Classifying asthma: disagreement among specialists.

OBJECTIVE: The asthma practice guidelines developed by the National Institutes of Health include a system for classifying asthma severity. The goal of the present study was to assess the interrater reliability of this classification system by measuring agreement among pediatric asthma specialists.
DESIGN: A survey containing eight case summaries was mailed to 24 board-certified pediatric allergists and pulmonologists, who were asked to classify each case according to the national guidelines. The case summaries included the patient's medical history, physical examination, and chest radiograph and pulmonary function test results. Physicians were also asked to interpret the pulmonary function tests, to indicate the main factors used to classify each case (daytime symptoms, nighttime symptoms, pulmonary function testing, or various combinations), and to make treatment recommendations. kappa statistics were used to measure agreement.
RESULTS: Fourteen of 24 surveys mailed (58%) were completed and returned. Agreement was poor for classifying asthma (kappa = 0.29; 95% confidence interval [CI], 0.25 to 0.33) and for the main factors used to make the classifications (kappa = 0.19; 95% CI, 0.14 to 0.23). Specialists exhibited higher agreement in their interpretation of pulmonary function tests (no asthma, kappa = 0.66; asthma on baseline, kappa = 0.53; exercise-induced asthma, kappa = 0.65). While physicians' treatment recommendations were consistent with their severity classifications, the low level of agreement in those classifications led to substantial variability in the treatments recommended.
CONCLUSIONS: The low level of agreement among pediatric asthma specialists in classifying asthma severity suggests the need to refine the classification system used in the national guidelines to help ensure the consistent application of those guidelines.