OBJECTIVES:

The Aronson rule is a point-based clinical decision rule for the identification of febrile infants ≤60 days of age at low risk of invasive bacterial infection (IBI) in the emergency department. This rule uses variables of temperature, age, urinalysis, and absolute neutrophil count. We sought to externally validate this decision rule.

METHODS:

We conducted a secondary analysis of a multicenter prospective cohort of febrile infants ≤60 days old presenting to the emergency department between December 2008 and May 2013. Infants were excluded if they had clinical sepsis or chronic conditions or were missing any laboratory components of the Aronson score. Our outcome was IBI (bacteremia and/or bacterial meningitis). We assessed the accuracy of the Aronson rule by reporting metrics of diagnostic accuracy with 95% confidence intervals (CIs) at different point thresholds.

RESULTS:

Of 4130 included patients (780 <21 days of age; 2362 boys), 87 (2.1%) had an IBI, including 65 with isolated bacteremia and 22 with meningitis. Using an Aronson cutoff score of 2 resulted in a sensitivity of 93.1% (95% CI 85.6%–97.4%), specificity of 26.6% (95% CI 25.3%–28.0%), and negative predictive value of 99.4% (95% CI 98.8%–99.8%). Six patients with IBI (3 with bacterial meningitis) were misclassified as low risk when using a threshold of 2.

CONCLUSIONS:

The Aronson rule demonstrates metrics of diagnostic accuracy that are comparable to the derivation study. Our findings suggest that the rule may be generalizable for the risk stratification of well-appearing febrile infants.

Febrile young infants are at risk for serious bacterial infections (SBIs) including urinary tract infections (UTIs), bacteremia, and bacterial meningitis. Some recent research has been focused on the identification of invasive bacterial infections (IBIs), a definition limited to bacteremia and bacterial meningitis,1  given the ability of point-of-care urine testing to reliably and rapidly diagnose UTI2,3  and the lower severity of UTI compared with the former conditions. Whereas overall SBI prevalence ranges from 9% to 14%,4,5  the prevalence of IBI is considerably lower (bacteremia 2%, meningitis 0.9%)4,6  but requires more accurate diagnosis and treatment to prevent unnecessary workup and empirical treatment.

Historic criteria711  based on both clinical and laboratory data have been used to identify well-appearing infants at low risk of SBI. These criteria, although easy to use, have suffered from high rates of false-positives, resulting in preventable excess lumbar punctures, hospitalizations, antimicrobial use, and parental anxiety.7,1214  Several recent criteria (Step-by-Step, Pediatric Emergency Care Applied Research Network [PECARN])5,15  have been developed to risk-stratify young febrile infants without routine use of lumbar puncture by incorporating new laboratory testing such as procalcitonin. One recently published rule by Aronson et al16  using data from the multicenter Febrile Young Infant Research Collaborative does not require routine use of lumbar puncture and is unique in that it does not require procalcitonin testing. The lack of procalcitonin in this decision rule underscores its significance: in a study in which authors evaluated 61 acute care settings in 2017 in Massachusetts, only 15% of institutions had access to on-site procalcitonin; in another 2019 evaluation of 16 Canadian pediatric hospitals, researchers reported only 2 (13%) having access to this test.17,18 

The Aronson rule was derived by using data from a case-control study of infants presenting to 11 participating emergency departments (EDs) during a 5-year study period.16  In that study, investigators discovered 4 predictors associated with the presence of IBI: age, highest temperature recorded in the ED, abnormal urinalysis result, and absolute neutrophil count (ANC). Optimal cut points for each predictor with corresponding points were determined by using logistic regression. In the derivation population, an Aronson score of ≥2 (between a possible range of 0–10) had a sensitivity of 98.8% and specificity of 31.3% for IBI. This score had high precision for identifying IBI in non–ill-appearing febrile infants, comparable to other low-risk criteria (eg, Boston, Philadelphia, and Rochester), without the use of cerebrospinal fluid (CSF) testing.

To assess the robustness of decision rules in health care settings, external validation, in which the rule is applied to a population distinct from the derivation population, is an essential step.19,20  Our objective was to externally validate the Aronson rule by using data from a large multicenter prospective study of febrile infants.

