The Antipsychotic Problem in Elderly Care

Antipsychotic medications prescribed to elderly patients — particularly those with dementia — carry serious risks including increased mortality. Despite FDA warnings dating back nearly two decades, hundreds of thousands of elderly Americans continue to receive these drugs, often without a psychiatric diagnosis that would justify their use. CMS specifically tracks this metric as a quality measure, and our risk model incorporates it as a key signal of potentially problematic prescribing.

The FDA Black Box Warning

In 2005, the FDA took the extraordinary step of requiring a black box warning — the most serious warning the agency can mandate — on all atypical antipsychotic medications when prescribed to elderly patients with dementia-related psychosis. The warning was based on a meta-analysis of 17 placebo-controlled trials involving drugs like olanzapine, aripiprazole, risperidone, and quetiapine. These trials showed that elderly dementia patients taking atypical antipsychotics had roughly 1.6 to 1.7 times the risk of death compared to those taking a placebo.

The causes of death varied — heart failure, sudden cardiac death, infections, particularly pneumonia — but the pattern was consistent across drugs in the class. In 2008, the FDA extended the black box warning to conventional (first-generation) antipsychotics like haloperidol, after observational studies suggested the mortality risk was at least as high, and potentially higher, with the older drugs.

It is important to understand what this warning means in practice: no antipsychotic medication is FDA-approved for the treatment of dementia-related behavioral symptoms. Every prescription of an antipsychotic for agitation, aggression, or psychosis in an elderly dementia patient is, by definition, off-label use. This does not make it automatically inappropriate — physicians prescribe off-label for many conditions — but it does mean that each prescription should reflect a careful weighing of risks against benefits, with informed consent from the patient or their surrogate decision-maker.

The Nursing Home Problem

The black box warning was issued in 2005, but antipsychotic use in nursing homes barely budged. By 2011, studies found that roughly one in four nursing home residents with dementia was receiving an antipsychotic medication. For most of these patients, there was no documented diagnosis of schizophrenia, bipolar disorder, or another condition for which antipsychotics are approved.

The problem is rooted in a difficult reality. Dementia patients can exhibit behaviors that are genuinely distressing and dangerous — hitting, screaming, wandering, resisting care. These behaviors are exhausting for caregivers and disruptive to other residents. Antipsychotics sedate patients effectively, and in understaffed facilities, the pressure to prescribe is enormous.

This practice has a name: chemical restraint. Rather than physical restraints — strapping a patient to a bed or wheelchair, which fell out of favor after reforms in the 1980s and 1990s — facilities turned to medications that achieved a similar outcome. A sedated patient does not wander, does not hit, does not scream. The patient is "managed."

Federal regulations under the Nursing Home Reform Act of 1987 (OBRA '87) explicitly prohibit the use of chemical restraints for the convenience of staff. Yet enforcement has historically been inconsistent. Surveyors may note antipsychotic use in facility records, but the line between "medically necessary" and "convenience prescribing" is often blurry in practice, and facilities have learned to document justifications that satisfy inspectors without necessarily reflecting clinical reality.

CMS and the National Partnership

In 2012, CMS launched the National Partnership to Improve Dementia Care in Nursing Homes, an ambitious initiative with a specific, measurable goal: reduce antipsychotic use in nursing homes. The Partnership brought together state survey agencies, nursing home operators, medical directors, and advocacy groups. It emphasized non-pharmacological interventions — music therapy, structured activities, individualized care plans, staff training in dementia care — as alternatives to medication.

The results were significant. Between 2012 and 2020, the national rate of antipsychotic use in long-stay nursing home residents declined by roughly 40 percent. CMS began publicly reporting facility-level antipsychotic use rates on its Nursing Home Compare website (now Care Compare), creating transparency and competitive pressure among facilities.

