Validation of self-reported asthma medication use in Sweden and Finland

<bold>Background:</bold> Epidemiological surveys are a common tool used in asthma research, but little is known about the agreement between self-reported asthma medication use and register-confirmed asthma medication dispensations. <bold>Aim:</bold> To assess the validity of, and agreement between, self-reported asthma medication use and register-confirmed asthma medication dispensations among adults in Sweden and Finland. <bold>Methods:</bold> Survey data were collected in 2016 from 5,466 adults aged 20-69 years in northern Sweden, 24,534 in western Sweden, 1,761 in Helsinki, and 1,849 in Seinäjoki-Vaasa, Finland. These data were linked to national register data on asthma medication dispensations. Sensitivity and specificity were calculated using the survey data as gold standard, along with Cohen’s kappa as a measure of agreement. Meta-analysis was used to calculate pooled total and subgroup estimates, among all and stratified by self-reported physician-diagnosis of asthma. <bold>Results:</bold> Self-reported asthma medication use compared with registered asthma medication dispensations yielded a pooled sensitivity of 67%, specificity of 97%, and a kappa of 0.68 among all subjects. Results were consistent across sex, smoking status, and education levels. In subjects with physician-diagnosed asthma, pooled results showed a sensitivity of 73%, specificity of 90% and a kappa of 0.56. Conversely in subjects without physician-diagnosed asthma, sensitivity was 51%, specificity 97% and kappa 0.46. <bold>Conclusions:</bold> The overall agreement between self-reported asthma medication use and registered asthma medication dispensations was good. Assessment of asthma medication use was highly specific in all subjects, both among those with and without a physician-diagnosis of asthma.