Abstract
Introduction Early disease morbidity has been associated with asthma persistence in wheezing preschoolers; however, whether asthma control trajectories shortly after diagnosis could influence remission is unknown. We examined the association between asthma control trajectories 2 years post-diagnosis in preschoolers and subsequent disease remission.
Methods We conducted a multicentre population-based retrospective cohort study consisting of 48 687 children with asthma diagnosed before 5 years old and born between 1990 and 2013 in four Canadian provinces who had prolonged disease activity post-diagnosis. Prolonged disease activity was defined as one or more medical visits or medications for asthma every 6-month period for at least four of the six periods post-diagnosis. Follow-up began at 3 years post-diagnosis (at cohort entry). Remission was defined as 2 consecutive years without drug claims or medical visits for asthma or asthma-like conditions following cohort entry. Asthma control trajectories, ascertained over four 6-month periods following diagnosis using a validated index, were classified as: “controlled throughout”, “improving control”, “worsening control”, “out of control throughout” and “fluctuating control”. Adjusted Cox models estimated associations between asthma control trajectories and time to remission. A random effects meta-analysis summarised province-specific hazard ratios (HRs).
Results The pooled remission rate was 8.91 (95% CI 8.80–9.02) per 100 person-years. Compared with children controlled throughout, poorer asthma control was associated with incrementally lower hazard ratios of remission in four other trajectories: improving control (HR 0.89, 95% CI 0.82–0.96), fluctuating control (HR 0.78, 95% CI 0.71–0.85), worsening control (HR 0.68, 95% CI 0.62–0.75) and out of control throughout (HR 0.52, 95% CI 0.45–0.59).
Conclusions Asthma control trajectories 2 years following a diagnosis in preschoolers were associated with remission, highlighting the clinical relevance of documenting control trajectories in early life.
Abstract
In this multicentre population-based cohort study, the worse the asthma control trajectory shortly following diagnosis in preschoolers, the lower the likelihood of remission https://bit.ly/3lHVsNf
Introduction
The burden of asthma is important among preschoolers, who display a 2- to 3-fold higher rate of emergency department visits compared with other age groups [1, 2]. Although 50–60% of preschoolers outgrow asthma-like symptoms by school age [3], preschoolers with recurrent wheeze may exhibit permanent deficits in lung function by 6 years [4]. The multiple-hit hypothesis suggests that preschool asthma has its origins in defective barrier function of the airway epithelium combined with an impaired innate response to viral respiratory infections, leading to allergic sensitisation and asthma-like exacerbations [5]. Recurrent exacerbations with insufficient repair time may further increase susceptibility to respiratory viruses, air pollution and allergens, worsening control and lung injury. Recurrent exacerbations have been associated with lung function decline, possibly via airway remodelling; the latter could lead to irreversible lung damage and a permanent shift toward a more severe disease trajectory or persistent phenotype (i.e. phenotype programming) [6–8]. Better characterisation of the relationship between the frequency (and severity) of early events and disease evolution in preschoolers may help support or dispute this hypothesis. Moreover, it could suggest orienting the management goal toward prevention versus treatment of exacerbations, which currently raises considerable uncertainty in experts, physicians and parents, particularly in exacerbation-prone preschoolers with little interim symptoms [9–11]. In other words, could better asthma control during the preschool period lead to any long-term benefits?
Multiple birth or early-life cohort studies have explored predictors of disease persistence to identify “high-risk” children, who could be targeted for preventive management; these predictors were summarised in a 2017 systematic review [12]. In addition to child (sex, atopy and age) and parental (socioeconomic status, asthma and allergy) characteristics, several disease-related features (episode frequency and severity) were associated with asthma persistence beyond 6 years of age [12]. Whereas child and parental characteristics suggest susceptibility due to genetic and/or environment factors, the disease-related features are closely linked to asthma control; however, given the time-varying nature of asthma control, no study has evaluated whether the trajectory of asthma control shortly after diagnosis could be associated with long-term disease evolution.
We, therefore, aimed to assess if asthma control trajectories shortly after diagnosis in preschoolers were associated with subsequent disease remission (primary outcome) and recurrence (secondary outcome). Assuming that the 2-year period following an asthma diagnosis in preschoolers could represent a critical window for phenotype programming [13], we hypothesised that asthma control trajectories measured during this period would be associated with remission and recurrence.
