Antipsychotics and Dementia: A Last Resort Still Prescribed to Many

Antipsychotic medicines are linked to higher rates of stroke, pneumonia and death in people with dementia, and guidelines say they should be a last resort. Yet in the countries with good data, anywhere from under one in ten to more than one in four people with dementia are prescribed them, and the COVID-19 pandemic pushed prescribing up in all six countries one international study examined^[1][2].

Distress, agitation, hallucinations and delusions can be among the hardest parts of dementia, both for the person living with it and for the people around them. Often that distress has a cause that can be found and eased: pain, an infection, a noisy ward, boredom, a care routine that does not fit the person. Where staff are stretched or specialist support is missing, a prescription can stand in for that work. The figures below show how often that happens, where rates have fallen and where they have not, and what the evidence says would help.

The harms are well documented: more pneumonia, strokes and deaths

Concern about antipsychotics in dementia rests on two decades of consistent evidence. A 2005 meta-analysis of 15 placebo-controlled trials of newer (“atypical”) antipsychotics, mostly lasting 10 to 12 weeks, found that 3.5% of people with dementia given the drugs died during the trials, compared with 2.3% given a placebo: an odds ratio of 1.54^[5]. The same year, the US Food and Drug Administration warned that atypical antipsychotic use in dementia was associated with increased mortality^[6], and similar safety regulations followed worldwide^[7].

In a study of 90,786 US veterans aged 65 and over with dementia, starting an antipsychotic was associated with an absolute increase in the risk of death over 180 days of between 2.0% (quetiapine) and 3.8% (haloperidol), compared with similar people not given one. Put another way, for every 26 people started on haloperidol, or every 27 on risperidone, there was one additional death^[8]. A UK trial that followed people with Alzheimer’s disease living in care facilities found that 46% of those randomised to continue their antipsychotic were alive at 24 months, against 71% of those switched to a placebo^[9].

The harms go beyond mortality. A 2024 study of 173,910 adults with dementia in English primary-care records found that in the 90 days after a prescription, antipsychotic use was associated with more than twice the rate of pneumonia, and higher rates of acute kidney injury, blood clots, stroke, fractures, heart attack and heart failure, than among matched non-users. In the first 90 days, 4.48% of new users developed pneumonia, against 1.49% of non-users^[10]. The range of harms was wider than regulatory alerts had highlighted, the authors noted.

England’s clinical guideline adds that prescribing antipsychotics for distress in dementia is an off-label use for almost all of these drugs, and that for people with dementia with Lewy bodies or Parkinson’s disease dementia, “antipsychotics can worsen the motor features of the condition”^[11].

From under one in ten to more than one in four: how many people with dementia take antipsychotics

There is no single international measure, so the best picture comes from national registers, claims data and large studies, each counting differently. The table gives the most recent figure we could verify for each country. Levels are not directly comparable: care-home figures run higher than figures for everyone with a diagnosis, and a year of prescribing counts more people than six weeks.

Two things stand out. First, rates are higher where people live in care homes. In the Japanese study, using institutional long-term care was associated with 2.34 times the odds of antipsychotic use^[12]. In Sweden, a review of nursing-home studies found use ranging from 6% to 38%^[17]. In Australia, the median length of a first course of antipsychotics for people with dementia in residential aged care was 555 days^[14]; England’s guideline, by comparison, asks for the need for the medicine to be reassessed at least every six weeks^[11]. Second, England’s figure of 8.6% hides a split: among people on the register with no recorded diagnosis of psychosis, 7.0% had an antipsychotic prescription; among the much smaller group with psychosis recorded, 54.0% did^[1]. Our study of dementia in care homes looks at life in residential care more widely.

Warnings and national targets brought rates down in some countries, but not in others

Where governments acted, prescribing often fell. Among US veterans with dementia, 17.7% were using an antipsychotic in 1999; use was already declining before the 2005 warning, and the warning was followed by a significant acceleration in the decline^[6]. In Denmark, national register data show use among people with dementia aged 65 and over falling from 31.3% in 2000 to 20.4% in 2012^[7]. Sweden’s increasingly restrictive national directives coincided with a falling trend in nursing homes^[17].

