PURPOSE: To assess recent and to predict future time trends in the incidence of osteoporosis using routine databases in The Netherlands in the period 1993-2015 and to compare estimations based on hip fractures versus a proxy based on pharmacy and hspitalisation data. METHODS: The incidence of hip fractures was estimated over the period 1986-2002 using information from the Dutch Medical Registry, covering more than 99% of all Dutch hospitalisations. Additionally, a proxy (hospitalisation for osteoporosis or osteoporotic fractures, treatment with glucocorticosteroids or treatment with anti-osteoporosis drugs) was constructed in order to identify osteoporotic patients in the PHARMO database. Age and gender specific incidences of hip fractures and osteoporosis were calculated and extrapolated to The Netherlands. Results of both studies were extrapolated till 2015 using a power function. RESULTS: The incidence of hip fractures decreased slightly (270 per 100 000 in 1993, 260 per 100 000 in 2002). Theincidence of osteoporosis using the constructed proxy decreased from 870 per 100 000 in 1993 to 700 per 100 000 in 2002.The incidence of hip fractures and osteoporosis remained fairly constant when modelled till 2015. Both studies showed the same time trends. CONCLUSION: Both the estimations based on the hospitalisations for hip fracture and on our proxy for osteoporosis showed that the increase in the incidence of osteoporosis as observed in the 1990s is levelling off. Due to ageing of the populationthe absolute number of hip fractures will however increase. Our definition of osteoporosis resulted in a higher estimation of the incidence of osteoporosis and may be used in future studies to follow developments in osteoporosis prevalence and incidence.Copyright (c) 2006 John Wiley & Sons, Ltd.