OBJECTIVES:

We sought to determine predictors of hospitalization for children presenting with croup to emergency departments (EDs), as well as predictors of repeat ED presentation and of hospital readmissions within 18 months of index admission. We also aimed to develop a practical tool to predict hospitalization risk upon ED presentation.

METHODS:

Multiple deterministically linked health administrative data sets from Ontario, Canada, were used to conduct this population-based cohort study between April 1, 2006 and March 31, 2017. Children born between April 1, 2006, and March 31, 2011, were eligible if they had 1 ED visit with a croup diagnosis. Multivariable logistic regression was used to determine factors associated with hospitalization, subsequent ED visits, and subsequent croup hospitalizations. A multivariable prediction tool and associated scoring system were created to predict hospitalization risk within 7 days of ED presentation.

RESULTS:

Overall, 1811 (3.3%) of the 54 981 eligible children who presented to an Ontario ED were hospitalized. Significant hospitalization predictors included age, sex, Canadian Triage and Acuity Scale score, gestational age at birth, and newborn distress. Younger patients and boys were more likely to revisit the ED for croup. Our multivariable prediction tool could forecast hospitalization up to a 32% probability for a given patient.

CONCLUSIONS:

This study is the first population-based study in which predictors of hospitalization for croup based on demographic and historical factors are identified. Our prediction tool emphasized the importance of symptom severity on ED presentation but will require refinement before clinical implementation.

Croup, a widespread pediatric viral illness characterized by hoarseness, barking cough, and stridor,1,2  is the most common form of airway obstruction in young children.3,4  It is most often caused by parainfluenza virus,2  but many other viruses cause a similar clinical presentation.5  Croup affects ∼80 000 children yearly in Canada6  and represents between 3% and 5% of all emergency department (ED) visits in children <2 years of age,7  posing a significant burden of illness on the health care system.

Most children with croup have short-lived symptoms.6  However, severe upper airway obstruction, although rare, can occur.8  In the ED setting, treatment of croup includes comfort measures, oral steroids, and inhaled epinephrine,1,9  which, for most children, result in rapid symptom improvement. However, a minority of children will continue to deteriorate. On average, 1.5% to 6% of croup ED visits result in hospitalization.4,7,10  Despite the wide prevalence of croup, there is a paucity of guidelines to help physicians determine which children are at higher risk of deteriorating after initial presentation and therefore more likely to require hospitalization.

Researchers in previous studies have commonly identified male sex,2,1113  age <2 years2,3,11,12,14  and late autumn and winter presentation2,3,15  as risk factors for severe croup and hospitalization. Other potential risk factors for severe disease include prematurity,5,12  history of intubation,16,17  trisomy 21,5,18  asthma,5  and bronchiolitis.13  However, many of these studies are outdated, and the authors only examined a limited number of factors. Furthermore, no previous studies have developed a weighted scale to determine which variables are most predictive of croup hospitalization.

We determined predictors of hospitalization for children presenting to Ontario EDs with a croup diagnosis, using population-based health administrative data from Ontario, Canada’s most populous province. We also examined predictors of re-presentation to the ED, as well as readmissions to hospital, for children previously hospitalized with croup. We developed a practical tool to determine children’s risk for hospitalization when they first present to the ED with croup.

This was a retrospective cohort study of all children <6 years of age born in Ontario, Canada, identified by using provincial health administrative data. Approval from the Institutional Research Ethics Board was obtained before study initiation.

