Fluid Balance Is Associated with Clinical Outcomes and Extravascular Lung Water in Children with Acute Asthma Exacerbation.

RATIONALE: The effects of fluid administration during acute asthma exacerbation are likely unique in this patient population: highly negative inspiratory intrapleural pressure resulting from increased airway resistance may interact with excess fluid administration to favor the accumulation of extravascular lung water, leading to worse clinical outcomes. OBJECTIVES: Investigate how fluid balance influences clinical outcomes in children hospitalized for asthma exacerbation. METHODS: We analyzed the association between fluid overload and clinical outcomes in a retrospective cohort of children admitted to an urban children's hospital with acute asthma exacerbation. These findings were validated in two cohorts: a matched retrospective and a prospective observational cohort. Finally, ultrasound imaging was used to identify extravascular lung water and investigate the physiological basis for the inferential findings. MEASUREMENTS AND MAIN RESULTS: In the retrospective cohort, peak fluid overload [(fluid input - output)/weight] is associated with longer hospital length of stay, longer treatment duration, and increased risk of supplemental oxygen use (P values < 0.001). Similar results were obtained in the validation cohorts. There was a strong interaction between fluid balance and intrapleural pressure: the combination of positive fluid balance and highly negative inspiratory intrapleural pressures is associated with signs of increased extravascular lung water (P < 0.001), longer length of stay (P = 0.01), longer treatment duration (P = 0.03), and increased risk of supplemental oxygen use (P = 0.02). CONCLUSIONS: Excess volume administration leading to fluid overload in children with acute asthma exacerbation is associated with increased extravascular lung water and worse clinical outcomes.

Equilibration Time Required for Respiratory System Compliance and Oxygenation Response Following Changes in Positive End-Expiratory Pressure in Mechanically Ventilated Children.

OBJECTIVES: Increases in positive end-expiratory pressure are implemented to improve oxygenation through the recruitment and stabilization of collapsed alveoli. However, the time it takes for a positive end-expiratory pressure change to have maximum effect upon oxygenation and pulmonary compliance has not been adequately described in children. Therefore, we sought to quantify the time required for oxygenation and pulmonary system compliance changes in children requiring mechanical ventilation. DESIGN: Retrospective analysis of continuous data. SETTINGS: Multidisciplinary ICU of a pediatric university hospital. PATIENTS: Mechanically ventilated pediatric subjects. INTERVENTIONS: A case was eligible for analysis if during a 90-minute window following an increase in positive end-expiratory pressure, no other changes to the ventilator were made, ventilator and physiologic data were continuously available and a positive oxygenation response was observed. Time to 90% (T90) of the maximum change in oxygenation and compliance was computed. Differences between oxygenation and compliance T90 were compared using a paired t test. The effect of severity of illness (by oxygen saturation index) upon oxygenation and compliance was analyzed. MEASUREMENTS AND MAIN RESULTS: A total of 200 subjects were enrolled and 1,150 positive end-expiratory pressure change cases were analyzed. Of these, 54 subjects with 171 positive end-expiratory pressure change case were included in the analysis (67% were responders).Changes in dynamic compliance (T90 = 38 min) preceded changes in oxygenation (T90 = 71 min; p < 0.001). Oxygenation response differed depending on severity of illness quantified by oxygen saturation index; lung dysfunction was associated with a longer response time (p = 0.001). CONCLUSIONS: T90 requires 38 and 71 minutes for dynamic pulmonary compliance and oxygenation, respectively; the latter was directly observed to be dependent upon severity of illness. To our knowledge, this is the first report of oxygenation and compliance equilibration data following positive end-expiratory pressure increases in pediatric mechanically ventilated subjects.

Body Composition in Children with Chronic Illness: Accuracy of Bedside Assessment Techniques.

