What this page covers
The harms of antipsychotics in dementia, non-drug approaches to changed behaviours, and the law on chemical restraint are covered in detail in the dementia care section, linked at the foot of this page. This page focuses on what happens when these medicines are reviewed and reduced.
How much use varies
Use differs widely between homes. In 2019 data from 15,442 residents of 342 Australian aged care homes who had a medication review, fewer than 14% of residents received a potentially inappropriate antipsychotic, by the Beers criteria, in the lowest-using fifth of homes, and more than 31% in the highest-using fifth. For potentially inappropriate benzodiazepines the figures were under 19% and over 45%. After adjusting for the resident and home characteristics they could measure, the authors found residents in the highest-using homes had about seven times the odds of receiving such an antipsychotic, and about eight times the odds for a benzodiazepine 19.
Variation this wide suggests that local prescribing habits, as well as residents’ needs, play a part.
The evidence
Stopping a long-term antipsychotic when behaviour has settled
Supported by trial evidenceFor most residents with dementia on long-term antipsychotics, stopping did not worsen behaviour on average. People who had responded well to treatment for psychosis or severe agitation were more likely to relapse.
The evidence
A Cochrane review included 10 randomised trials with 632 participants, eight of them in nursing homes, in people with dementia who had taken an antipsychotic for at least three months. Stopping may make little or no difference to overall behavioural and psychological symptoms (7 trials, 519 participants), adverse events, quality of life or cognition, with low-quality evidence. In two trials limited to people with psychosis, agitation or aggression who had responded to treatment, stopping was followed by more relapse. Subgroup data suggested people with milder symptoms may become less agitated after stopping, and those with more severe symptoms may worsen. The effect on mortality could not be determined 16.
In practice
The reviewers noted that most behavioural symptoms of dementia come and go and often do not last longer than three months. The certainty of the evidence is low and the number of trials small 16.
Home-wide programs to reduce antipsychotic and benzodiazepine use
Promising, not yet provenTwo Australian programs combining education, review and audit reduced use without other sedatives being substituted. Neither was a randomised trial.
The evidence
In the HALT study, a single-arm study in 23 Sydney nursing homes, 139 residents on regular antipsychotics without a primary psychotic illness or severe symptoms took part in a deprescribing protocol, with education for GPs, pharmacists and nurses. Over 12 months the number on regular antipsychotics fell by 81.7%, with no rise in as-needed antipsychotics or benzodiazepines, and no change in behavioural symptoms, falls, hospital admissions or cognitive decline 17. RedUSe, a non-randomised program across 150 Australian homes with 12,157 residents, used audit and feedback, staff education and case review. Over six months the proportion of residents on antipsychotics fell from 21.6% to 18.9%, and on regular benzodiazepines from 22.2% to 17.6%, without substitution by sedating antidepressants or as-needed psychotropics 18.
In practice
HALT recruited a selected group and excluded residents with severe symptoms, so its result does not apply to everyone 17.
Chemical restraint: the consent rules in brief
Under the Aged Care Act 2024 and the Aged Care Rules 2025, chemical restraint means using a medicine or chemical substance to influence a person’s behaviour, rather than to treat a physical illness or condition or a mental disorder, or for end-of-life care. It is one of the regulated restrictive practices 20.
A restrictive practice may be used only as a last resort to prevent harm, in the least restrictive form and for the shortest time necessary, after best-practice alternatives have been considered or used and recorded in the person’s behaviour support plan, and with informed consent, including consent to how it is to be used. If the person cannot consent, a restrictive practices substitute decision-maker appointed under state or territory law may give consent 20.
For chemical restraint, the provider must be satisfied that a medical or nurse practitioner has assessed the person as posing a risk of harm, assessed that the restraint is necessary, prescribed the medicine for that use and obtained informed consent to prescribing it 20. These rules are stated as at October 2026 and do change.
The national Clinical Care Standard for psychotropic medicines in cognitive impairment points the same way in clinical terms: non-drug strategies first and alongside any medicine, informed consent documented before use, and regular review, with tapering and stopping considered 21.
Related pages on this site
- Medicines in dementia Dementia care section: antipsychotic harms, chemical restraint and the law, and a deprescribing sequence.
- Changed behaviours Dementia care section: understanding and responding to distress without medicines first.
- Restraint, bed rails and alarms Falls and frailty section: physical restraint and the restrictive practice rules.
Sources cited on this page
- 3 Kalisch Ellett LM, Kassie GM, Pratt NL, Kerr M, Roughead EE. Prevalence and duration of use of medicines recommended for short-term use in aged care facility residents. Pharmacy (Basel). 2019;7(2):55. (Department of Veterans’ Affairs claims data, 14,237 residents, 2015–2016; observational.) View source
- 16 Van Leeuwen E, Petrovic M, van Driel ML, et al. Withdrawal versus continuation of long-term antipsychotic drug use for behavioural and psychological symptoms in older people with dementia. Cochrane Database Syst Rev. 2018;3:CD007726. (10 randomised trials, 632 participants.) View source
- 17 Brodaty H, Aerts L, Harrison F, et al. Antipsychotic deprescription for older adults in long-term care: the HALT study. J Am Med Dir Assoc. 2018;19(7):592–600.e7. (Single-arm study, 23 Sydney nursing homes; not randomised.) View source
- 18 Westbury JL, Gee P, Ling T, et al. RedUSe: reducing antipsychotic and benzodiazepine prescribing in residential aged care facilities. Med J Aust. 2018;208(9):398–403. (Prospective non-randomised intervention, 150 homes.) View source
- 19 Sawan MJ, Clough AJ, Hillen J, et al. Variation in the potentially inappropriate use of antipsychotic and benzodiazepine medications among individuals receiving residential medication management reviews. Int J Pharm Pract. 2025;33(3):332–9. (15,442 residents in 342 Australian homes, 2019; cross-sectional.) View source
- 20 End of Life Law for Clinicians, Australian Centre for Health Law Research, Queensland University of Technology. Restrictive Practices in Aged Care (factsheet on the Aged Care Act 2024 and Aged Care Rules 2025). November 2025. View source
- 21 Australian Commission on Safety and Quality in Health Care. Psychotropic Medicines in Cognitive Disability or Impairment Clinical Care Standard. Sydney: ACSQHC. View source
See every source cited across the medicines and deprescribing section →
Written by Dr Josh Hatch, FRACGP. General information only, reflecting his reading of the published evidence as at October 2026. It is not individual medical advice, does not create a doctor–patient relationship, and must not be used to start, stop or change any treatment. Evidence and Australian regulatory, PBS and program arrangements change — always confirm current advice with the treating GP, pharmacist or specialist.
This is general education, not a clinical protocol. The doctors who practise with Umbrella Aged Care are independent general practitioners, and each makes their own clinical decisions with the resident and family. Your GP’s advice for an individual may differ from what is written here. If it does, please ask them to talk it through with you.