Methods
Design
We assembled four retrospective population-based birth cohorts of all Canadian children born between April 1997 and December 2014 (British Columbia), February 1990 and December 2015 (Saskatchewan), and January 1997 and November 2014 (Manitoba). In Quebec, the birth cohort was derived from a population-based sample of pregnant women with asthma matched 1:4 with randomly selected nonasthmatic pregnant women who gave birth between January 1990 and March 2010, with follow-up until December 2013 [14]. Only these four provinces provided access to both health administrative and drug claims data from birth, which were study requirements. Of note, the distribution of ethnicity, socioeconomic status indicators, air pollution and prevalence of paediatric asthma differed slightly between provinces (supplementary table S1). Permission to create linkage with de-identified data from governmental administrative databases was granted by each province's health regulatory agency. The Research Ethics Boards of the Centre Hospitalier Universitaire Sainte-Justine (Montreal, QC, Canada) and each collaborating centre approved the study. Further methodological details can be found in the supplementary material.
Cohorts and data sources
Healthcare administrative databases included information on sociodemographic variables of the mother and child (perinatal data available in British Columbia only after April 2000) [15]; hospitalisations derived from validated hospital discharge summaries; physician visits for services reimbursed on a fee-for-service basis in all provinces as well as for services provided by salaried physicians with “shadow billing” in Saskatchewan and Manitoba; and pharmaceutical (drug claims) services. Services billed in the emergency department were only available for British Columbia and Quebec. Hospitalisations and physician visits, delivered through the Canadian publicly funded healthcare system, are free-of-charge to all residents; prescribed drugs were free-of-charge for children with parents receiving social assistance and, since 1997, for Quebec children with parents covered by public drug insurance.
Study population
Children were included if they: 1) had a physician-confirmed asthma diagnosis (supplementary material), 2) were <5 years of age at diagnosis and 3) had prolonged disease activity. A physician-confirmed asthma diagnosis was defined as one hospitalisation or two physician outpatient visits for asthma (codes 493 and J45 from the International Statistical Classification of Diseases and Related Health Problems, Ninth and Tenth Revisions (ICD-9 and ICD-10), respectively) at least 15 days but no more than 730 days apart (a Canadian database definition validated in children by To et al. [16]). Children were considered to have prolonged disease activity if they had one or more clinical markers of asthma activity, i.e. a medical visit (in the emergency department, inpatient or outpatient settings) or dispensed short-acting β2-agonist (SABA), controller therapies or oral corticosteroid (OCS) for asthma, every 6-month period, for at least four of the six (consecutive or nonconsecutive) periods following diagnosis (supplementary material). We implemented this criterion to specifically exclude children with infrequent wheezing episodes, whose symptoms typically resolve within 2–3 years following an initial wheezing event [17]. Patients were excluded if there was no/incomplete access to their healthcare services or drug data in the 12 months before and 36 months following diagnosis, or if they had cystic fibrosis.
To assess the prolonged disease activity criterion (prolonged disease activity inclusion criterion assessment window), cohort entry (CE1) began at 3 years post-diagnosis for the remission outcome, whereas asthma control was measured in the 2 consecutive years post-diagnosis (exposure assessment window). Cohort entry for recurrence (CE2) began at the date of asthma remission. Outcomes were evaluated from cohort entry until the end of follow-up (outcome assessment windows), i.e. the outcome date, date of death, emigration or last day of drug insurance coverage and/or data availability, whichever occurred first (figure 1a).