OBJECTIVE: To evaluate the accuracy of estimated fat mass and fat-free mass from bedside methods compared with reference methods in children with chronic illnesses. STUDY DESIGN: Fat mass and fat-free mass values were obtained by skinfold, bioelectrical impedance analysis (BIA), dual-energy x-ray absorptiometry (DXA), and deuterium dilution method in children with spinal muscular atrophy, intestinal failure, and post hematopoietic stem cell transplantation (HSCT). Spearman's correlation and agreement analyses were performed between (1) fat mass values estimated by skinfold equations and by DXA and (2) fat-free mass values estimated by BIA equations and by DXA and deuterium dilution methods. Limits of agreement between estimating and reference methods within ±20% were deemed clinically acceptable. RESULTS: Fat mass and fat-free mass values from 90 measurements in 56 patients, 55% male, and median age of 11.6 years were analyzed. Correlation coefficients between the skinfold-estimated fat mass values and DXA were 0.93-0.94 and between BIA-estimated fat-free mass values and DXA were 0.92-0.97. Limits of agreement between estimated and DXA values of fat mass and fat-free mass were greater than ±20% for all equations. Correlation coefficients between estimated fat-free mass values and deuterium dilution method in 35 encounters were 0.87-0.91, and limits of agreement were greater than ±20%. CONCLUSION: Estimated body composition values derived from skinfold and BIA may not be reliable in children with chronic illnesses. An accurate noninvasive method to estimate body composition in this cohort is desirable.

High-Frequency Oscillatory Ventilation in Pediatric Acute Lung Injury: A Multicenter International Experience.

OBJECTIVE: We aim to describe current clinical practice, the past decade of experience and factors related to improved outcomes for pediatric patients receiving high-frequency oscillatory ventilation. We have also modeled predictive factors that could help stratify mortality risk and guide future high-frequency oscillatory ventilation practice. DESIGN: Multicenter retrospective, observational questionnaire study. SETTING: Seven PICUs. PATIENTS: Demographic, disease factor, and ventilatory and outcome data were collected, and 328 patients from 2009 to 2010 were included in this analysis. INTERVENTIONS: None. MEASUREMENT AND MAIN RESULTS: Patients were classified into six cohorts based on underlying diagnosis. We used univariate analysis to identify factors associated with mortality risk and multivariate logistic regression to identify independent predictors of mortality risk. An oxygenation index greater than 35 and immunocompromise exhibited the greatest predictive power (p < 0.0001) for increased mortality risk, and respiratory syncytial virus was associated with lowest mortality risk (p = 0.003). Differences in mortality risk as a function of oxygenation index were highly dependent on primary underlying condition. A trend toward an increase in oscillator amplitude and frequency was observed when compared with historical data. CONCLUSIONS: Given the number of centers and subjects included in the database, these findings provide a robust description of current practice regarding the use of high-frequency oscillatory ventilation for pediatric hypoxic respiratory failure. Patients with severe hypoxic respiratory failure and immunocompromise had the highest mortality risk, and those with respiratory syncytial virus had the lowest. A means of identifying the risk of 30-day mortality for subjects can be obtained by identifying the underlying disease and oxygenation index on conventional ventilation preceding the initiation of high-frequency oscillatory ventilation.

Metabolic assessment and individualized nutrition in children dependent on mechanical ventilation at home.

OBJECTIVE: To evaluate the nutritional and metabolic status and body composition of children on long-term mechanical ventilation using a home-based model. STUDY DESIGN: Children on home mechanical ventilation, for at least 12 hours a day, were eligible. We performed anthropometry, bioelectrical impedance analysis (BIA), actual energy intake (AEI), and indirect calorimetry in the subject's home. Agreement between measured energy expenditure (MEE) from indirect calorimetry, and estimated energy expenditure by the Schofield equation and a novel volumetric carbon dioxide production-based equation was examined. Agreement between fat mass estimates from anthropometry and BIA was examined and compared with population norms. RESULTS: We enrolled 20 children, 11 (55%) male; mean age 8.4 years (SD 4.8). Mean weight for age z-score was -0.26 (SD 1.48); 9/20 had z-scores <-1 or >+1. Thirteen were underfed (AEI:MEE <90%) or overfed (AEI:MEE >110%); 11 of 19 had protein intake that was less than recommended by guidelines. Fifteen subjects were hypo- or hypermetabolic. Mean (SD) fat mass % was 33.6% (8.6) by anthropometry, which was significantly greater than matched population norms (mean 23.0%, SD 6.1, P < .001). The estimated energy expenditure by a volumetric carbon dioxide production-based equation was in stronger agreement with the MEE than the Schofield equation (mean bias 0.06%, limits -15.98% to 16.16% vs mean bias -1.31%, limits -74.3% to 72%, respectively). BIA and anthropometric fat mass values were not in agreement. CONCLUSION: A majority of children on home ventilation are characterized by malnutrition, altered metabolic status, and suboptimal macronutrient intake, in particular low protein intake. A multidisciplinary home-based model facilitates individualized energy and protein delivery and may improve outcomes in this cohort.

