Research finds persistent prescribing of risky medications to older adults with dementia

Despite years of clinical guidelines warning against the practice, one in four Medicare beneficiaries with dementia is prescribed brain-altering medications linked to falls, confusion, and hospitalization, according to new research to be published January 12 in the peer-reviewed journal JAMA.

While prescriptions for these medications fell from 20% to 16% over the nine-year study period among all Medicare beneficiaries, they continue to be prescribed to individuals with cognitive impairment who are particularly susceptible to these ill effects.

While this decline was encouraging, over two-thirds of patients receiving these prescriptions lacked a documented clinical indication in 2021, the end of the study period, suggesting high levels of potentially inappropriate and harmful prescribing. Compared with patients with normal cognition, we also found higher levels of prescribing among older adults with cognitive impairment, who face a higher risk of adverse effects from these drugs. These results underscore substantial opportunities to improve the quality and safety of care for millions of older Americans."

Dr. John N. Mafi, senior author, associate professor-in-residence of medicine, division of general internal medicine and health services research, David Geffen School of Medicine at UCLA

For this study, the researchers used survey data from the Health and Retirement Study linked to Medicare fee-for-services claims to trace central nervous system (CNS)-active prescribing patterns between January 1, 2013 through December 31, 2021 of possibly inappropriate CNS-active medications given to older adults with 1) normal cognition, 2) cognitive impairment without dementia, and 3) dementia.

The CNS-active medications studied included five drug classes: antidepressants with strong anticholinergic properties, antipsychotics, barbiturates, benzodiazepines, and non-benzodiazepine hypnotics.

While CNS medications were prescribed for 17% of older adults with normal cognition, nearly 22% of those with cognitive impairment but without dementia were given the drugs. About 25% of the group with dementia were prescribed the CNS-active medications.

Among all Medicare fee-for-service beneficiaries, prescription trends by medication class were as follows:

• Benzodiazepines declined by 11.4% to 9.1%
• Nonbenzodiazepine hypnotics, or sleep drugs, fell from 7.4% to 2.9%

• Antipsychotic medication prescriptions rose from 2.6% to 3.6%
• Prescriptions for anticholinergic antidepressants remained at 2.6% through the study period
• Barbiturate prescriptions fell slightly from 0.4% to 0.3%

Furthermore, clinically justified prescriptions fell from 6% in 2013 to 5.5% in 2021, while likely inappropriately prescribed CNS-active medications saw a significant decline from 15.7% to 11.4%.

The improvement was driven largely by reductions in prescriptions for benzodiazepines and sleep medications as well as in inappropriate prescriptions.

The findings are limited by unavailable Medicare Advantage data, possibly missing clinical information such as agitation, and a focus on prescribing prevalence rather than cumulative exposure.

"While CNS-active prescriptions may be appropriate in some cases, it is important for older patients or their caregivers to work closely with their physicians to ensure that these medications are appropriate to their cases. When inappropriate, patients and their care teams should consider alternative treatments and consider whether it might be safe to taper or stop the medication," said Dr. Annie Yang, a scholar in the National Clinician Scholars Program at Yale University who led this study as a UCLA internal medicine resident.

Study co-authors are Mei Leng, Dr. Dan Ly, Chi-Hong Tseng, Dr. Catherine Sarkisian, and Nina Harawa of UCLA; Cheryl Damberg of RAND, and Dr. A. Mark Fendrick of University of Michigan. Ly and Sarkisian are also affiliated with VA Greater Los Angeles Healthcare System.

The National Institutes of Health/National Institute on Aging (R01AG070017-01) funded the study.