Timeline for measurements of exposures and outcomes. CE: cohort entry; PPACI: Pediatric Pharmacoepidemiology Asthma Control Index; SABA: short-acting β2-agonist; OCS: oral corticosteroid. a) In children with an asthma diagnosis before 5 years of age and prolonged disease activity (i.e. four consecutive or nonconsecutive 6-month periods with at least one clinical marker of asthma activity during the prolonged disease activity inclusion criterion assessment window), asthma control was assessed with the PPACI at four consecutive 6-month periods over 24 months following the date of diagnosis (i.e. the exposure assessment window). For the primary (remission) and secondary (recurrence) outcomes, children entered the cohort 3 years following diagnosis (CE1) and at the date of asthma remission (CE2), respectively; they were followed until the occurrence of the outcome (i.e. the outcome assessment window), date of lost to follow-up or end of data coverage and/or availability in the provincial databases, whichever occurred first. A 12-month look-back period prior to diagnosis (or since birth for those diagnosed prior to 1 year of age) served to accurately assess the PPACI after diagnosis if a SABA had been dispensed prior to diagnosis. b) Based on the validated PPACI, asthma control was classified over each 6-month period as: “controlled”(<4 SABA doses per week and no OCS, no emergency department visit and no hospitalisation for asthma; “partly controlled” (4– <7 SABA doses per week and no OCS, no emergency department and no hospitalisation), “poorly controlled” (≥7 SABA doses per week or ≥1 OCS course or ≥1 emergency department visit) and no hospitalisation) and “very poorly controlled” (≥1 hospitalisation). We summarised asthma control trajectories over the 2 years following diagnosis as: “controlled throughout” (controlled or partially controlled during each of the four 6-month periods), “improving control” (poorly controlled or very poorly controlled in the first 6 months and controlled or partially controlled in the last 6 months), “worsening control” (controlled or partially controlled in the first 6 months and poorly controlled or very poorly controlled in the last 6 months), “out of control throughout” (poorly controlled or very poorly controlled throughout) and “fluctuating control” (any other combination).
Outcomes
The primary end-point was asthma remission, defined as the absence of a hospitalisation or physician visit for asthma or asthma-like symptoms (ICD-9: 493, 466, 490; ICD-10: J45, J20, J21, J40) and no dispensed asthma medication for 2 consecutive years of follow-up accrued after cohort entry (CE1). Recurrence, the secondary end-point, was defined only in children who first experienced a remission, provided they fulfilled the asthma diagnosis definition again within a 2-year period after CE2 (figure 1a) [18].
Exposure: asthma control trajectories
Asthma control was ascertained using the validated four-level Pharmacoepidemiologic Pediatric Asthma Control Index (PPACI) over four consecutive 6-month periods post-diagnosis (exposure assessment window) (figure 1a) [19]. Briefly, this index was developed for use in population studies and validated against physician assessment in children aged 0–17 years; it categorises asthma control over 6 months in one of four levels: “controlled”, “partly controlled”, “poorly controlled” and “very poorly controlled”, based on the average doses per week of SABA computed with a validated algorithm [19, 20], OCS, emergency department visits and hospitalisations for asthma (figure 1b; left panel). OCS use was considered only if served within 14 days of a physician visit or hospitalisation for asthma to exclude OCS dispensations for indications other than asthma. Because of the substantial global variation regarding how and when physicians diagnose asthma in preschoolers [21, 22], the PPACI also includes OCS, emergency department and hospitalisation events related to asthma-like episodes, i.e. wheezing, bronchitis, bronchiolitis and bronchospasm (ICD-9: 466, 490; ICD-10: J20, J21, J40), provided they occurred after asthma diagnosis. We then grouped controlled with partly controlled and poorly controlled with very poorly controlled categories in each 6-month period of the exposure assessment window and summarised asthma control trajectories (figure 1b; right panel) as: “controlled throughout”, “improving control”, “fluctuating control”, “worsening control” and “out of control throughout”. Trajectories not defined as controlled throughout the exposure assessment period were considered to be suboptimal (supplementary material).
Confounders and covariates
Data were available on important baseline covariates (at diagnosis) related to the child (age at diagnosis, sex and atopy), mother (social assistance (or, if unavailable, lowest neighbourhood income quintile) at diagnosis), disease (diagnosis based on a hospitalisation or two physician visits for asthma, season and year of diagnosis) [16] and controller therapy dispensed at or in the year before diagnosis. Atopy was defined as having at least one of the following conditions from the asthma diagnosis date until cohort entry for remission: eczema, atopic, contact, food and other related dermatitis, urticaria or allergic rhinitis (ICD-9: 691, 692, 708, 477; ICD-10: L20, L23, L50, J30). For recurrence, the child's age of remission was also considered as a covariate. Perinatal covariates, including prematurity (<37 weeks of gestation), low birthweight (<2500 g) and the mother's age at birth, were also of interest, but these data were only consistently available in Quebec and Manitoba.
