Studies have found infrequent interventions after croup admission. Our objectives were to achieve 25% reduction in (1) admission rate and (2) neck radiograph utilization among patients presenting to the emergency department.
At our tertiary children’s hospital, we implemented clustered interventions including education, guideline, and orderset integration. We included patients 3 months to 8 years old with an emergency department, observation, or inpatient encounter for croup. We excluded patients with direct or ICU admissions, complex chronic conditions, or concurrent asthma, pneumonia, or bronchiolitis. We reviewed a random sample of 60% of encounters from baseline (October 1, 2017 to September 30, 2019) and implementation (October 1, 2019 to September 30, 2020) periods. We conducted a posthoc analysis from October 1, 2017 to December 1, 2021 to assess sustainment during coronavirus disease 2019. Interrupted time series analysis was used to evaluate changes in outcome, process, and balancing measures.
There were 2906 (2123 baseline and 783 implementation) encounters included. Extrapolating preintervention trend estimates, the baseline admission rate of 8.7% decreased to 5.5% postintervention (relative decrease 37% [95% confidence interval: 8 to 66]) and sustained over 26 months after implementation. Admission rate in patients receiving 2 or fewer racemic epinephrine was significantly lower in implementation (1.7%) compared with baseline (6.3%), relative decrease of 72% (95% confidence interval: 68 to 88). There were no significant changes in neck radiographs, length of stay, or revisits.
Croup quality improvement interventions were associated with a significant decrease in hospital admissions with no increase in revisits.
Croup is a common viral respiratory illness among young children. Treatment includes systemic corticosteroids and administration of racemic epinephrine (RE) in more severe cases.1 Although the majority of croup treatments are administered in the emergency department (ED), hospitalization in young children is common, with a median admission rate of 9.1% across US children’s hospitals.2 ∼80% of patients do not need further RE or airway interventions, such as oxygen or intubation, after admission.3–5
There are notable opportunities for improving the value of care for patients presenting to the ED with croup. In 1 study, admitted patients who did not require additional RE or airway intervention had a median length of stay (LOS) of only 16.5 hours. However, their total costs were 3.1 times higher compared with ED discharged patients.3 Despite the low rate of intervention after admission, recommendation for hospitalization after 1 or more RE doses is cited in textbooks with little evidence of clinical benefit or value to the healthcare system.6–9
Our organization had an admission rate of approximately 9.0% for croup. We found that only 20.8% of patients had received 3 or more doses of RE before admission, suggesting an opportunity to refine admission guidance.3 We have previously used local clinical guidelines to drive improvements in patient outcomes and applied similar strategies to croup.10–13
The primary prespecified aim of our quality improvement (QI) initiative was to reduce admission for croup by 25% after implementation of a clinical guideline. Additionally, we aimed to achieve a 25% reduction in neck radiographs within a similar period.
Methods
Context
This QI initiative was conducted at a tertiary children’s healthcare organization in the midwestern United States with 2 free-standing hospitals that see approximately 2000 ED croup encounters annually.3 Patients requiring observation are admitted to the hospital from the ED as there is no separate observation area. There was no pre-existing organizational croup guideline. In July 2017, a multidisciplinary croup QI workgroup, including representatives from ED and hospital medicine, was formed. They reviewed existing literature and used the Model for Improvement14 to develop a key driver diagram (Fig 1), explore potential measurements, and propose interventions. Gaps in literature were identified as a key driver, and the workgroup focused on first conducting a retrospective study to better understand croup outcomes at our organization.3
Interventions
The croup workgroup identified several key interventions through the driver diagram (Fig 1). While these were originally structured as Plan-Do-Study-Act (PDSA) cycle, challenges with obtaining data prospectively precluded the ability to monitor main outcome variables in real time or conduct rapid-cycle changes. Thus, we considered our QI initiative with a single group of clustered interventions, including education, development, and integration of a clinical guideline and orderset. Education to ED and hospitalist providers included a didactic plus discussion session (ED) or newsletter (hospitalist), reviewing key literature, and the results of our previous study on croup outcomes within our population.3
The main intervention (October 1, 2019) was implementation of an evidence-based clinical guideline and orderset for croup (Supplemental Fig 1) embedded in the electronic health record (EHR). The guideline encourages early initiation of systemic steroids, recommends up to a 2 hour ED observation time after each RE dose with consideration for admission after 3 total doses (including those received at outside facilities). The guideline also notes avoiding imaging but does recommend further workup if a patient fails to respond to RE in an effort to not miss croup mimicking conditions (eg, bacterial tracheitis or foreign body aspiration). The guidelines and ordersets were reviewed in staff meetings or e-mail updates between August 1, 2019 and November 20, 2019. We elected to use October 1, 2019 as the implementation date because recommendations were formalized and published in clinical decision support tools on that day.