Impact of Individualized Diet Intervention on Body Composition and Respiratory Variables in Children With Respiratory Insufficiency: A Pilot Intervention Study.

OBJECTIVES: Diet modification may improve body composition and respiratory variables in children with respiratory insufficiency. Our objective was to examine the effect of an individualized diet intervention on changes in weight, lean body mass, minute ventilation, and volumetric CO2 production in children dependent on long-term mechanical ventilatory support. DESIGN: Prospective, open-labeled interventional study. SETTING: Study subjects' homes. PATIENTS: Children, 1 month to 17 years old, dependent on at least 12 hr/d of transtracheal mechanical ventilatory support. INTERVENTIONS: Twelve weeks of an individualized diet modified to deliver energy at 90-110% of measured energy expenditure and protein intake per age-based guidelines. MEASUREMENTS AND MAIN RESULTS: During a multidisciplinary home visit, we obtained baseline values of height and weight, lean body mass percent by bioelectrical impedance analysis, actual energy and protein intake by food record, and measured energy expenditure by indirect calorimetry. An individualized diet was then prescribed to optimize energy and protein intake. After 12 weeks on this interventional diet, we evaluated changes in weight, height, lean body mass percent, minute ventilation, and volumetric CO2 production. Sixteen subjects, mean age 9.3 years (SD, 4.9), eight male, completed the study. For the diet intervention, a majority of subjects required a change in energy and protein prescription. The mean percentage of energy delivered as carbohydrate was significantly decreased, 51.7% at baseline versus 48.2% at follow-up, p = 0.009. Mean height and weight increased on the modified diet. Mean lean body mass percent increased from 58.3% to 61.8%. Minute ventilation was significantly lower (0.18 L/min/kg vs 0.15 L/min/kg; p = 0.04), and we observed a trend toward lower volumetric CO2 production (5.4 mL/min/kg vs 5.3 mL/min/kg; p = 0.06) after 12 weeks on the interventional diet. CONCLUSIONS: Individualized diet modification is feasible and associated with a significant decrease in minute ventilation, a trend toward significant reduction in CO2 production, and improved body composition in children on long-term mechanical ventilation. Optimization of respiratory variables and lean body mass by diet modification may benefit children with respiratory insufficiency in the ICU.

Mechanical ventilation guided by electrical impedance tomography in experimental acute lung injury.

