The history of Broadmoor involves the management of mentally abnormal offenders, or ‘criminal lunatics’ as they were called, over the last 180 years (Partridge 1953; Walker and McCabe 1973; Allderidge 1977). Before 1800 there were no special facilities for such patients, and they were incarcerated in local prisons where, John Howard wrote, the conditions were ‘crowded and offensive because the rooms which were designed for prisoners are occupied by the insane. Where these are not kept separate they disturb and terrify other prisoners. No care is taken of them although it is probable that by medicines and proper regimen some of them might be restored to their senses and usefulness in life’.

Little information is available about how long patients stay in the Special Hospitals. This paper describes what is believed to be the first systematic study of this topic.

Many activities undertaken by the forensic psychiatrist have nothing to do with dynamic psychotherapy, and rightly so. Likewise, the main thrust of dynamic psychotherapy, is independent of forensic issues. Nevertheless, there is a preserve of clinical responsibility within forensic psychiatry which, on occasion, justly invokes the aid of dynamic psychotherapy. This may be when the scope of the patient's treatment, already initiated by the forensic psychiatrist, needs to be enlarged to include dynamic psychotherapy, or when assessment of intrapsychic processes calls for prolonged engagement with the patient at depth, within the proximity provided in a therapeutic setting. For example, the complex task of disentangling amnesia or pseudo-amnesia from repression is one which no one experienced in either field would regard as easy.

Psychological assessment at Broadmoor in the 20 years lifetime of the Psychology Department has been directed at the description of patterns of response indicative of abnormalities associated with diagnostic and offender categories and suggestive of treatment needs. The results have also frequently been accumulated for research purposes, particularly to identify any common patterns associated with treatment outcome and with discharge prediction and outcome.