SCCM Resource Library-30-Flow Index: A Novel Method to Report the Respiratory Support for Children on High-Flow Nasal Cannula

Video Transcription

My name is Sandeep Tripathi and I am an Associate Professor of Pediatrics at the University of Illinois College of Medicine at Peoria.  I will be presenting my research on the description and validation of a novel score as a clinical  indicator of the level of respiratory support to children on high flow nasal cannabis. We call this score as the flow index.  My co-investigators in this project are Jeremy, who did the statistical analysis, Dr. Sheikh,  who assisted with the design of the project, and Logan, who did the automated data extraction and validation. We have no conflict of interest to disclose.  High flow nasal cannula is increasingly being utilized to provide non-invasive support for children with acute respiratory failure.  It provides a higher flow rate, which provides PEEP for lung expansion and a defined FiO2 for oxygenation.  The level of respiratory support by high flow nasal cannula is dependent on the size of the patient.  Thus, to accurately describe the respiratory support, all three have to be stated. We want to propose a new score, which takes into account the three variables.  This equation includes flow rate in liters per minute, multiplied by FiO2 as a percentage, and divided by patient's weight in kilograms.  Our intent in developing this score was for it to be used as a communication tool as well as to discriminate based on the score which patients are at risk of escalation to ICU,  intubation, etc. The specific aims of this research project were to identify the associations of the first  and the maximum flow index of a patient on high flow nasal cannula with the hospital length of stay.  The first flow index was the first value of FiO2 and flow rate charted for the patient, while the maximum flow index was any point of time during patient's treatment with high  flow nasal cannula when the product of the FiO2 and flow rate were maximum. Our secondary objective was to identify the sensitivity and specificity of the first and  maximum flow index to predict escalation of care to the ICU and rapid response team activation, which is called PERT, or Pediatric Emergency Response Team at our institution.  To achieve this objective, we did a single center retrospective chart review. All children less than 18 years of age who required high flow nasal cannula during their  hospitalization at our hospital from January 2015 to December 2019 were eligible in this study.  We excluded patients with congenital malformations of circulatory system, sepsis, chronic respiratory  failure, chronic obstructive sleep apnea, pulmonary hypertension, and tracheostomy, as titration of oxygen and flow rate may be different from titration in children not with  these conditions. Patients who died were also excluded as the primary outcome variable was hospital length of stay.  We used simple regression to determine the relationship between the hospital length of stay and the individual components of the flow index, as well as the flow index itself.  We further did a stepwise multiple regression to determine its significance after controlling for the confounding variables.  Then we grouped the flow index into mild, moderate, and severe categories, which we call as green, yellow, and red, based on the flow index quartiles.  We used logistic regression to identify the relationship between first and maximum flow index and the binary events of escalation to ICU and PERT activation.  A total of 2,209 patients met the inclusion criteria during the study period, out of which 672 patients were excluded and 1,537 patients were included in the analysis.  23% of these patients were directly admitted to the ICU from the emergency department, and 18.5% patients were initially admitted to general pediatrics and required escalation  to the ICU. The study sample was 57% male and 64% Caucasian with a median age of 20 months. The median hospital and ICU length of stay were 4.5 and 2.2 days.  The median duration of high flow nasal cannula was 45 hours. The whole population's median first flow index was 24.1 and maximum flow index was 38.1.  While not shown in this figure, 15% of patients had a admission diagnosis of asthma, 19% had bronchiolitis, and 7% had a diagnosis of pneumonia.  This table shows the results of simple linear regression between the log-transformed hospital  and ICU length of stay and the first and maximum recorded weight, FiO2, oxygen flow rate, FiO2 multiplied by the flow rate, and the flow index itself.  For hospital length of stay, weight was not significantly related to the length of stay.  However, all other variables showed a significant positive relationship with the length of stay,  indicating that an increase in these variables would be associated with an increased length of stay.  Both first and the maximum flow index had a stronger correlation coefficient than the variables used to calculate them.  The first flow index had a correlation coefficient of 0.25 and the maximum flow index had a correlation coefficient of 0.31.  For ICU admissions, weight had a significant negative association with the length of stay. Except for the first recorded FiO2, all other variables were significantly associated with  the ICU length of stay. As with the hospital length of stay, both first and maximum flow index had stronger correlation than the index-constituent variables.  While not shown in this table, the multivariate regression after adjusting for confounders of age, race, sex, and diagnostic categories, both first and maximum flow index remained  significantly associated with increased hospital length of stay. The first and maximum flow index were categorized into three subgroups based on quartiles.  The cutoff point for the first flow index was less than 13, 13 to 43, and more than 43. While the cutoff point for the maximum flow index was less than 20, 20 to 64, and more  than 64. Patients who had first and maximum flow index in the yellow or red category had a significantly  higher geometric mean length of stay compared to patients who had initial or maximum flow index in the green category.  This difference remained significant on multivariate regression. Because of the low number of adolescents in the high flow index category, above the  binary thresholds, adolescents were excluded from the analysis of the prediction of ICU transfer in PERT based on the flow index.  Patients with a first flow index more than 20 and max flow index more than 59.5 had a significantly higher likelihood of an ICU transfer with an odds ratio of 2.4 and 7.6  respectively. The first flow index had a negative association with the incidence of PERT. Patients with first flow index less than 31 were significantly more likely to have a PERT.  The maximum flow index did not have a significant association with PERT. These ROC curves show the area under the curve and the sensitivity and specificity of ICU  escalation based on the first flow index and the maximum flow index threshold described in the earlier slide.  For a cutoff of 20 for the first flow index, the area under the curve was 0.6 and a sensitivity of 72%.  While for a cutoff of 59.5 for maximum flow index, the AUC was 0.73 and a relatively higher specificity of 80%.  In this study, we have described a novel single metric to characterize the degree of respiratory support by high flow nasal canvas.  We have shown that this flow index is a valid measure to assess disease severity using hospital length of stay as a surrogate marker.  We have also described the cutoff of the flow index for mild, moderate, and severe degree of support.  We have also described the cutoff and sensitivity and specificity of predicting ICU and PERT using the flow index. Our study has several limitations.  Even though the flow index performed better than the individual components, the overall correlation of the flow index with the length of stay is small.  The flow index gives equal weight to the flow rate and the fraction of inspired oxygen. Technically, it cannot differentiate between the level of support provided to an asthmatic  versus that provided to a patient with ARDS. Our sensitivity and specificity of ICU escalation and PERT may have limited external validity  as these are hospital policy sensitive parameters. To conclude, flow index can be used as a metric for rapid assessment and communication of  the degree of respiratory support in children on high flow nasal cannula. Mild, moderate, and severe categories, as defined in our study, have significant association  with the hospital length of stay. The score would need to be validated in a separate cohort of patients. Thank you.

