When Agitation Is Pain: Chemical Restraint in Dementia Care

An older woman with dementia who pushes the aide away and cries out when she is turned is usually charted as agitated. The response is often a sedative, and when a drug is given to quiet behavior rather than to treat a medical symptom, federal nursing-home rules have a name for it: a chemical restraint. Chapter 16 of The Pained Brain, the video above, by Dr. Gurpreet Singh Padda, MD, MBA, MHP, and Dr. KrisJay Fucanan, MD, asks the question the medication list rarely asks. Does she hurt?

How much pain goes uncounted in nursing homes

The familiar up-to-80-percent figure comes from a 1999 review, and its catch is the lesson. Studies that asked residents or read their charts found pain in 49 to 83 percent. Studies that counted only who received a painkiller found 27 to 44 percent.

Modern federal data show where the undercount lives. Across 8.6 million admissions, pain was documented in 68.9 percent of residents with little or no cognitive impairment and 32.9 percent of those with severe impairment. When the two methods were compared, self-report recorded moderate to severe pain in 9.60 percent of residents and staff observation in 34.04 percent. Advanced dementia does not halve pain. It halves reporting, and the chart keeps only what is reported.

Signs of pain in dementia patients that get read as behavior

Pain that cannot be described surfaces as guarding, resisting care and calling out at night. The clearest test of that link is a Norwegian trial across 60 nursing-home units and 352 residents with dementia and agitation. An eight-week stepwise pain protocol cut agitation by 17 percent, and 69 percent of treated residents needed nothing stronger than acetaminophen. When the analgesics were withdrawn, the agitation drifted back.

When scheduled acetaminophen went to 95 residents selected for low quality of life rather than for pain, it did nothing. Treating pain works; treating everyone as if they hurt does not. Untreated pain also derails thinking. After a hip fracture, delirium was nine times as likely in clear-thinking patients with severe pain, and 5.4 times as likely in those receiving under 10 milligrams of morphine equivalents daily.

What a chemical restraint costs

In surveyor guidance, a chemical restraint is a drug given for discipline or for staff convenience when no medical symptom requires it, and a sedating drug given for a symptom nobody diagnosed fits. Among 90,786 veterans with dementia, one extra death within 180 days occurred for every 26 people given haloperidol and every 50 given quetiapine. In a randomized withdrawal trial, 24-month survival was 46 percent among patients who continued their antipsychotic against 71 percent among those switched to placebo.

The system measured one drug class, and the sedation moved. Antipsychotic use among long-stay residents has fallen 40.6 percent since late 2011, but the decline began before the national campaign, and mood-stabilizer use climbed to 20.1 percent of residents, mostly gabapentin. In 2021, 85.0 percent of gabapentin starts in nursing homes carried no approved indication. That is the economic and regulatory driver in plain view: a quality metric rewards a lower number on one line, thin staffing rewards a quiet floor, and neither pays for the time it takes to examine a resident who cannot talk.

Polypharmacy in the elderly and a body that clears drugs slowly

The aging liver clears most drugs 10 to 40 percent more slowly, kidney filtration falls 0.5 to 1.0 milliliters per minute each year, and body fat rises as muscle falls, so fat-soluble drugs accumulate. The aging brain is also more sensitive at any blood level. Polypharmacy of five or more prescriptions rose from 23.5 percent to 44.1 percent of older Americans over two decades.

Pain drugs belong on that list. In a matched Medicare arthritis cohort, starting an opioid instead of an anti-inflammatory carried 4.47 times the fracture risk. Gabapentinoids raised hip fracture odds 1.30-fold after adjustment, although a multinational self-controlled study found fracture risk highest in the 90 days before the drug started, which points at the pain rather than the pill. None of this is a reason to stop a medication alone. It is a reason to ask the prescriber for a review that begins with a blood pressure taken standing. One medication list, one person, one fall walks through that conversation.

What was never diagnosed

The fracture often sits on a scan already taken for something else. When two radiologists re-read 1,481 routine chest and abdominal CT scans, 11.5 percent of patients had a vertebral fracture, and only 14.7 percent of those fractures reached the report’s impression. On abdominal CT in patients over 60, radiologists reported 30 percent of compression fractures. Why a cough can break an osteoporotic spine explains the mechanism.

