Bronchiolitis is a leading cause of pediatric hospitalization. Treatment focuses on supportive care including supplemental oxygen for hypoxemia. High-flow nasal cannula (HFNC) has emerged as a modality to provide respiratory support with or without supplemental oxygen. At a freestanding children’s hospital, inappropriate supplemental oxygen was frequently used. This study aimed to decrease the proportion of patients started on supplemental oxygen without documented hypoxemia from ∼90% to <70% and the proportion of patients weaned from HFNC without supplemental oxygen to nasal cannula with supplemental oxygen from ∼23% to <10%.
A multidisciplinary taskforce was convened to develop an evidence-based protocol for HFNC usage. Data collection was obtained among patients aged <2 years admitted with bronchiolitis from September 2018 to September 2021. Institution-wide protocol changes occurred in November 2019 and October 2020, with ongoing education and evaluation. Data were summarized using statistical process control charts.
Following implementation of a revised protocol in October 2020, the percentage of patients without documented hypoxemia (defined as an oxygen saturation <90% on pulse oximetry) who were inappropriately started on supplemental oxygen decreased from a baseline of 90.2% to 57.2%. At the same time, the percentage of patients weaned from HFNC without nasal cannula oxygen decreased from a baseline of 23.1% to 4.7%.
Using supplemental oxygen in the absence of hypoxemia in bronchiolitis is an example of low-value care. Implementation of focused, standardized protocols with concurrent education can feasibly decrease inappropriate and unnecessary use of supplemental oxygen in children with bronchiolitis.
Bronchiolitis is a viral lower respiratory tract illness that commonly affects young children aged <2 years old and can lead to acute respiratory failure. Generally, bronchiolitis is self-limited and can be managed with outpatient supportive care. However, In the United States, 2% to 3% of children aged <12 months are hospitalized annually because of bronchiolitis, accounting for 57 000 to 172 000 admissions and $1.7 billion in annual costs.1,2 Rates of admission for bronchiolitis have increased from 1% to 3% of all infants, leading to even more medical expenditures.3
Disease-targeted therapies such as corticosteroids, β-agonists, chest physiotherapy, hypertonic saline, and antimicrobial agents have not demonstrated significant benefit in children admitted with bronchiolitis, and guidelines favor supportive care.4,5 After age, hypoxemia with peripheral oxygen saturation (SpO2) <90% is a predictor of hospitalization; therefore, supplemental oxygen is often additionally used in hospitalized children with bronchiolitis.6,7 High-flow nasal cannula (HFNC) has emerged as a newer option for supporting children admitted with bronchiolitis, although data suggest it may increase cost without decreasing length of stay or affecting overall patient outcomes.8,9 Supplemental oxygen can be provided through the same interface, allowing for simultaneous treatment of hypoxemia. However, prolonged or unnecessary use of supplemental oxygen may not only lead to cellular damage via free radical formation but also exacerbate atelectasis.10 There is a recommendation (Evidence Quality D; weak strength) not to administer supplemental oxygen if the SpO2 exceeds 90%.5 Additionally, unnecessary supplemental oxygen may lead to other low-value care, such as continuous pulse oximetry, which might otherwise not be used.11
At our institution, despite high usage rates of HFNC, there was not a standardized protocol for initiating, titrating, and discontinuing HFNC, leading to concerns about unnecessary variation and likely overuse. We hypothesized that by creating a comprehensive standardized protocol, we could decrease overuse of HFNC by outlining the process for initiating, escalating, and weaning support. In the initial needs assessment, however, it was additionally discovered that a large proportion of patients received supplemental oxygen with fraction of inspired oxygen (FiO2) >21% without having had documented hypoxemia (at least 1 desaturation <90% documented by nursing in the medical record). At our institution, patients are transitioned from HFNC to either room air or low-flow nasal cannula with flow rates < 4 L/min at 100% FiO2. We do not have the ability to use an oxygen blender with our low-flow nasal cannula. Thus, we developed specific aims to (1) decrease the proportion of patients started on supplemental oxygen without documented hypoxemia from ∼90% to <70% and (2) decrease the proportion of patients receiving HFNC without supplemental FiO2 who were weaned to oxygen by low-flow nasal cannula from ∼23% to <10%.
This quality improvement report evaluates the implementation of a multidisciplinary evidence-based protocol for initiation, escalation, and weaning of HFNC in children with bronchiolitis at a 255-bed freestanding children’s hospital with 6 specialized medical/surgical units. Most patients with bronchiolitis are admitted to 1 of 2 units specializing in infants and respiratory illnesses, respectively.
Data were obtained via retrospective chart review from September 2018 to September 2021. Initiation of the pilot protocol occurred in November 2019 followed by a revised version in October 2020, which specifically addressed the concern of unnecessary use of supplemental oxygen (Fig 1). Inclusion criteria were patients aged <2 years, admitted as inpatient or observation status with any International Classification of Diseases, Tenth Revision, diagnosis of bronchiolitis (J21.*) who received HFNC during their admission. Patients admitted to the ICU were excluded from chart review. If >40 patients met inclusion criteria in any given month, a computer-randomized sample of 40 patients, stratified by admission unit, was reviewed. Documentation regarding SpO2 and use of supplemental FiO2 was reviewed, in addition to duration of HFNC usage and length of stay. Hypoxemia was defined as at least 1 desaturation with SpO2 <90% that was documented by nursing in the medical record. Patients on HFNC are maintained on continuous pulse oximeter in our institution, and SpO2 data are manually entered by nursing staff.