Children born in Ontario from April 1, 2006, to March 31, 2011 (fiscal years 2006–2010), were identified through the Better Outcomes Registry and Network (BORN) database, and those eligible for universal health care through the Ontario Health Insurance Plan (OHIP) (>99% of the population) from birth until age 6 years were included. Ontario is Canada’s most populous province, with nearly 14 million residents.19  Children were eligible for the study if they had 1 ED visit with an International Classification of Diseases, 10th Revision (ICD-10) code corresponding to a diagnosis of croup (J050, J040, J041, J042) within the National Ambulatory Care Reporting System (NACRS) between fiscal years 2006 and 2016. Although these codes were not previously validated, researchers in a recent study20  showed high agreement and reliability in ICD-10 diagnostic codes used in Canadian Emergency Departments, with the J050 code showing a sensitivity of 1.00 (95% confidence interval [CI] = 0.88–1.00) and positive predictive value of 0.97 (95% CI = 0.83–1.00). Individuals’ first ED visit before age 6 was considered the index event to maintain independence of observations and to minimize the impact of potential recurrent croup episodes. Children were excluded if they had a suspected or known diagnosis of congenital airway or esophageal malformation, as identified through diagnostic codes or history of previous bronchoscopy. A list of all ICD-10 codes used for inclusion and exclusion criteria and study variables can be found in Supplemental Table 4.

ICES is an independent, nonprofit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyze health care and demographic data, without consent, for health system evaluation and improvement. Multiple data sets available for analysis at ICES were used in this project and were linked deterministically by using a unique identification number based on encrypted OHIP health card number. The OHIP data set includes information on all outpatient services, including physician visits and laboratory and diagnostic tests. The Ontario Registered Persons Database provides basic demographic information (age, sex, location of residence, date of birth, and date of death for deceased individuals) for those issued an OHIP number. The Canadian Institute of Health Information (CIHI) Discharge Abstract Database (DAD) contains details of all hospitalizations including associated ICD-10 diagnosis codes and Canadian procedural codes. The ICES-derived MOMBABY data set links DAD inpatient admissions for delivering mothers and their newborns. The CIHI NACRS contains data for all hospital-based and community-based ambulatory care including ED visits. The above ICES data sets were further linked by using mixed deterministic and probabilistic methods to the BORN, which contains information on every pregnancy, birth, and child in Ontario. Linkage between BORN and ICES data are 99.03% successful when using these methods.21,22  In Supplemental Table 5, we detail the list of data sets used in the study. Data were available in uncleaned and unedited format to the analyst (BDK).

The primary outcome was hospitalization after an individual’s index ED visit for croup, as ascertained through the CIHI-DAD. An individual was classified as having this outcome if a hospitalization record with a most responsible or main diagnosis of croup was found within 7 days of the croup index ED visit. This 7-day window was used to capture same-day hospitalizations as well as children who likely should have been hospitalized on initial ED, to include “bounce-back” ED visits. For children hospitalized with croup, secondary outcomes included (1) subsequent croup visit to ED within 18 months of the index admission and (2) subsequent croup hospital readmission within 18 months, determined by NACRS and DAD, respectively. An 18-month time horizon was chosen to encompass ≤2 full croup seasons for recurrent illness episodes. Most children enrolled in the study had a full 18-month period of observation, except for children whose index admission was between October 2015 and March 2016. These children were observed between 12 and 18 months post–index event because data were collected until March 2017. Croup seasons tend to coincide with winter, which we describe as October 1 to March 31, because this period corresponds to a higher incidence of acute upper respiratory tract viral infections.23 