Statistical analysis
Within each province, Cox regression models were used to estimate associations between asthma control trajectories and time to remission/recurrence, with children controlled throughout the 2 years post-diagnosis as the reference control category. Each model was adjusted for covariates, generating the ratio of the remission (or recurrence) hazard for each suboptimal control trajectory relative to the reference category. The proportional hazards assumption was assessed for all Cox models by visually inspecting the scaled Schoenfeld residuals plots of covariates with different time transformations (log, linear and squared). For covariates violating the assumption, interaction terms with the appropriate time transformation were included in the models. We summarised province-specific hazard ratio (HR) estimates and 95% confidence intervals for each control trajectory using a random effects meta-analysis, to allow for between-province variation [23], and displayed province-specific as well as summarised estimates in forest plots.
A priori sensitivity analyses were performed, where we: 1) assessed if the control achieved in the period closest to cohort entry, the fourth 6-month period, was clinically relevant for remission; 2) adjusted for available perinatal variables in Quebec and Manitoba to assess their potential influence on remission estimates; and 3) computed E-values, defined as the minimum strength of association that an unmeasured confounder would need to have with both the exposure and outcome to nullify the estimated hazard ratios (supplementary material) [24]. A posteriori sensitivity analyses included computing 3-year, 5-year and overall absolute risks and risk differences for remission using Austin's risk approximation method for Cox models [25] and stratifying all estimated hazard ratios by sex and atopy status. SAS version 9.4 (SAS Institute, Cary, NC, USA) was used for data management and province-specific analyses, while R version 3.4.3 (www.r-project.org) was used for meta-analyses.
Results
Of the 2 247 223 newborns across all provinces, 95 105 (4.2%) preschoolers with asthma were potentially eligible for inclusion; of these, only 48 687 (51.2%) had prolonged disease activity (figure 2). Most children were male (65.4%), diagnosed based on two physician visits for asthma (83.9%) and aged <3 years (69.2%) at diagnosis, with half (50.1%) atopic by cohort entry (table 1). Compared with included children, those excluded from not meeting the prolonged disease activity criterion were more often nonatopic and less likely to have an asthma diagnosis based on a hospitalisation or dispensed a controller therapy at diagnosis; after diagnosis, they were less often dispensed controller therapy with fewer clinical markers of asthma activity (supplementary tables S2 and S3). The average median (interquartile range (IQR)) follow-up time in the remission cohorts was 1476 (891–2458) days.
Flow diagram of the selection of patients in each provincial cohort: British Columbia (BC), Saskatchewan (SK), Manitoba (MB) and Quebec (QC). Whereas the proportions of exclusions due to not meeting the diagnostic criteria of asthma or because of other diagnoses were similar across provinces, a larger proportion of children were excluded due to no/inconsistent access to healthcare services or drug data in Quebec (10.6%) than in other provinces; this was likely due to children whose parents no longer had access to the governmental drug coverage because of a private drug insurance or the end of social assistance. Approximately half of children with asthma were removed after applying the prolonged disease activity criterion.
Characteristics of patients included in the remission cohort
Although many children (37.5–49.0%) were controlled in each period of the exposure assessment window (supplementary figure S1a), fewer than a third (31.0%) were controlled throughout the 2 years (supplementary figure S1b), with some between-province variation (supplementary figure S2). Saskatchewan had the highest proportion of children controlled throughout (45.0%), while Quebec had the highest proportion (19.9%) of those out of control throughout the 2 years post-diagnosis.
A total of 22 668 children experienced remission across 254 534 person-years at risk, with the peak age of remission occurring between 8 and 9 years (table 1). The pooled remission rate was 8.91 (95% CI 8.80–9.02) per 100 person-years and was similar across provinces. Asthma control trajectories were consistent predictors of remission across provinces (table 2). Compared with children who were controlled throughout, all suboptimal control trajectories were associated with an increasingly lower likelihood of remission, with a pooled adjusted hazard ratio for remission falling to 0.52 (95% CI 0.45–0.59) in those out of control throughout (figure 3a). This translated to overall absolute risks of remission varying from 0.53 (95% CI 0.47–0.59) for those controlled throughout to 0.34 (95% CI 0.27–0.42) for those out of control throughout, with similar patterns at 3 and 5 years of follow-up (supplementary table S4). When using the last 6-month period of the exposure assessment window in a sensitivity analysis, a similar trend was observed (supplementary figure S3), which may have been due, in part, to some overlap between asthma control categories in the last 6 months and asthma control trajectories (supplementary table S5). Conclusions remained unchanged after adjustment for perinatal variables (supplementary table S6). The E-value sensitivity analysis suggested that the estimated average hazard ratios and confidence intervals of remission for worsening control and out of control trajectories can only be moved to a null value if moderate-to-strong unmeasured confounding (risk ratio (RR) 1.94–2.52 for average estimates; RR 1.74–2.24 for confidence limits), conditional on measured covariates, exists (supplementary table S7). In contrast, weak-to-moderate unmeasured confounding (RR 1.39–1.66 for average hazard ratios; RR 1.20–1.48 for confidence limits) could have nullified the associations for improving and fluctuating control trajectories on remission, conditional on measured covariates.