OBJECTIVE: To utilize real-time electrical impedance tomography to guide lung protective ventilation in an animal model of acute respiratory distress syndrome. DESIGN: Prospective animal study. SETTING: Animal research center. SUBJECTS: Twelve Yorkshire swine (15 kg). INTERVENTIONS: Lung injury was induced with saline lavage and augmented using large tidal volumes. The control group (n = 6) was ventilated using ARDSnet guidelines, and the electrical impedance tomography-guided group (n = 6) was ventilated using guidance with real-time electrical impedance tomography lung imaging. Regional electrical impedance tomography-derived compliance was used to maximize the recruitment of dependent lung and minimize overdistension of nondependent lung areas. Tidal volume was 6 mL/kg in both groups. Computed tomography was performed in a subset of animals to define the anatomic correlates of electrical impedance tomography imaging (n = 5). Interleukin-8 was quantified in serum and bronchoalveolar lavage samples. Sections of dependent and nondependent regions of the lung were fixed in formalin for histopathologic analysis. MEASUREMENTS AND MAIN RESULTS: Positive end-expiratory pressure levels were higher in the electrical impedance tomography-guided group (14.3 cm H2O vs. 8.6 cm H2O; p < 0.0001), whereas plateau pressures did not differ. Global respiratory system compliance was improved in the electrical impedance tomography-guided group (6.9 mL/cm H2O vs. 4.7 mL/cm H2O; p = 0.013). Regional electrical impedance tomography-derived compliance of the most dependent lung region was increased in the electrical impedance tomography group (1.78 mL/cm H2O vs. 0.99 mL/cm H2O; p = 0.001). Pao2/FIO2 ratio was higher and oxygenation index was lower in the electrical impedance tomography-guided group (Pao2/FIO2: 388 mm Hg vs. 113 mm Hg, p < 0.0001; oxygentation index, 6.4 vs. 15.7; p = 0.02) (all averages over the 6-hr time course). The presence of hyaline membranes (HM) and airway fibrin (AF) was significantly reduced in the electrical impedance tomography-guided group (HMEIT 42% samples vs. HMCONTROL 67% samples, p < 0.01; AFEIT 75% samples vs. AFCONTROL 100% samples, p < 0.01). Interleukin-8 level (bronchoalveolar lavage) did not differ between the groups. The upper and lower 95% limits of agreement between electrical impedance tomography and computed tomography were ± 16%. CONCLUSIONS: Electrical impedance tomography-guided ventilation resulted in improved respiratory mechanics, improved gas exchange, and reduced histologic evidence of ventilator-induced lung injury in an animal model. This is the first prospective use of electrical impedance tomography-derived variables to improve outcomes in the setting of acute lung injury.

A Simplified Adult Expiratory Support Device: First Clinical Cases in a Low-Resource Setting.

PURPOSE: In low and middle-income countries, mechanical ventilators or commercially available devices used to offer continuous positive airway pressure are not readily affordable and available. In Ghana, nearly 10% of critically ill patients presenting to the emergency department require ventilator support. DESCRIPTION: We designed, built, and tested a simple expiratory positive airway pressure (EPAP) device to provide adult respiratory support in low resource environments with or without supplemental oxygen and without the need for electricity. EVALUATION: Laboratory tests demonstrated that the device is capable of delivering EPAP at levels expected to provide significant assistance to some patients. We present the first 2 cases where the use of this simple EPAP device provided critical respiratory support during weaning of patients from mechanical ventilation. CONCLUSIONS: A low-cost 3-dimensional printable adult respiratory support device could provide substantial benefit to patients suffering from respiratory distress through the delivery of appropriate levels of EPAP in a low-resource setting with limited infrastructure. Further clinical validation is needed for broader application in low-resource settings.

Monitoring Big Data During Mechanical Ventilation in the ICU.

The electronic health record allows the assimilation of large amounts of clinical and laboratory data. Big data describes the analysis of large data sets using computational modeling to reveal patterns, trends, and associations. How can big data be used to predict ventilator discontinuation or impending compromise, and how can it be incorporated into the clinical workflow? This article will serve 2 purposes. First, a general overview is provided for the layperson and introduces key concepts, definitions, best practices, and things to watch out for when reading a paper that incorporates machine learning. Second, recent publications at the intersection of big data, machine learning, and mechanical ventilation are presented.

2019 Year in Review: Neonatal Respiratory Support.

Respiratory support of the critically ill neonate has steadily shifted from invasive to noninvasive forms of support. There have recently been a number of important advances in our understanding of the changes to neonatal resuscitation practices as they pertain to clinically important outcomes, mechanisms of gas exchange for high-flow nasal cannula, and best use of noninvasive ventilation and predicting response. Although the proportion of infants requiring intubation and mechanical ventilation has decreased, the most severely ill often still require intubation and ventilation. Recently, volume-targeted ventilation, high-frequency ventilation, and different methods of assessing weaning and extubation have been investigated. This review summarizes a number of important advances that have been made in the management of prematurity and neonatal respiratory distress syndrome.