Several patient factors were identified as predictors of interest, through a thorough review of the literature. They comprised both demographic and clinical factors, with the date of an individual’s index event (first ED croup visit) as the baseline reference point. Predictors included demographic factors such as age, sex, income quintile, and rural or urban household (using the Rurality Index for Ontario [RIO] scores24 ). The RIO score is ascribed to a community based on its population and population density, as well as travel time to nearest basic and advanced referral centers. The RIO score ranges from 0 to 100, with higher numbers corresponding to more rural areas. Birth order was used as a proxy measure for increased exposure to infections in early life. Birth characteristics such as each child’s gestational age at birth and birth complications were ascertained through BORN records collected at delivery. Comorbidities that could increase croup hospitalization risk included congenital lung disease, congenital heart disease, newborn distress (defined as respiratory distress syndrome, intraventricular hemorrhage, or persistent fetal circulation), previous adenotonsillectomy, trisomy 21, gastroesophageal reflux disease (GERD), previous acute wheeze (encompassing asthma, bronchiolitis and wheeze), and previous intubation. Comorbidities were defined by using ICD-10 diagnosis codes and Canadian Classification of Health Interventions codes from previous acute health care visit records. (Supplemental Table 1). Within this list of comorbidities, diagnostic codes for bronchiolitis25  and asthma were previously validated.26  Encounter characteristics included linear continuous distance (in kilometers) from the hospital, as well as the Canadian Triage and Acuity Scale (CTAS) score on ED presentation. The CTAS is a triage tool used in Canadian hospitals to define a patient’s need for timely care based on symptom severity and acuity at triage.27  The CTAS score has 5 levels; level 1: resuscitation; level 2: emergent; level 3; urgent; level 4; less urgent; level 5: nonurgent. Levels 1 and 2, as well as levels 4 and 5, were grouped together. For hospitalized children, admission to the PICU was recorded as a marker for severe disease.

For the primary analysis, multivariable logistic regression was used to determine factors associated with hospitalization upon presentation to the ED. This model included all covariates identified a priori (see above). Collinearity between covariates was assessed in the model; only 1 variable from any pair of collinear variables was maintained in the model. Interactions between the exposure and (1) previous adenotonsillectomy and (2) chronic lung disease were tested. In the secondary analysis, among children initially hospitalized for croup, logistic regression was used to predict which factors led to (1) subsequent ED visits and (2) subsequent hospitalizations for croup. Variables considered for inclusion in secondary models were the same as those in the primary model. Results of all regression analyses were presented as odds ratios (ORs) with 95% confidence limits. All analyses were performed by using SAS Enterprise guide 7.15 (SAS Institute, Inc, Cary, NC).

The methodology outlined by Sullivan et al28  was used to create a multivariable prediction tool and associated scoring system, to predict the risk of hospitalization within 7 days of presenting to the ED with croup. Continuous variables from the primary model were categorized, and each variable was assigned a reference level. A logistic regression model was employed using these variables for 1000 bootstrapped samples. Regression coefficients from each bootstrapped model were used to create a CI for each OR, which was used to select which variables were maintained in the scoring system. Only variables in which the bootstrapped CI did not overlap 1.0 were included. The average regression coefficient for each variable was standardized to a constant, reflecting the number of regression units corresponding to 1 point in the scoring system. The effect of a 6-month increase in age was chosen as this constant. Each variable in the model was assigned a number of points (rounded to the nearest integer) by dividing its regression coefficient by the constant. This allowed for the calculation of point totals for each included individual. Finally, the estimated risk of hospitalization was calculated for each possible point total within this framework.

In Figure 1, we illustrate the cohort creation for this project, outlining inclusion and exclusion criteria. Overall, this study included 54 981 children who presented to an ED for croup before age 6 during the study period, with 1811 (3.3%) children hospitalized after ED presentation. Of these admitted children, 64 (3.5%) were admitted to the PICU at some point during their hospitalization.

In Table 1, we present descriptive characteristics of our cohort, comparing those who were hospitalized to those who were not. Presented in Table 2 are adjusted ORs of admission for each predictor of admission identified in Table 1. Significant predictors of hospitalization included age, sex, CTAS score, gestational age at birth, and newborn distress. Of note, intubation was removed from the model as it was collinear with gestational age.

A total of 461 (25.5%) of the admitted cohort of patients subsequently presented to the ED for croup within 18 months of the index admission, and 59 (3.2%) were hospitalized again within the same 18-month period. Younger patients (OR per year = 0.86, 95% CI = 0.77–0.96) and boys (OR = 1.66, 95% CI = 1.29–2.12) were more likely to revisit the ED for croup. Both younger patients and boys also tended to be more likely to be rehospitalized for croup within 18 months; however, this association was not statistically significant (OR per year = 0.77, 95% CI = 0.58–1.02, OR for boys = 1.67; 95% CI = 0.89–3.13).