Video Summary

In this video presentation, Dr. Sandeep Tripathi discusses his research on a new score called the flow index, which is a clinical indicator of respiratory support in children on high flow nasal cannula. The flow index takes into account the flow rate, FiO2, and patient's weight. The study aimed to determine the associations between the flow index and hospital length of stay, as well as its ability to predict escalation of care to the ICU. The results showed that higher flow index scores were associated with longer hospital stays, and certain flow index thresholds were predictive of ICU transfers. The flow index can be a useful tool for assessing respiratory support and may help in communication and decision-making for patient care.

Asset Subtitle

Pediatrics, Pulmonary, 2021

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2021 Critical Care Congress held virtually from January 31-February 12, 2021.

Sandeep Tripathi

Introduction/Hypothesis: High flow nasal cannula (HFNC) provides support to breathing by flow rate and FiO2. The weight of the patient impacts the degree of support. We devised a scoring system to report the degree of support using one number (flow rateÃ—FiO2/weight). This retrospective chart review was conducted to validate this scoring system (flow Index, FI) in its ability to quantify and predict disease severity and outcome.

Methods: Children who were managed with HFNC in the hospital from 01/2015 to 11/2019 were identified from electronic medical records. Patients demographics and values of initial and max FI and index just prior to the ICU transfer were extracted from EMR. Patients that were missing in data points required to calculate the FI were excluded. Simple regression was used to determine the relationship between the outcome (length of stay, LOS) and individual components of the index and index itself. Multiple regression was performed to adjust for confounding variables. Statistical analysis was performed using R (v 4.0.0).

Results: 1537 patients met study criteria. The median age was 20 months, and the median duration on HFNC was 45 hours. 80.6% had a respiratory diagnosis. Median, initial, and max FI were 24.1 and 38.1, respectively. On simple linear regression, first and Max value of the score showed a significant correlation with LOS (r 0.25 and 0.31, p <0.001). The correlation for the index was stronger than the coefficients of the variables used to calculate them. FI remains significantly associated with increased LOS after controlling for other variables, including diagnoses and age. Patients in the 3rd and 4th quartile of the initial FI (24.1 â€“ 42.9 and > 42.9) and patients in 2nd /3rd and 4th quartile for the max FI (20.6 â€“ 38.1, 38.1 â€“ 64.7 and > 64.7) had significantly higher LOS compared to patients on the 1st quartile. There was a significant -ve interaction between the FI and age (b= -0.004, p <0.001), indicating the difference in LOS for patients in the higher FI group decreases with age. Of the 195 patients who met the inclusion criteria for escalation of care to ICU, there was no significant correlation of FI with ICU length of stay.

Conclusions: Flow index (FI) is a valid measure to assess the severity of illness. Respiratory support can be categorized based on FI Quartiles.

Meta Tag

Content Type Presentation

Knowledge Area Pediatrics

Knowledge Area Pulmonary

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Oxygenation

Tag Ventilation

Year 2021

Keywords

flow index

respiratory support

high flow nasal cannula

hospital length of stay

ICU transfers

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.