Study of the Interventions
We conducted a prospective QI initiative with a formal retrospective analysis of patient data to determine if the aims were achieved.
We included children 3 months to 8 years old with an ED, observation or inpatient encounter with an International Classification of Disease, Tenth Revision (ICD-10) discharge diagnosis of croup (J05.×, J38.×,) between October 1, 2017 and September 30, 2020. We collected age and race and ethnicity to assess for differences in study groups that might impact disposition decision making. We used Emergency Severity Index scores as a marker of disease severity. We excluded patients if they were directly admitted from an outside facility, admitted initially to the ICU, had a complex chronic medical condition,15 concurrent diagnosis of asthma, pneumonia, bronchiolitis, or if general consent to use research data were not signed. Exclusions were applied during data extraction from the EHR and were confirmed upon chart review.
Measures
The study period was divided into a 24 month baseline (October 1, 2017 to September 30, 2019), and a 12 month intervention period (October 1, 2019 to September 30, 2020). The primary outcome measures were percentage of patients admitted after an ED encounter for croup and percentage of patients with neck radiographs. Our secondary outcome measure was average encounter hospital charges (inflation adjusted to April 2021 United States dollars using the Consumer Price Index16 ).
The process measure was admission rate among patients who received 2 or fewer RE doses before or during ED encounter, because this subpopulation is the target for whom admissions could be reduced.
Balancing measures included rate of inpatient intervention, such as additional RE dosing or airway intervention (oxygen or heliox therapy, intubation, or ICU admissions3 ), that we a priori assumed might increase as we shifted admission recommendations. We also measured ED and total hospital LOS and all-cause ED or hospital revisits within 72 hours of discharge.
Analysis
Demographics, clinical presentation, medical management, and measures were compared during baseline and intervention periods using Pearson χ-Square test and Welch’s t test. To account for potential trends in the baseline period, interrupted time series analysis were used to evaluate changes in outcome, process, and balancing measures after the start of the intervention in October of 2019. Specifically, a level and slope change model was assumed for all models and the baseline trend line was extrapolated into the intervention period to compare baseline and postintervention estimates.17,18 Logistic and linear regression was used for binary and continuous outcomes, respectively, and bootstrap methods were used to estimate standard errors.19
Given the disruption of the healthcare system by coronavirus disease 2019 (COVID-19), we conducted a post hoc analysis to assess the sustainability of our intervention. We pulled data directly from the EHR for encounters from October 1, 2017 to December 1, 2021 of patients 3 months to 8 years old with an ED, observation or inpatient encounter with an ICD-10 discharge diagnosis of croup (J05.×, J38.×,). We excluded patients if they were directly admitted from an outside facility, admitted initially to the ICU, had a complex chronic medical condition,15 concurrent ICD-10 diagnosis of asthma, pneumonia, or bronchiolitis. We were unable to conduct chart review for this cohort, so it likely includes encounters that would have been excluded using our original methodology. We compared outcomes available from the EHR including admission, ED LOS, and all-cause revisit across baseline period (October 2017 to September 2019), implementation period before COVID-19 pandemic (October 2019 to March 2020), and implementation period after start of COVID-19 pandemic (March 2020 to December 2021).
For all analyses we set the significance level at 5% and used Stata, R, and/or SAS for analyses.
Ethical Considerations
This study was approved by the organization’s Institutional Review Board.
Results
There were 5497 eligible croup encounters in the study period. We randomly selected 3304 (60%) for chart review, of which 2906 (2123 baseline, 783 implementation) were included in analysis, (Fig 2). Median age was 23 months and 63.3% of patients were male (Table 1). There was a significant difference in race and ethnicity between study periods, with a higher proportion of Non-Hispanic White and Asian patients evaluated for croup during implementation compared with baseline (50.5 versus 46.9% and 9.1 versus 7.4%, P = .003).
