Table 1

Criteria for the search strategy and rules devised to operationalise inclusion/exclusion criteria

Search strategyDefinitionRules to operationalise
Children aged 5–12 years with doctor-diagnosed asthma, across all severities and degrees of controlWe included studies with a wider range of ages if results for children aged 5–12 years were reported separately or if more than 50% of the children were within this age range.
We excluded:
  • Studies in which the diagnosis of asthma was not doctor diagnosed (eg, parent reported).

  • Epidemiological studies in which the denominator was the general population (eg, asthma admissions from deprived/rural/urban communities) as opposed to a population of children with asthma.

Any setting
Severe asthma attacks as defined by the ATS/ERS Task Force14: asthma symptoms and/or objective evidence of obstruction outside the normal variation for the patient necessitating (A) a short course (at least 3 days) of oral corticosteroids, and/or (B) a hospitalisation or emergency department visit requiring systemic corticosteroidsWe excluded studies assessing risk factors for:
  • Attacks defined only by use of rescue bronchodilator medication.

  • Poor control/severe asthma (eg, as defined in GINA6) unless ‘attacks’ were listed as a separate outcome.

  • Failure to recover, relapse, early readmission from an existing attack.

  • Fatality/intubation/ICU as specific outcomes (though these may be included within the definition of severe asthma).

Risk factors/predictorsAny feature of the child, family or social context, clinical history, examination finding or test routinely available in clinical practice which increased the propensity of the child to have an asthma attackWe excluded:
  • Tests currently only available in the context of research (eg, novel biomarkers, genomics).

  • Triggers for attacks, for example, respiratory tract infections, exposure to airborne allergens, exercise.

Any routine clinical care (eg, attendance (or not) at routine reviews) and disease-related behaviours (eg, adherence (or not) to preventer medication)We excluded:
  • Trials of interventions (eg, inhaled steroids, self-management) proven to reduce exacerbations because we anticipated confounding by indication.

  • Complex/multifaceted interventions (eg, enrolled in a local disease management initiative) because the impact is unlikely to be generalisable.

Study designsCohort, case-control studies and cross-sectional studies (but not case studies and case series)We included controlled trials if they included an analysis that provided an estimate of risk of an attack (eg, cohort or cross-sectional studies from baseline data).
  • ATS, American Thoracic Society; ERS, European Respiratory Society; GINA, Global Initiative for Asthma; ICU, intensive care unit.