Article Text

Canadian Anaphylaxis Network-Predicting Recurrence after Emergency Presentation for Allergic REaction (CAN-PREPARE): a prospective, cohort study protocol
  1. Waleed Alqurashi1,2,
  2. Marcus Shaker3,
  3. George A Wells4,5,
  4. Gary Stephen Collins6,
  5. Matthew Greenhawt7,
  6. Janet A Curran8,
  7. Roger Zemek1,2,
  8. Suzanne Schuh9,
  9. Anne Ellis10,
  10. Jennifer Gerdts11,
  11. Cheryl Kreviazuk2,
  12. Andrew Dixon12,
  13. Mohamed Eltorki13,
  14. Stephen B Freedman14,
  15. Jocelyn Gravel15,
  16. Naveen Poonai16,
  17. Margitta Worm17,
  18. Amy C Plint1,2
  1. 1Department of Pediatrics and Emergency Medicine, University of Ottawa Faculty of Medicine, Ottawa, Ontario, Canada
  2. 2Children's Hospital of Eastern Ontario Research Institute, Ottawa, Ontario, Canada
  3. 3Section of Allergy and Clinical Immunology, Children's Hospital at Dartmouth-Hitchcock, Lebanon, New Hampshire, USA
  4. 4Cardiovascular Research Methods Centre, University of Ottawa Heart Institute, Ottawa, Ontario, Canada
  5. 5School of Epidemiology and Public Health, University of Ottawa, Ottawa, Ontario, Canada
  6. 6Centre for Statistics in Medicine, University of Oxford Nuffield Department of Orthopaedics Rheumatology and Musculoskeletal Sciences, Oxford, UK
  7. 7Section of Allergy and Clinical Immunology, Children's Hospital Colorado, Aurora, Colorado, USA
  8. 8Pediatrics, IWK Health Centre, Halifax, Nova Scotia, Canada
  9. 9Pediatrics, The Hospital for Sick Children, Toronto, Ontario, Canada
  10. 10Division of Allergy and Immunology, Queen's University, Kingston, Ontario, Canada
  11. 11Food Allergy Canada, Toronto, Ontario, Canada
  12. 12Department of Pediatrics, University of Alberta, Edmonton, Alberta, Canada
  13. 13McMaster Children's Hospital, Hamilton, Ontario, Canada
  14. 14Departments of Pediatrics and Emergency Medicine, Alberta Children’s Hospital, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada
  15. 15Centre Hospitalier Universitaire Sainte-Justine, Universite de Montreal, Montreal, Québec, Canada
  16. 16Departments of Paediatrics, Internal Medicine, Epidemiology & Biostatistics, Western University, London, Ontario, Canada
  17. 17Division of Allergy and Immunology, Department of Dermatology and Allergy, Charite Universitatsmedizin Berlin, Berlin, Germany
  1. Correspondence to Dr Waleed Alqurashi; walqurashi{at}


Introduction Anaphylaxis is a severe, potentially fatal multiorgan system manifestation of an allergic reaction. The highest incidence of anaphylaxis is in children and adolescents. Biphasic anaphylaxis (BA) is defined as the recurrence of allergic symptoms after resolution of an initial reaction. It has been reported to occur in 10%–20% of cases within 1–48 hours from the onset of the initial reaction. The dilemma for physicians is determining which patients with resolved anaphylaxis should be observed for BA and for how long. Guidelines for duration of postanaphylaxis monitoring vary, are based on limited evidence and can have unintended negative impacts on patient safety, quality of life and healthcare resources. The objectives of this study are to derive a prognostic model for BA and to develop a risk-scoring system that informs disposition decisions of children who present to emergency departments (ED) with anaphylaxis.

Methods and analysis This prospective multicentre cohort study will enrol 1682 patients from seven paediatric EDs that are members of the Paediatric Emergency Research Canada network. We will enrol patients younger than 18 years of age with an allergic reaction meeting anaphylaxis diagnostic criteria. Trained ED research assistants will screen, obtain consent and prospectively collect study data. Research assistants will follow patients during their ED visit and ascertain, in conjunction with the medical team, if the patient develops BA. A standardised follow-up survey conducted following study enrolment will determine if a biphasic reaction occurred after ED disposition. Model development will conform to the broad principles of the PROGRESS (Prognosis Research Strategy) framework and reporting will follow the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis Statement.