. | Baseline (N = 2123) . | Implementation (N = 783) . | . | ||
---|---|---|---|---|---|
Demographic . | Frequency . | Column % . | Frequency . | Column % . | P . |
Age in months, median (IQR) | 23 (1–40) | — | 23 (13–40) | — | .86 |
Sex, female | 779 | 36.7 | 288 | 36.8 | .97 |
Race and ethnicity | .003 | ||||
Asian | 157 | 7.4 | 71 | 9.1 | |
Hispanic | 217 | 10.2 | 63 | 8.1 | |
Black or African-American | 388 | 18.3 | 115 | 14.7 | |
Multi-race | 196 | 9.2 | 59 | 7.5 | |
Other or unknown | 169 | 8.0 | 80 | 10.2 | |
Non-Hispanic White | 996 | 46.9 | 395 | 50.5 | |
Primary payer | .13 | ||||
Private or military | 1088 | 51.3 | 431 | 55.0 | |
Public | 912 | 43.0 | 304 | 38.8 | |
Other or unknown | 123 | 5.8 | 48 | 6.1 | |
Medical history | |||||
Significant medical historya | 663 | 31.2 | 264 | 33.7 | .20 |
Prearrival and ED encounter intervention | |||||
Seen at outside facility within 24 h | 355 | 16.7 | 117 | 14.9 | .25 |
Racemic epinephrine provided at outside facility | 147 | 6.9 | 45 | 5.8 | .44 |
Steroid given at outside facility within 24 h | 195 | 9.2 | 68 | 8.7 | .68 |
Steroids given in ED | 1944 | 91.6 | 709 | 90.5 | .43 |
Median time to steroid given in ED, min, median (IQR) | 58 | (38–99) | 62 | (40–106) | .18 |
Emergency severity index triage score | .65 | ||||
Level 1–2 | 612 | 28.8 | 232 | 29.6 | |
Level 3 | 711 | 33.5 | 248 | 31.7 | |
Level 4–5 | 800 | 37.7 | 303 | 38.7 |
. | Baseline (N = 2123) . | Implementation (N = 783) . | . | ||
---|---|---|---|---|---|
Demographic . | Frequency . | Column % . | Frequency . | Column % . | P . |
Age in months, median (IQR) | 23 (1–40) | — | 23 (13–40) | — | .86 |
Sex, female | 779 | 36.7 | 288 | 36.8 | .97 |
Race and ethnicity | .003 | ||||
Asian | 157 | 7.4 | 71 | 9.1 | |
Hispanic | 217 | 10.2 | 63 | 8.1 | |
Black or African-American | 388 | 18.3 | 115 | 14.7 | |
Multi-race | 196 | 9.2 | 59 | 7.5 | |
Other or unknown | 169 | 8.0 | 80 | 10.2 | |
Non-Hispanic White | 996 | 46.9 | 395 | 50.5 | |
Primary payer | .13 | ||||
Private or military | 1088 | 51.3 | 431 | 55.0 | |
Public | 912 | 43.0 | 304 | 38.8 | |
Other or unknown | 123 | 5.8 | 48 | 6.1 | |
Medical history | |||||
Significant medical historya | 663 | 31.2 | 264 | 33.7 | .20 |
Prearrival and ED encounter intervention | |||||
Seen at outside facility within 24 h | 355 | 16.7 | 117 | 14.9 | .25 |
Racemic epinephrine provided at outside facility | 147 | 6.9 | 45 | 5.8 | .44 |
Steroid given at outside facility within 24 h | 195 | 9.2 | 68 | 8.7 | .68 |
Steroids given in ED | 1944 | 91.6 | 709 | 90.5 | .43 |
Median time to steroid given in ED, min, median (IQR) | 58 | (38–99) | 62 | (40–106) | .18 |
Emergency severity index triage score | .65 | ||||
Level 1–2 | 612 | 28.8 | 232 | 29.6 | |
Level 3 | 711 | 33.5 | 248 | 31.7 | |
Level 4–5 | 800 | 37.7 | 303 | 38.7 |
—, not applicable; IQR, interquartile range.
