Application of Step-by-Step and Paediatric Emergency Care Applied Research Network (PECARN) Clinical Decision Aids in the management of young febrile infants in a UK cohort

TL;DR

This study tested two clinical decision tools for identifying low-risk febrile infants in the UK, finding they correctly identified all infants with serious infections but missed many cases.

Contribution

Validated high-sensitivity clinical decision aids for febrile infants in a UK cohort using Procalcitonin data.

Findings

PECARN and Step-by-Step CDAs had 100% and 96% sensitivity respectively for detecting invasive bacterial infections.

Both tools had low specificity (14-15%), identifying only 14% of infants as low-risk.

Step-by-Step misclassified one infected infant as low-risk while PECARN did not.

Abstract

Young febrile infants are at high risk of invasive bacterial infections (IBIs). Clinical Decision Aids (CDA) such as the Step-by-Step and Paediatric Emergency Care Applied Research Network (PECARN) use Procalcitonin (PCT), limiting their application in settings without PCT access. This study aimed to test the performance of these CDAs in a UK cohort.

This was a planned analysis of the Febrile Infant Diagnostic Assessment and Outcome Study, a large, prospective multicentre observational study conducted across over 30 sites in the UK. Febrile infants (0–90 days of age) with complete biomarker data, who also underwent PCT testing, were included. Two CDAs, PECARN and Step-by-Step, were applied to the cohort, using their recommended low-risk criteria. The diagnostic performance of the CDAs was analysed.

Of the 1527 infants who completed biomarker testing in the main study, 442 had PCT testing and were included, 22 (5%) were diagnosed with an IBI. PECARN and Step-by-Step CDAs demonstrated sensitivities of 1.00 (95% CI: 0.85 to 1.00) and 0.96 (95% CI: 0.77 to 1.00) respectively. The PECARN CDA performed with a specificity of 0.14 (95% CI: 0.11 to 0.18) identifying 14% of the participants as low-risk and did not misclassify any infants. The Step-by-Step CDA performed with a specificity of 0.15 (95% CI: 0.12 to 0.19) identifying 14% of the participants as low-risk and misclassifying one participant with IBI as low-risk.

Both PECARN and Step-by-Step CDAs demonstrated high sensitivity for detecting IBI in our cohort. While specificity was relatively low, these tools could potentially identify a subset of low-risk infants suitable for less intensive management.

Figures