Decreasing Unnecessary Antibiotic Usage in Patients Admitted With Bronchiolitis

We conducted a retrospective chart review of infants aged ≤2 years hospitalized at our tertiary care center with a primary discharge diagnosis of bronchiolitis identified by using International Classification of Disease, Ninth Revision and International Classification of Diseases, 10th Revision codes during 2 subsequent respiratory seasons: November 1, 2017, to March 31, 2018, and November 1, 2018, to March 31, 2019. Our organization is a free-standing, 358-bed tertiary care children’s hospital serving 12 counties in central California.

Electronic medical charts were reviewed to collect data, including age, comorbid conditions, indication and duration of antibiotic therapy, and outcomes. Outcomes included transfer to the PICU, readmission within 30 days, and mortality. Patients who were aged ≤2 years, admitted with the diagnosis of bronchiolitis and without comorbid conditions, met the inclusion criteria. Significant comorbid conditions, including congenital heart disease, cystic fibrosis, tracheostomy, neurologic impairment, and immunocompromise, were excluded from the study. The ASP team consisted of 1 infectious disease physician (dedicated 0.25 full-time equivalent), a hospitalist, and an ASP pharmacist (dedicated 0.5 full-time equivalent).

Between the 2 study periods, the interventions performed included a single session of didactic review of AAP bronchiolitis guidelines with attending physicians during October 2018 and reinforcement of these guidelines from our ASP team through prospective audit and timely feedback during weekday in-person rounds as well as expansion of ASP service to include all medical floors and PICUs with more focus targeting patients with respiratory tract infections. Antibiotics were considered to be appropriate by the ASP team if they met the following criteria: (1) change in clinical condition requiring fluid resuscitation and increase in respiratory support with or without identification of another source of infection (suspicion of superimposed bacterial infection), (2) change in laboratory investigations with increased white cell count or increase in inflammatory markers like C-reactive protein associated with clinical changes, and (3) abnormal localized chest radiograph findings suspicious of superimposed bacterial pneumonia (SIBP).

Data were analyzed by using Statistical Package for the Social Sciences version 22.0. Descriptive statistics were used to describe the characteristics of the participants. χ² test and Fisher’s exact test (as appropriate) were used to compare patient demographics and antibiotic usage between the 2 respiratory seasons. Comparison of age between the 2 respiratory seasons was analyzed by using the Mann-Whitney U test. Percent change in antibiotics was calculated as the difference in percent of patients on antibiotics in the intervention period divided by percentage of patients on antibiotics preintervention.

We identified a total of 1193 patients who were hospitalized with the diagnosis of bronchiolitis during the 2 respiratory seasons (Fig 1). Patients who did not meet the criteria were excluded (n = 39). Of 1154 patients, 549 were identified in the preintervention period and 605 in the postintervention period. The overall median age was 7.2 months (interquartile range: 4.3–12.0), and 80% of patients did not have comorbid conditions (Table 1). Two hundred twenty-seven (20%) children received concurrent antibiotics throughout the study, 137 (25%) of those in the preintervention and 90 (15%) in the postintervention period.

Overall, the percentage decrease in antibiotic use was 40% between the 2 study periods (25% vs 15%, P < .001). Common indications for antibiotic use in both the pre- and postintervention periods included otitis media (38% vs 52%), pneumonia (49% vs 40%), suspected sepsis (8% vs 3%), Streptococcal pharyngitis (0.7% vs 0%), urinary tract infection (4% vs 3%) and bacteremia (0% vs 1%). Overall, 147 patients were deemed to be on appropriate antibiotics. There was no statistically significant difference in appropriate antibiotic use in the pre- and postintervention periods (63% vs 67%, P = .63).

Among patients who received antibiotics, 103 (45%) were started because of concerns for SIBP (Table 1). Among those with SIBP, 30 (45%) in the preintervention period and 14 (39%) in the postintervention period were determined to be appropriate by the ASP team.

There was no difference in outcome in the pre-and postintervention periods among patients with respect to PICU transfer (17% vs 10%) and readmission (0.7% vs 2%). There was 1 death in the preintervention period, but it was not related to antibiotic prescribing.

Acute viral bronchiolitis contributes to numerous hospitalizations in pediatrics. Efforts to decrease inappropriate antibiotic use in this patient population can have a significant impact on overall hospital antibiotic use.¹²  The majority of efforts to decrease antibiotic use in bronchiolitis have been focused on children who receive antibiotics for potential SIBP.¹³  The AAP practice guideline recommendation regarding the judicious use of chest radiographs is specifically aimed at decreasing antibiotic use based on atelectasis, which could potentially influence clinicians’ antibiotic prescribing in the absence of other objective clinical data that antibiotics are necessary.¹ 

Through the Sharing Antimicrobial Reports for Pediatric Stewardship collaborative, we identified inappropriate antibiotic usage among 27% of patients admitted with suspected or confirmed lower respiratory tract infection. In our study, we have determined that antibiotic usage for this indication can be impacted with easy-to-implement, specific antibiotic stewardship interventions. We were able to achieve a 40% decrease in overall antibiotic use in patients admitted with bronchiolitis during the study period.

Antibiotics are often prescribed in clinical bronchiolitis for concomitant otitis media. The coinfection rate of otitis media with bronchiolitis has been reported to be as high as 53%, which is comparable to our study.¹⁴  For the management of otitis media in bronchiolitis, the guideline recommends use of the AAP otitis media practice parameter. Patients with bronchiolitis are aged <2 years and often have a fever >39°C. Because of this, the watchful waiting recommendation for otitis media frequently does not apply and these patients are prescribed systemic antibiotics, therefore contributing to overall antibiotic use in viral bronchiolitis. Additional efforts with watchful waiting are needed to decrease unnecessary antibiotic use in bronchiolitis patients with concomitant otitis media.

In a study involving US emergency departments, researchers investigated antibiotic prescribing between 2007 and 2015 for pediatric patients with acute bronchiolitis and demonstrated that 25.6% of the patients received antibiotics (an estimated ∼83 000 prescriptions annually).¹⁵  However, 70% of those patients had no documented bacterial coinfection. Rates of bacterial coinfection with bronchiolitis are rare after universal Haemophilus influenza and pneumococcal vaccination. Although bacterial coinfection has been reported in children with respiratory failure secondary to respiratory syncytial virus infection and those requiring ICU admission, overall rates are low.^3,16  In our study, only 1% of patients had documented bacterial coinfection confirmed by cultures.

There is also evidence that physicians’ perception of their prescribing practice is not consistent with their actual prescribing. Moreover, physicians are also interested in improving their prescribing practices. Providing physicians with feedback on their prescribing practices and the appropriateness of the prescribing has been revealed to improve their prescribing practice.¹⁷  With the interventions outlined in this study, prescribing clinicians can receive objective, real-time feedback on their prescribing practices to decrease unnecessary antibiotic use. Based on our results, these interventions can improve appropriate antibiotic prescribing in viral bronchiolitis.

Our study has some limitations. The interventions were performed at a single site over 2 subsequent bronchiolitis seasons. PICU transfer has institutional variability as well as variability based on provider comfort level. We did not have substantial physician or nursing turnover between the 2 seasons, but providers would have an additional year of experience between the seasons, which could potentially account for some differences in prescribing practices.

In summary, targeted stewardship interventions with timely feedback related to antibiotic prescribing in viral bronchiolitis can impact overall use as well as the appropriateness of prescribing.