Abstract
Assessing the cost-effectiveness of long-term treatment for osteoporosis requires use of mathematical models to estimate health effects and costs for competing interventions. The primary motivations for model-based analyses include the lack of long-term clinical trial outcome data and the lack of data comparing all relevant treatments within randomized clinical trials. We report on specific modeling challenges that arose in the development of a model of the natural history of postmenopausal osteoporosis that is suitable for assessing the cost-effectiveness of osteoporosis interventions among various population subgroups in diverse countries. These include choice of modeling changes in bone mineral density (BMD) or in fracture rate, definition of health states, modeling mortality and costs of long-term care following fracture, incorporation of health utility, and model validation. This report should facilitate future postmenopausal osteoporosis model development and provide insight for decisionmakers who must evaluate model-based economic analyses of postmenopausal osteoporosis interventions.