AIM Responses to respiratory questionnaires are often used to identify individuals with asthma symptoms and may also be used to identify asymptomatic individuals. This study investigates the repeat responses over four years to such a questionnaire in a population of adult New Zealanders. METHODS Seven hundred and twenty three asthmatics were sent two almost identical questionnaires in three areas of New Zealand, separated by approximately four years. All of them had answered yes to at least one of the three questions under study in the first survey. RESULTS Following the second asthma questionnaire only 487 (67.4%) answered yes to at least one of the survey questions. Similarly, 51.1% of those who had reported having nocturnal shortness of breath in the first survey did so in the second survey, 69.9% of those who reported having had an asthma attack in the first survey did so in the second survey, and finally 74.8% of those who reported using asthma medication in the first survey did so in the second survey. CONCLUSION Even in a previously identified symptomatic asthmatic group, a large proportion did not report respiratory symptoms and asthma medication use four years later. This implies that the true prevalence pool of susceptibles is likely to be far greater than is identified in surveys of the 12-month period prevalence of asthma symptoms. This has implications not only for the design of epidemiological studies (e.g., it poses problems for the selection of a control group of non-asthmatics in prevalence case-control studies), but also for the planning of health services and educational programmes for people with asthma.