Reference: Umana E, et al. Performance of clinical decision aids for the care of young febrile infants: A multicenter prospective cohort study. eClinicalMedicine Lancet December 2024

Date: March 6, 2025

Dr. Demetris Athanasiou

Guest Skeptic: Dr. Demetris Athanasiou is a paediatric registrar based in London and enrolled in the PEM MSc program through Queen Mary University in London.

Case: A 6-week-old boy is brought by his parents to your emergency department (ED) for fever. His older sister has been sick with upper respiratory symptoms for the past week but seems to be recovering. Today, while his father was feeding him a bottle, he noticed that the baby was feeling warm and took his temperature, which was 38.2°C (100.7 °F). The boy has otherwise been feeding and acting normally. You examine the baby with an astute medical trainee. As you discuss the next steps in management, she asks you, “I know there’s a bunch of guidelines or decision tools to help risk stratify which babies are low risk for bacterial infections, but I can never keep them straight. Is there one you prefer?”

Background: Back in the day, we were performing lumbar punctures (LP) on febrile infants up to 3 months of age because there was concern for bacterial infections. We used to lump urinary tract infections, bacteremia, and meningitis under one umbrella term, “serious bacterial infection” or SBI. Recently, we’ve been told to stop using that term and be more specific about what we are referring to. Bacteremia and meningitis have been termed invasive bacterial infections (IBI) and, fortunately, are rare, occurring in 1-4%.

There have been several guidelines and clinical decision tools, such as those developed by the National Institute for Health and Care Excellence (NICE), the American Academy of Pediatrics (AAP), and others that offer strategies to identify low-risk infants who might avoid invasive procedures like a lumbar puncture.

These clinical decision tools have been developed to stratify febrile infants into high- and low-risk categories to balance the risk of under-treatment and over-treatment. Several of these tools have been reviewed on the SGEM.

The hot new test is procalcitonin. Unfortunately, it’s expensive, and not all EDs have access to it or can receive the results promptly to help with decision making. Some are still using other inflammatory markers like C-reactive protein (CRP).

With ongoing research and evolving guidelines, the clinical utility of these decision tools continues to be refined. Understanding their strengths, limitations, and applicability in various healthcare systems remains a crucial aspect of evidence-based emergency medicine.


Clinical Question: How well do various clinical decision aids perform in identifying febrile infants at low risk for invasive bacterial infection?


Reference: Umana E, et al. Performance of clinical decision aids for the care of young febrile infants: A multicenter prospective cohort study. eClinicalMedicine Lancet December 2024

  • Population: Infants from birth to 90 days of age from across 35 paediatric EDs and paediatric assessment units across the UK and Ireland with fever ≥38°C
    • Excluded: Guardians who declined or withdrew consent
  • Intervention: Application of clinical decision aids (CDA) [American Academy of Pediatrics (AAP), British Society Antimicrobial Chemotherapy (BSAC), National Institute for Health and Care Excellence (NICE) NG143, Aronson]
  • Comparison: Against each other and “treat all” approach
  • Outcome:
    • Primary Outcome: Diagnostic accuracy of CDAs
    • Secondary Outcomes: Etiology of IBI, clinical predictors of IBI, and mean cost per patient
  • Trial: Prospective multicenter cohort study

Guest Author : Dr. Etimbuk Umana (Timbs) is a consultant in emergency medicine and lead author of the FIDO study.

Authors’ Conclusions: The AAP and BSAC CDAs are highly sensitive at excluding IBI, with a cost saving to hospital services when compared to a treat all approach. The substitution of CRP for PCT made no difference to the performance of the AAP CDA in this cohort and was more costly.”

Quality Checklist for Observational Study:

  1. Did the study address a clearly focused issue? Yes
  2. Did the authors use an appropriate method to answer their question? Yes
  3. Was the cohort recruited in an acceptable way? Yes
  4. Was the exposure accurately measured to minimize bias? Yes
  5. Was the outcome accurately measured to minimize bias? Unsure 
  6. Have the authors identified all-important confounding factors? Unsure
  7. Was the follow up of subjects complete enough? Yes
  8. How precise are the results? Pretty precise
  9. Do you believe the results? Yes
  10. Can the results be applied to the local population? Yes, to the UK and Ireland pediatric populations
  11. Do the results of this study fit with other available evidence? Yes
  12. Funding of the Study: No financial conflicts of interest

Results: There were 1,821 infants included with a median age of 46 days, 61% male, 14% had comorbidities present, and 58% appeared unwell. There were 67 (3.7%) infants who were diagnosed with IBI.

  • 62 had bacteremia
  • 9 had bacterial meningitis
  • 4 infants had both bacteremia and meningitis

Key Results: BSAC and AAP CDAs had the highest sensitivities, while NICE NG143 and Aronson CDAs had the highest specificities.


The AAP and BSAC CDAs each misclassified one infant as low risk who was diagnosed with IBI. Both were in the 29-to-60-day age range and presented very early after fever onset.

Tune into the podcast to hear Dr. Umana answer our questions.

Public and Patient Involvement

We love that there was involvement from the public continuously throughout this study right from the start.

