Predicting Failure of Non-Invasive Ventilation With RAM Cannula in Bronchiolitis

Maamari, Mia; Niño, Gustavo; Bost, James E.; Cheng, Yao I.; Sochet, Anthony A.; Sharron, Matthew

doi:10.1177/0885066620979642

Cited by 12 publications

(6 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When evaluating our model's feature importance through Shapley additive explanations, we found that features representing physiologic vitals data and F i O 2 parameters were considered important. These findings are consistent with previous reports demonstrating the importance of persistent vital sign abnormalities and other comorbidities (i.e., concurrent bacterial pneumonia, persistent hypercarbia, and history of prematurity) as likely indicators of NIV non-response ( 10 , 13 , 40 – 42 ). Despite their poor performance predicting flow rate escalation in independent models, the ROX and ROX-HR features within our prediction model were highly valuable.…”

Section: Discussionsupporting

confidence: 93%

Predicting Flow Rate Escalation for Pediatric Patients on High Flow Nasal Cannula Using Machine Learning

Krachman

Patricoski

et al. 2021

Front. Pediatr.

View full text Add to dashboard Cite

Background: High flow nasal cannula (HFNC) is commonly used as non-invasive respiratory support in critically ill children. There are limited data to inform consensus on optimal device parameters, determinants of successful patient response, and indications for escalation of support. Clinical scores, such as the respiratory rate-oxygenation (ROX) index, have been described as a means to predict HFNC non-response, but are limited to evaluating for escalations to invasive mechanical ventilation (MV). In the presence of apparent HFNC non-response, a clinician may choose to increase the HFNC flow rate to hypothetically prevent further respiratory deterioration, transition to an alternative non-invasive interface, or intubation for MV. To date, no models have been assessed to predict subsequent escalations of HFNC flow rates after HFNC initiation.Objective: To evaluate the abilities of tree-based machine learning algorithms to predict HFNC flow rate escalations.Methods: We performed a retrospective, cohort study assessing children admitted for acute respiratory failure under 24 months of age placed on HFNC in the Johns Hopkins Children's Center pediatric intensive care unit from January 2019 through January 2020. We excluded encounters with gaps in recorded clinical data, encounters in which MV treatment occurred prior to HFNC, and cases electively intubated in the operating room. The primary study outcome was discriminatory capacity of generated machine learning algorithms to predict HFNC flow rate escalations as compared to each other and ROX indices using area under the receiver operating characteristic (AUROC) analyses. In an exploratory fashion, model feature importance rankings were assessed by comparing Shapley values.Results: Our gradient boosting model with a time window of 8 h and lead time of 1 h before HFNC flow rate escalation achieved an AUROC with a 95% confidence interval of 0.810 ± 0.003. In comparison, the ROX index achieved an AUROC of 0.525 ± 0.000.Conclusion: In this single-center, retrospective cohort study assessing children under 24 months of age receiving HFNC for acute respiratory failure, tree-based machine learning models outperformed the ROX index in predicting subsequent flow rate escalations. Further validation studies are needed to ensure generalizability for bedside application.

show abstract

Section: Discussionsupporting

confidence: 93%

Predicting Flow Rate Escalation for Pediatric Patients on High Flow Nasal Cannula Using Machine Learning

Krachman

Patricoski

et al. 2021

Front. Pediatr.

View full text Add to dashboard Cite

show abstract

“…While previous trials have compared HFNC and NIV use, the present study is the first multicenter analysis to evaluate risk factors for NIV failure in bronchiolitis. 40,41 Similar to prior single-center work, 42,43 we found higher severity of illness was associated with higher odds of HFNC and NIV failure. Characterizing which patients are likely to be successfully treated with NIV remains an important goal of PICU precision medicine.…”

Section: Discussionsupporting

confidence: 80%

Respiratory Support Practices for Bronchiolitis in the Pediatric Intensive Care Unit

