Regional epidemiology and research on regional health care variations both focus on geographical aspects of health. The link between the two disciplines, however, is not a strong one. Numerous reasons for this are discussed in relation to the concepts of need, demand and use of health care, and in relation to different data sets on mortality and morbidity. The variations in use of health care (hospital use) are usually explained as a function of need, supply and medical practice style and clinical judgement. Epidemiology may play an important role in exploring the basis for regional differences in need. Further, medical geography is brought into the discussion. Possible contributions from the medical geographer into health-related items are outlined: small-area analysis, comparison of health conditions in geographical units of appropriate aggregation level, etc. A pilot study in three Norwegian hospital areas, carried out to test the association between morbidity and mortality of ischaemic heart disease, shows a negative association between observed hospital discharge rate and mortality rate. However, the expected rates influence this association. The results are interpreted in the context of data quality and methodological pitfalls.