Rationale, aims and objectives Multiple treatments are available for osteoporosis; however, little is known about treatment change patterns and associated factors. Osteoporosis treatment change patterns, discontinuation and factors associated with treatment change in members of a large national health plan were examined. Methods A retrospective cohort study was conducted in 7315 commercial and 34 146 Medicare Advantage Prescription Drug (MAPD) members newly initiated on an osteoporosis medication between 2006 and 2008. Osteoporosis treatment change, discontinuation and re-initiation patterns were assessed. Multivariate logistic regression was used to examine factors associated with treatment change. Commercial and MAPD members were assessed separately because of differences in demographics and insurance benefits. Results Approximately 12% of members had a change in index therapy within 12 months. Almost 60% of members discontinued the index medication at least once, based on a 90-day refill gap. Over 40% of members discontinued all osteoporosis medications by the end of 12 months post-index. Among MAPD and commercial members, women and those with risedronate, ibandronate or calcitonin at index, index therapy in 2008 and an osteoporosis diagnosis were more likely to have a treatment change while members with health plans other than health maintenance organizations and generic alendronate at index were less likely to have a treatment change. Conclusions Osteoporosis treatment change occurred in approximately 12% of members, while a greater proportion of members discontinued treatment completely within 12 months. Member characteristics may be used to predict therapy change for evaluation and quality initiatives within a health plan. © 2011 Blackwell Publishing Ltd.