Ethics and dissemination Ethics approval has been received from all participating centres. Our dissemination plan focuses on informing clinicians, policy makers and parents of the results through publication in peer-reviewed journals and broadcasting on multiple media platforms.

Trial registration number NCT05135377.

  • Paediatric A&E and ambulatory care

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Largest prospective cohort study on paediatric biphasic anaphylaxis conducted to date.

  • Sample size calculation and statistical analysis plan are based on the highest methodological standard for prediction modelling research.

  • We established an international, multidisciplinary expert team encompassing paediatrics, emergency medicine, allergy/immunology, research methodology and statistics, and knowledge translation.

  • We instituted an advisory council of external parents, youth and clinicians end-users and community partners to monitor milestones, identify potential barriers and enablers for future implementation, and guide future decision aid tools.

  • This study is not designed to generalise our findings to settings outside of an academic paediatric emergency department; this limitation may be mitigated when we yield a clinically useful and statistically sensitive model that may be externally validated.


Anaphylaxis is the most severe form of allergic reaction that rapidly affects multiple body systems and can be fatal.1 2 The highest incidence is in children and adolescents.3–8 In Canada, approximately every 10 min, there is an emergency department (ED) visit for food allergy.9 10 Up to 80% of anaphylactic reactions in children are triggered by food,11 and 8% of allergy-related ED visits are due to anaphylactic shock.3

According to the Canadian Institute for Health Information, the rate of children visiting Ontario and Alberta EDs for anaphylaxis more than doubled between 2007 and 2014.3 Among 13–17 yeas, ED visits increased significantly (from 23/100 000 in 2007 to 59/100 000 in 2014). The highest annual rate of ED visits was among children aged 4 and younger.3 Similarly, the Cross-Canada Anaphylaxis Registry reported a steady increase in paediatric ED visits: from 1.8/1000 in 2011 to 4.5/1000 in 2015.10 12 These estimates are higher than data from the USA and Europe.13 14

As the volume of anaphylaxis-related ED visits continues to rise,10 12 ambiguity in how physicians manage anaphylaxis increases the healthcare burden and may contribute to ED crowding. Current Canadian and international guidelines recommend that all patients with anaphylaxis present to the ED, and after initial reactions have been treated, remain there for a prolonged period to be monitored for biphasic anaphylaxis (BA, also called delayed or late-phase anaphylaxis).15–17 BA is a second wave of symptoms after initial resolution.18 19 The reported incidence of this potentially serious phenomenon varies from 10% to 20%; the majority occur within 1–24 hours from onset of the initial reaction.16–47 However, these studies vary considerably in their design (prospective vs retrospective), enrolled population (adults vs children or mixed), settings (ED vs outpatient allergy clinics), and definition and severity of anaphylaxis and biphasic reaction. Recent systematic review and meta-analyses48–50 underline these epidemiological factors that explain the significant clinical heterogeneity between previous observational studies. This inconsistency of the literature creates dilemma for ED physicians in deciding which patients should be observed and the optimum duration of observation.51 As a result, guidelines for postanaphylaxis care vary,1 16 17 are based on poor or little evidence, and have negative impacts on patient safety and quality of life.18 36 52 53 This clinical uncertainty originates from the lack of validated clinical predictors for BA. Consequently, many children are hospitalised or undergo prolonged monitoring in the ED after resolution of initial anaphylaxis.53 54

In the USA, ED care and hospitalisations are the largest drivers of annual direct medical costs (US$1.9 billion) for food allergic children.55 The incremental cost of extended ED observation of resolved anaphylaxis (6 hours vs 1 hour) is US$62 374 per case of BA identified (US$68 411 from the societal perspective). ED monitoring beyond 6 hours of patients who quickly stabilise after treatment is associated with an incremental cost-effectiveness ratio of US$230 202 per case observed (societal perspective).56 As ED crowding and visits for anaphylaxis increase, current postanaphylaxis clinical practice is neither sustainable nor cost-effective.29

Providing the best evidence-based value care at the lowest cost is critical to optimise resource stewardship and eliminate wasteful spending in healthcare. In alignment with national and international research priorities,57–61 our goal is to derive a prognostic clinical prediction model that identifies children with anaphylaxis who are at heightened risk of BA. This model will address a gap in current knowledge and practice, with anticipated benefits for patient care and health system efficiency worldwide.