2019 Year in Review: Asthma.

Asthma is an obstructive airway disease affecting children and adults throughout the world. It is a heterogeneous disease with a variety of causes and treatments. Research in the diagnosis, treatment, and management of asthma is ongoing, and there were > 8,000 publications on asthma in 2019. This paper reviews several research articles about asthma from 2019 that are most relevant for practicing respiratory therapists caring for patients with asthma.

Local Effects of Two Intravenous Formulations of Pulmonary Vasodilators on Airway Epithelium.

BACKGROUND: Intravenous formulations of epoprostenol are frequently delivered via nebulizer to treat pulmonary hypertension in acutely ill patients. Although their efficacy as pulmonary vasodilators has been shown to be comparable to inhaled nitric oxide, the local effects of these formulations within the airways have not been determined. We hypothesized that the alkaline diluents of these compounds would lead to increased airway epithelial cell death and ciliary cessation. METHODS: Human bronchial epithelial cells were exposed to epoprostenol in glycine and arginine diluents or control fluid. Ciliary beat frequency, lactate dehydrogenase, and total RNA levels were measured before and after exposure. Results were compared between exposure and control groups. RESULTS: Ciliary beat frequency ceased immediately after exposure to epoprostenol with both diluents. Lactate dehydrogenase levels increased by 200% after exposure to epoprostenol and glycine diluent (P = .002). Total RNA levels were undetectable after exposure to epoprostenol and arginine, indicating complete cell death and lysis (P = .015). Ciliary beat frequency ceased after 30 s of exposure to epoprostenol and glycine (P = .008). There was no difference between cells exposed to epoprostenol and those exposed only to diluent. CONCLUSIONS: Exposure to intravenous formulations of epoprostenol in glycine and arginine caused increased cell death and ciliary cessation in bronchial epithelial cells. These findings suggest that undesired local effects may occur when these compounds are delivered as inhaled aerosols to patients.

Mortality Risk Factors in Preterm Infants Treated with High-Frequency Jet Ventilation.

BACKGROUND: High-frequency jet ventilation (HFJV) has been used in conjunction with conventional ventilation for premature infants with respiratory failure. We sought to identify parameters that were associated with mortality in subjects who underwent HFJV. METHODS: Subjects were enrolled if birthweight was ≤ 2,000 g and they were ≤ 34 weeks gestational age. Subjects were excluded if they received HFJV at the time of admission because we aimed to study subjects who failed conventional ventilation. Subject demographics, ventilator parameters, and laboratory data were extracted and analyzed. The Mann-Whitney U-test was used to assess differences in continuous variables, and the chi-square and Fisher exact tests were used for categorical variables between the groups. To assess variables that were predictive of mortality, we used both univariate and multivariate logistic regression analysis. Independent predictors of mortality were identified and used to create a multivariate risk score. Receiver operating characteristic curves were constructed to evaluate the predictive accuracy of the multivariate risk score. RESULTS: A total of 53 premature subjects (n = 37 male) were studied, of whom 39 (74%) survived to discharge or transfer back to referring hospital. In the univariate model, female sex, older gestational age, higher birthweight, HFJV peak inspiratory pressure at 1 h, and oxygen saturation index at 4 h were associated with mortality. In the final multivariate logistic regression model, female sex (odds ratio 4.1, 95% CI 1.2-19.8, P = .044), closed ductus arteriosus (odds ratio 7.7, 95% CI 1.3-39.5, P = .016), and oxygen saturation index > 5.5 (odds ratio 6.0, 95% CI 1.5-28.3, P = .02) were independent predictors of mortality. CONCLUSIONS: We identified that oxygen saturation index > 5.5 after 4 h of HFJV, female sex, and closed ductus arteriosus were independent predictors of mortality.

Physiologic Effects of Nasal Aspiration and Nasopharyngeal Suctioning on Infants With Viral Bronchiolitis.

