Introduction February epidemics of asthma exacerbations in children are becoming a well-recognised phenomenon in the state of New South Wales (NSW), Australia. These epidemics can be identified by examination of routinely collected hospital admission data. Little is known about the extent and consistency of these epidemics, and causal explanations remain speculative. Other techniques for evaluating spatial variation in disease incidence, such as standardised incidence ratios, cannot be used to evaluate these epidemics because the background incidence of asthma hospitalisation against which these epidemics occur is not constant across geographic areas in NSW. The differences in background incidence may arise from systematic geographic variation in hospital admission protocols, population access to hospital services, asthma prevalence and prevailing environmental exposures that trigger asthma exacerbations. I used an alternative technique to assess the intensity of the epidemics and used mapping software to present the results. Methods Using a database of NSW hospital admissions organised in Australian financial years (July to June) from July 1991 to June 2002, I prepared daily counts of hospital admissions for asthma in children aged I to 14 years for geographic units used in the Australian Census. To assess the intensity of the epidemic within each year and geographic unit, I compared the observed count of admissions in February with that expected from the average daily count in the remainder of the year. The intensity was defined as the number of standard deviations from the expected count for the geographic unit, assuming a Poisson distribution in daily counts of admissions. I used a geographic information system to map the results for each year. Results Intensity of asthma admissions in February, compared with expectation, ranged from −3.6 to 14.9 standard deviations over the 11 years and 41 geographic units. Epidemics with a median intensity across all geographic units of greater than 1 standard deviation above expectation occurred in February of 1993 (median standard deviations = 3.0), 1994 (1.8), 1995 (2.3), 1996 (3.4), 1999 (2.6) and 2001 (3.2). Maps showed that in all epidemic years, epidemic activity was concentrated in, but not exclusive to, the densely populated coastal urban areas of NSW. Additional pockets of high intensity occurred in other areas, particularly the northern and southern coastal and near-coastal regions. Conclusions The method provided a useful way of visualising the geographic and annual variation of the epdiemic phenomenon. The concentration of the epidemics in the most densely populated regions is consistent with an infectious cause. Ecological evaluation of variables spatially and temporally associated with epidemic intensity may provide further insights into the cause of the phenomenon.