We used a total of 5 variables (age, sex, CTAS score, gestational age, newborn distress) that were found to be both clinically and statistically significant in the primary model to develop a multivariable prediction tool. This prediction tool yielded a 26-point scoring system that identifies children most at risk for hospitalization on index ED croup presentation (Table 3). Outlined in Figure 2 is the likelihood of admission based on each possible score.

In this large Ontario population-based study, we identified several patient characteristics associated with an increased risk of hospitalization at presentation to the ED with croup.

As described in previous studies,2,3,1114  our results showed that male sex and younger age were associated with an increased risk of hospitalization. Although male sex and younger age continued to be significantly associated with an increased risk of subsequent croup ED visit, this association was no longer observed for readmission within 18 months of initial presentation, likely because of the small sample size of this particular outcome. Although there is no clear physiologic basis for male sex to increase the likelihood of admission, this finding is consistent across studies. A similar phenomenon is seen in asthma,2931  where a smaller airway size in young boys has been postulated as a cause for the increased severity in presentation.29  Not surprisingly, a lower CTAS score (more urgent) was strongly associated with a greater likelihood of hospitalization, representing a sicker group of patients on ED presentation. We did not find any increased hospitalization risk related to seasonality, although late autumn and winter season presentations have previously been reported as risk factors for severe disease.2,3,15 

Consistent with previous reports,5,12  we found prematurity, particularly extreme prematurity (<28 weeks gestational age), to be strongly associated with hospitalization. Given the retrospective nature of the study, it is not clear whether this higher likelihood of admission is due to a sicker group of patients or whether physicians have a lower threshold of admission for infants born prematurely, because those children are often perceived as more fragile. Additionally, infants and children born prematurely often have other comorbid conditions impacting their clinical presentation and risk of decompensation; they may also have been intubated for prolonged periods of time and have some degree of subglottic stenosis, potentially resulting in more severe croup symptoms.

Although previous diagnoses of asthma5  and bronchiolitis13  have been reported in single studies as risk factors for croup admission, this was not corroborated in our study by using health administrative data. We used a broad definition, namely wheeze, to capture these diagnoses, because young children may not have yet received a diagnosis of asthma by the time they first present with a wheezing episode. Asthma diagnosis in young children is challenging,32  and our use of a broad wheezing definition may have diluted positive associations with asthma. Interestingly, an association between croup early in life and a later asthma diagnosis has been noted in certain studies13,33  but not in others.34  In previous reports,5,17  authors have described an association between trisomy 21 and more-severe croup presentations. Most children with trisomy 21 are reported to have some subglottic stenosis, despite being asymptomatic,35  and would therefore be expected to be sicker from their croup episode. However, this was not noted in our population. Children with GERD similarly did not have a higher likelihood of admission in our study as compared with their counterparts. GERD has been shown to be prevalent in patients with recurrent croup,16  whereas in our study, we focused on children with a first croup presentation.

Previous research in Canada has revealed that relationships between rurality and health service use are variable and multifaceted, but, as a general rule, rural residents tend to have higher relative risks of hospitalization.36  In our study, children living in remote areas seemed less likely to be hospitalized than those living in nonrural areas, as determined by the RIO score (Table 1). This association disappeared when the RIO score was included in our multivariable model (Table 2), suggesting the presence of confounders. It is interesting to note that children with a higher RIO score (therefore living in more rural areas) had in general a higher CTAS score on presentation (lower acuity) (Table 2). Although the reason for this difference cannot be established with certainty, it may reflect decreased access to primary health care providers, resulting in ED visits earlier in the course of illness.

Our multivariable prediction tool allowed us to forecast hospitalization up to a 32% probability for a given croup patient, as compared with a 3% to 5% pretest probability. Although this suggests that predictive variables for croup hospitalization are not fully captured through the use of health administrative databases, our prediction tool can be considered a first step toward the development of a larger predictive model, which could incorporate additional clinical and physiologic variables. This would allow for the development of a more standardized yet personalized approach to care. Interestingly, the variables with the largest contribution to our prediction tool were low CTAS score, corresponding to children exhibiting worse symptoms on presentation to the ED, young age, and prematurity. These factors are likely to have an impact on physicians’ decision to hospitalize, and therefore correlate with current clinical practice.