Methods and analysis

Study design

We will conduct a prospective multicentre cohort study. Prospective data collection is necessary to minimise research waste in prediction modelling, accurately assess the risk and impact of BA on patients and the healthcare system, and derive a clinically useful prediction rule. Our design ensures consistency and precision of data collection of all clinically relevant potential predictors and enables accurate assessment of critical outcomes. Our methods follow established guidelines for developing clinical prediction rules.62–71 We conform to the PROGRESS (Prognosis Research Strategy) methods of prediction modelling.69 72–74

Study population

All children aged 0‒17 years who present to a participating ED will be screened for study enrollment based on the following criteria:

Inclusion criteria

  1. Age <18 years.

  2. Presenting to ED with an allergic reaction that matches diagnostic criteria for anaphylaxis as defined by the World Allergy Organization (WAO) in 2019.75 Anaphylactic reaction is a multisystem allergic reaction characterised by one or more clinical features involving the respiratory or cardiovascular systems and associated with one or more clinical features involving the skin or gastrointestinal tract. These criteria are universally accepted and endorsed by most international allergy and emergency medicine organisations.15 57 76 The 2019 WAO guidelines clarify the involvement of two organ systems is not always requisite for diagnosis: ‘Although the diagnosis of anaphylaxis usually depends on the involvement of multiple organ systems, anaphylaxis may present as an acute cardiac or respiratory event as the only manifestation of anaphylaxis.’75 Thus, an individual with isolated hypotension, bronchospasm, or upper airway obstruction after exposure to a known or potential trigger will be deemed to have anaphylaxis, even if typical skin features are absent.75 77

  3. Language proficiency in English or French

Exclusion criteria

  1. Anaphylactic reaction that occurred in the context of a suicidal attempt or intoxication.

  2. Anaphylactic reaction that began in hospital and managed outside the ED (inpatient or outpatient unit).

  3. Inability or unwillingness of individual and/or caregiver to complete the follow-up surveys post -ED discharge.

Study setting

Between April 2022 and June 2024, the study will enrol participants in EDs from seven hospitals: CHU Sainte-Justine, Children’s Hospital of Eastern Ontario, Hospital for Sick Children, McMaster Children’s Hospital, Children’s Hospital-London Health Sciences Centre, Alberta Children’s Hospital and Stollery Children’s Hospital. These EDs are members of Paediatric Emergency Research Canada (PERC; Research staff will follow site-specific Research Ethics Boards’ guidelines for approaching potential participants and families for research studies, screening for eligibility and obtaining consent.


The primary outcome is development of BA. As per the recently published consensus definition,79 to be classified as BA, an anaphylactic reaction must meet three criteria: (1) initial anaphylactic reaction followed by resolution of all initial manifestations for≥1 hour, with no new symptoms or treatment administered in that time; (2) second phase of new or recurrent symptoms or signs that meet the consensus definition of anaphylaxis occurring within 1–48 hours from complete resolution of initial symptoms or signs and (3) new or recurrent symptoms or signs not caused by antigen re-exposure.35 We will capture any new or recurrent symptoms or signs, but only clinical manifestations that meet diagnostic criteria for anaphylaxis will be defined as anaphylactic biphasic responses. This definition focuses on clinically important or major biphasic reactions.29 30 Mild symptoms that involve only the skin (eg, urticarial rash) will be captured and classified as minor biphasic responses, but they do not meet our case definition for BA.

Data collection in ED

A research assistant (RA) or research nurse (RN) in the ED will approach potential participants to screen for eligibility and provide a study overview. When the prescreen has been completed, the RA/RN will consult with the attending physician to confirm that the symptoms are consistent with anaphylaxis. If the attending physician considers the signs and symptoms to be more in line with another diagnosis (eg, gastroenteritis), the patient will be excluded. After confirming participant eligibility, the RA/RN will obtain written informed consent (and assent as appropriate) and proceed with data collection. Table 1 lists the independent variables that will be collected.