However, critics have raised concerns about how much of this decline represents genuine improvement versus diagnostic gaming. Some facilities began adding diagnoses of schizophrenia or bipolar disorder to patient records — conditions that exclude patients from the quality measure — without clear clinical justification. A 2021 HHS Office of Inspector General report found that the rate of schizophrenia diagnoses among newly admitted nursing home residents had increased suspiciously during the same period that antipsychotic quality measures were being tracked.

Despite these concerns, the National Partnership represents one of the most successful quality improvement initiatives in Medicare's history. The combination of transparency, measurement, and education did move the needle — even if the picture is more complicated than the headline numbers suggest.

What Our Data Shows

The Medicare Part D prescriber dataset includes a specific field — antipsychGE65Claims — that counts the number of antipsychotic prescriptions each provider wrote for beneficiaries aged 65 and older. This field allows us to identify providers with unusually high volumes of elderly antipsychotic prescribing relative to their peers in the same specialty.

Across the dataset, the distribution of this field is heavily skewed. Most providers have zero or very few antipsychotic claims for elderly patients. A small number of providers account for a disproportionate share of total prescriptions. This pattern is consistent with what we would expect: a small number of clinicians serving nursing homes or long-term care facilities where antipsychotic use is concentrated.

When a provider's antipsychotic prescribing to patients 65+ significantly exceeds the norm for their specialty, our model assigns risk points. The threshold is not a simple cutoff — it accounts for specialty type, total patient volume, and the statistical distribution of prescribing within that specialty. A psychiatrist treating patients in a state psychiatric hospital will naturally have more antipsychotic claims than a family medicine physician, and our model accounts for this.

Which Specialties Prescribe the Most

Antipsychotic prescribing to elderly patients is not evenly distributed across medical specialties. As one would expect, psychiatry accounts for the highest per-provider rates. Psychiatrists are the specialists most likely to treat schizophrenia, bipolar disorder, and treatment-resistant psychosis in elderly populations, and many of their prescriptions are clinically appropriate.

More concerning are elevated rates among internal medicine physicians, family medicine physicians, and nurse practitioners — particularly those who serve as attending providers in nursing homes and long-term care facilities. These providers may be under pressure from facility staff to prescribe antipsychotics for behavioral management, and they may lack the psychiatric training to fully evaluate whether the medication is appropriate or whether non-pharmacological alternatives have been adequately tried.

Nurse practitioners have become an increasingly significant presence in nursing home care, partly due to physician shortages in long-term care settings. While many NPs provide excellent care, the combination of autonomous prescribing authority, limited psychiatric training, and the high-pressure nursing home environment creates conditions where antipsychotic overprescribing can occur. Our data reflects this — NPs appear among flagged providers at rates that warrant attention.

The Human Cost

The statistics about increased mortality risk can feel abstract, so it is worth stating plainly what they mean. For every 100 elderly dementia patients treated with antipsychotics for a typical course of 10-12 weeks, the evidence suggests that roughly one to two additional patients will die compared to a group receiving a placebo. These are excess deaths — deaths that would not have occurred without the medication.

Beyond mortality, antipsychotics carry a range of serious side effects in elderly patients. They increase the risk of stroke and cerebrovascular events. They cause sedation that leads to falls, which in frail elderly patients can result in hip fractures and subsequent decline. They can worsen cognitive impairment, accelerating the very dementia they are sometimes prescribed to manage. They cause metabolic effects including weight gain and diabetes. And they can cause extrapyramidal symptoms — involuntary movements, rigidity, tremor — that reduce quality of life.

For patients who are already in the final stages of life, sedation may sometimes be compassionate. But for the many elderly patients who are prescribed antipsychotics chronically — months or years without reassessment — the cumulative harm is substantial. Studies suggest that many nursing home residents could have their antipsychotics safely tapered and discontinued, with behavioral symptoms managed through non-pharmacological approaches, yet inertia and convenience keep prescriptions in place.