Significant past medical history was defined as history of any of the following: prematurity (<37 wk gestation), asthma, croup, tracheal anomalies, or intubation.
The primary outcome of admission rate was significantly lower in the intervention period, 5.5%, compared with the baseline period, 10.2% (Table 2, P < .001); the relative decrease of 46% (95% confidence interval [CI]: 29 to 63) met our aim of a 25% decrease. Interrupted time series analysis extrapolated a slight downward preintervention trend (Fig 3, Supplemental Table 1 for parameter estimates) and estimated a baseline rate of 8.7% for a relative decrease of 37% (95% CI: 8 to 66). Posthoc analysis showed that the improvements were sustained in a 26-month period following implementation; including 6 months before the COVID-19 pandemic (relative decrease of 38% from baseline, 95% CI 23 to 53) and 20 months during the COVID-19 pandemic (relative decrease of 49%, 95% CI 40 to 59, Supplemental Table 3).
. | Baseline, N = 2123 Mean (SD)or % (n) . | Implementation, N = 783, Mean (SD) or % (n) . | Pa . |
---|---|---|---|
Rate of admission | 10.22 (217) | 5.49 (43) | <.001 |
Rate of admission, RE ≤2 dosesb | 6.3 (127/2013) | 1.7 (13/747) | <.001 |
Rate of admission, RE >2 dosesb | 81.8 (90/110) | 83.3 (30/36) | .99 |
Neck radiograph obtained | 8.7 (184) | 7.7 (60) | .39 |
Length of stay | |||
Hours if ED discharge | 2.5 (1.5) | 2.6 (1.4) | .23 |
Hours if admission | 20.7 (14.7) | 19.7 (11.8) | .62 |
All cause revisit or readmission within 72 h | 2.9 (62) | 1.8 (14) | .09 |
Transfer to ICU after admission | 0.09 (2) | 0.13 (1) | .99 |
Total charges,c 1000 dollars | 1572.8 (2376.94) | 1422.24 (1700.64) | .06 |
. | Baseline, N = 2123 Mean (SD)or % (n) . | Implementation, N = 783, Mean (SD) or % (n) . | Pa . |
---|---|---|---|
Rate of admission | 10.22 (217) | 5.49 (43) | <.001 |
Rate of admission, RE ≤2 dosesb | 6.3 (127/2013) | 1.7 (13/747) | <.001 |
Rate of admission, RE >2 dosesb | 81.8 (90/110) | 83.3 (30/36) | .99 |
Neck radiograph obtained | 8.7 (184) | 7.7 (60) | .39 |
Length of stay | |||
Hours if ED discharge | 2.5 (1.5) | 2.6 (1.4) | .23 |
Hours if admission | 20.7 (14.7) | 19.7 (11.8) | .62 |
All cause revisit or readmission within 72 h | 2.9 (62) | 1.8 (14) | .09 |
Transfer to ICU after admission | 0.09 (2) | 0.13 (1) | .99 |
Total charges,c 1000 dollars | 1572.8 (2376.94) | 1422.24 (1700.64) | .06 |
Welch’s t test for continuous variables. Pearson χ2 or Fischer’s Exact Test for categorical variables.
Racemic epiephrine (RE) doses given before or during ED encounter.
Inflation-adjusted for April 2021.
Rates of neck radiographs were comparable in the intervention and baseline periods (7.7% versus 8.7%, P = .39, Table 2). The relative decrease of 12% (95% CI: −13 to 37) did not meet our aim of a 25% decrease. Interrupted time series analysis produced similar results (Supplemental Fig 2). For the secondary outcome of charges, the mean total charges decreased, but the difference was not statistically significant ($1573 baseline versus $1422 implementation, P = .06, Table 2).
Among 2760 (95%) patients who received 2 or fewer doses of RE before or during their ED encounter, the process measure of admission rate was significantly lower in the intervention period, 1.7%, compared with the baseline period, 6.3%, for a relative decrease of 72% (95% CI: 68 to 88). Interrupted time series analysis estimated a baseline rate of 4.0% for a relative decrease of 57% (95% CI: 24 to 89; Fig 4, Supplemental Table 1). In contrast, admission rates among patients who received more than 2 doses of RE were considerably higher throughout and remained comparable in both periods (81.8 versus 83.3, P = .99, Table 2).