Table 1

Data collection variables

The RA/RN will review the physical exam findings with the clinical team (treating ED physician/bedside nurse). Because anaphylaxis is a clinical diagnosis, participants or caregivers will be asked about the spectrum of symptoms and signs experienced before and on arrival in the ED. The RN/RA will verbally administer a structured questionnaire to participants or caregivers to collect demographics, medical history, risk factors, reaction characteristics, and symptoms. Information from the participant and from the medical record about treatment before and after ED arrival, and BA events during the ED monitoring period, will be captured by the RN/RA. Missing data will be obtained by questioning the participant, caregiver or treating ED team. To capture all BA events and ascertain symptom recurrence while participants are being monitored, the research RN/RA will follow the participant/caregiver throughout the ED visit. Events occurring outside study team hours will be captured in the follow-up questionnaire.

First follow-up after ED discharge or hospital admission

Published data have reported symptom recurrence up to 48 hours from anaphylaxis onset.28 44 We will contact participants by telephone or email 2–5 days after enrolment to complete a standardised questionnaire that will capture the nature and timing of new and recurrent symptoms or signs, follow-up with health providers, return ED visits and treatments received. Events that took place in-hospital, but were not previously captured by the study team (eg, outside study team hours), will be verified from the participant’s medical chart.

Second follow-up after ED discharge or hospital admission

Participants whose anaphylaxis trigger or culprit allergen was unknown at the time of study enrolment will be contacted 6–9 months after enrolment. We will determine if the participant had been assessed by an allergy specialist in the interim, and if so, whether an allergic agent had been identified.

Strategies for retention

For the follow-up survey, the families of participants will be asked: (1) their preferred mode of contact (email or telephone) and (2) the best time to reach them and contact number. Based on their preferences, we will send the follow-up questionnaire as an automated REDCap survey to the parent/caregiver email address or administer the survey by telephone. If the e-survey is not completed within 24 hours, a second email will be sent. If there is no response to a second email, experienced staff will contact the participant for a telephone interview. A similar schedule of repeat calls will be used to reach those who selected telephone follow-up.

Sample size

Based on our research,35 48 49 estimates from prospective ED studies28 44 45 80 and published data from large adult and paediatric studies,81 82 10% is a conservative estimate of the population-wide event rate of BA. Our systematic reviews of potential predictors48 and other relevant studies identified 19 potential predictive variables.83 Recent BMJ and Stat in Med articles offer practical guidance for calculating the sample size required for the development of clinical prediction models.84 85 Following these guidelines, we considered sample size from four perspectives, with the largest being selected as the sample size needed. The four calculations are based on: the approximate 95% CI for the overall outcome proportion 0.10 in the study population (calculated sample size needed n=139); the mean absolute prediction error of the average error in the model’s outcome (n=274); achievement of an expected uniform shrinkage factor of ≤10% (n=1529); and ensuring a small, expected optimism in the apparent proportion of overall variation explained R2 (n=719). Details of these calculations with the selection of the parameter estimates and sensitivity considerations are provided in online supplemental material A. Taking the largest sample size that meets all four criteria, we need to enrol 1529 participants with anaphylaxis. Based on previous studies by our network, we anticipate 10% lost to follow-up.86 87 Thus, our estimated sample size is 1682 participants.

Dependent predictors selection for analysis

Table 2 lists the 19 candidate-dependent predictors that we will include in the analysis. We chose these 19 variables based on clinical studies of predictors of BA by our team and by others,16–47 two systematic reviews,48 50 the meta-analysis from the 2020 anaphylaxis practice parameter49 and clinical experience. These predictors encompass recently published BA predictors from the European Anaphylaxis Registry retrospective data.88 Given the direct association between initially severe anaphylaxis and subsequent BA, we also include risk factors of severe anaphylaxis.89

Table 2

Candidate-dependent predictors that will be included in the analysis

Data analysis

The statistical analysis will be performed using R statistical software V.4.0.5 (R Core Team, Vienna, Austria).90 Descriptive analysis will be used to summarise baseline participant demographics, anaphylaxis clinical manifestations and management characteristics. Although race and indigenous status will be collected as demographic characteristics, we will not perform race-based analysis; these variables will be used as descriptors to demonstrate the diversity and representativeness of our sample.

Multivariable regression analysis will be used to derive a predictive model for BA. As recommended by Royston et al,91 our modelling strategy will follow six steps.

  1. Evaluate data quality. Predictors found to be complete (<10% missing data) will be used in a full model approach. Missing data will be considered Missing at Random. If any potential predictor has>10% missing value, a multiple imputation procedure will be followed to replace these values.64 67 If >50% data are missing, the variable will be omitted from the analysis.

