Although asthma has been recognised for millennia,^1,2 early writers made no attempt to estimate its prevalence. Salter³ declared that asthma “cannot, in this country, be said to be by any means rare, and I believe that all who direct their attention to it will find it to be much commoner than is imagined”. He did not, however, venture any estimate of its prevalence, and it was another 60 years before epidemiological studies began to appear.⁴

These early studies were bedevilled by the lack of an agreed definition of asthma, reluctance to diagnose a chronic untreatable condition that in the Oslerian tradition was widely regarded as psychoneurotic,⁵ and failure to appreciate that upper respiratory tract infection was a major trigger of asthma attacks, so that wheeze triggered by infection was usually diagnosed as bronchitis. Thus, when Collins⁶ described morbidity patterns in 9000 American families in 1935, he reported that, in children aged 5–9 years, bronchitis and chest colds affected 5.6% whereas asthma affected only 0.5%. A similar numerical relationship between asthma and bronchitis was still being reported from English general practices over 20 years later.⁷ How many of these children with bronchitis and chest colds would now be offered a diagnosis of asthma is a moot point, but there can be little doubt that the prevalence of asthma was underestimated on both sides of the Atlantic.

In the 1950s several papers on the epidemiology of childhood asthma appeared from Scandinavia, reporting prevalences of 0.6–1.4%.^8,9,10,11 These studies relied on identifying cases that were known to school doctors, nurses or other professional …