(148) Potentially Harmful Medication Prescribing to Nursing Home Residents with Dementia: The Role of Prescriber Specialization in US Nursing Homes

Publication Number: 822

Background: Potentially harmful medication (PHM) use is associated with particularly poor outcomes in patients with dementia (ADRD), who represent over half of nursing home (NH) residents in the US. We know that physician practice in US NHs has been changing rapidly with the growth of a subset of NH physicians and advanced practitioners (e.g., nurse practitioners) who focus most of their practice on NH patients (NH specialists).

Objectives: Our objective was to compare PHM prescribing for long-stay NH patients with ADRD under the care of NH specialists vs. non-NH specialists.

Methods: This was a retrospective cohort study of a 20% random sample of Medicare beneficiaries with ADRD who resided in 12,269 US NHs in 2017. Long-stay NH patients with ADRD were identified using MDS, Medicare Part A (facility) and Part B (professional) claims. Patients < 65 years old and those without continuous Part D (medication) coverage were excluded. Physicians in generalist specialties and advanced practitioners with at least 90% of Part B claims for NH care were considered NH specialists. Patients were assigned to one of two groups - NH specialists vs. non-NH-specialists based on plurality of Part D claims submitted for that patient by clinicians in one of the two groups. Our primary outcomes were any PHM use (defined using the Beers Criteria) and proportion of NH days on any PHM. Our secondary outcomes were the same variables by medication class. Outcomes were modeled using GEE with binary distribution for any PHM use and multinomial outcome for the proportion of NH days on a PHM (<=25%, 25-75%, or >75% of days exposed). Models included patient demographics, clinical characteristics, measures of cognitive and functional status, behavioral assessments and facility characteristics. Standard errors were adjusted to account for clustering of patients within facilities.

Results: Of the 54,713 patients in the sample, 27.9% received prescriptions predominantly from NH specialists and 72.1% from non-NH-specialists. There was no statistically significant difference in any PHM use between the groups (odds of receiving PHM in the NH specialist group: OR=0.99, 95% CI 0.94-1.05) but lower odds of being on a PHM more than 75% of NH days (OR=0.87, 95% CI 0.80-0.95) for patients in the NH specialist group compared to the patients in the non-NH-specialist group. The same pattern of findings was observed by medication class, with the exception of any benzodiazepine use which was higher in the NH specialist group (OR=1.07, 95% CI 1.01-1.14).

Conclusions: Long-term care NH residents under the care of prescribers who specialize in NH practice are less likely to receive PHMs over longer periods of time. However, they are more likely to experience occasional benzodiazepine use than NH residents under the care of non-NH-specialists.