Background Tonsillopharyngitis is a common presentation to paediatric emergency departments (PEDs). FeverPAIN (FP) and Centor scoring systems are recommended in the UK to help delineate bacterial aetiology, despite being primarily evidenced in adult populations. We investigate how the use of FP or Centor compares to actual clinician practice in guiding antibiotic prescription rates in PED. We establish current national practice in English PEDs.

Methods We performed a retrospective cohort study of tonsillopharyngitis presentations to a tertiary PED in January–February 2020. Investigators retrospectively assigned each patient FP and Centor scores using documented symptoms. We compared antibiotic prescription rates guided by FP/Centor against the actual rate prescribed by clinicians, and assessed agreement between these strategies using kappa analysis. We contacted 153 English emergency departments to establish national practice.

Results We identified 632 consecutive patients aged 6 months to 15 years. Actual clinician-prescribed antibiotics numbered 116 (18.4%). Prescriptions predicted by FP score numbered 124 (19.6%) and Centor 112 (17.7%). Kappa (K) analysis indicated only moderate agreement between clinician choice versus FP (K=0.434) and clinician choice versus Centor (K=0.476). This was similar for cohorts aged under and over 3 years.

National practice was reportedly heterogeneous, with 70 emergency departments (45.8%) not employing a specific system.

Conclusion Current guidance is variably interpreted and inconsistently implemented in paediatric populations. FeverPAIN and Centor scoring systems may not rationalise antibiotics as much as previously reported compared with judicious clinician practice. Producing clear paediatric-specific national guidelines, especially for under-5s who are omitted from NICE sore throat guidance, may help further rationalise and standardise antibiotic use in paediatric tonsillopharyngitis.