The rest of the terrain is read just as loosely. Sarcopenia raised the odds of chronic pain 1.52-fold. In Swedish nursing homes, 82 percent of residents were vitamin D deficient, though whether deficiency causes pain is inconclusive, and megadoses carry their own fall and fracture risk. Among Americans over 65, 29.2 percent have diabetes and 48.8 percent prediabetes. Nursing-home days ran at a median of 54 lux, with 10.5 minutes above the 1,000 lux that helps set the body clock, and pain control runs on that clock. The old body also signals quietly: 12.6 percent of heart attacks arrived without chest pain, in patients who were older and more often women. Less reported pain is not evidence of less disease.

The procedure is not the aggressive option

The frailty objection holds for elective surgery: among 1,284 people aged 90 or older given a knee replacement, 18 percent had a major complication against 5.2 percent in their sixties. It does not hold for image-guided procedures. Across 16,638 consecutive fluoroscopically guided epidurals there were no immediate major adverse events. In patients over 65 receiving an epidural, grip strength predicted a good response, an odds ratio of 1.142 per kilogram, while psoas muscle size did not. For a broken hip, nerve blocks across 49 trials cut pain on movement by 2.5 points within thirty minutes, prevented one case of delirium for every 12 patients treated, and got patients out of bed 10.8 hours sooner.

Humility belongs here. Vertebroplasty is contested: one sham-controlled trial of fractures under six weeks old found 44 percent of treated patients below a pain score of 4 at two weeks against 21 percent, while the Cochrane pooling found 0.7 points of benefit at one month, below the threshold that matters. Deprescribing alone does not reliably prevent falls in the community. And stopping drugs cuts both ways: among older adults, rapid opioid tapering lowered overdose but raised all-cause mortality 1.28-fold, while escalation raised it 1.51-fold. Stewardship is harm reduction, the lowest dose that preserves function, not a race to zero. Why an abrupt taper is also harm makes that case in full.

What families can ask for

If a parent with dementia has become agitated, ask whether pain was assessed by observation rather than by a question she cannot answer. Ask whether any recent scan, even one for a cough, was reviewed for a fracture. Ask for a standing blood pressure, a medication review, a vitamin D level and a hemoglobin A1c, and ask what each sedating drug is treating. Then ask what would replace the sedative: scheduled acetaminophen, a topical anti-inflammatory, a targeted procedure where a pain generator is found, and daylight plus movement, because the body clock needs light and muscle needs load. Work every medication change through the prescriber. What repeated steroid injections cost is the prior argument, and why a dose is not a diagnosis comes next. The studies behind every number are in the Technical Supplement to Chapter 16, with what each one does and does not show.

Frequently asked questions

What is a chemical restraint in a nursing home?

Federal rules give nursing-home residents the right to be free from chemical restraints, meaning a drug given for discipline or staff convenience rather than for a medical symptom that requires it. A drug with a sedating effect that is not treating a diagnosed symptom counts. Antipsychotics used for agitation in dementia carry a boxed mortality warning, which makes the diagnosis behind the prescription essential. Why stacked sedating drugs get weaned is explained here.

How can you tell if someone with dementia is in pain?

Watch behavior rather than waiting for words: wincing or guarding when moved, resisting care, crying out during turning, new restlessness at night, or pulling back from walking and meals. Observational pain scales built for dementia exist, yet in one national dataset only about a quarter of residents with severe cognitive impairment were assessed that way. In hospital, 57 percent of people with dementia showed pain on movement at least once. How chronic pain relates to dementia risk is covered here.

Why is polypharmacy dangerous for older adults?

Aging slows liver and kidney clearance, so drugs build up, and the brain grows more sensitive to sedating effects. Adults over 65 visit emergency departments for medication harm at 12.1 per 1,000 people, against 5.0 in younger adults. Opioids, gabapentinoids, sedatives and blood-pressure drugs each add fall or fracture risk, and the risk climbs when they are stacked. A review with the prescriber is the safe route. Antidepressants showed limited benefit for chronic pain in older adults.

Can a compression fracture cause agitation in someone with dementia?

It can. A person who cannot describe back pain may show it by resisting care, pushing staff away or calling out when turned. Vertebral fractures are often missed: on routine CT scans read for other reasons, only 14.7 percent of fractures reached the report’s impression. Treating pain in agitated residents with dementia reduced agitation by 17 percent in a randomized trial, so a fracture search belongs in the workup. What vertebroplasty involves and who it helps is explained here.

