Backgroundantipsychotic drugs are regularly prescribed as first-line treatment for neuropsychiatric symptoms in persons with dementia although guidelines clearly prioritise non-pharmacological interventions.Objectivewe investigated a person-centred care approach, which has been successfully evaluated in nursing homes in the UK, and adapted it to German conditions.Designa 2-armed 12-month cluster-randomised controlled trial.Settingnursing homes in East, North and West Germany.Methodsall prescribing physicians from both study arms received medication reviews for individual patients and were offered access to 2 h of continuing medical education. Nursing homes in the intervention group received educational interventions on person-centred care and a continuous supervision programme. Primary outcome: proportion of residents receiving at least one antipsychotic prescription after 12 months of follow-up. Secondary outcomes: quality of life, agitated behaviour, falls and fall-related medical attention, a health economics evaluation and a process evaluation.Resultsthe study was conducted in 37 nursing homes with n = 1,153 residents (intervention group: n = 493; control group: n = 660). The proportion of residents with at least one antipsychotic medication changed after 12 months from 44.6% to 44.8% in the intervention group and from 39.8 to 33.3% in the control group. After 12 months, the difference in the prevalence was 11.4% between the intervention and control groups (95% confidence interval: 0.9–21.9; P = 0.033); odds ratio: 1.621 (95% confidence interval: 1.038–2.532).Conclusionsthe implementation of a proven person-centred care approach adapted to national conditions did not reduce antipsychotic prescriptions in German nursing homes.Trial registrationClinicalTrials.gov NCT02295462.