We performed a secondary analysis of a multicenter prospective observational study of patients enrolled at 26 geographically diverse EDs in the PECARN between December 2008 and May 2013. We used the public use data set for this study (www.pecarn.org) that was last revised on September 28, 2015,21  use of which was considered nonhuman research by our institutional review board.

The PECARN study was performed as a convenience study in which well-appearing infants aged ≤60 days with fever (defined as rectal temperature ≥38°C in the ED or 24 hours before presentation) and who received a blood culture test were approached for enrollment. Patients with clinical sepsis, indwelling hardware, recent antibiotic use, or comorbid conditions (including congenital abnormality, inborn error metabolism, and immunodeficiency) were excluded from the PECARN study. For the purpose of this secondary analysis, patients were excluded if they did not have blood culture results and at least 1 of the following: CSF culture or, if no lumbar puncture was performed, successful documentation of telephone contact or medical record review done at 7 days after ED presentation in accordance with the study protocol to ensure no subsequent evaluation for bacterial meningitis. Positive culture results were classified as contaminants, pathogenic, or indeterminate. Indeterminate cultures underwent a secondary round of review. Patients with cultures of indeterminate significance after all review cycles were excluded from the current study. Patients with missing components of the Aronson score (urinalysis, complete blood cell [CBC] count) were excluded.

Our outcome of interest was IBI, defined as bacteremia and/or bacterial meningitis, determined by the presence of positive culture results. For patients with multiple blood or CSF culture results, any single positive culture result was used to define the presence of IBI. We extracted variables from the PECARN database to calculate the score and characterize false-positives. This included demographics (age, sex), enrolling temperature and location of temperature measurement (in ED, in previous health care setting, or at home within 24 hours), clinical assessment (Yale Observation Scale [YOS] score, suspicion for SBI), laboratory data (white blood cell [WBC] count, ANC, urinalysis results), and bacterial culture results (urine, blood, and CSF). The YOS score was calculated from an attending or fellow physician’s assessment of 6 domains (quality of cry, reaction to parents, state variation, color, hydration, response), each assigned 1, 3, or 5 points, for a possible range of 6 for the most well-appearing infant to 30 for the most ill-appearing infant.22,23  Suspicion for SBI was classified as the attending or fellow physician suspicion for SBI, assessed before awareness of laboratory results on the basis of history and physical examination. Patients in the PECARN data set had a single enrolling temperature, along with documentation of the place where the elevated temperature was recorded. If there was a fever during the ED assessment, this was used as the elevated temperature. As such, we defined encounters with an enrolling temperature outside the ED (ie, in an outpatient setting or at home within 24 hours of presentation) as afebrile in the ED.24 

We calculated the Aronson score, ranging from 0 to 10 points, for all included patients, using the criteria derived by Aronson et al.16  Patients were assigned points on the basis of the following factors: age <21 days (1), temperature in the ED ranging from 38.0°C to 38.4°C (2), temperature in the ED of ≥38.5°C (4), abnormal urinalysis (3), and ANC ≥5185 cells per μL (2). Abnormal urinalysis was defined as urine dipstick with positive leukocyte esterase result, positive nitrites, or urine microscopy with >5 WBCs per high-power field. The derivation study concluded that infants with an Aronson score <2 were at low probability of IBI, and infants with scores of 2 to 4 were at moderate probability of IBI.

We calculated sensitivity, specificity, negative and positive predictive values, and negative and positive likelihood ratios and their 95% confidence intervals (CIs) for the Aronson score at cut points of 1 to 4. We identified patients with IBI who were missed by an Aronson score ≥2 and reported demographic, clinical, and laboratory data for each case. Analyses were performed by using R version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria).

We planned 3 exploratory analyses. First, we planned an analysis using an outcome of SBI. For this, patients missing urine cultures were also excluded. Second, we performed an analysis wherein patients with missing laboratory data were not excluded, reporting metrics of diagnostic accuracy using this modified inclusion criteria. Third, we performed an analysis excluding infants defined as ill appearing by a YOS score of >10.