Strengths of the study include the large population-based sample size, the ability to link between data sets and therefore follow our population over time, and generalizability to the Ontario population. Study limitations include the lack of individual information on treatment received and severity of illness, because clinical and drug data were not available from the available health administrative databases. Some children hospitalized with croup do not receive any further treatment12  and, therefore, may not have required admission. However, lack of information on drugs administered within hospital precluded our ability to discern these milder episodes. On the contrary, mild croup episodes may be diagnosed as unspecified respiratory illnesses in the ED and may therefore have been missed. Either scenario could have led to misclassification bias, although it is difficult to determine if this would have overestimated or underestimated risk of admission. An additional limitation is the slightly shorter observation period for development of secondary outcomes (ED or inpatient admission) for children whose index event was recorded between October 2015 and March 2016. These children would have been observed between 12 and 18 months post–index event, as opposed to the 18 months of observation recorded for the remainder of the cohort, and therefore would not have been captured if they re-presented to the ED or had a repeat inpatient admission between April and October 2017. This number, however, is likely to be small, because croup most often occurs between October and March.37 

In conclusion, this study is the first large population-based study in which predictors of hospitalization for croup based on demographic and historical factors are identified. Although our prediction tool does not have immediate direct clinical applicability, it is the first step toward the development of a broader model that would include clinical and physiologic variables not available in health administrative data and could lead to the delivery of more precise medicine for this patient population. An improved prediction tool could guide physicians, particularly in centers with less experience with croup, and might significantly impact clinical decision-making.

Dr Pound conceptualized and designed the study, contributed to interpret the data, drafted the initial manuscript, and reviewed and revised the manuscript; Mr Knight contributed to the design of the study and to the interpretation of the data, conducted the analyses, critically reviewed the manuscript for important intellectual content, and revised the manuscript; Dr Webster contributed to the design of the study, the interpretation of the data, and the analyses and critically reviewed the manuscript for important intellectual content; Dr Benchimol contributed to interpretation of the data and critically reviewed the manuscript for important intellectual content; Dr Radhakrishnan contributed to the design of the study and to the interpretation of the data and critically reviewed the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Dr Benchimol was supported by a New Investigator Award from the Canadian Institutes of Health Research, Canadian Association of Gastroenterology, and Crohn’s and Colitis Canada. Dr Benchimol was also supported by the Career Enhancement Program of the Canadian Child Health Clinician Scientist Program. This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. Parts of this material are based on data and information compiled and provided by the Ministry of Health and Long-Term Care and the Canadian Institute for Health Information. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred. This study is based in part on data provided by Better Outcomes Registry and Network, part of the Children’s Hospital of Eastern Ontario. The interpretation and conclusions contained herein do not necessarily represent those of Better Outcomes Registry and Network Ontario.