Meta-analyses of asthma control trajectories# 2 years post-diagnosis as a predictor of remission and recurrence
Forest plots depicting the adjusted likelihood of a) remission and b) recurrence for each asthma control trajectory compared with children who were controlled throughout the 2 years following diagnosis in British Columbia, Saskatchewan, Manitoba and Quebec. Province-specific hazard ratios are represented by a square with a 95% confidence interval (horizontal line). A diamond represents the summary estimate for each asthma control trajectory derived from the meta-analysis using random effects with hazard ratio values displayed on the right. The weight of each province, estimated by the inverse of the variance, is depicted in the far right column. The vertical line represents the line of identity (HR=1): values on the left of the line indicate a reduced likelihood of remission or recurrence; values on the right of the line indicate a higher likelihood of remission or recurrence
Older age at diagnosis, atopy and an asthma diagnosis based on a hospitalisation or occurring in seasons other than winter were consistently associated with a lower likelihood of remission in all provinces; yet, in most cases, the magnitude of association was smaller than the worsening and out of control trajectories (supplementary table S8).
Of the 22 668 children achieving remission, 3547 suffered a recurrence (15.6%) over 149 330 person-years at risk, with an average median (IQR) follow-up time of 1810 (1026–2868) days. The pooled recurrence rate was 2.38 (95% CI 2.30–2.45) per 100 person-years at risk. Suboptimal control trajectories were also associated with a higher likelihood of recurrence (HR 1.17, 95% CI 1.11–1.24) compared with children controlled throughout (figure 3b).
When stratified by sex or atopy, estimated hazard ratios for remission did not appear to differ significantly between trajectories, although there were some differences for recurrence (supplementary tables S9–S12). Among children experiencing remission, females tended to have higher recurrence rates than males, particularly in Quebec, and children who were not atopic early in life were generally more likely to experience a recurrence than atopic children.
Discussion
In this large multicentre population-based birth cohort study, asthma control trajectories in the 2 years following an asthma diagnosis in preschoolers with prolonged disease activity were associated with remission; worse asthma control trajectories were associated with incrementally lower likelihoods of remission. The direction and magnitude of associations were consistently observed across provinces. In children experiencing remission, an association between asthma control trajectories and recurrence was also observed.
We reported overall remission proportions ranging from 34% to 53%, which aligns with those reported in previous birth cohorts of preschoolers with asthma or persistent wheeze, varying from 59% at 6 years of age [3], 48.6% at 12 years of age [18] to 22% by 10–14 years of age [26]. Although not directly comparable because of various follow-up durations (2–6 years), data sources (symptoms, healthcare resource utilisation and drug claims) and age at which remission was assessed, the lower proportions observed in previous studies documenting remission in adolescence (compared with early school age) suggests that some children with early remission may experience subsequent recurrence, as observed in our study and others [18, 27–29].
Despite heterogeneity, the consistency in the direction and strength of associations across provinces and, importantly, the increasingly lower likelihood of remission with poorer asthma control trajectory, suggest the clinical predictive value of asthma control indicators based on healthcare resource utilisation following a preschool diagnosis. To the best of our knowledge, no other studies have investigated the association between asthma control trajectories following a preschool diagnosis and remission. However, many birth cohort studies have shown that the frequency and severity of preschool wheezing/asthma-like events by a given age were important predictors of wheeze and/or asthma persistence at school age, reporting findings consistent with our study [3, 12, 30, 31]. In particular, a large Canadian birth cohort study including preschoolers diagnosed with asthma who were followed until 11 years of age showed that children hospitalised or with four or more physician visits for asthma in the year following diagnosis had a 2.5-fold higher risk of persistent asthma compared with their counterparts [18].
