The Mediterranean osteoporosis (MEDOS) study was carried out in 14 centres from six countries in Southern Europe to determine the incidence rates and risk factors associated with hip fracture over the age of 50 years. This paper discusses both the validity and relevance of the data, that is, whether the number of collected cases of hip fracture and the size and age distribution of the population are representative of the population as a whole, and whether the incidence measures used in the study are suitable for comparing the risk of hip fracture between populations and for predicting future risk within populations. Five measures of risk were assessed at each centre: crude incidence over the age of 50 years; age-standardised incidence; risk increase/fracture doubling time by age; computed incidence at 50 years; and excess morbidity. Three standardised populations were used for comparison: the MEDOS population (incidences standardised to the overall age and sex distribution of all the participating centres); the weighted MEDOS population (only including the age range 60 to 84 years); and the Swedish population at the start of the study. The MEDOS study showed that the incidence of hip fracture increased exponentially with age in both sexes at all centres. The regression slope of incidence against age was affected by the age distribution of the population, but not by the absolute size of the population. Methods used to define the population of the catchment areas did not introduce errors of a greater magnitude than would the occasional addition or removal of single fracture cases.(ABSTRACT TRUNCATED AT 250 WORDS)