The United States set an explicit national target for nursing homes. The National Partnership to Improve Dementia Care, launched by the Centers for Medicare & Medicaid Services (CMS), cut the share of long-stay residents receiving antipsychotics from 23.8% at the end of 2011 to 19.8% within 21 months, a 17.1% reduction, and then set goals of a 25% reduction by the end of 2015 and a 30% reduction by the end of 2016, against the 2011 baseline^[4]. CMS data published in September 2026 put the national figure at 15.4%^[3]. In Australia, prolonged use (more than 90 days) among people with dementia in residential aged care edged down from 24.5% in 2011–12 to 21.8% in 2015–16^[14].

Elsewhere, warnings did not hold. In Spain, antipsychotic prescribing to people with newly diagnosed dementia in primary care almost doubled, from 12.6% in 2002 to 23.8% in 2018, driven by quetiapine, which rose from 0.2% to 16.4%; the authors found Spain’s prevalence was double the UK’s^[15]. In Taiwan, overall use among outpatients with dementia rose slightly, from 25.5% to 26.5% between 2005 and 2013^[16]. And among older Americans with dementia living at home, outside the reach of the nursing-home target, prescribing rose from 9.4% in 2010–12 to 16.0% in 2016–18, despite recommendations against first-line use for agitation^[13].

A lower antipsychotic rate does not always mean fewer psychotropic medicines

Targets measure one class of medicine, and others can take its place. The US Office of Inspector General found that about 80% of Medicare’s long-stay nursing-home residents were prescribed a psychotropic drug in each year from 2011 to 2019. While CMS focused on antipsychotics, use of anticonvulsants rose, and overall psychotropic use stayed constant^[18].

The same report raised a second concern. The CMS antipsychotic measure excludes residents reported to have schizophrenia, so a home that recorded more schizophrenia diagnoses would look better on the measure. Between 2015 and 2019, the number of residents reported with schizophrenia who lacked a matching diagnosis in Medicare claims and encounter data rose by 194%, and this unsupported reporting was concentrated in 99 nursing homes. The Inspector General presented the link to the measure as a risk, not a proven motive^[18].

Canada shows a similar pattern. In Ontario, a study of more than 70,000 residents each quarter between 2010 and 2019 found antipsychotic use falling by 0.70 percentage points a year, while antidepressant use rose by 0.89 points and anticonvulsant use by 1.06 points a year. Recorded delusions, one of the symptoms used to justify antipsychotic prescribing, rose from 3.5% to 10.2% of residents between 2014 and 2019, while recorded schizophrenia stayed stable^[19]. A follow-up study of 13,306 residents admitted on an antipsychotic found that those whose antipsychotic was stopped started the antidepressant trazodone at a rate 82% higher than those who continued^[20].

None of this means reduction efforts were wrong. It means a single prescribing rate is too narrow a measure of good care.

COVID-19 pushed prescribing up, and in many places it stayed up

In England, the share of people on the dementia register prescribed an antipsychotic had held steady at between 9.28% and 9.47% throughout 2018 and 2019. It rose to 9.99% in April 2020. Part of that rise reflected the deaths of many people with dementia early in the pandemic, which shrank the register by 3.60% between March and April 2020. The authors suggested most of the increase was probably a response to distress made worse by restrictions, such as “care-home residents confined to their bedrooms, cessation of communal activities and family visits”^[21].

A study using eight health databases in France, Germany, Italy, South Korea, the UK and the US, covering 857,238 people with dementia aged 65 and over, found that the rate of antipsychotic prescribing rose in all six countries. Compared with the same month in 2019, it was 2.11 times as high in May 2020 in one South Korean database and 1.96 times as high in June 2020 in the UK. Rates “did not decrease to prepandemic levels” after the acute phase had ended^[2]. In three Canadian provinces, people admitted to long-term care during the pandemic had higher odds of antipsychotic use than those admitted before it, and lower odds of having the medicine stopped at follow-up^[22].

The picture is not uniform. In Wales, linked records for 57,396 people with dementia showed only a slight increase, from 1,253 to 1,305 prescriptions per 10,000 person-months between March 2019 and September 2020^[23]. England’s most recent figure, 8.6% in June 2026, is lower than in 2020, though NHS England changed the format of its dementia data in 2026, so the two figures may not be strictly comparable^[1]. Our study of end-of-life care in dementia looks at care in the last months of life.

What the evidence asks of us

England’s NICE guideline says that, “as initial and ongoing management”, services should offer psychosocial and environmental interventions to reduce distress, after a structured assessment of possible causes such as pain or delirium. Antipsychotics should be offered only to people at risk of harming themselves or others, or experiencing agitation, hallucinations or delusions causing them severe distress; at the lowest effective dose, for the shortest possible time, with reassessment at least every six weeks; and stopped where there is no clear ongoing benefit. People should still have access to non-drug support while taking them and after they stop^[11]. The gap is not in knowing what good care looks like. It is in resourcing it.

