In their article, Hansen et al. [1] found a scant overlapping between parental report of doctor diagnosis, the diagnosis from a population-based hospitalization registry and anti-asthmatic drug use from a population-based prescription registry.
The described picture is expected, and the differences are greater in the preschool children, when all asthma phenotypes are present (persistent or transient wheezers and their subgroups) [2], and only a few will end up with asthma diagnosis. Moreover, the difficulty to diagnose asthma in the preschoolers is one of the main reasons for choosing a population ≥ 6 years old in many epidemiological studies, like study by Moth el al [3].
Moreover, all children with at least one R03 prescription in their first 7 years of age were considered as asthmatics. This choice contributed to the lack of overlap, since many cases may have received a few occasional prescriptions for an acute episode (e.g. viral wheezing) or for clarifying a diagnosis. Criteria by Moth el al have been tested in a 6-14 years population and for one year observation period. [3] The same criteria cannot be applied to a 7 year period. We think that considering the degree of anti-asthma therapy utilization (i.e. number of packages prescribed per year), might be a manner to identify actual asthma cases, and to reduce discrepancy between asthma prevalence and anti-asthma drugs prescription.
Finally, the discrepancy between hospitalization and asthma prevalence (actual, or estimated by prescription databases) is expected, since history of hospitalizations is one of the predictor of severe exacerbations [4] (not a predictor of persistent symptoms), and less than all it could be a proxy of the disease prevalence. Furthermore, hospitalization diagnosis could not be accurate. However, it should be underlined that an agreement of 90% was found between cases identified both in hospitalization and in prescription databases.
We analysed anti-asthmatic drug prescriptions (with the exclusion of nebulised suspensions) dispensed during 2008 to 1,047,241 subjects 6-17 year olds living in Lombardy Region, Italy, and collected into an administrative prescription database. An agreement of 76.3% and kappa 0.77 (0.73–0.81) were found between cases identified by anti-asthmatic drug prescriptions and cases with a diagnosis made by the family paediatrician. [5] We estimated an asthma prevalence based on drug prescriptions of 3.9%, very similar to the prevalence based on paediatrician’s diagnoses (4.0%). In our sample, the hospitalization rate for asthma was 0.43 per 1000 inhabitants, and 1.1% of children 6-17 year old receiving an asthmatic drug prescription were hospitalized.
In conclusion, we agree that healthcare administrative databases are a useful tools to estimate asthma prevalence, and that integrating different data can result in a more accurate estimates. However, we think that when using drug use as a proxy of asthma, age and quantity and quality of drug prescriptions should be taken into account.

Marina Bianchi, Antonio Clavenna, Maurizio Bonati
Laboratory for Mother and Child Health, Department for Public Health, "Mario Negri" Institute for Pharmacological Research, Milan, Italy

References
1. Hansen S, et al. Comparison of Three Methods to Measure Asthma in Epidemiologic Studies: Results from the Danish National Birth Cohort. PloS one 2012; 7(5):e36328
2. Martinez F. New insights into the natural history of asthma: Primary prevention on the horizon. JACI 2011: 128:939
3. Moth G, et al. Identification of asthmatic children using prescription data and diagnosis. Eur J Clin Pharmacol. 2007 Jun;63(6):605-11
3. Wu AC, et al Predictors of symptoms are different from predictors of severe exacerbations from asthma in children. Chest 2011;140(1):100-7
4. Bianchi M, et al. Asthma diagnosis vs. analysis of anti-asthmatic prescriptions to identify asthma in children. Eur J Clin Pharmacol. 2011;67(9):967-8