1
Gates
A
,
Gates
M
,
Vandermeer
B
, et al
.
Glucocorticoids for croup in children
.
Cochrane Database Syst Rev
.
2018
;
8
(
8
):
CD001955
2
Petrocheilou
A
,
Tanou
K
,
Kalampouka
E
,
Malakasioti
G
,
Giannios
C
,
Kaditis
AG
.
Viral croup: diagnosis and a treatment algorithm
.
Pediatr Pulmonol
.
2014
;
49
(
5
):
421
429
3
Segal
AO
,
Crighton
EJ
,
Moineddin
R
,
Mamdani
M
,
Upshur
RE
.
Croup hospitalizations in Ontario: a 14-year time-series analysis
.
Pediatrics
.
2005
;
116
(
1
):
51
55
4
Yang
WC
,
Lee
J
,
Chen
CY
,
Chang
YJ
,
Wu
HP
.
Westley score and clinical factors in predicting the outcome of croup in the pediatric emergency department
.
Pediatr Pulmonol
.
2017
;
52
(
10
):
1329
1334
5
Maalouli
WM
,
Hodges
JS
.
Croup admissions: can we shrink the elephant in the room? [published online ahead of print March 28, 2019]
Pediatr Emerg Care
.
6
Bjornson
CL
,
Johnson
DW
.
Croup in children
.
CMAJ
.
2013
;
185
(
15
):
1317
1323
7
Rosychuk
RJ
,
Klassen
TP
,
Metes
D
,
Voaklander
DC
,
Senthilselvan
A
,
Rowe
BH
.
Croup presentations to emergency departments in Alberta, Canada: a large population-based study
.
Pediatr Pulmonol
.
2010
;
45
(
1
):
83
91
8
Johnson
DW
.
Croup
.
BMJ Clin Evid
.
2014
;
2014
:
0321
9
Bjornson
C
,
Russell
K
,
Vandermeer
B
,
Klassen
TP
,
Johnson
DW
.
Nebulized epinephrine for croup in children
.
Cochrane Database Syst Rev
.
2013
;(
10
):
CD006619
10
Tyler
A
,
McLeod
L
,
Beaty
B
, et al
.
Variation in inpatient croup management and outcomes
.
Pediatrics
.
2017
;
139
(
4
):
e20163582
11
Denny
FW
,
Murphy
TF
,
Clyde
WA
 Jr
,
Collier
AM
,
Henderson
FW
.
Croup: an 11-year study in a pediatric practice
.
Pediatrics
.
1983
;
71
(
6
):
871
876
12
Elder
AE
,
Rao
A
.
Management and outcomes of patients presenting to the emergency department with croup: can we identify which patients can safely be discharged from the emergency department?
J Paediatr Child Health
.
2019
;
55
(
11
):
1323
1328
13
Lin
SC
,
Lin
HW
,
Chiang
BL
.
Association of croup with asthma in children: a cohort study
.
Medicine (Baltimore)
.
2017
;
96
(
35
):
e7667
14
Chan
PW
.
Risk factors associated with severe viral croup in hospitalised Malaysian children
.
Singapore Med J
.
2002
;
43
(
3
):
124
127
15
Marx
A
,
Török
TJ
,
Holman
RC
,
Clarke
MJ
,
Anderson
LJ
.
Pediatric hospitalizations for croup (laryngotracheobronchitis): biennial increases associated with human parainfluenza virus 1 epidemics
.
J Infect Dis
.
1997
;
176
(
6
):
1423
1427
16
Hiebert
JC
,
Zhao
YD
,
Willis
EB
.
Bronchoscopy findings in recurrent croup: a systematic review and meta-analysis
.
Int J Pediatr Otorhinolaryngol
.
2016
;
90
:
86
90
17
Narayanan
S
,
Funkhouser
E
.
Inpatient hospitalizations for croup
.
Hosp Pediatr
.
2014
;
4
(
2
):
88
92
18
Hilton
JM
,
Fitzgerald
DA
,
Cooper
DM
.
Respiratory morbidity of hospitalized children with Trisomy 21
.
J Paediatr Child Health
.
1999
;
35
(
4
):
383
386
20
Peng
M
,
Eastwood
C
,
Boxill
A
, et al
.
Coding reliability and agreement of International Classification of Disease, 10th revision (ICD-10) codes in emergency department data
.
Int J Popul Data Sci
.
2018
;
3
:
14
21
Fell
DB
,
Wilson
K
,
Ducharme
R
, et al
.
Infant respiratory outcomes associated with prenatal exposure to maternal 2009 A/H1N1 influenza vaccination
