Objectives Bronchiolitis is the commonest reason for hospital admission in infants in the first year of life. ¹ Bronchiolitis, a viral chest infection, is a clinical diagnosis with no effective treatment beyond supportive care. International guidelines are consistent in their recommendations against the use of routine chest xrays in bronchiolitis.^{2 3 4} In 2017, the Paediatric and Child Health Division of the Royal Australasian College of Physicians, nominated use of CXR in bronchiolitis amongst the top 5 ‘not to do’ practices in paediatrics.⁵ CXR in bronchiolitis do not discriminate well between bronchiolitis and pneumonia, with a Canadian study, demonstrating that 133 CXRs need to be performed to find one alternative diagnosis.⁶ In addition, the rate of antibiotics increased five-fold following an x-ray. We aimed to assess rates of CXR use in our institution and the impact of CXR on antibiotic prescribing.

Methods Our institution is a major specialist paediatric hospital. The Electronic Medical Record was introduced in April 2016. We extracted data from the EMR between April 2016 and June 2018 to assess rates of CXR orders and antibiotic prescription in infants diagnosed with bronchiolitis in the Emergency Department. A smaller chart review on 98 CXR orders for patients with a diagnosis of bronchiolitis ordered a CXR over Winter 2017 was performed to assess indication and compare radiologist reports with decision to prescribe antibiotics.

Results Amongst 3897 infants, we found 11.2% (n=439) were ordered a CXR. A smaller chart review audit of 98 CXR ordered for infants with a diagnosis of bronchiolitis in Winter 2017 demonstrated 65% of these were ordered to rule out pneumonia (65%). When comparing antibiotic use with radiologist report of consolidation, we found no significant different in treatment decisions between those whose CXR was suggestive of pneumonia and those who had normal reported findings (p=0.40). Of those with reported consolidation on CXR who did not receive antibiotics (n= 8), there were no adverse events. The overall rate of antibiotic prescription was 5.1% but rose to 28% in those who were ordered a CXR and 59% in those whose indication for CXR was to rule out pneumonia. The rate of antibiotics prescription in those not ordered a CXR, was 2.1% which was 10-fold lower and statistically significant (p <0.0001)

Conclusions CXR are difficult to interpret in bronchiolitis due to an overlap in the findings on CXR in viral and bacterial infections. In addition, it is difficult to obtain a true inspiratory CXR in infants and suboptimal inspiration can mimic lung pathology. As with previous studies in this area, we demonstrate an ongoing high rate of CXR ordering to exclude pneumonia with subsequent over-use of antibiotics.This is concerning, both for the unnecessary side effects for our infants, but also for antibiotic stewardship.