Our study was not designed to investigate how asthma control trajectories exert their effects on remission, such as via controller therapy use or changes in environmental factors, due to a lack of relevant clinical information. In a recent trial, 2 years of daily inhaled corticosteroid in preschoolers was insufficient to prevent recurrence of symptoms, which were comparable to the control group in the year following drug cessation [32]. Thus, mediating factors linking control trajectories to disease evolution remain to be explored.
As for other potential predictors of remission, atopy was consistently associated with persistence [29, 33]; in line with the literature [18], older age at diagnosis, male sex and season other than winter at diagnosis were associated with a lower likelihood of remission in most provinces. Of interest, most covariates had a lower magnitude of association on remission than the three worst asthma control trajectories.
The strengths of our study include the generalisability of findings to a large, diverse, population-based sample of children with onset of asthma symptoms during preschool and the use of reliable healthcare services and prescription data.
We acknowledge the following study limitations. First, the Quebec cohort included only patients covered by the public drug insurance plan, resulting in an over-representation of lower socioeconomic status families and shorter follow-up due to drug insurance interruption. Nevertheless, Quebec findings were consistent with other provinces. Second, because diagnosis of asthma was based on recorded ICD-9/10 codes in health administrative data without access to triggers and symptoms, we cannot exclude the possibility that some children diagnosed as asthma would have been diagnosed as viral-associated wheeze in other countries. The asthma definition used in this database study has been validated in Canadian children aged 0–17 years with a specificity of 72% [16], which could result in up to 28% of children potentially misclassified as having well-controlled asthma, when they could have had mild infrequent wheezing. We thus purposely restricted eligibility to children with prolonged disease activity to minimise the proportion of “infrequent or transient wheezers”. With this restriction, a further 52% of preschoolers initially fulfilling the asthma definition were excluded, which is similar to the 59% of preschoolers reported to be never/infrequent wheezers in the Avon Longitudinal Study of Parents and Children birth cohort [34]. Third, the validated database index used to assess control did not include symptoms and exercise limitation, and the absence of emergency department data for visits not billed as fee-for-service in Saskatchewan and Manitoba may have under-represented severity and overestimated control, possibly leading to an overestimation of remissions and underestimation of recurrences. However, in a previous sensitivity analysis conducted in Quebec, the impact of removing emergency department visits from the control trajectory definitions resulted in only slight overestimations of control. Fourth, while we adjusted for variables available in administrative databases, other potentially important confounders such as parental history of asthma, environmental factors (e.g. smoke and air pollution), behavioural variables linked to nonadherence, eosinophil counts and ethnicity were unmeasured [35–43]. Nevertheless, our sensitivity analysis suggests that only moderate-to-strong unmeasured confounding could have nullified the estimated associations between worsening and/or out of control trajectories and remission. We cannot rule out the possibility that the presence of weak-to-moderate unmeasured confounding (e.g. from unmeasured tobacco smoke) may have led to the associations found between the fluctuating/improving control trajectories and remission. Fifth, in the absence of important clinical variables, our study did not evaluate underlying mediating pathways that could explain the associations between early-life control trajectories and remission. Sixth, we recognise that remission defined as 2 years of disease inactivity may not equate to cure; ∼16% still experienced recurrence. Finally, while all children were required to have an asthma diagnosis before 5 years of age, those diagnosed at 4 or 5 years were no longer preschoolers at the beginning of follow-up given the 3-year lag between diagnosis and cohort entry.