Of 6014 infants enrolled in the PECARN study, 4130 were included for this analysis (Fig 1). This study included 2362 (57.2%) male infants, and 780 (18.9%) were <21 days of age. Among 87 infants (2.1%) with IBI, 65 had bacteremia without meningitis, 12 had bacterial meningitis without bacteremia, and 10 had bacterial meningitis with concomitant bacteremia. Of the 4130 included infants, 2625 (63.6%) had an enrolling temperature in an ED, whereas 1505 (36.4%) had enrolling temperatures at an office, other facility, or at home.

FIGURE 1

Patient inclusion.

FIGURE 1

Patient inclusion.

Close modal

Test characteristics at different cutoffs are provided in Table 1. The sensitivity and specificity of a score ≥2 were 93.1% (95% CI 85.6%–97.4%) and 26.6% (95% CI 25.3%–28.0%), respectively. IBI scores of ≥3 and ≥4 had higher specificities (49.6% and 56.5%) but lower sensitivities (83.9% and 73.6%).

TABLE 1

Assessment of Diagnostic Accuracy of the Aronson Rule Using Differing Score Cutoffs

CutoffSensitivity, % (95% CI)Specificity, % (95% CI)PPV, % (95% CI)NPV, % (95% CI)+LR (95% CI)−LR (95% CI)
AS ≥1 95.4 (88.6–98.7) 22.7 (21.4–24.0) 2.6 (2.1–3.2) 99.6 (98.9–99.9) 1.23 (1.17–1.30) 0.20 (0.08–0.53) 
AS ≥2 93.1 (85.6–97.4) 26.6 (25.3–28.0) 2.7 (2.1–3.3) 99.4 (98.8–99.8) 1.27 (1.19–1.35) 0.26 (0.12–0.56) 
AS ≥3 83.9 (74.5–90.9) 49.6 (48.0–51.1) 3.5 (2.7–4.3) 99.3 (98.8–99.6) 1.66 (1.51–1.83) 0.32 (0.2–0.53) 
AS ≥4 73.6 (63–82.4) 56.5 (55.0–58.1) 3.5 (2.7–4.5) 99.0 (98.5–99.4) 1.69 (1.49–1.93) 0.47 (0.33–0.66) 
CutoffSensitivity, % (95% CI)Specificity, % (95% CI)PPV, % (95% CI)NPV, % (95% CI)+LR (95% CI)−LR (95% CI)
AS ≥1 95.4 (88.6–98.7) 22.7 (21.4–24.0) 2.6 (2.1–3.2) 99.6 (98.9–99.9) 1.23 (1.17–1.30) 0.20 (0.08–0.53) 
AS ≥2 93.1 (85.6–97.4) 26.6 (25.3–28.0) 2.7 (2.1–3.3) 99.4 (98.8–99.8) 1.27 (1.19–1.35) 0.26 (0.12–0.56) 
AS ≥3 83.9 (74.5–90.9) 49.6 (48.0–51.1) 3.5 (2.7–4.3) 99.3 (98.8–99.6) 1.66 (1.51–1.83) 0.32 (0.2–0.53) 
AS ≥4 73.6 (63–82.4) 56.5 (55.0–58.1) 3.5 (2.7–4.5) 99.0 (98.5–99.4) 1.69 (1.49–1.93) 0.47 (0.33–0.66) 

AS, Aronson score; NPV negative predictive value; PPV, positive predictive value; −LR, negative likelihood ratio; +LR, positive likelihood ratio.

The clinical features of infants with IBI missed by the Aronson score ≥2 cutoff are summarized in Table 2. Six patients (3 with isolated bacteremia, 2 with meningitis, 1 with bacteremia and meningitis) with IBI were missed at the cutoff of ≥2. Of the 4 patients with an Aronson score of 0, 1 patient had bacterial meningitis. This patient was a 42-day-old male infant without leukocytosis with an enrolling temperature before ED presentation, noted to have low clinical suspicion (1%–5% risk) for SBI, and ultimately diagnosed with Escherichia coli meningitis on CSF culture. Of the 2 patients with an Aronson score of 1, one had isolated Klebsiella oxytoca meningitis, and the other had concomitant Enterobacter cloacae bacteremia and meningitis. Both were <21 days, had an enrolling temperature before ED presentation, and had low clinical suspicion for SBI.

