Drug-Drug Interactions in Nursing Home Residents With vs Without Dementia

OBJECTIVES: To examine whether nursing home (NH) residents with Alzheimer's disease and related dementias (ADRD) are more likely than those without ADRD to be exposed to potential drug-drug interactions (DDIs) and to experience longer durations of exposure to DDIs. DESIGN: Retrospective observational study. SETTING AND PARTICIPANTS: Long-stay US NH residents. METHODS: We leveraged Medicare Fee-for-Service enrollment and claims data linked to Minimum Data Set 3.0 clinical assessments from 2018 to 2020 and identified NH residents aged ≥66 years with observable part D prescription drug data. We identified exposure to 98 potential DDIs during the duration of their NH stay. We described the number of DDIs with a clinically meaningful difference in prevalence and/or duration between groups, defined as a ≥1% difference in prevalence and ≥7 days' duration, respectively. We used a log-binomial model to estimate the association between ADRD and exposure to any DDI as an adjusted prevalence ratio with 95% CIs and predicted the marginal prevalence of exposure to any DDI for those with vs without ADRD, standardized to population covariate distributions. RESULT: We identified 485,251 eligible NH residents (average age, 84.6 [SD, 8.1] years; 67.1% female; 70.1% with ADRD; 56.6% with a potential DDI). A diagnosis of ADRD was associated with a higher prevalence of DDI exposure (prevalence ratio, 1.09 [95% CI, 1.08, 1.09]; marginal predicted prevalence for those with vs without ADRD: 58.0% vs 53.3%). Twelve DDIs had different prevalences between groups, and 34 DDIs had different differences in duration. For example, 29.6% of residents with vs 21.2% without ADRD were exposed to ≥3 central nervous system-active drugs. CONCLUSIONS AND IMPLICATIONS: NH residents with ADRD had a higher prevalence and duration of potential DDIs. Future research should evaluate whether the interactions we identified have causal effects on adverse outcomes such as falls.