared to patients who were initially treated with BiPAP and underwent subsequent intubation and IMV (34.8% vs 26.3%, p < 0.0001, OR 1.49, 95% CI 1.26,1.76). Patients first treated with HFNC who underwent subsequent intubation and IMV had a significantly increased mortality compared to patients who underwent immediate intubation and IMV (34.8% vs. 21.5%, p ≤ 0.0001, OR 1.94, 95% CI 1.67, 2.27). Stratified based on ICD-10 diagnosis, patients with a diagnosis of COPD exacerbation or heart failure treated with HFNC and subsequent intubation and IMV had higher mortality as compared to those treated with immediate IMV alone. This trend did not hold true for patients with a diagnosis of pneumonia. Conclusion: In a real-world retrospective analysis, use of HFNC was associated with increased mortality as compared to BiPAP and IMV alone. Further study is needed to confirm these associations.

Sign in / Sign up

Export Citation Format

Share Document