Fifty-four psychiatric wards in Germany treating voluntary and involuntary patients were randomly allocated to either an intervention or to a waiting list condition. The intervention consisted of the implementation of three out of 12 suggested recommendations as selected by the ward teams, supported by external study workers. As the primary outcome measure, the number of coercive measures used per bed and month in the final 3 months of the intervention period was determined. Secondary outcomes were the cumulative duration of coercive measures used per bed and months and assaults per bed and month. Achieved guideline adherence was measured by a fidelity scale developed for this purpose during a pilot study for the PreVCo Rating Tool. After a 3-month baseline collection period under routine conditions, randomisation was done after matching wards pairwise according to frequency of coercive measures used and scores on the PreVCo Rating Tool at baseline.