TABLE 2

Infants With IBI Missed by Aronson Score ≥2

UAANC, Cells per μLWBC, ×103 Cells per μLED Temperature, °CClinical Suspicion for SBI, %YOSCulture Results
Aronson score of 0        
 37-d-old male Negative 1575 9.0 N/A 1–5 16 E coli bacteremia 
 54-d-old female Negative 1090 2.8 N/A 1–5 14 Enterococcus faecalis bacteremia 
 42-d-old male Negative 1430 6.5 N/A 1–5 E coli meningitis 
 28-d-old female Negative 4210 6.9 N/A 11–50 14 Group B Streptococcus bacteremia 
Aronson score of 1        
 11-d-old femalea Negative 1420 6.1 N/A 1–5 K oxytoca meningitis 
 15-d-old femalea Negative 4160 6.4 N/A 1–5 E cloacae bacteremia and meningitis 
UAANC, Cells per μLWBC, ×103 Cells per μLED Temperature, °CClinical Suspicion for SBI, %YOSCulture Results
Aronson score of 0        
 37-d-old male Negative 1575 9.0 N/A 1–5 16 E coli bacteremia 
 54-d-old female Negative 1090 2.8 N/A 1–5 14 Enterococcus faecalis bacteremia 
 42-d-old male Negative 1430 6.5 N/A 1–5 E coli meningitis 
 28-d-old female Negative 4210 6.9 N/A 11–50 14 Group B Streptococcus bacteremia 
Aronson score of 1        
 11-d-old femalea Negative 1420 6.1 N/A 1–5 K oxytoca meningitis 
 15-d-old femalea Negative 4160 6.4 N/A 1–5 E cloacae bacteremia and meningitis 

All infants had an enrolling temperature before ED presentation. N/A, not applicable; UA, urinalysis result.

a

Points allocated to the Aronson score.

In our analysis using an outcome of SBI, 48 patients missing urine culture data were removed, leaving a total of 4082 patients. Of these, 398 (9.8%) patients had a SBI, including 343 patients with a UTI, 75 patients with bateremia, and 22 patients with bacterial meningitis (38 patients had 2 concomitant infections and 2 patients had all 3 forms of SBI). By using a cutoff score of 2, this provided higher sensitivity (97.7% [95% CI 95.8%–99.0%]) and specificity (28.8% [95% CI 27.3%–30.3%]) for predicting SBI compared with the primary analysis predicting IBI (Supplemental Table 3). We then performed an analysis including infants with missing CBC count and/or urinalysis data using an outcome of IBI. In addition to the 4130 included in the primary analysis, 262 patients with missing CBC count or urinalysis data were included. Ninety-three patients had an IBI (69 with bacteremia alone, 13 with meningitis alone, and 11 with meningitis and concomitant bacteremia). Sensitivity and specificity were similar. A cutoff score of 2, for example, resulted in a sensitivity of 93.5% (95% CI 86.5%–97.6%) (Supplemental Table 4). Lastly, we performed an analysis excluding patients deemed ill-appearing on the basis of a YOS score >10, again using an outcome of IBI. This analysis retained 3639 encounters, of which 64 (1.8%) had IBIs. Sensitivity and specificity were slightly increased in this group (Supplemental Table 5).

We performed an external validation of a recently published decision rule for IBI using a data set from a multicenter prospective study. At a cutoff of ≥2, the Aronson rule had similar sensitivity and specificity in our study population (sensitivity 93.1%, specificity 26.6%) than in the derivation population. Our findings support the generalizability of this clinical decision rule for well-appearing febrile infants ≤60 days old.

Compared with the derivation study (sensitivity 98.8%, specificity 31.3%), our external validation revealed a small decrease in sensitivity and specificity, which is expected when applying a decision rule to an external population.25  In a recent validation performed in a smaller subset of these data,26  the sensitivity of the Aronson rule appeared to be higher. The present analysis incorporates the entire available PECARN data set with minimal exclusions. Overall, the score is highly sensitive, and the CIs in this validation (sensitivity 85.6%–97.4%, specificity 25.3%–28.0%) overlap with those in the derivation study (sensitivity 95.8%–99.9%, specificity 26.3%–36.6%). Patients in this study were similar to those in the derivation study with respect to demographics (age, sex) and clinical characteristics (ie, febrile, well appearing, and absence of chronic conditions). Altogether, this reveals comparable performance of the Aronson rule to the previously published decision rule in our population and that the score may generalize to clinical practice.

