Implementation of guidelines on prevention of coercion and violence: baseline data of the randomized controlled PreVCo study.

Introduction: The PreVCo study examines whether a structured, operationalized implementation of guidelines to prevent coercion actually leads to fewer coercive measures on psychiatric wards. It is known from the literature that rates of coercive measures differ greatly between hospitals within a country. Studies on that topic also showed large Hawthorne effects. Therefore, it is important to collect valid baseline data for the comparison of similar wards and controlling for observer effects. Methods: Fifty five psychiatric wards in Germany treating voluntary and involuntary patients were randomly allocated to an intervention or a waiting list condition in matched pairs. As part of the randomized controlled trial, they completed a baseline survey. We collected data on admissions, occupied beds, involuntarily admitted cases, main diagnoses, the number and duration of coercive measures, assaults and staffing levels. We applied the PreVCo Rating Tool for each ward. The PreVCo Rating Tool is a fidelity rating, measuring the degree of implementation of 12 guideline-linked recommendations on Likert scales with a range of 0-135 points covering the main elements of the guidelines. Aggregated data on the ward level is provided, with no patient data provided. We performed a Wilcoxon signed-rank-test to compare intervention group and waiting list control group at baseline and to assess the success of randomization. Results: The participating wards had an average of 19.9% involuntarily admitted cases and a median 19 coercive measures per month (1 coercive measure per occupied bed, 0.5 per admission). The intervention group and waiting list group were not significantly different in these measurements. There were 6.0 assaults per month on average (0.3 assaults per occupied bed and 0.1 per admission). The PreVCo Rating Tool for guideline fidelity varied between 28 and 106 points. The percentage of involuntarily admitted cases showed a correlation with coercive measures per month and bed (Spearman's Rho = 0.56, Discussion: Our findings that coercion varies widely within a country and mainly is associated with involuntarily admitted and aggressive patients are in line with the international literature. We believe that we included a sample that covers the scope of mental health care practice in Germany well.