Patterns of oral bisphosphonate deprescribing in older nursing home residents with dementia

Joshua D Niznik,Sherrie L Aspinall,Laura C Hanson,Meredith A Gilliam,Xintong Li,Casey J Kelley,Carolyn T Thorpe

DOI: https://doi.org/10.1007/s00198-021-06141-9

Abstract:In a national sample of Medicare nursing home residents with dementia treated with bisphosphonates, 20% had bisphosphonates deprescribed. Residents with clinical characteristics representing decreased likelihood for long-term benefit were more likely to have bisphosphonates deprescribed. Future studies are needed to evaluate outcomes of deprescribing bisphosphonates in this population. Introduction: To determine incidence of deprescribing bisphosphonates among nursing home (NH) residents with dementia and identify factors associated with deprescribing. Methods: 2015-2016 Medicare claims, Part D prescriptions, Minimum Data Set (MDS) 3.0, and Nursing Home Compare for non-skilled NH residents aged 65 + with dementia and prescriptions for oral bisphosphonates overlapping the first 14 days of the stay. Our primary definition for deprescribing was a 90-day gap in medication supply; we also explored the reliability of different deprescribing definitions (30-, 90-, 180-day gaps). We estimated associations of NH, provider, and resident characteristics with deprescribing bisphosphonates using competing risks regression models. Results: Most NH residents with dementia treated with bisphosphonates (n = 5312) were ≥ 80 years old (72%), white (81%), and female (90%); about half were dependent for transfers (50%) or mobility (45%). Using a 90-day gap in supply, the 180-day cumulative incidence of deprescribing bisphosphonates was 14.8%. This increased to 32.1% using a 30-day gap and decreased to 11.7% using a 180-day gap. Factors associated with increased likelihood for bisphosphonate deprescribing were age ≥ 90 years, newly admitted (vs. prevalent stay), dependent for mobility, swallowing difficulty, > 1 hospitalization in the prior year, CCRC facility, and nurse practitioner primary provider (vs. physician). Cancer and western geographic region were associated with reduced likelihood for deprescribing. Conclusion: In a national sample of NH residents with dementia, bisphosphonate deprescribing was uncommon, and associated with clinical characteristics signifying poor prognosis and decreased likelihood for long-term benefit. Future studies should evaluate clinical outcomes of deprescribing bisphosphonates in this population.