The Aronson rule performs similarly to other recently derived decision rules that do not require lumbar puncture. The PECARN rule had a slightly higher sensitivity and specificity (sensitivity 97.7%, specificity 60.0%) but used SBI, rather than the more narrowly defined IBI, as the outcome of interest.5  The sensitivity of this external validation for an outcome of SBI, performed as an exploratory analysis, appears to be comparable to that reported in the PECARN investigation. Notably, when the investigators of the PECARN rule limited their outcome measure to only IBI, the overall sensitivity was lower (96.7% [95% CI 83.3%–99.4%]). The Step-by-Step rule, which used IBI as its outcome, had similar sensitivity (92.0% [95% CI 84.3%–96.0%]) and higher specificity (46.9% [95% CI 44.8%–49.0%]) when compared with the Aronson rule.15 

The Aronson rule achieves high sensitivity without the use of serum procalcitonin, a relatively new laboratory test that is not routinely available in many North American EDs.17,18  Although the availability of procalcitonin is increasing in certain hospital settings, the Aronson rule has particular value because it can be used to identify febrile infants at low risk of IBI in hospitals or in other settings without access to procalcitonin testing. Older decision rules (Philadelphia, Rochester) that were modified in a recent validation study to be usable without lumbar puncture had similar sensitivity (modified Philadelphia 91.9%) and specificity (34.5%) to the Aronson rule in the present investigation.27  Compared to existing decision rules, the Aronson rule is a validated score with utility in a variety of clinical settings to reduce unnecessary lumbar punctures and hospitalizations.

Six IBI cases, including 3 bacterial meningitis cases, were misclassified as low risk by the Aronson core cutoff of ≥2. All of these patients with IBI had fever on history alone and normal laboratory values and all but one had low clinical suspicion (1%–5% risk) of IBI. Two of the 3 meningitis cases were in patients ≤28 days, a population with higher prevalence of meningitis than in infants 29 to 60 days.6  Incorporation of an age variable in the Aronson score gave these patients a score of 1, but still they would not be detected by our cutoff of ≥2, pointing to the importance for high clinical suspicion in this population. Notably, one of the patients misclassified by the PECARN rule (30-day-old male infant with E cloacae bacteremia with procalcitonin of 0.14 ng/mL) had an Aronson score of 2 and thus was appropriately classified as not low risk; this patient would also have been missed on the application of the Step-by-Step rule as well. However, as a rule with lower sensitivity, the Aronson score should be used in conjunction with clinical judgment and shared decision-making to determine the likelihood of IBI and need for further testing. One potential strategy may be to use a lower cutoff of the Aronson (≥1) to minimize the false-negative rate, although the improvement in sensitivity with this approach was small (93.1%–95.4%).

Our study has several limitations. The Aronson rule uses the highest temperature for a patient recorded in the ED. Because patients in the PECARN study had only 1 enrolling temperature, we were unable to identify infants who had rising fever after ED enrollment. The PECARN study did not have data on the administration of antipyretics, either outside or within the ED. Furthermore, we considered those with enrolling temperatures recorded outside the ED (ie, as outpatients or at home; present in 1505 [36.4%] infants) as afebrile, potentially missing some patients who were initially afebrile at ED presentation but subsequently spiked a fever during their ED stay. Although it has been shown that infants febrile on history alone have a lower risk of developing SBI compared with infants febrile in the ED,24  the overall risk of IBI is similar.28  Secondly, a small proportion of patients in this study did not have a CBC count or urinalysis test performed and were thus excluded; however, test characteristics were unchanged after sensitivity analysis. Thirdly, the data set from this study was derived from a convenience sample, and we were unable to report on the outcomes of patients who were screened for study eligibility but not enrolled. The overall rate of IBI is comparable to other literature, suggesting these findings are not subject to systematic sampling bias.4  Although we used a positive culture result as an outcome measure, we do not have information on treatments provided for these infants. Despite these limitations, this study, in which we use a large, prospectively acquired population, provides a useful reference point to assess the generalizability of the Aronson rule.