.
PLoS One
.
2016
;
11
(
8
):
e0160342
22
Fell
DB
,
Platt
RW
,
Basso
O
, et al
.
The relationship between 2009 pandemic H1N1 influenza during pregnancy and preterm birth: a population-based cohort study
.
Epidemiology
.
2018
;
29
(
1
):
107
116
23
Eccles
R
.
An explanation for the seasonality of acute upper respiratory tract viral infections
.
Acta Otolaryngol
.
2002
;
122
(
2
):
183
191
24
Kralij
B
.
Measuring Rurality-RIO2008_BASIC: Methodology and Results
.
Toronto, Canada
:
Ontario Medical Association Economics Department
;
2009
25
Pisesky
A
,
Benchimol
EI
,
Wong
CA
, et al
.
Incidence of hospitalization for respiratory syncytial virus infection amongst children in Ontario, Canada: a population-based study using validated health administrative data
.
PLoS One
.
2016
;
11
(
3
):
e0150416
26
To
T
,
Dell
S
,
Dick
PT
, et al
.
Case verification of children with asthma in Ontario
.
Pediatr Allergy Immunol
.
2006
;
17
(
1
):
69
76
27
Murray
M
,
Bullard
M
,
Grafstein
E
;
CTAS Natioanal Working Group
CEDIS National Working Group
.
Revisions to the Canadian emergency department triage and acuity scale implementation guidelines
.
CJEM
.
2004
;
6
(
6
):
421
427
28
Sullivan
LM
,
Massaro
JM
,
D’Agostino
RB
 Sr
.
Presentation of multivariate data for clinical use: the Framingham Study risk score functions
.
Stat Med
.
2004
;
23
(
10
):
1631
1660
29
Chen
Y
,
Stewart
P
,
Johansen
H
,
McRae
L
,
Taylor
G
.
Sex difference in hospitalization due to asthma in relation to age
.
J Clin Epidemiol
.
2003
;
56
(
2
):
180
187
30
McConnochie
KM
,
Russo
MJ
,
McBride
JT
,
Szilagyi
PG
,
Brooks
AM
,
Roghmann
KJ
.
Socioeconomic variation in asthma hospitalization: excess utilization or greater need?
Pediatrics
.
1999
;
103
(
6
). Available at: www.pediatrics.org/cgi/content/full/103/6/e75
31
Säynäjäkangas
O
,
Valmari
P
,
Tuuponen
T
,
Keistinen
T
.
Trends in hospitalization for childhood asthma in Finland in 1996-2004
.
Acta Paediatr
.
2007
;
96
(
6
):
919
923
32
Ducharme
FM
,
Dell
SD
,
Radhakrishnan
D
, et al
.
Diagnosis and management of asthma in preschoolers: a Canadian Thoracic Society and Canadian Paediatric Society position paper
.
Can Respir J
.
2015
;
22
(
3
):
135
143
33
Nafstad
P
,
Brunekreef
B
,
Skrondal
A
,
Nystad
W
.
Early respiratory infections, asthma, and allergy: 10-year follow-up of the Oslo Birth Cohort
.
Pediatrics
.
2005
;
116
(
2
). Available at: www.pediatrics.org/cgi/content/full/116/2/e255
34
Sherman
CB
,
Tosteson
TD
,
Tager
IB
,
Speizer
FE
,
Weiss
ST
.
Early childhood predictors of asthma
.
Am J Epidemiol
.
1990
;
132
(
1
):
83
95
35
Hamilton
J
,
Yaneza
MM
,
Clement
WA
,
Kubba
H
.
The prevalence of airway problems in children with Down’s syndrome
.
Int J Pediatr Otorhinolaryngol
.
2016
;
81
:
1
4
36
Pong
RW
,
DesMeules
M
,
Heng
D
, et al
.
Patterns of health services utilization in rural Canada
.
Chronic Dis Inj Can
.
2011
;
31
(
suppl 1
):
1
36
37
Rosychuk
RJ
,
Klassen
TP
,
Voaklander
DC
,
Senthilselvan
A
,
Rowe
BH
.
Seasonality patterns in croup presentations to emergency departments in Alberta, Canada: a time series analysis
.
Pediatr Emerg Care
.
2011
;
27
(
4
):
256
260

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