The assessment of dangerous and severe personality disorder: lessons from a randomised controlled trial linked to qualitative analysis

Abstract

Randomised controlled trials are difficult to carry out in high security prisons and very few have succeeded. We describe here a randomised controlled trial of early versus late assessment for the pilot phase of the new DSPD programme for dangerous and severe personality disorder, which assessed prisoners (n = 75) at baseline, then six months, and then one year after randomisation. The trial enjoyed 100% success in getting records and obtained useful qualitative data that helped to explain the findings, but the trial was compromised by repeated protocol violations on grounds that were seldom acknowledged openly but which we conclude were primarily due to ignorance of the purpose of such trials. This led to such contamination of the two arms of the trial that no clear conclusions could be drawn from the trial itself, except that relative costs showed expected differences. However, the trial also showed that the assessment programme was associated with better quality of life in terms of social relationships (p = .03), with an increase in aggression (p = .01), and with worse social functioning in those with less severe personality disorder (p < .01), with the qualitative data suggesting that frustration and unfulfilled expectations lay behind these findings. Suggestions are made about revisions to the assessment process and the changes necessary for successful trials to be mounted in the future.