The evidence points to five things services, commissioners, regulators and governments can do:

• Fund the alternative, not just the review. In a UK trial in 16 nursing homes, medication review halved antipsychotic use, but on its own it was followed by worse neuropsychiatric symptoms. Paired with a social interaction programme, it was associated with lower mortality^[24]. Cutting prescriptions without putting time and activity in their place shifts the problem.
• Expect implementation to be hard. A German trial that adapted a person-centred care programme for 37 nursing homes did not reduce antipsychotic prescribing^[25]. Programmes need sustained support and evaluation.
• Measure more than one drug class. Monitoring should track all psychotropic medicines and audit the diagnoses used to justify them, given the substitution and coding shifts seen in the US and Canada^[18][19].
• Track duration, not just prevalence. Australian researchers built a national indicator of use for more than 90 days for exactly this reason^[14].
• Publish comparable data. Shared definitions would let dementia organisations hold every system to the same standard.

Families and carers have a part in this too, and NICE asks that benefits and harms be discussed with them before an antipsychotic is started^[11]. That conversation works only if there is a real alternative on offer. More evidence is in our Dementia & Health library.

Language: this piece uses person-first language ("people living with dementia", "people with dementia"), as most dementia organisations and people with lived experience prefer, and uses clinical terms such as "Alzheimer's disease" or "mild cognitive impairment" only where a study, dataset or diagnosis refers to that specific condition.

Why we're publishing this

Antipsychotics have a place for some people at some times. But when they are used because there is no time, no staff or no specialist support to do anything else, people living with dementia carry the risk of a gap in care that is not of their making. We have gathered the international evidence in one place so that dementia charities, carers’ groups and campaigners can make the case for closing it with figures that are sourced, dated and clear about what they measure.

How to read this data

Different measures. Each country’s figure counts something different: everyone with a recorded diagnosis or only care-home residents; a prescription in the last six weeks, any use over a year, or use for more than 90 days. Compare trends within a source rather than levels between sources.

Derived figure. England’s 8.6% is our own calculation from NHS England’s June 2026 sub-ICB data file, adding people with and without a recorded psychosis diagnosis who had an antipsychotic prescription in the six weeks to 30 June 2026. NHS England changed its dementia data publication in 2026, so this may not match earlier monthly series exactly.

Study types. The harm figures come from a meta-analysis of short trials, a long-term withdrawal trial and large observational cohorts. Observational studies show associations; people given antipsychotics may differ from those who are not in ways the studies cannot fully adjust for, though the English cohort found no excess of an unrelated control outcome. Some national data (Denmark, Japan, Taiwan, Australia) are now several years old.

Gaps. We found little comparable recent data for low- and middle-income countries, and no international body publishes a standard indicator of antipsychotic use in dementia.

Use this data

Free to cite with attribution to Understanding Aging. Charities, carers' groups, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research and statistics for general information. It is not medical, care or financial advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's memory or health, please speak to a GP or another qualified health professional.