Legitimate vs Concerning Use

Not every antipsychotic prescription for an elderly patient represents poor care. There are clear clinical scenarios where these medications are appropriate:

• Schizophrenia and bipolar disorder — These conditions do not disappear at age 65. Patients with lifelong psychiatric illness continue to need their medications, and discontinuing antipsychotics in a stable schizophrenia patient would be harmful.
• Severe agitation with psychosis — When a patient is experiencing hallucinations or delusions that cause significant distress or danger, and non-pharmacological interventions have been tried and failed, short-term antipsychotic use may be the most humane option.
• Delirium — In acute hospital settings, low-dose antipsychotics are sometimes used for hyperactive delirium, though evidence for this practice is mixed.
• Parkinson's disease psychosis — Pimavanserin (Nuplazid) is specifically approved for this indication, and quetiapine and clozapine are used off-label with reasonable evidence.

The prescriptions that raise concern are those that look like:

• Chemical restraints — Sedating patients primarily for the convenience of facility staff rather than for the patient's clinical benefit
• First-line dementia management — Prescribing antipsychotics before adequately trying behavioral interventions, environmental modifications, or caregiver training
• Chronic prescriptions without review — Continuing antipsychotics indefinitely without periodic reassessment and attempts at dose reduction or discontinuation
• Multiple antipsychotics simultaneously — Polypharmacy with two or more antipsychotics, which compounds risks without clear evidence of benefit

How We Flag This

Antipsychotic prescribing to patients 65+ is one of the components in our specialty-adjusted risk model. When a provider's volume of elderly antipsychotic claims is statistically elevated relative to their specialty peers, the model assigns risk points proportional to the degree of deviation.

This metric is not evaluated in isolation. A provider who appears elevated on antipsychotic elderly prescribing but is otherwise unremarkable will receive a modest risk score. A provider who is elevated on antipsychotic elderly prescribing and shows other concerning patterns — high opioid volume, unusual brand-name prescribing, geographic outlier status — will receive a higher composite score that reflects the convergence of multiple signals.

We emphasize that a flag does not indicate wrongdoing. Many flagged providers are psychiatrists or geriatric specialists serving high-acuity populations where antipsychotic use, while elevated compared to the average, is clinically justified. The flag is an invitation to look more closely — not a verdict.

Policy Implications

The persistence of antipsychotic overprescribing in elderly care, despite nearly two decades of warnings and a major national initiative, suggests that the problem is structural rather than merely educational. Providers know about the risks. The issue is that the incentives and conditions in long-term care settings continue to favor prescribing over alternatives.

Meaningful reform likely requires several approaches working in concert: adequate staffing ratios in nursing homes so that behavioral interventions are feasible, not just aspirational; robust enforcement of existing chemical restraint prohibitions by state survey agencies; prescriber education that goes beyond awareness and includes practical training in non-pharmacological dementia care; and transparency tools like OpenPrescriber that make prescribing patterns visible to families, regulators, and the public.

The data also points to a need for better diagnostic integrity. If quality measures are driving diagnostic gaming — adding schizophrenia diagnoses to avoid antipsychotic quality flags — then the measures need to be redesigned to account for this. CMS has begun exploring diagnosis-adjusted metrics, but progress has been slow.

What You Can Do

If you have a family member in a nursing home or long-term care facility, you have the right to ask questions about their medications. Specifically:

• Ask whether they are receiving antipsychotics — and if so, for what diagnosis. If the answer is "agitation" or "behavioral management" without a qualifying psychiatric diagnosis, ask whether non-pharmacological alternatives have been tried.
• Request a gradual dose reduction — Federal regulations require facilities to attempt dose reductions of antipsychotics unless clinically contraindicated. You can ask about this process and whether it has been attempted.
• Check the facility's quality ratings — CMS Care Compare reports antipsychotic use rates for every Medicare-certified nursing home. Facilities with rates significantly above the national average deserve scrutiny.
• Use OpenPrescriber — Search for the prescribing provider on our platform to see how their antipsychotic prescribing compares to their peers. Context matters, but data is a starting point for informed conversations.