Patterns of treatment among a cohort of older low-income adults starting new medications for osteoporosis.

Abstract

Among 125,954 new users of osteoporosis (OP) medications, 77% of subjects stopped OP medications, and 23% of subjects added or started a new OP medication during follow-up, with the first addition or start of a new OP medication occurring in a mean of 739days after original OP treatment. We described patterns and predictors of OP medication use, focusing on treatment changes over time. We analyzed health and pharmacy insurance claims for a large cohort of low-income Medicare beneficiaries with a drug benefit for the years 1998-2008. Study subjects had documented Medicare claims and no receipt of OP medications (i.e., bisphosphonate, raloxifene, calcitonin, teriparatide, or hormonal therapy) during a baseline of 180days. Subjects were then required to start an OP medication. Baseline patient and prescriber characteristics were assessed in multivariable Cox regression models to identify correlates of adding or starting a new OP medication. Fractures, bone mineral density testing, and visits with endocrinologists or rheumatologists occurring after baseline were also examined as correlates. We included 125,954 new users of OP medications with a mean age of 78years, 97% female, and 92% white. OP medication prescribers included specialists (i.e., endocrinologists or rheumatologists) (6.2%), orthopedic surgeons (1.0%), primary care providers (64.9%), other physicians (3.7%), and missing (24.1%). Seventy-seven percent of subjects stopped OP medications, and 23% of subjects added or started a new OP medication during follow-up, with the first addition or start of a new OP medication occurring in a mean of 739days after original OP treatment; 4% added or started a new OP medication more than once. In fully adjusted models, many baseline variables correlated with starting a second OP medication. Post-baseline fractures [hazard ratio (HR) 1.76, 95% confidence interval (CI) 1.71-1.82] and bone mineral density testing (HR 2.94, 95% CI 2.86-3.03) were strong predictors. Approximately one quarter of patients starting an OP medication added or started a new OP medication during follow-up. Long-term sequential treatment strategy trials would inform optimal medication treatment for OP.