  2. Handle and model continuous predictors. To maximise the predictive ability of the regression model, we will maintain continuous variables such as age.92 93 A multivariable fractional polynomial procedure will be used to identify and model nonlinear continuous variables. Our a priori categorisation of some originally continuous predictors, such as ‘time to epinephrine treatment,’ is based on plausible clinical and basic science research48 94 95 and recent regression analysis.39

  3. Develop final model (predictor selection). Predictors that match the above two criteria will be entered in a ‘full model’ that contains the main effects of all candidate predictors. The objective of predictors reduction is to find the best combinations of variables for accurate prediction (low mean squared error) in a model that is easy for clinicians to use and that contains as few variables as possible. Therefore, we will assess for collinearity and use shrinkage technique as a method of variable reduction.71 Collinearity between predictors will be evaluated with correlation coefficient (r) and variance inflation factors (VIFs), which measure the degree to which collinearity degrades the precision of estimate coefficients. Strongly correlated predictors (r>0.8 or VIF>10) will be combined in a single variable. In accordance with Harrell and Steyerberg, we will use Penalised maximum likelihood (PML) estimation to perform shrinkage reduction (reduction of the regression coefficients to improve prediction quality). Maximising a modified Akaike’s information criterion will be used to choose the optimal penalty factor for PML and select the best model. This approach includes a penalty against large models to deal with the trade-off between overfitting and model simplicity.91 The added benefit of this approach is that we could use more penalty factor if we found significant interaction.

  4. Assess model performance with three measures67 96 97

    1. Calibration refers to the accuracy of absolute risk estimates.96 Model calibration will be assessed by calibration slope, and graphically, by locally weighted scatterplot smoothing (LOESS) plots of observed vs predicted probabilities of the outcome. The slope of the calibration curve is a measure of overoptimism of the model predictions.

    2. Discrimination will be assessed by the receiver operating characteristics curve and the concordance (C) index, which measures how well the model discriminates between participants with and without BA.

    3. Clinical usefulness of the prediction model will be assessed using net benefit as a decision analytic.97 98 The derived prediction rule will be cross-validated by comparing the classification of each participant with their actual primary outcome status.

  5. Validate model

    • Internal validation. Recruiting from geographically separated sites enhances generalisability and supports internal validation of the model.64 99 100 To correct for overfitting and quantify optimism in model performance, our model will be validated internally using bootstrapping through the following steps92 101 102: (1) After developing the prediction model using the entire original sample and determining apparent performance, we will generate a bootstrap sample by sampling individuals with replacement from the original sample; (2) Develop a model using the bootstrap sample (applying the same modelling and predictor selection in step 3 above); (3) Determine the apparent bootstrap performance of this model (performance of bootstrap model in the original sample and calculate the optimism as the difference between bootstrap performance and test performance); (4) Repeat steps 1 through 3 at least 500 times and (5) Average the estimates of optimism in step 4, and subtract the value from the apparent performance obtained in step 1 to obtain an optimism-corrected estimate of performance.

    • External validation: Before broad clinical implementation, our derived rule requires external validation. Lack of external validation is a limitation of many clinical prediction models.48 103 For two reasons, this proposal focuses only on model derivation: (1) Requesting funding for external validation may be premature. Before embarking on external validation, we need proof that our a priori risk factors yield a clinically useful and statistically sensitive model and (2) The validation phase should be broader, in different settings, with other participant, and with different clinicians.104 105 Our ultimate goal is to validate our model and risk score in an international setting. Such validation is feasible because PERC is a member of the Paediatric Emergency Research Networks (PERN), and member networks have a history of collaboration.86 87 106

  6. Present model. As described by Sullivan et al107: we will use the regression coefficient in our final fitted model to generate a clinical decision rule that enables point-of-care risk assessment of BA. To develop a points score system, we will follow the steps described in a recent BMJ paper108: (1) Multiply and round regression coefficients of binary predictors; (2) Search for score for continuous predictors to determine the difference in regression units; (3) Estimate multiplication factor for the scores; (4) Use decision curve analysis to assign participants to risk groups and quantify any deterioration in discriminative performance and (5) Present accompanying table of probabilities to allow points score to be translated into a predicted risk. The anticipated stoplight scoring system (green=low→discharge; yellow=moderate→monitor in ED/preference-sensitive care; red=high→admit to hospital) will inform evidence-based disposition decisions by clinicians and anticipatory guidance to families.

