What is new? Comparison of parental and teenager questionnaire reports showed a moderate agreement for rhinitis and eczema and very high agreement for environmental factors. The methodological change from parent to index subject as questionnaire respondent did not affect the prevalence estimates or risk factor analyses in the longitudinal study. Teenagers aged 13 to 14 years were able to complete a questionnaire about allergic diseases and environmental factors in an adequate way.
Most epidemiological studies of asthma and other allergic disorders among children rely on questionnaire reports from their parents [1], [2]. In longitudinal studies of children, the participants reach an age when they become a more reliable source of information about symptoms rather than their parents; therefore, a change in methodology along the way is necessary. The consequences of this methodological change in the study of allergic diseases are mainly unknown. Reliable reports from the study respondents are essential for correct conclusions of prevalence, incidence, and risk factor patterns [3].
Even though several longitudinal cohort studies of asthma and allergic conditions have been carried out from early childhood throughout adolescence [2], [4], [5], [6], only a few studies have compared parental and teenager reports about asthma symptoms [7], [8], [9], [10], [11]. These comparative studies are predominantly based on cross-sectional studies and not evaluations of a methodological change within a longitudinal study [7], [8], [10], [11]. To evaluate the agreement between two observers, a common statistical method to use is the kappa coefficient [12]. In comparative studies focusing on asthma, the agreement between parental and teenager questionnaire reports was good—the kappa estimates were 0.7–0.9—whereas the prevalence of current respiratory symptoms was significantly higher in the reports from the teenagers [7], [8], [9], [10], [11]. Results from comparative studies of allergic rhinitis and eczema are very limited. Two available studies reported a fair agreement between parental and teenager reports; kappa estimates were 0.2–0.4 [8], [10]. However, the difference in prevalence between the parents' and the teenagers' reports was greater for symptoms of rhinitis than the corresponding difference of disease (i.e., hay fever).
Regarding risk factors for asthma and allergic diseases, such as heredity [13], [14] and environmental factors (for instance, indoor dampness [15] and environmental tobacco smoke [16], [17]), comparative studies between different sources of information are also uncommon. Two available studies found very good agreement between parental and teenager reports of parental smoking; kappa values ranged from 0.7 to 0.9 [18], [19].
The aim of the present study was to examine whether the methodological change from parent to index subject as questionnaire respondent affected the prevalence estimates and risk factor patterns for allergic diseases, mainly rhinitis and eczema, in a longitudinal study.
The Obstructive Lung Disease in Northern Sweden (OLIN) pediatric study of asthma, allergy, rhinitis, and eczema started in 1996. The aims and methods have been described previously [2]. All 3,525 children aged 7 to 8 years in three municipalities were invited, and 3,430 (97%) participated. An annual questionnaire was distributed, which included the International Study of Asthma and Allergies in Childhood (ISAAC) questionnaire [1] and additional questions, including respiratory symptoms,
Comparison between the study sample and the entire cohort was made to evaluate selection bias. There were no statistically significant differences in the prevalence of symptoms between the reports from the randomly selected teenagers in the comparison study and all the teenagers participating in the larger questionnaire study (Table 1). Therefore, the sample was considered to be representative. There were no statistically significant differences between the parents and the teenagers in the
We have compared questionnaire reports from teenagers and their parents to evaluate a methodological change within a longitudinal study. The study showed that the prevalence estimates of allergic diseases, symptoms, and environmental factors were not affected by who responded to the questionnaire. However, because the agreement between parental and teenager reports of rhinitis and eczema was fair to moderate according to the kappa statistics, the parents and teenagers, to some extent,
Aina Jonsson, SRN, Kerstin Kemi Björnström, SRN, Lena Gustafsson, SRN, and Berit Lindberg, SRN, are acknowledged for data collection. This work was supported by The Swedish Heart-Lung Foundation, The Swedish Foundation for Health Care Science and Allergy Research (Vårdal), The Swedish Asthma-Allergy Foundation, and U.S. National Institute of Allergy and Infectious Disease.
The same studies showed a similar association with pulmonary function for child and parental reports and suggested that self-reports from children and parental reports were of equivalent quality [18,19]. Studies on allergic diseases show results ranging from poor to moderate and good agreement between parent and child self-report [20–23]. Research suggests that children from age 6–7 years old can report on health experiences [13,24].
The level of agreement for symptoms regarding eczema was moderate to good in the present study. We observed a higher level of agreement than previous studies in comparable age groups [6–8], which might be explained by our longitudinal study design. Parents and children included in a prospective birth cohort might be more aware of skin symptoms than participants sampled randomly from the population for the previous cross-sectional studies [8,21].
The reliability of the smoking reports in the present study is supported by the prevalence of smoking being similar to the Swedish average.24 Further, a comparative study between questionnaires completed by the parents and the teenagers in this cohort show very good agreement for the questions about parental smoking.29 In the present study, self-reported physician-diagnosed asthma was not validated against objective methods.