BACKGROUND: There is limited evidence supporting an optimum method for removing mucus from the airways of hospitalized infants with bronchiolitis. This study was designed to evaluate short-term physiologic effects between nasal aspiration and nasopharyngeal suctioning in infants. METHODS: Sixteen infants requiring hospitalization for supportive management of bronchiolitis were instrumented with transcutaneously measured partial pressure of carbon dioxide ([Formula: see text]) and [Formula: see text] monitoring. Electrical impedance tomography (EIT) was used to estimate changes in inspiratory and end-expiratory lung volume loss and recovery. Subjects were suctioned with both nasal aspiration and nasopharyngeal suctioning methods in a randomized order (8 received nasal aspiration followed by nasopharyngeal suctioning, and 8 received nasophayrgeal suctioning followed by nasal aspiration). Noninvasive gas exchange and EIT measurements were obtained at baseline (pre-suction) and at 10, 20, and 30 min following each suctioning intervention. Sputum mass was obtained following suctioning, and clinical respiratory severity scores, before and after suctioning, were computed. RESULTS: There were no differences in inspiratory EIT (P = .93), change in end-expiratory lung impedance (ΔEELI; P = .53), [Formula: see text] (P = .41), [Formula: see text] (P = .88), heart rate (P = .31), or breathing frequency (P = .15) over the course of suctioning between nasal aspiration and nasopharyngeal suctioning. Sputum mass (P = .14) and clinical respiratory score differences before and after suctioning (P = .59) were not different between the 2 suctioning interventions. Sputum mass was not associated with ΔEELI at 30 min for nasal aspiration (ρ = 0.11, P = .69), but there was a moderate positive association for nasopharyngeal suctioning (ρ = 0.50, P = .048). CONCLUSIONS: Infants with viral bronchiolitis appeared to tolerate both suctioning techniques without adverse short-term physiologic effects, as indicated by the unchanged gas exchange and estimated lung volumes (EIT). Nasopharyngeal suctioning recovered 36% more sputum than did nasal aspiration and there was moderate correlation between sputum mass and end-expiratory lung impedance change at 30 minutes post-suction with nasopharyngeal that was not present with nasal aspiration. It is possible that a subset of patients may benefit from one type of suctioning over another. Future research focusing on important outcomes for suctioning patients with bronchiolitis with varying degrees of lung disease severity is needed.

Distribution of Ventilation Measured by Electrical Impedance Tomography in Critically Ill Children.

BACKGROUND: Electrical impedance tomography (EIT) is a noninvasive, portable lung imaging technique that provides functional distribution of ventilation. We aimed to describe the relationship between the distribution of ventilation by mode of ventilation and level of oxygenation impairment in children who are critically ill. We also aimed to describe the safety of EIT application. METHODS: A prospective observational study of EIT images obtained from subjects in the pediatric ICU. Images were categorized by whether the subjects were on intermittent mandatory ventilation (IMV), continuous spontaneous ventilation, or no positive-pressure ventilation. Images were categorized by the level of oxygenation impairment when using [Formula: see text]/[Formula: see text]. Distribution of ventilation is described by the center of ventilation. RESULTS: Sixty-four images were obtained from 25 subjects. Forty-two images obtained during IMV with a mean ± SD center of ventilation of 55 ± 6%, 14 images during continuous spontaneous ventilation with a mean ± SD center of ventilation of 48.1 ± 11%, and 8 images during no positive-pressure ventilation with a mean ± SD center of ventilation of 47.5 ± 10%. Seventeen images obtained from subjects with moderate oxygenation impairment with a mean ± SD center of ventilation of 59.3 ± 1.9%, 12 with mild oxygenation impairment with a mean ± SD center of ventilation of 52.6 ± 2.3%, and 4 without oxygenation impairment with a mean ± SD center of ventilation of 48.3 ± 4%. There was more ventral distribution of ventilation with IMV versus continuous spontaneous ventilation (P = .009), with IMV versus no positive-pressure ventilation (P = .01) cohorts, and with moderate oxygenation impairment versus cohorts without oxygenation impairment (P = .009). There were no adverse events related to the placement and use of EIT in our study. CONCLUSIONS: Children who had worse oxygen impairment or who received controlled modes of ventilation had more ventral distribution of ventilation than those without oxygen impairment or the subjects who were spontaneously breathing. The ability of EIT to detect changes in the distribution of ventilation in real time may allow for distribution-targeted mechanical ventilation strategies to be deployed proactively; however, future studies are needed to determine the effectiveness of such a strategy.

