As a physician and medical director in the post-acute and long-term care setting, I encounter dementia-related behaviors almost every day. Agitation, aggression, or resistance to care are rarely “just behaviors” — they’re often the patient communicating discomfort, fear, or confusion in the only way they can. When a nurse calls me about a resident striking out during bathing or becoming restless at night, my first instinct isn’t to reach for a prescription pad. Instead, I pause and ask: What has changed for this person? I look at recent labs, review pain control, check for constipation, infection, or dyspnea, and ask whether their usual routines — music, lighting, or sleep patterns — have been disrupted. Many times, small adjustments like restoring a familiar evening routine, dimming harsh lights, or breaking hygiene tasks into smaller steps can calm a patient without medication.

That said, there are times when non-drug approaches simply aren’t enough, and the behavior creates suffering or a real safety risk for staff or the patient. In those situations, I believe medications are appropriate — but they must be chosen thoughtfully and documented clearly. I try to avoid benzodiazepines whenever possible because of their risks for falls, delirium, and paradoxical agitation. If medication is needed, I typically consider an atypical antipsychotic such as low-dose quetiapine, risperidone, or olanzapine, or an SSRI if anxiety or mood changes are driving the behavior. Mood stabilizers such as valproate (Depakote) can be helpful in specific situations — particularly when the patient’s dementia is accompanied by marked impulsivity, disinhibition, or mood lability that does not respond well to antipsychotics alone. I’ve found them occasionally useful in frontotemporal dementia (FTD), where behaviors are often driven by irritability and poor impulse control rather than classic psychosis, and in some mixed vascular dementias where agitation seems tied to emotional reactivity rather than hallucinations or delusions. When I choose valproate, I document the behavioral target (e.g., “refractory mood lability with verbal outbursts in FTD”), monitor LFTs and platelets, and note the rationale for avoiding antipsychotics if poorly tolerated or ineffective.

Regardless of the medication I use, I ensure that the chart reflects a clear and thoughtful decision-making process. For example: “Resident with advanced Alzheimer’s has severe late-day agitation and striking during care. Non-drug measures (music, calm lighting, pre-care positioning) were tried without relief. After risk/benefit discussion with family, will initiate low-dose quetiapine 12.5 mg PO at bedtime for comfort and safety; will monitor sedation, falls, and metabolic effects.” Or: “Patient with frontotemporal dementia and severe impulsivity unresponsive to low-dose quetiapine; initiating valproate 125 mg BID, will monitor liver function and platelets.” It would also be beneficial to get the Depakote level if the patient is on higher doses or there is concern for sedation - or back off on the dose as you would with any sedating medication. This kind of documentation and approach reassures surveyors that our prescribing is symptom-driven, risk/benefit balanced, and carefully monitored.

Over time, I’ve learned that an individualized, stepwise approach — ruling out reversible causes, minimizing unnecessary medications, and tailoring interventions to the patient’s history and comfort — not only meets regulatory expectations but also feels right for the patient and family. Our goal in dementia care is always dignity and quality of life. Medications play a crucial role when safety or suffering is at stake, but they should be administered following careful evaluation, clear documentation, and ongoing monitoring to ensure we’re truly helping rather than sedating.