In conclusion, the asthma control trajectory in the 2 years following diagnosis in preschoolers appears as an important predictor of both remission and recurrence; the worse the control, the lower the likelihood of remission. Long-term randomised controlled trials are needed to assess if our findings have any causative implications for remission; nevertheless, documenting asthma control trajectories shortly after diagnosis could be used for risk stratification or to inform machine learning models to further improve remission risk prediction in preschoolers with newly diagnosed asthma. Further research on mediating factors underlying asthma control trajectories following a preschool age diagnosis is warranted to determine potential interventions and identify children at risk of recurrent exacerbations. Meanwhile, it seems reasonable to aim for achieving rapid and sustained optimal control, targeting the prevention of exacerbations, through evidence-based management and close follow-up of preschoolers with asthma.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-01897-2020.SUPPLEMENT
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01897-2020.Shareable
Acknowledgements
We acknowledge the Manitoba Centre for Health Policy for use of data contained in the Manitoba Population Research Data Repository under project HIPC#2014/2015-28, the Saskatchewan Health Quality Control, Population Data BC and the Régie de l'Assurance Maladie du Québec. We are also grateful to the Commission d'Accès à l'Information du Québec for authorising the study. We are indebted to Christine Lord for study coordination, and Jacques Lacroix, Martha McKinney and Sze Man Tse at the Centre Hospitalier Universitaire Sainte-Justine (Montreal, QC, Canada) for their editorial comments. This research is based, in part, on de-identified data provided by the Ministries of Health of British Columbia, Manitoba, Quebec and Saskatchewan. The inferences, interpretation, conclusions and implications contained herein are those of the authors, and do not necessarily reflect the opinions or policies of the Data Stewards, Government or Ministries of Health of British Columbia, Manitoba, Quebec and Saskatchewan.
Footnotes
This article has supplementary material available from erj.ersjournals.com
Author contributions: F.M. Ducharme and C. Longo had full access to the Quebec data, M. Brownell had full access to the Manitoba data, J.M. Quail had full access to the Saskatchewan data, and M. Sadatsafavi had full access to the British Columbia data; these authors take joint responsibility for the integrity of the data and the accuracy of the data analysis. C. Longo, L. Blais, M. Brownell, J.M. Quail, M. Sadatsafavi, R.W. Platt and F.M. Ducharme contributed to the conception of the research question and the design. C. Longo, L. Blais, M. Brownell, J.M. Quail, M. Sadatsafavi, R.W. Platt, F.M. Ducharme, A. Forget, M-A. Turcot, Y. Nie, W. Li, H. Tavakoli, Q. Tan and Y. Fan acquired, analysed or interpreted the data. C. Longo and F.M. Ducharme drafted the manuscript, and all authors critically revised and approved the manuscript.
Conflict of interest: C. Longo has nothing to disclose.
Conflict of interest: L. Blais reports grants, personal fees and research contracts from AstraZeneca, grants from Teva, outside the submitted work.
Conflict of interest: M. Brownell has nothing to disclose.
Conflict of interest: J.M. Quail has nothing to disclose.
Conflict of interest: M. Sadatsafavi has nothing to disclose.
Conflict of interest: A. Forget has nothing to disclose.
Conflict of interest: M-A. Turcot has nothing to disclose.
Conflict of interest: Y. Nie has nothing to disclose.
Conflict of interest: W. Li has nothing to disclose.
Conflict of interest: H. Tavakoli has nothing to disclose.
Conflict of interest: Q. Tan has nothing to disclose.
Conflict of interest: Y. Fan has nothing to disclose.
Conflict of interest: R.W. Platt reports personal fees from Biogen, Amgen, Merck, AbbVie, Pfizer, Elli Lilly and Analysis Group, outside the submitted work.
Conflict of interest: F.M. Ducharme reports grants from GlaxoSmithKline Canada and MedTeq, grants and personal fees for lectures from Covis Pharma and Thorasys Inc., unrestricted donations from Novartis and Trudell Medical International, outside the submitted work.
Support statement: This study was funded by the Canadian Respiratory Research Network (CRRN). The CRRN is supported by grants from the Canadian Institutes of Health Research (CIHR)–Institute of Circulatory and Respiratory Health (132213), Canadian Lung Association (CLA)/Canadian Thoracic Society (CTS), British Columbia Lung Association, and Industry Partners Boehringer Ingelheim Canada Ltd, AstraZeneca Canada Inc. and Novartis Canada Ltd. We also acknowledge the Fonds de Recherche du Québec–Santé for the fellowship award to C. Longo, and the infrastructure support provided to the Research Centre of the CHU Sainte-Justine, Hôpital du Sacré-Cœur de Montréal, McGill University Health Centre and the Lady Davis Institute of the Jewish General Hospital in Montreal. The funders had no role in the study design, data collection and analysis or preparation of the manuscript. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received May 20, 2020.
- Accepted November 11, 2020.
- Copyright ©ERS 2021. For reproduction rights and permissions contact permissions{at}ersnet.org
References
