Can schoolchildren provide valid answers about their respiratory health experiences in questionnaires? Implications for epidemiological studies.

To evaluate the relative validity of information on children's respiratory experience given by different informants, we examined and compared the relationship between low ventilatory function (defined as more than 1 standard deviation below the corresponding mean) and schoolchildren's respiratory symptoms or illnesses reported separately by the children and their parents, using a standard respiratory questionnaire. A total of 1,963 children aged 8-12 years from 12 primary schools in three districts of Hong Kong provided parent-completed and self-completed questionnaires, as well as acceptable spirometric measurements. Prevalence of low forced expiratory volume ratio (FEV1/FVC) and low forced expiratory flow rate between 25-75% of FVC (FEF25-75) were higher among those with either parent or child-reported symptoms/illnesses. Child-reported cough and phlegm performed better than the corresponding parent-reported symptoms in predicting low FEV1/FVC. The contrary was true for wheezing and bronchitis. For low FEF25-75, parent-reported wheezing, asthma, and bronchitis performed better, while the opposite was true for cough. Subgroup analysis by age showed that for older children (age 10 or above), child-reported symptoms/illnesses performed better in general in the prediction of low FEV1/FVC. On the other hand, parent-reported symptoms/illnesses seemed to have an advantage over child-reported ones in predicting low FEF25-75. Subgroup analysis by sex did not reveal any clear pattern. Overall, there was little difference between respiratory illness experiences reported by schoolchildren and their parents in terms of their associations with low ventilatory function. In a population-based study in which schoolchildren are subjects, it would be appropriate for respiratory questionnaires to be administered to the children themselves, especially if they have reached age 10. By doing so, higher response rates, and perhaps also better yields of correct information, may be obtained. Copyright 2004 Wiley-Liss, Inc.