Sources

1. NHS England. Primary Care Dementia Data, June 2026 (release page, data dictionary and sub-ICB data file). NHS England Digital, 2026. digital.nhs.uk/…/primary-care-dementia-data/june-2026
2. Luo H, Lau WCY, Chai Y, et al. Rates of antipsychotic drug prescribing among people living with dementia during the COVID-19 pandemic. JAMA Psychiatry, 2023. europepmc.org/article/MED/36696128
3. Centers for Medicare & Medicaid Services. State US Averages (nursing home quality measures), September 2026. CMS Provider Data Catalog, 2026. data.cms.gov/provider-data/dataset/xcdc-v8bm
4. Centers for Medicare & Medicaid Services. Data show National Partnership to Improve Dementia Care exceeds goal to reduce use of antipsychotic medications in nursing homes (fact sheet). CMS, 2014. cms.gov/newsroom/fact-sheets/…
5. Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA, 2005. europepmc.org/article/MED/16234500
6. Kales HC, Zivin K, Kim HM, et al. Trends in antipsychotic use in dementia 1999–2007. Archives of General Psychiatry, 2011. europepmc.org/article/MED/21300946
7. Nørgaard A, Jensen-Dahm C, Gasse C, Hansen HV, Waldemar G. Time trends in antipsychotic drug use in patients with dementia: a nationwide study. Journal of Alzheimer’s Disease, 2016. europepmc.org/article/MED/26444790
8. Maust DT, Kim HM, Seyfried LS, et al. Antipsychotics, other psychotropics, and the risk of death in patients with dementia: number needed to harm. JAMA Psychiatry, 2015. europepmc.org/article/MED/25786075
9. Ballard C, Hanney ML, Theodoulou M, et al. The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. The Lancet Neurology, 2009. europepmc.org/article/MED/19138567
10. Mok PLH, Carr MJ, Guthrie B, et al. Multiple adverse outcomes associated with antipsychotic use in people with dementia: population based matched cohort study. BMJ, 2024. europepmc.org/article/MED/38631737
11. National Institute for Health and Care Excellence. Dementia: assessment, management and support for people living with dementia and their carers (NG97), Recommendations 1.7.1–1.7.9. NICE, 2018. nice.org.uk/guidance/ng97/chapter/Recommendations
12. Kuroda N, Hamada S, Sakata N, et al. Antipsychotic use and related factors among people with dementia aged 75 years or older in Japan: a comprehensive population-based estimation using medical and long-term care data. International Journal of Geriatric Psychiatry, 2019. europepmc.org/article/MED/30478985
13. Yang AW, Leng M, Arbanas JC, et al. Trends in antipsychotic prescribing among community-dwelling older adults with dementia, 2010–2018. Health Affairs Scholar, 2025. europepmc.org/article/MED/40040648
14. Sluggett JK, Caughey GE, Air T, et al. National surveillance using a clinical quality indicator for prolonged antipsychotic use among older Australians with dementia who access aged care services. International Journal of Geriatric Psychiatry, 2024. europepmc.org/article/MED/38676658
15. Sánchez-Soblechero A, Gil M, Rojo JM, et al. Trends of antipsychotic prescribing for people with dementia in Spain. Journal of Alzheimer’s Disease, 2025. europepmc.org/article/MED/40676897
16. Chiu YW, Huang CW, Chen PJ, et al. Trends and factors associated with antipsychotic use among elderly patients with dementia in Taiwan from 2005 to 2013: a population-based study. International Clinical Psychopharmacology, 2017. europepmc.org/article/MED/28542035
17. Karlsson S, Rahm Hallberg I, Midlöv P, Fagerström C. Trends in treatment with antipsychotic medication in relation to national directives, in people with dementia: a review of the Swedish context. BMC Psychiatry, 2017. europepmc.org/article/MED/28705243
18. US Department of Health and Human Services, Office of Inspector General. Long-term trends of psychotropic drug use in nursing homes (OEI-07-20-00500). HHS OIG, 2022. oig.hhs.gov/reports/all/2022/…
19. Harris DA, Maclagan LC, Iaboni A, et al. Potential unintended consequences of antipsychotic reduction in Ontario nursing homes. Journal of the American Medical Directors Association, 2022. europepmc.org/article/MED/35143749
20. Harris DA, Maclagan LC, Pequeno P, et al. Antipsychotic discontinuation and new trazodone use in Ontario nursing homes: evidence of medication substitution. Journal of the American Medical Directors Association, 2024. europepmc.org/article/MED/38944053
21. Howard R, Burns A, Schneider L. Antipsychotic prescribing to people with dementia during COVID-19. The Lancet Neurology, 2020. europepmc.org/article/PMC/PMC7577650
22. Turcotte LA, Heckman GA, McArthur C, et al. Antipsychotic medication use among newly admitted long-term care residents during the COVID-19 pandemic in Canada. Health Services Insights, 2024. europepmc.org/article/MED/39619866
23. Schnier C, McCarthy A, Morales DR, et al. Antipsychotic drug prescribing and mortality in people with dementia before and during the COVID-19 pandemic: a retrospective cohort study in Wales, UK. The Lancet Healthy Longevity, 2023. europepmc.org/article/MED/37543047
24. Ballard C, Orrell M, YongZhong S, et al. Impact of antipsychotic review and nonpharmacological intervention on antipsychotic use, neuropsychiatric symptoms, and mortality in people with dementia living in nursing homes: a factorial cluster-randomized controlled trial by the WHELD program. American Journal of Psychiatry, 2016. europepmc.org/article/MED/26585409
25. Richter C, Berg A, Langner H, et al. Effect of person-centred care on antipsychotic drug use in nursing homes (EPCentCare): a cluster-randomised controlled trial. Age and Ageing, 2019. europepmc.org/article/MED/30806453