One barrier to receiving adequate asthma care is inaccurate estimations of symptom severity. To interview parents of children with asthma in order to: (1) describe the range of reported illness severity using three unstructured methods of assessment; (2) determine which assessment method is least likely to result in a "critical error" that could adversely influence the child's care; and (3) determine whether the likelihood of making a "critical error" varies by sociodemographic characteristics. A total of 228 parents of children with asthma participated. Clinical status was evaluated using structured questions reflecting National Asthma Education and Prevention Panel (NAEPP) criteria. Unstructured assessments of severity were determined using a visual analogue scale (VAS), a categorical assessment of severity, and a Likert scale assessment of asthma control. A "critical error" was defined as a parent report of symptoms in the lower 50th centile for each method of assessment for children with moderate-severe persistent symptoms by NAEPP criteria. Children with higher severity according to NAEPP criteria were rated on each unstructured assessment as more symptomatic compared to those with less severe symptoms. However, among the children with moderate-severe persistent symptoms, many parents made a critical error and rated children in the lower 50th centile using the VAS (41%), the categorical assessment (45%), and the control assessment (67%). The likelihood of parents making a critical error did not vary by sociodemographic characteristics. All of the unstructured assessment methods tested yielded underestimations of severity that could adversely influence treatment decisions. Specific symptom questions are needed for accurate severity assessments.
