Methodological issues in measuring coercion

Abstract

Background Despite the widespread and controversial use of coercion in psychiatry, little is known about the impact of coercion on outcome. One reason for this lack of knowledge is problems related to measuring coercion in a valid way. Objective: To explore predictors of perceived coercion and compare different measures of coercion. Methods 928 patients, aged 18–60, admitted to acute psychiatric wards in the Nordic countries were interviewed focusing on their experiences during admission. We used a questionnaire where both the MacArthurPerceived Coercion Scale (MPCS) and "The Coercion Ladder" (CL; a ten step visual analogue scale) were included. Information from medical records was obtained and participants were assessed by GAF and BPRS. Results MPCS and CL scores showed strikingly similar patterns between the countries, both regarding score distribution and levels of perceived coercion. The correlation between MPLS- and CL scores was fairly good (Pearson's R 0,648 and Kappa 0,404), but the MPCS seemed to be more sensitive to low impact coercion. Being exposed to "process exclusion" and "negative pressures" predicted high levels of perceived coercion, as did the patients' opinion on their own legal status. These predictors explained the same proportion of variance in the perceived coercion scores (approximately 60%). GAF, BPRS scores and the formal legal status had little impact on perceived coercion. Conclusion MPCS and CL seem to measure the same phenomena and scores are predicted by the same variables. The somewhat surprising finding that perceived coercion is not correlated to legal status (or as shown in other studies the use of coercive measures) raises questions about the validity of the measures used for coercion in clinical studies.