Industry Buzz
There’s a very broad array of tools that you can do before reaching for these risky drugs.
—John N. Mafi, MD
New research published January 12, 2026, in JAMA lands on an uncomfortable truth many clinicians already recognize in practice: A substantial proportion of patients with dementia are still being prescribed medications that guidelines have long flagged as risky.[]
Using linked Medicare claims and Health and Retirement Study data, researchers estimate that about one in four traditional Medicare beneficiaries with dementia received at least one potentially inappropriate, brain-altering drug between 2013 and 2021.
The researchers looked at five broad categories of medications—including antidepressants, antipsychotics, and barbiturates—that are associated with sedation, confusion, falls, and, in some cases, increased mortality in older adults with cognitive impairment.
Notably, people with dementia were prescribed drugs included on the Beers Criteria—a list maintained by the American Geriatrics Society—more often than any other cognitive group.
These are “not trivial drugs,” John N. Mafi, MD, MPH, an internist at David Geffen School of Medicine at UCLA, told Washington Post, noting that some carry FDA black box warnings because they nearly double mortality risk in patients with dementia.[]
Why the stakes are so high
Age-related changes in pharmacokinetics and pharmacodynamics, polypharmacy, and altered neurotransmitter sensitivity all raise the stakes.
Drugs that modulate dopamine, serotonin, GABA, or acetylcholine can tip vulnerable patients into delirium, worsen gait instability, or accelerate functional decline.
The researchers stratified prescribing by cognitive status—normal cognition, mild impairment, and dementia—and found a clear gradient: These drugs were prescribed an average of 17% of the time to cognitively healthy older adults, compared with nearly 22% of the time across cognitively impaired individuals and 25% of the time for people with dementia. Even after adjusting for age, comorbidities, and socioeconomic factors, the pattern persisted.[]
“That’s a very high number,” Dr. Mafi said. []
Are these prescriptions always inappropriate?
As Annie Yang, MD, the study’s lead author, emphasized, there are legitimate indications for these medications in older adults, including schizophrenia, bipolar disorder, or other forms of psychosis.[] Behavioral and psychological symptoms of dementia—agitation, aggression, or combativeness—can also justify cautious, time-limited use when the risk-benefit balance favors treatment.
The study differentiated appropriate from potentially inappropriate use, finding stable prescribing when a clinical indication was documented and a modest decline when it was not. Even so, patients without a recorded indication still outnumbered those with one nearly two to one. Anupam Jena, MD, PhD, of Harvard Medical School cautioned that claims data often miss behavioral symptoms that influence real-world prescribing.[]
“That sort of presumes that the doctors don’t know that these drugs have risks,” he told The Washington Post, arguing that many clinicians are already exercising restraint and making nuanced trade-offs.[]
The guideline gap: nonpharmacologic care
Still, most professional guidelines converge on one clear point: Before medication, try nonpharmacologic interventions. This is where the study’s most important—and most sobering—message lands.
“There’s a very broad array of tools that you can do before reaching for these risky drugs,” Dr. Mafi said. But they often require resources that many people don’t have.[]
So, what are these tools?
Systematic assessment of unmet needs: Agitation in dementia often reflects pain, hunger, thirst, constipation, urinary retention, fear, or overstimulation. Treating the underlying trigger can reduce behaviors without a prescription.
Environmental and routine modifications: Consistent daily schedules, reduced noise, adequate lighting, and familiar caregivers can meaningfully decrease distress.
Behavioral strategies and caregiver coaching: Training caregivers to redirect, reassure, and de-escalate can be as effective as medications for many symptoms.
Meaningful activity and engagement: Boredom and isolation worsen behavioral symptoms. Structured activities tailored to a patient’s interests can help.
Addressing sleep, sensory, and mobility issues: Hearing aids, vision correction, pain management, and sleep hygiene often reduce nighttime agitation and daytime confusion
Why these tools are hard to implement
For clinicians, none of this is controversial. The challenge is execution. Nonpharmacologic care is time-intensive, often poorly reimbursed, and heavily dependent on caregiver availability, social support, and access to interdisciplinary teams.
Many practices lack embedded social workers, dementia care specialists, or behavioral health resources. Families may be overwhelmed, understaffed nursing facilities under-resourced, and community programs unevenly distributed or nonexistent.
In that context, prescribing a medication—despite its risks—can feel like the only actionable option in a 15-minute visit.
Related: 5 breakthroughs redefining dementia care in 2025Clinical context
For practicing physicians, the takeaway is not to eliminate these drugs altogether, but to pause before defaulting to them.
Each prescription should trigger a brief checklist:
Have we ruled out pain or infection?
Has constipation been addressed?
Has the caregiver been given concrete behavioral strategies?
Is there a way to mobilize community or interdisciplinary support?
Even incremental shifts in taking medication—shorter durations, clearer stop dates, regular reassessment—can reduce harm.
The study underscores a structural problem more than an individual failure. Clinicians largely know what the guidelines recommend; what they often lack are the resources, time, and system-level support to follow them consistently.
Until those gaps are addressed, the tension between best practice and real-world care will persist—and so will the reliance on medications we all wish we used less.