There were no significant changes between baseline and implementation in balancing measures including rate of inpatient airway intervention (23.0% versus 27.9%, P = .49), ED LOS (2.51 versus 2.58 hours, P = .23), admission LOS (20.74 versus 19.7 hours, P = .62), or 72 hour all-cause ED revisit (2.9% versus 1.8%, P = .09). Interrupted time series analyses were similar (Supplemental Figs 2–4). Posthoc analysis found no change in 72 hour all cause revisits. However, we did identify a small increase in ED LOS (absolute difference 0.5 hours [95% CI 0.4 to 0.6]) in the 20 month implementation period following the onset of the COVID-19 pandemic (Supplemental Table 3).
Discussion
In this study we describe a QI initiative to reduce admissions for croup at a large tertiary children’s hospital. We achieved a 37% relative reduction in hospital admission following ED encounters for croup. Improvements were sustained over 26 months after implementation. There was no reduction in neck radiographs. Notably, there was no significant change in ED revisits or readmissions following the initiative.
Although the demographics and initial severity were relatively similar between baseline and implementation periods, there was a higher proportion of Non-Hispanic White and Asian patients during implementation period. As race and ethnicity is a social construct, we would not expect these differences to be related to disease susceptibility, rather there may have been different patterns of health system use. Previous studies found sharper declines in ED visits for non-Hispanic Black children compared with non-Hispanic White children during early 2020, coinciding with our implementation period.20
We suspect that implementation of our local clinical guideline was the key strategy for improvement. Though implementation of clinical guidelines for other pediatric conditions have led to improved patient outcomes and reduced healthcare costs,11,21,22 there is surprisingly little literature surrounding the use of croup guidelines to improve resource utilization and patient outcomes. One study used a pre and post design to examine the impact of a croup guideline in an Australian hospital. Although the study found a 34.9 percentage point decrease in admission, generalizability is limited because of the small sample size and a substantially higher baseline admission rate (52.9%) than the United States (9.1%).7
Our clinical guideline focused on changing the number of RE doses prompting admission, a key element related to disposition determination in patients with croup. In a previous study, 71.3% of patients seen in the ED with receipt of 2 doses of RE were hospitalized for further monitoring, yet only 16.9% of those patients required additional RE.3 Our guideline recommended monitoring ED patients up to 2 hours after their second RE dose to determine if a third dose of RE was needed and admission was recommended. Our results show that we reduced the admission rate among patients who received 2 or fewer RE doses by over 50%. Having local data to reinforce our new guideline recommendations as well as support from the ED medical director to potentially increase this ED observation time were keys for success.
Although we noted change in admission rate before the COVID-19 pandemic began, we likely saw ongoing impact of COVID-19 on encounter volumes and admission decision making. Previous studies have shown decreased respiratory infections in children during the latter half of 2020, likely because of the increase in social-distancing and hygiene measures.23,24 As with many other organizations, we also saw a decline in overall and croup-related encounters during the COVID-19 pandemic. Whereas COVID-19 has been described as potentially being associated with croup in a small case series, it was not recognized as a common presenting symptom during our study period.25–31 However, recent work has suggested that the most recent ο variant may be associated with increased croup prevalence and severity.32–34
We suspect that during the study period, COVID-19 had a mixed impact on disposition of patients with croup. Some providers may have elected to admit patients with respiratory illness if they felt that COVID-19 was possible. Alternately, some providers may have been less likely to admit patients to prevent potential hospital exposures or because of caregiver preferences. Because of this dynamic impact of COVID-19 on our primary outcome, we did a post hoc analysis to include an additional 14 months of data, for a total of 26 months postimplementation. We found reductions in admission for croup in the 6 month implementation period before the onset of the pandemic, and these reductions were sustained through for over a year and a half into the pandemic. However, we cannot entirely disentangle the impact of our intervention from the effect of the COVID-19 pandemic.
