Nearly 30% of nursing home residents with dementia took three or more CNS-active drugs

Design

Retrospective cohort using Medicare claims linked to Minimum Data Set assessments

Population

485,251 long-stay US nursing home residents aged 66 or over

Primary outcome

Prevalence and duration of potential drug-drug interactions

Effect

Any interaction PR 1.09 (1.08 to 1.09); ≥3 CNS-active drugs 29.6% vs 21.2%

This retrospective study, published in August 2026, linked Medicare claims and clinical assessment data for 485,251 long-stay US nursing home residents aged 66 or over between 2018 and 2020, 70% of whom had dementia. It screened for 98 potential drug-drug interactions.

More than half of residents, 56.6%, were exposed to at least one potential interaction. Dementia was associated with higher exposure (prevalence ratio 1.09, 95% CI 1.08 to 1.09; 58.0% vs 53.3%) and longer exposure for 34 of the interactions. The most striking difference: 29.6% of residents with dementia were on three or more CNS-active drugs, against 21.2% of those without.

Combinations of antipsychotics, antidepressants, benzodiazepines, opioids and antiepileptics in a person with dementia raise the risk of falls, sedation and delirium. The study cannot show these interactions caused harm, but three or more CNS-active drugs is a recognised marker of risk and an easy one to check. Indian data on care homes are sparse, but the same prescribing patterns are common wherever behavioural symptoms are managed with medication.

• Count CNS-active drugs at every medication review for residents with dementia.
• Review any regimen with three or more CNS-active drugs, aiming to reduce the number.
• Look for and treat pain, constipation and infection before adding drugs for behaviour.
• Taper benzodiazepines and antipsychotics gradually where they are no longer needed.
• Use a structured interaction check when starting any new drug in a care home resident.

Why it matters

The residents least able to report side effects are the ones most often given interacting brain-active drugs.

Don't overread it

Claims data show exposure to potential interactions, not that they caused falls or other harm.

The statistics, in plain English

A prevalence ratio of 1.09 means about 9% more residents with dementia were exposed to an interaction; the narrow interval reflects the very large sample. The bigger practical difference is in CNS-active drugs, 29.6% against 21.2%, which is about one extra resident in twelve.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.