Examining the Effect of Medication Adherence on Risk of Subsequent Fracture Among Women with a Fragility Fracture in the U.S. Medicare Population.

BACKGROUND In the United States, osteoporosis affects approximately 10 million people, of whom 80% are women, and it contributes a significant clinical burden to the community. Poor adherence to osteoporosis medications adds to the overall burden of illness. OBJECTIVE To examine the association of osteoporosis medication adherence and the risk of a subsequent fracture among Medicare-enrolled women with a previous fragility fracture. METHODS This study was a retrospective observational analysis of U.S. administrative claims data among female Medicare beneficiaries who had a nontrauma closed fragility fracture between January 1, 2011, and December 31, 2011. Patients were required to have continuous medical and pharmacy enrollment 12 months pre- and postfracture date. In addition, patients were required to have an osteoporosis medication prescription for a bisphosphonate (alendronate, risedronate, pamidronate, etidronate, zoledronate, and tiludronate), calcitonin, denosumab, raloxifene, or teriparatide during the follow-up period. Adherence was calculated using cumulative medication possession ratio (MPR) from the treatment initiation date in 30-day increments. MPR was stratified into high adherence (MPR ≥ 80%), moderate adherence (50% ≤ MPR > 80%), and low adherence (MPR < 50%). Outcomes included first subsequent fracture after treatment initiation; patients were censored at treatment discontinuation, or end of the 12-month period posttreatment initiation. Covariates included demographics, comorbidities, osteoporosis medications, medications associated with falls, and health care utilization. Cox regression was used to model subsequent fractures with time-dependent cumulative MPR. RESULTS Of the 1,292,248 Medicare enrollees who had a fracture in 2011, a total of 103,852 (8.0%) women aged ≥ 65 years with a fragility fracture were identified. Overall, 27,736 (26.7%) patients were treated with osteoporosis medication within 12 months of the fragility fracture (mean time to treatment initiation was 85.0 ± 84.6 days). Over half of the patients were highly adherent (MPR ≥ 80%) to osteoporosis medications during the follow-up (n = 14,112; 50.9%). Almost a third of the patients had low adherence (MPR < 50%; n = 9,022, 32.5%), followed by patients with moderate adherence (50% ≤ MPR > 80%; n = 4,602, 16.6%). After adjusting for demographics and clinical characteristics, patients with low and moderate adherence to osteoporosis medications were 33% (hazard ratio [HR] = 1.33; 95% CI = 1.17-1.50, P < 0.001) and 19% (HR = 1.19; 95% CI = 1.02-1.38, P = 0.026) more likely to have a subsequent fracture, respectively, compared with patients with high adherence. Low adherence patients had a 32% and 34% increased risk for a hip/pelvis/femur fracture (HR = 1.32; 95% CI = 1.09-1.59, P = 0.005) and a clinical vertebral fracture (HR = 1.34; 95% CI = 1.09-1.63, P = 0.005), respectively, compared with high adherence patients. CONCLUSIONS Medicare-enrolled women with low and moderate adherence to osteoporosis medications had a higher risk of a subsequent fracture compared with high adherence patients. These results highlight the importance of improving osteoporosis medication adherence among women enrolled in Medicare. DISCLOSURES This study was funded by Eli Lilly. Xie, Keshishian, and Baser are employees of STATinMED Research, a paid consultant to Eli Lilly in connection with the study design, data analysis, and development of the manuscript for this study. Boytsov, Burge, Lombard, and Zhang are employees and stock owners of Eli Lilly. At the time of research, Krohn was an employee of Eli Lilly. Study concept and design were contributed by Burge and Lombard, along with the other authors. Xie, Baser, and Keshishian took the lead in data collection, assisted by the other authors. Data interpretation was performed by Krohn and Zhang, with assistance from the other authors. The manuscript was written by Keshishian and Boytsov, along with the other authors, and revised by Boytsov, Keshishian, and Burge, along with the other authors.