Variation in Evaluation and Management of Febrile Infants 61 to 90 Days Old

Key Takeaways
- Management patterns shifted over time, with less urine, blood, and cerebrospinal fluid testing, fewer hospitalizations and parenteral antibiotics, and more respiratory viral testing and procalcitonin use.
- Hospital-level variation was substantial in diagnostic testing and hospitalization during 2022 to 2023, while invasive bacterial infection prevalence stayed low and similar across hospitals.
- The authors concluded that decision support tools are needed to standardize management and reduce practice variation.
Investigators conducted a retrospective cohort study of noncritically ill, previously healthy febrile infants aged 61 to 90 days with temperatures of at least 38°C. The trend analysis included 9041 infants across 12 Pediatric Emergency Care Applied Research Network Registry emergency departments from 2017 to 2023, and the variation analysis included 4448 infants across 17 emergency departments during 2022 to 2023. Outcomes included urine, blood, cerebrospinal fluid, and respiratory viral testing, antibiotic administration, hospitalization, and invasive bacterial infections.
From 2017 to 2023, urine testing decreased from 53.8% to 45.7%, blood cultures decreased from 35.0% to 18.9%, and cerebrospinal fluid testing decreased from 3.8% to 1.1%, with P less than .01 for each trend. Hospitalizations also decreased from 22.6% to 16.9%, and parenteral antibiotics decreased from 10.2% to 5.8%, again with P less than .01. Over the same period, respiratory viral testing increased from 13.4% to 63.3%, and procalcitonin use increased from 1.6% to 14.2%, with P less than .01 for both measures.
During 2022 to 2023, median urine testing across hospitals was 46.3%, with a range of 32.2% to 73.9%, and median blood culture use was 21.1%, with a range of 8.0% to 51.1%. Median hospitalization was 16.2%, ranging from 7.9% to 26.1%, while median respiratory viral testing was 71.8%, ranging from 34.7% to 95.5%. Median procalcitonin use was 13.4%, ranging from 0% to 37.0%, and median invasive bacterial infection prevalence was 0.3%, with a range of 0% to 1.2%, remaining low and similar across hospitals.
The authors concluded that decision support tools are needed to standardize management and reduce practice variation.