A comparison of childhood asthma case definitions based on parent-reported data.

PURPOSE

The purpose of this study was to compare research definitions of childhood asthma based on parent-reported data.

METHODS

We conducted a multicenter, prospective cohort study of 921 infants hospitalized for bronchiolitis. Follow-up was conducted via biannual parent interviews. Asthma definitions were developed using parent-reported data: clinician diagnosis by the age of 5 years ("broad definition"), clinician diagnosis by the age of 5 years with either asthma medication use or asthma symptoms during the age of 4-4.9 years ("epidemiologic definition"), clinician diagnosis by the age of 5 years with either long-term inhaled corticosteroid use or asthma symptoms during the age of 4-4.9 years ("strict definition"), and a "flexible definition" met by any two of the three criteria in the epidemiologic definition. Asthma outcome definitions were evaluated using unadjusted associations with known major asthma risk factors and validated against the medical record in a subset (n = 116).

RESULTS

Asthma prevalence for the broad definition was 294 of 875 (34%); epidemiologic definition, 235 of 859 (27%); strict definition, 229 of 859 (27%); and flexible definition, 364 of 826 (44%). Risk factors had similarly strong associations with definitions that required clinician diagnosis and weaker associations with the flexible definition. The epidemiologic and strict definitions had the highest specificity (96%) and positive predictive value (92%).

CONCLUSIONS

The parent report of clinician-diagnosed asthma correlates well with known asthma risk factors.