We performed an external validation of a decision rule for IBI recently published by Aronson et al16  using a data set from a multicenter prospective study in PECARN EDs. At a cutoff of ≥2, the Aronson score had similar sensitivity and specificity (sensitivity 93.1%, specificity 26.6%) as in the derivation study. Our findings support the generalizability of this clinical decision rule for well-appearing febrile infants ≤60 days old.

Ms Tsai designed the study, conducted the analyses, and drafted and revised the manuscript; Dr Ramgopal designed and conceptualized the study and reviewed the manuscript; and all authors approved the final manuscript as submitted.

FUNDING: No external funding.

1
Woll
C
,
Neuman
MI
,
Aronson
PL
.
Management of the febrile young infant: update for the 21st century
.
Pediatr Emerg Care
.
2017
;
33
(
11
):
748
753
2
Tzimenatos
L
,
Mahajan
P
,
Dayan
PS
, et al
;
Pediatric Emergency Care Applied Research Network (PECARN)
.
Accuracy of the urinalysis for urinary tract infections in febrile infants 60 days and younger
.
Pediatrics
.
2018
;
141
(
2
):
e20173068
3
Velasco
R
,
Benito
H
,
Mozun
R
, et al
;
Group for the Study of Febrile Infant of the RiSEUP-SPERG Network
.
Using a urine dipstick to identify a positive urine culture in young febrile infants is as effective as in older patients
.
Acta Paediatr
.
2015
;
104
(
1
):
e39
e44
4
Greenhow
TL
,
Hung
YY
,
Herz
AM
,
Losada
E
,
Pantell
RH
.
The changing epidemiology of serious bacterial infections in young infants
.
Pediatr Infect Dis J
.
2014
;
33
(
6
):
595
599
5
Kuppermann
N
,
Dayan
PS
,
Levine
DA
, et al
;
Febrile Infant Working Group of the Pediatric Emergency Care Applied Research Network (PECARN)
.
A clinical prediction rule to identify febrile infants 60 days and younger at low risk for serious bacterial infections
.
JAMA Pediatr
.
2019
;
173
(
4
):
342
351
6
Powell
EC
,
Mahajan
PV
,
Roosevelt
G
, et al
;
Febrile Infant Working Group of the Pediatric Emergency Care Applied Research Network (PECARN)
.
Epidemiology of bacteremia in febrile infants aged 60 days and younger
.
Ann Emerg Med
.
2018
;
71
(
2
):
211
216
7
Baker
MD
,
Bell
LM
,
Avner
JR
.
Outpatient management without antibiotics of fever in selected infants
.
N Engl J Med
.
1993
;
329
(
20
):
1437
1441
8
Baskin
MN
,
O’Rourke
EJ
,
Fleisher
GR
.
Outpatient treatment of febrile infants 28 to 89 days of age with intramuscular administration of ceftriaxone
.
J Pediatr
.
1992
;
120
(
1
):
22
27
9
Jaskiewicz
JA
,
McCarthy
CA
,
Richardson
AC
, et al
;
Febrile Infant Collaborative Study Groups
.
Febrile infants at low risk for serious bacterial infection–an appraisal of the Rochester criteria and implications for management
.
Pediatrics
.
1994
;
94
(
3
):
390
396
10
Herr
SM
,
Wald
ER
,
Pitetti
RD
,
Choi
SS
.
Enhanced urinalysis improves identification of febrile infants ages 60 days and younger at low risk for serious bacterial illness
.
Pediatrics
.
2001
;
108
(
4
):
866
871
11
Lyons
TW
,
Garro
AC
,
Cruz
AT
, et al
;
Herpes Simplex Virus Study Group of the Pediatric Emergency Medicine Collaborative Research Committee (PEM CRC)
.
Performance of the modified Boston and Philadelphia criteria for invasive bacterial infections
.
Pediatrics
.
2020
;
145
(
4
):
e20193538
12
Aronson
PL
,
Thurm
C
,
Alpern
ER
, et al
;
Febrile Young Infant Research Collaborative
.
Variation in care of the febrile young infant <90 days in US pediatric emergency departments
. [