Were there any changes that you made based on their insights and experiences?

Included/Excluded Patients

The FIDO study included many different existing CDAs. Among the patients included in the study, close to 60% were categorized as “unwell appearing.”

At least some of these existing guidelines, like the AAP guidelines, are meant to be applied to “well-appearing” febrile infants.

CDAs are not meant to supersede clinician judgement so if we encounter an infant with a fever who looks sick, they have already fallen off any existing algorithm, and we are likely performing the full workup [1].

Can you comment on the decision to include the unwell-appearing febrile infants?

Some of the infants were excluded due to missing data. How do you think that could have biased the results?

Clinical Risk Factors

The study identified four independent clinical risk factors for IBI. Among them was the clinical opinion of IBI likely (p < 0.001). Previous studies on clinical observations such as the Yale Observation Scale Score demonstrate poor reliability in identifying infants with IBI [2].

Why do you think that the clinical opinion of IBI in this study was significant? Do you think it has anything to do with the inclusion of unwell-appearing infants?

Missed Cases

It’s mentioned that because the standard of care was followed, not all infants underwent blood testing, and it was assumed that the data would have been in the normal range.

There was a group of participants who did not have cultures or PCR testing done. It was assumed that these infants did not have IBI if they were not found to have been diagnosed with it within seven days of discharge on checking hospital records.

Is it possible this method may have missed some children who did not re-present to the hospital or presented a center that was not among the 35 included in the study?

Role of Viral Testing

Some of the infants were tested for respiratory syncytial virus, influenza, and SARS-CoV-2 [3].

Around three-quarters (76.6%) of infants had respiratory viral testing. Of those tested, close to a quarter (24.5%) were positive for one of the viruses we mentioned. Of the infants who had a positive viral test, 1.5% had IBI compared to 3.8% who had a negative viral test. When you looked at infants 29 days or older, the rate of IBI was 0.7% for infants with a positive viral test compared to 3.2% with a negative viral test. Both differences were statistically significant.

We talk about the difference between statistical significance and clinical significance on the SGEM quite a bit. Do you think this difference of 0.6% overall or 2.5% in the older than 29 days group is clinically significant?

Bonus Question: It looks like there was a portion of the patients who did not have urine testing (19%) and/or blood testing (11%). We also noticed that your study included ages up to 90 days. The AAP guidelines go up to 60 days.  There is a lot of variation in workup for infants greater than 2 months (specifically in the 2 to 6 month range) [4].

Can you comment on why you included up to 90 days. Did the infants who did not receive blood or urine testing tend to be >60 days?

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We agree with the authors’ conclusion.


SGEM Bottom Line: There are many CDAs to help risk stratify the care of febrile infants. Of the ones included in this study, AAP and BSAC are the most sensitive. Use these CDAs in conjunction with your clinical judgment.


Case Resolution: You tell your trainee that there are many CDAs out there to help us risk stratify febrile infants. This is because it is often difficult to rely solely on our clinical exam to be accurate. We do not want to miss infections such as UTIs, bacteremia, and meningitis. Currently, many CDAs recommend testing blood and urine. Inflammatory markers such as C-reactive protein and procalcitonin can be used to guide decision making about whether a lumbar puncture should be performed in slightly older infants. Unfortunately, procalcitonin is not available where you practice.

After discussion, you use the AAP CDA with CRP to help guide the work up.

Clinical Application:

Given the evidence from the FIDO study, Dr. Umana would use AAP or BSAC to risk stratify febrile infants in the UK and Ireland. This is more cost effective than a treat all approach and could limit LP and antibiotics in up to 20% of cases.

What Do I Tell My Trainee? Great question! There are indeed many different CDAs that have been created to help us risk stratify febrile infants. This is because relying solely on our clinical exam at this young age group may not be accurate. This baby could just have a viral illness, but we must think about other infections like urinary tract infections, bacteremia, and meningitis. Most CDAs will recommend that we should at least obtain blood and urine testing. This is an older infant of 6 weeks, so we can use inflammatory markers to help guide us on whether we need to perform a lumbar puncture. It’s important to remember that CDAs should be used in conjunction with our clinical judgement. They should never dictate care.

We use the AAP CDA here but do not have procalcitonin testing, so we will use C-reactive protein instead.


Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.


 References:

  1. Gomez B, Mintegi S, Bressan S, et al. Validation of the “step-by-step” approach in the management of young febrile infants. Pediatrics. 2016;138(2):e20154381.
  2. Nigrovic LE, Mahajan PV, Blumberg SM, et al. The yale observation scale score and the risk of serious bacterial infections in febrile infants. Pediatrics. 2017;140(1):e20170695.
  3. Evans JUmana EWaterfield T FIDO Study Group in collaboration with PERUKI Respiratory viral testing for young febrile infants presenting to emergency care: a planned secondary analysis of the Febrile Infants Diagnostic assessment and Outcome (FIDO) prospective observational cohort study. 
  4. Green RS, Sartori LF, Lee BE, et al. Prevalence and management of invasive bacterial infections in febrile infants ages 2 to 6 months. Ann Emerg Med. 2022;80(6):499-506.