We did not observe a significant decrease in neck radiographs after implementation of our guideline. This result is likely caused by a low percentage of neck radiographs ordered in the baseline period (9%), which may leave little room for improvement but may also reflect a lack of dedicated QI resources and focused PDSA cycles on this metric. In a recent study including over 6000 children with croup across United States children’s hospitals, the median percentage of patients with a lateral neck radiograph was 21% (range of 8% to 51%).35
We found a modest reduction in total charges of $150 per encounter on average, though this did not reach statistical significance. Notably, our ED does not have a dedicated observation area, thus patients who are in the hospital on “observation” status (eg, watching overnight or for more than several hours) would typically be under the care of the hospitalist team in a short-stay unit or medical-surgical unit. Thus, patients may incur additional charges from professional fees or room and board.
Importantly, there were no significant changes in balancing measures, including inpatient airway interventions, croup-related revisits or LOS. We a priori assumed inpatient airway interventions might increase, as patients with milder illness were being admitted less frequently. However, this did not reach statistical significance, likely given the low number of admitted patients having these interventions.
We were surprised that ED LOS did not increase during implementation. Our data of the early phase of the pandemic suggest providers may have modified their approach to the management of children with croup. For instance, ED providers may have been more likely to forego RE in mild cases to minimize aerosol generating procedures. Additionally, as this was a period of low ED volumes, there may have been improved throughput related to census. Posthoc analysis found a small increase by 0.5 hours on average in ED LOS, which may reflect stabilization of ED processes over the course of the COVID-19 pandemic.
Our findings have implications for value in croup care. Fewer hospital admissions for croup could lessen the burden to caregivers and families in missed work or school, and the stress of staying in the hospital. Previous studies have estimated there are ∼18 287 annual hospitalizations for croup in the United States. Hospitalizations associated with parainfluenza virus alone (a common cause of croup) cost an estimated 58 million dollars annually.36 Even moderate reductions in hospitalization rate and encounter charges, paired with no increase in readmissions, would have the potential for millions of dollars of annual healthcare savings and improving value-based care delivery.
There were several limitations to our study. The experience of a single children’s healthcare organization may not be generalizable to the United States population. Lack of timely data for targeted audit and feedback affected the ability to conduct defined prospective PDSA cycles.37 We did not examine alternate reasons for admission, such as caregiver preferences or dehydration. Our organization does not use standardized croup respiratory scores (eg, Westley score), which limited our ability to compare illness severity across implementation periods.38 To account for this, we analyzed Emergency Severity Index scores. Additionally, we excluded patients who were initially admitted to the ICU and thus, findings should not be applied to critically ill patients.
Given unanticipated shifts in resource allocation related to COVID-19, we were not able to accomplish several interventions identified on our key driver diagram (eg, nurse-driven steroid administration protocol). Additional strategies and interventions may result in future improvements as resources allow.
Conclusions
A croup guideline, limiting hospital admission until 3 doses of RE are needed, led to a 37% reduced rate of hospital admission without significant increase in revisits or readmissions. These improvements were seen in 6 months before COVID-19 and were sustained in an additional 20-month postimplementation period during COVID-19. Clinical guidelines may be 1 strategy to reduce unneeded resource use for patients with croup.
Acknowledgments
We thank Timothy Barnes, PhD, MPH; Jodi O’Neill, APRN-CNP; Gloria Swanson, MD; Tracey McGuinn, MD; Paul Zenker, MD; Gigi Chawla, MD; Robert Sicoli, MD; Julia Kancans; Kylie Burgess; Angela Bornhoft; Megan Udoeyop; Brooke Hall; and Greta Goetz.
Dr Hester conceptualized and designed the study, assisted with data collection and interpretation of data, and drafted the initial manuscript; Dr Watson assisted with the study design, analyzed data, assisted with interpretation of data, and built the control charts; Ms Nickel assisted with study design, managed the database and acquisition of data, assisted with data analysis and interpretation of data, and prepared tables and figures; Drs Maalouli and Bergmann assisted with study design, data acquisition, and interpretation of data; and all authors critically reviewed and revised the manuscript, approve of the final manuscript as submitted, and agree to be accountable for all aspects of the work.
Accepted for publication.
FUNDING: This study was funded by an organizational internal grant program to support chart review by trained research assistants. No external funding.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest to disclose.
Comments