Noninvasive Ventilation Is Interrupted Frequently and Mostly Used at Night in the Pediatric Intensive Care Unit.

BACKGROUND: Noninvasive ventilation (NIV) is commonly used to support children with respiratory failure, but detailed patterns of real-world use are lacking. The aim of our study was to describe use patterns of NIV via electronic medical record (EMR) data. METHODS: We performed a retrospective electronic chart review in a tertiary care pediatric ICU in the United States. Subjects admitted to the pediatric ICU from 2014 to 2017 who were mechanically ventilated were included in the study. RESULTS: The median number of discrete device episodes, defined as a time on support without interruption, was 20 (interquartile range [IQR] 8-49) per subject. The median duration of bi-level positive airway pressure (BPAP) support prior to interruption was 6.3 h (IQR 2.4-10.4); the median duration of CPAP was 6 h (IQR 2.1-10.4). Interruptions to BPAP had a median duration of 6.3 h (IQR 2-15.5); interruptions to CPAP had a median duration of 8.6 h (IQR 2.2-16.8). Use of NIV followed a diurnal pattern, with 44% of BPAP and 42% of CPAP subjects initiating support between 7:00 pm and midnight, and 49% of BPAP and 46% of CPAP subjects stopping support between 5:00 am and 10:00 am. CONCLUSIONS: NIV was frequently interrupted, and initiation and discontinuation of NIV follows a diurnal pattern. Use of EMR data collected for routine clinical care allowed the analysis of granular details of typical use patterns. Understanding NIV use patterns may be particularly important to understanding the burden of pediatric ICU bed utilization for nocturnal NIV. To our knowledge, this is the first study to examine in detail the use of pediatric NIV and to define diurnal use and frequent interruptions to support.

Adding Continuous Vital Sign Information to Static Clinical Data Improves the Prediction of Length of Stay After Intubation: A Data-Driven Machine Learning Approach.

BACKGROUND: Bedside monitors in the ICU routinely measure and collect patients' physiologic data in real time to continuously assess the health status of patients who are critically ill. With the advent of increased computational power and the ability to store and rapidly process big data sets in recent years, these physiologic data show promise in identifying specific outcomes and/or events during patients' ICU hospitalization. METHODS: We introduced a methodology designed to automatically extract information from continuous-in-time vital sign data collected from bedside monitors to predict if a patient will experience a prolonged stay (length of stay) on mechanical ventilation, defined as >4 d, in a pediatric ICU. RESULTS: Continuous-in-time vital signs information and clinical history data were retrospectively collected for 284 ICU subjects from their first 24 h on mechanical ventilation from a medical-surgical pediatric ICU at Boston Children's Hospital. Multiple machine learning models were trained on multiple subsets of these subjects to predict the likelihood that each of these subjects would experience a long stay. We evaluated the predictive power of our models strictly on unseen hold-out validation sets of subjects. Our methodology achieved model performance of >83% (area under the curve) by using only vital sign information as input, and performances of 90% (area under the curve) by combining vital sign information with subjects' static clinical data readily available in electronic health records. We implemented this approach on 300 independently trained experiments with different choices of training and hold-out validation sets to ensure the consistency and robustness of our results in our study sample. The predictive power of our approach outperformed recent efforts that used deep learning to predict a similar task. CONCLUSIONS: Our proposed workflow may prove useful in the design of scalable approaches for real-time predictive systems in ICU environments, exploiting real-time vital sign information from bedside monitors. (ClinicalTrials.gov registration NCT02184208.).