Pelletier,

Maholtz,

Hanson

et al. 2024

JAMA Netw Open

View full text Add to dashboard Cite

ImportanceAdmissions to the pediatric intensive care unit (PICU) due to bronchiolitis are increasing. Whether this increase is associated with changes in noninvasive respiratory support practices is unknown.ObjectiveTo assess whether the number of PICU admissions for bronchiolitis between 2013 and 2022 was associated with changes in the use of high-flow nasal cannula (HFNC), noninvasive ventilation (NIV), and invasive mechanical ventilation (IMV) and to identify factors associated with HFNC and NIV success and failure.Design, Setting, and ParticipantsThis cross-sectional study examined encounter data from the Virtual Pediatric Systems database on annual PICU admissions for bronchiolitis and ventilation practices among patients aged younger than 2 years admitted to 27 PICUs between January 1, 2013, and December 31, 2022. Use of HFNC and NIV was defined as successful if patients were weaned to less invasive support (room air or low-flow nasal cannula for HFNC; room air, low-flow nasal cannula, or HFNC for NIV).Main Outcomes and MeasuresThe main outcome was the number of PICU admissions for bronchiolitis requiring the use of HFNC, NIV, or IMV. Linear regression was used to analyze the association between admission year and absolute numbers of encounters stratified by the maximum level of respiratory support required. Multivariable logistic regression was used to analyze factors associated with HFNC and NIV success and failure (defined as not meeting the criteria for success).ResultsIncluded in the analysis were 33 816 encounters for patients with bronchiolitis (20 186 males [59.7%]; 1910 patients [5.6%] aged ≤28 days and 31 906 patients [94.4%] aged 29 days to &lt;2 years) treated at 27 PICUs from 2013 to 2022. A total of 7615 of 15 518 patients (49.1%) had respiratory syncytial virus infection and 1522 of 33 816 (4.5%) had preexisting cardiac disease. Admissions to the PICU increased by 350 (95% CI, 170-531) encounters annually. When data were grouped by the maximum level of respiratory support required, HFNC use increased by 242 (95% CI, 139-345) encounters per year and NIV use increased by 126 (95% CI, 64-189) encounters per year. The use of IMV did not significantly change (10 [95% CI, −11 to 31] encounters per year). In all, 22 381 patients (81.8%) were successfully weaned from HFNC to low-flow oxygen therapy or room air, 431 (1.6%) were restarted on HFNC, 3057 (11.2%) were escalated to NIV, and 1476 (5.4%) were escalated to IMV or extracorporeal membrane oxygenation (ECMO). Successful use of HFNC increased from 820 of 1027 encounters (79.8%) in 2013 to 3693 of 4399 encounters (84.0%) in 2022 (P = .002). In all, 8476 patients (81.5%) were successfully weaned from NIV, 787 (7.6%) were restarted on NIV, and 1135 (10.9%) were escalated to IMV or ECMO. Success with NIV increased from 224 of 306 encounters (73.2%) in 2013 to 1335 of 1589 encounters (84.0%) in 2022 (P &lt; .001). In multivariable logistic regression, lower weight, higher Pediatric Risk of Mortality III score, cardiac disease, and PICU admission from outside the emergency department were associated with greater odds of HFNC and NIV failure.Conclusions and RelevanceFindings of this cross-sectional study of patients aged younger than 2 years admitted for bronchiolitis suggest there was a 3-fold increase in PICU admissions between 2013 and 2022 associated with a 4.8-fold increase in HFNC use and a 5.8-fold increase in NIV use. Further research is needed to standardize approaches to HFNC and NIV support in bronchiolitis to reduce resource strain.

show abstract

“…The distinct advantages of the study in comparison to prior studies are: (a) the continuous risk-score evaluation allows for continuous monitoring of patient state; (b) EPP enables early alert for healthcare providers to prescribe an appropriate treatment regimen based on the patient's physiological state at that time; and (c) the proposed risk stratification method would allow for identification of different risk groups after observed EPP, which can help identify a high-risk subgroup that has an extremely high likelihood of intubation. Some of the common risk factors associated with failure of non-invasive ventilation in children reported by prior studies [59][60][61][62][63][64][65] include higher FiO2, SpO2/FiO2 ratio, respiratory rate oxygenation index (ratio of SpO2/FiO2 to respiratory rate), heart rate, respiratory rate, younger age, and blood pH. The global variable importance of Model A also suggests corroborating evidence that SpO2, FiO2, age, respiratory rate, tachypnea, etc.…”

Section: Research Contributions and Conclusionsupporting

confidence: 62%

A data-driven model for early prediction of need for invasive mechanical ventilation in pediatric intensive care unit patients

et al. 2023

View full text Add to dashboard Cite

Rationale Acute respiratory failure is a life-threatening clinical outcome in critically ill pediatric patients. In severe cases, patients can require mechanical ventilation (MV) for survival. Early recognition of these patients can potentially help clinicians alter the clinical course and lead to improved outcomes. Objectives To build a data-driven model for early prediction of the need for mechanical ventilation in pediatric intensive care unit (PICU) patients. Methods The study consists of a single-center retrospective observational study on a cohort of 13,651 PICU patients admitted between 1/01/2010 and 5/15/2018 with a prevalence of 8.06% for MV due to respiratory failure. XGBoost (extreme gradient boosting) and a convolutional neural network (CNN) using medication history were used to develop a prediction model that could yield a time-varying "risk-score"—a continuous probability of whether a patient will receive MV—and an ideal global threshold was calculated from the receiver operating characteristics (ROC) curve. The early prediction point (EPP) was the first time the risk-score surpassed the optimal threshold, and the interval between the EPP and the start of the MV was the early warning period (EWT). Spectral clustering identified patient groups based on risk-score trajectories after EPP. Results A clinical and medication history-based model achieved a 0.89 area under the ROC curve (AUROC), 0.6 sensitivity, 0.95 specificity, 0.55 positive predictive value (PPV), and 0.95 negative predictive value (NPV). Early warning time (EWT) median [inter-quartile range] of this model was 9.9[4.2–69.2] hours. Clustering risk-score trajectories within a six-hour window after the early prediction point (EPP) established three patient groups, with the highest risk group’s PPV being 0.92. Conclusions This study uses a unique method to extract and apply medication history information, such as time-varying variables, to identify patients who may need mechanical ventilation for respiratory failure and provide an early warning period to avert it.

show abstract

Predicting Failure of Non-Invasive Ventilation With RAM Cannula in Bronchiolitis

Cited by 12 publications

References 27 publications

Predicting Flow Rate Escalation for Pediatric Patients on High Flow Nasal Cannula Using Machine Learning

Predicting Flow Rate Escalation for Pediatric Patients on High Flow Nasal Cannula Using Machine Learning

Respiratory Support Practices for Bronchiolitis in the Pediatric Intensive Care Unit

A data-driven model for early prediction of need for invasive mechanical ventilation in pediatric intensive care unit patients

Contact Info

Product

Resources

About