The data routinely collected about the use of accident and emergency services do not enable planners to make appropriate comparisons of the demand arising from different populations. This study describes how simple cross-sectional survey data can be used to calculate age, sex, and social class-specific utilisation rates and to assess the importance of other factors influencing the pattern of demand for a given area. The application of multiple regression techniques to provide an explanation of the differences observed in clinical casemix in terms of demographic factors and general practitioner availability; to compare the level and type of demand observed from a given population; and to derive a population weighting factor for planning and resource allocation purposes is also illustrated. It is suggested that, in future, these factors need to be taken into account by those responsible for planning and monitoring the performance of these services. Further studies of the relationship between patients' perceptions of the need for hospital or primary care, and the availability of general practitioner care; the influence of geographical distance, the organisation of hospital services and attitudes of staff providing accident and emergency care in different locations and the subsequent use of these services are required. Planners will need this additional information before they are able to recommend strategies to influence the demand for accident and emergency care from a given population.