Compliance with asthma medication is recognised to be a problem. Acquisition of medication is the first step towards compliance. Factors predicting poor collection of prophylactic medication were investigated.A case/control study was conducted. Cases were children who had had at least two consultations for poorly controlled asthma in 1 yr and collected prescriptions of prophylactic medication irregularly. Controls were children whose prescriptions were collected as instructed.Levels of knowledge about asthma and asthma medication were high in both groups. Parents of cases were more likely to perceive their child's asthma to be moderate or severe and more likely to report that their child's asthma was not well controlled. They reported more night-time symptoms, exercise symptoms and school absence. Parents of cases were less likely to report that administering inhalers was part of the evening routine. They were less likely to perceive their child's prophylactic medication to be very effective and more reluctant to administer prophylactic medication.Some parents may decide to undertreat their children, although lack of organised routine may contribute to poor compliance. Parents need guidance on interpreting symptoms and support in establishing routines for the administration of medication.

Assessment tools are a vital component of chronic disease management. The Royal College of Physicians has developed a patient-focused outcome measure for the treatment of asthma, the 'Three Key Questions'. However, in a study investigating the goals of people with asthma, several issues related to the tool emerged. Forty-seven adults of a range of ages and asthma severity but with no significant co-morbidity were interviewed. It emerged that the outcome measure may be subject to recall bias. Also, symptom reports may be conflated if daytime symptoms also occur with activity. 'Interference with activity' is a subjective term the interpretation of which varies considerably. Changes in the level of activity undertaken may be reported rather than changes in symptom severity. The 'Three Key Questions' are not fully patient-centred because they assess the presence of symptoms rather than their importance to the individual. The use of the 'Three Key Questions' as an outcome measure may not allow valid comparisons to be made between settings.