published correction appears in Pediatrics. 2015;135(4):775
].
Pediatrics
.
2014
;
134
(
4
):
667
677
13
De
S
,
Tong
A
,
Isaacs
D
,
Craig
JC
.
Parental perspectives on evaluation and management of fever in young infants: an interview study
.
Arch Dis Child
.
2014
;
99
(
8
):
717
723
14
Mintegi
S
,
Benito
J
,
Astobiza
E
,
Capapé
S
,
Gomez
B
,
Eguireun
A
.
Well appearing young infants with fever without known source in the emergency department: are lumbar punctures always necessary?
Eur J Emerg Med
.
2010
;
17
(
3
):
167
169
15
Gomez
B
,
Mintegi
S
,
Bressan
S
,
Da Dalt
L
,
Gervaix
A
,
Lacroix
L
;
European Group for Validation of the Step-by-Step Approach
.
Validation of the “step-by-step” approach in the management of young febrile infants
.
Pediatrics
.
2016
;
138
(
2
):
e20154381
16
Aronson
PL
,
Shabanova
V
,
Shapiro
ED
, et al
;
Febrile Young Infant Research Collaborative
.
A prediction model to identify febrile infants ≤60 days at low risk of invasive bacterial infection
.
Pediatrics
.
2019
;
144
(
1
):
e20183604
17
Fisher
KA
,
Landyn
V
,
Lindenauer
PK
,
Walkey
AJ
.
Procalcitonin test availability: a survey of acute care hospitals in Massachusetts
.
Ann Am Thorac Soc
.
2017
;
14
(
9
):
1489
1491
18
Burstein
B
,
Gravel
J
,
Aronson
PL
,
Neuman
MI
;
Pediatric Emergency Research Canada (PERC)
.
Emergency department and inpatient clinical decision tools for the management of febrile young infants among tertiary paediatric centres across Canada
.
Paediatr Child Health
.
2019
;
24
(
3
):
e142
e154
19
Laupacis
A
,
Sekar
N
,
Stiell
IG
.
Clinical prediction rules. A review and suggested modifications of methodological standards
.
JAMA
.
1997
;
277
(
6
):
488
494
20
Altman
DG
,
Vergouwe
Y
,
Royston
P
,
Moons
KG
.
Prognosis and prognostic research: validating a prognostic model
.
BMJ
.
2009
;
338
:
b605
21
Mahajan
P
,
Ramilo
O
,
Kuppermann
N
.
Application of Transcriptional Signatures for Diagnosis of Febrile Infants Within the Pediatric Emergency Care Applied Research Network (PECARN).
Salt Lake City, Utah
:
University of Utah School of Medicine
;
2012
22
McCarthy
PL
,
Sharpe
MR
,
Spiesel
SZ
, et al
.
Observation scales to identify serious illness in febrile children
.
Pediatrics
.
1982
;
70
(
5
):
802
809
23
McCarthy
PL
,
Lembo
RM
,
Baron
MA
,
Fink
HD
,
Cicchetti
DV
.
Predictive value of abnormal physical examination findings in ill-appearing and well-appearing febrile children
.
Pediatrics
.
1985
;
76
(
2
):
167
171
24
Ramgopal
S
,
Janofsky
S
,
Zuckerbraun
NS
, et al
.
Risk of serious bacterial infection in infants aged ≤60 days presenting to emergency departments with a history of fever only
.
J Pediatr
.
2019
;
204
:
191
195
25
Green
SM
,
Schriger
DL
,
Yealy
DM
.
Methodologic standards for interpreting clinical decision rules in emergency medicine: 2014 update
.
Ann Emerg Med
.
2014
;
64
(
3
):
286
291
26
Ramgopal
S
,
Horvat
CM
,
Yanamala
N
,
Alpern
ER
.
Machine learning to predict serious bacterial infections in young febrile infants
.
Pediatrics
.
2020
;
146
(
3
):
e20194096
.
27
Aronson
PL
,
Wang
ME
,
Shapiro
ED
, et al
;
Febrile Young Infant Research Collaborative
.
Risk stratification of febrile infants ≤60 days old without routine lumbar puncture
.
Pediatrics
.
2018
;
142
(
6
):
e20181879
28
Mintegi
S
,
Gomez
B
,
Carro
A
,
Diaz
H
,
Benito
J
.
Invasive bacterial infections in young afebrile infants with a history of fever
.
Arch Dis Child
.
2018
;
103
(
7
):
665
669

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data