We initially tracked the progress of our project using run charts, monitoring for shifts, trends, or runs to indicate special cause variation. Once enough data had been collected, the quality improvement team transitioned to statistical process control (SPC) p-charts with upper and lower control limits to determine if our intervention led to a special cause signal. Our SPC charts were created by using Microsoft Excel QI Macros (version 2021.07). We used Montgomery rules for special cause variation to interpret the SPC charts.
Our primary intervention was to create a revised HFNC protocol that clearly stated (1) “Start FiO2 at 21% unless documented desaturation <90%” when initiating HFNC and (2) “Transition straight to room air if SpO2 ≥90% on 21% FiO2” when discontinuing HFNC (Fig 1). We began our education intervention by providing standardized education to all respiratory therapists (RTs) and nurses who provide care to patients on the inpatient units, which was completed by the lead RT and nurse educators, respectively. Additionally, nurses on our 2 primary inpatient units received additional education on how to wean from the HFNC system. Patients on other units were weaned only by RTs to better ensure protocol adherence. Finally, we provided standardized education of this protocol to all residents and hospitalist faculty members who primarily care for these patients. This education was provided during a monthly house staff meeting (October 2020) for residents and monthly morning report presentations from November 2020 to March 2021 for residents and faculty on the 2 primary inpatient units. The protocol was also posted in all respiratory therapy, resident physician workrooms, and primary respiratory unit nursing stations for easy access and review.
Overall, 724 patients' charts were reviewed during the study duration with a geometric mean length of stay of 60.4 hours and geometric mean HFNC duration of 33.9 hours. A total of 597 patients were included in the initiation analysis, with 127 excluded for lacking oxygen saturation data before HFNC, and 594 were included in the inappropriate wean data, with 130 patients excluded either for having had desaturations or for being weaned directly from higher settings (>5 L/min or >21% FiO2). During the initial data review period (September 2018–October 2019), 90.2% of patients were inappropriately started on FiO2 >21% despite no recorded hypoxemia. Over the same period, 23.1% of patients receiving HFNC without supplemental FiO2 were weaned inappropriately to FiO2 >21% via low-flow nasal cannula. No change in either measure was seen after the initial HFNC protocol implementation. Subsequently, the HFNC protocol was revised in October 2020 to specifically address inappropriate use of supplementary oxygen (Fig 1). After implementing the updated protocol, inappropriately initiating supplemental FiO2 despite no documented desaturation occurred for 57.2% of patients, with a downward shift starting in September 2020 and inappropriate weaning to nasal cannula supplemental oxygen decreasing to 4.7%, with a downward shift starting in November 2020 (Figure 2).
At our institution, implementation and education of oxygen-therapy specific changes to a standardized hospital-wide HFNC protocol were associated with decreased inappropriate use of supplemental oxygen for patients admitted with bronchiolitis. Supplemental oxygen appears to be an underrecognized area of overuse in bronchiolitis in the age of widespread HFNC usage. Not only was supplemental oxygen frequently initiated despite no documented hypoxemia, but it was also started when weaning off HFNC with an FiO2 of 21%. Implementing the revised protocol concentrated on education and on specific language, prompting providers initiating HFNC to start at FiO2 21% unless there is a documented SpO2 <90% and to wean to room air unless the patient is already receiving supplemental oxygen.
Our study had limitations of note. This was a study of a single institution, which may limit generalizability, particularly given the cohorts of patients on 2 of the 6 medical/surgical units. Because of the nature of retrospective chart review, we may have included some patients with only transient, not clinically relevant, desaturation that was documented in the chart as hypoxemia. Conversely, we may have missed some patients with hypoxemia in which a desaturation was not documented in the nursing charting. Improved documentation of SpO2 may have affected the proportion of patients receiving supplemental oxygen, although we believe this had a minimal effect on the changes in our outcome measures. Finally, although the COVID-19 pandemic affected the incidence of patients admitted with bronchiolitis at the start of this project, we were able to capture the delayed surge of patients admitted during summer 2021.
In conclusion, supplemental oxygen in the absence of hypoxemia in bronchiolitis is an example of low-value care. Implementing focused, standardized protocols with concurrent education can feasibly decrease inappropriate and unnecessary use of supplemental oxygen in children with bronchiolitis. More research is needed to determine the impact of inappropriate supplemental oxygen use on the duration of illness, length of stay, and acute decompensation.
We acknowledge the Le Bonheur Children’s Hospital High-Flow Nasal Cannula Taskforce for aiding with the design and implementation of this quality improvement effort.
Dr Robinson conceptualized and designed the study, collected data, drafted the initial manuscript, and revised the manuscript. Dr Winer conceptualized and designed the study, coordinated, collected, and analyzed data, reviewed the manuscript, and revised the manuscript. Dr Bettin conceptualized and designed the study, reviewed the manuscript, and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no potential conflicts of interest to disclose.