Patient and public involvement

Patients and/or the public were involved in the design and dissemination plans for this research. To promote uptake of our results, potential knowledge users have been and will be engaged throughout the project.109 We have a multiphase approach to maximise collaboration and opportunities for diverse knowledge users to interact at various research phases.110 Our multisite team includes ED clinicians as typical end-users and champions for future implementation. We have established an advisory council of external end-users (parents, youth, ED clinicians) and community partners (Food Allergy Canada, Canadian Society of Allergy & Clinical Immunology) to monitor milestones, identify potential barriers and enablers for future implementation, and guide future decision aids study. The leadership team at Food Allergy Canada has reviewed and supports this proposal. To improve study operation and minimise the burden on patients and families, we sought feedback from the Patients and Families Advisory Committee at the Children’s Hospital of Eastern Ontario Research Institute.

Ethics and dissemination


Ethics approval has been received from all recruiting centres Written informed consent, and/or assent when appropriate, will be obtained from all participants or legal guardians.

The study is registered at (NCT05135377). Results information from this study will be submitted to

End-of-grant KT (knowledge translation)

ED personnel, providers, allergists, clinical researchers, administrators and government policy makers can use our study outputs to improve healthcare delivery. KT will focus on informing clinicians, other key user groups, and parents and participants. Our plan has three goals: increase knowledge awareness, inform/change practice and inform future research.111 112

We have a powerful infrastructure to disseminate our results. Study investigators are senior members of PERC and PERN, networks that include paediatric ED researchers worldwide (>100 hospitals across 6 PERN networks),113 practicing clinicians, medical educators and healthcare administrators. PERC is closely tied to the TRranslating Emergency Research Knowledge for Kids Network of Centres of Excellence,114 a partnership for knowledge exchange between general EDs and PERC sites. Our reporting/publication of the study results will conform to the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis checklist.102

Ethics statements

Patient consent for publication


We are grateful to the Patients and Family Advisory Committee at the Children’s Hospital of Eastern Ontario Research Institute, and Food Allergy Canada for the insightful comments and feedback on the recruitment approach and data collection forms. We are also grateful to the following research coordinators and research managers for the operational support of the study at their institutions: Candice McGahern, Jena Shank, Judith Sweeney, Kamary Coriolano DaSilva, Patricia Candelaria, Redjana Carciumaru and Stéphanie Pellerin.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Collaborators The Pediatric Emergency Research Canada (PERC) Network members include the following: Waleed Alqurashi, MD, MSc, Roger Zemek, MD, and Amy Plint, MD, MSc (Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada); Janet Curran, PhD, RN (IWK Health Centre, Halifax, Nova Scotia, Canada); Suzanne Schuh, MD (The Hospital for Sick Children, Toronto, Ontario, Canada); Andrew Dixon, MD (Stollery Children’s Hospital, Edmonton, Alberta, Canada); Mohamed Eltorki MBChB (McMaster Children’s Hospital, Hamilton, Ontario, Canada); Stephen B. Freedman, MDCM, MSc (Alberta Children’s Hospital, Calgary, Alberta, Canada); Jocelyn Gravel, MD, MSc (Sainte-Justine Pédiatrie, Montreal, Quebec, Canada); and Naveen Poonai, MD, MSc (Western University, London, Ontario, Canada).

  • Contributors WA and ACP conceived the study idea. WA, ACP and MS wrote the protocol with input from GAW, GSC, MG, JAC, RZ, SS, AE, JG, CK, AD, ME, SF, JG, NP and MW. All authors provided input into the methodology and analysis plan. All authors approved the final protocol manuscript. ACP and GAW are the supervisors of the study.

  • Funding This work is supported by the Canadian Institutes of Health Research (CIHR), grant number PJT-175057. This work is also supported by the Children’s Hospital Academic Medical Organization (CHAMO) Innovation Fund, grant number N/A.

  • Competing interests All authors have read and understood BMJ policy on declaration of interests and have no relevant interest to declare. ACP is supported by a Tier I University of Ottawa Research Chair. SF is supported by the Alberta Children’s Hospital Professorship in Child Health and Wellness.

  • Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.