BackgroundUp to a quarter of people in the UK with a diagnosis of dementia are prescribed an antipsychotic in any year. The potential risks of such treatment are becoming clearer, but the benefits remain uncertain. Concern about the frequency and quality of such prescribing was expressed in the National Dementia Strategy for England in 2009.AimsTo provide an estimate of the prevalence of antipsychotic use for dementia in secondary mental health services in the UK and to collect data relevant to quality improvement initiatives for such prescribing practice.MethodIn the context of a UK quality improvement programme, relevant clinical audit data were collected for patients with dementia under the care of specialist older people's mental health services.ResultsFifty-four mental health National Health Service (NHS) trusts submitted data on 10 199 patients. Of those patients without comorbid psychotic illness, 1620 (16%) were prescribed an antipsychotic; the common clinical indications for such medication were agitation, psychotic symptoms, aggression and distress. Multivariable regression found younger age, care home or in-patient setting, vascular or Parkinson's disease dementia and greater severity of dementia to be all significantly associated with being prescribed antipsychotic medication. Of the 1001 (62%) patients prescribed treatment for more than 6 months, only three-quarters had a documented review of therapeutic response in the previous 6 months.ConclusionsThe data reveal areas of relatively good current practice, including consideration of alternatives to antipsychotic medication and clear documentation of target symptoms. They also suggest areas for improvement, such as the frequency and quality of review of long-term medication. Strategies to reduce antipsychotic use should take account of the demographic and clinical variables predicting increased likelihood of antipsychotic prescription.