Modelling the use of health services by local populations can inform the allocation of central resources to larger regions. At aggregate, e.g. regional, level, use is largely driven by supply and such analysis can be misleading. At small area level, however, populations effectively compete for services and their different relative use of services, after allowing for varying availability, reflects variations in their relative need. Empirical quantitative estimates can thus be made of the net relation of factors such as mortality or social deprivation to relative need for health services. These estimates can then be used in weighted capitation formulae for allocating resources to larger regions.