Are pain procedures safe for people over 80?

Image-guided procedures and elective surgery carry very different risks. Across 16,638 fluoroscopically guided epidurals there were no immediate major adverse events, and hip fracture nerve blocks prevented one case of delirium for every 12 patients treated. Knee replacement in people aged 90 or older carried far higher complication rates. In adults over 65, grip strength predicted epidural response better than age. Leg pain when walking that eases when leaning forward is often spinal stenosis.

Sources

1. Fox, P. L., Raina, P., & Jadad, A. R. (1999). Prevalence and treatment of pain in older adults in nursing homes and other long-term care institutions: a systematic review. CMAJ, 160(3), 329–333. https://pubmed.ncbi.nlm.nih.gov/10065074/
2. Dube, C. E., Morrison, R. A., Mack, D. S., Jesdale, B. M., Nunes, A. P., Liu, S.-H., & Lapane, K. L. (2020). Prevalence of Pain on Admission by Level of Cognitive Impairment in Nursing Homes. Journal of Pain Research, 13, 2663–2672. https://doi.org/10.2147/JPR.S270689
3. Husebo, B. S., Ballard, C., Sandvik, R., Nilsen, O. B., & Aarsland, D. (2011). Efficacy of treating pain to reduce behavioural disturbances in residents of nursing homes with dementia: cluster randomised clinical trial. BMJ, 343, d4065. https://doi.org/10.1136/bmj.d4065
4. Morrison, R. S., Magaziner, J., Gilbert, M., Koval, K. J., McLaughlin, M. A., Orosz, G., Strauss, E., & Siu, A. L. (2003). Relationship between pain and opioid analgesics on the development of delirium following hip fracture. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 58(1), 76–81. https://doi.org/10.1093/gerona/58.1.m76
5. Maust, D. T., Kim, H. M., Seyfried, L. S., Chiang, C., Kavanagh, J., Schneider, L. S., & Kales, H. C. (2015). Antipsychotics, other psychotropics, and the risk of death in patients with dementia: number needed to harm. JAMA Psychiatry, 72(5), 438–445. https://doi.org/10.1001/jamapsychiatry.2014.3018
6. Ballard, C., Hanney, M. L., Theodoulou, M., Douglas, S., McShane, R., Kossakowski, K., Gill, R., Juszczak, E., Yu, L. M., Jacoby, R., & DART-AD investigators (2009). The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. The Lancet Neurology, 8(2), 151–157. https://doi.org/10.1016/S1474-4422(08)70295-3
7. Gerlach, L. B., Zhang, L., Montoya, A., Krein, S. L., Zullo, A. R., Bynum, J. P. W., & Maust, D. (2026). Off-Label Initiation of Gabapentin and Valproic Acid Among Long-Stay Nursing Home Residents With and Without Dementia. Journal of the American Geriatrics Society. e70574 (online ahead of print, 3 July 2026). https://doi.org/10.1111/jgs.70574
8. Peter, D. H., Jakob, S., Jasminka, I., Nikolaus, S., Regina, R., & Michael, F. (2026). Recognition and reporting of vertebral fractures in thoracic and abdominal computed tomography images – have we learnt our lesson? Osteoporos Int. online ahead of print, 2026-08-31. https://doi.org/10.1007/s00198-026-08204-1
9. Guay, J., & Kopp, S. (2020). Peripheral nerve blocks for hip fractures in adults. Cochrane Database of Systematic Reviews, 11(11), CD001159. https://doi.org/10.1002/14651858.CD001159.pub3
10. Yang, Y., Prajapati, P., Ramachandran, S., Bhattacharya, K., Bazzazzadehgan, S., Maharjan, S., Eriator, I., & Bentley, J. P. (2025). Opioid Tapering and Opioid Overdose, Opioid Use Disorder, and Mortality Among Older Adults: A Nested Case-Control Study. Journal of General Internal Medicine, 40(12), 2896–2905. https://doi.org/10.1007/s11606-025-09492-9

Medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP, board certified in anesthesiology, pain medicine, interventional pain management, addiction medicine and obesity medicine. Adapted from Chapter 16 of The Pained Brain by Dr. Padda and Dr. KrisJay Fucanan, MD. Last reviewed September 2026.

Dr. Gurpreet Singh Padda, MD, MBA, MHP