Sir: Not surprisingly, I welcome the paper by Harvey Rees (Psychiatric Bulletin, April 2000, 24, 124-126).

Cognitive—analytic therapy (CAT) was always intended to offer a NHS-relevant model of psychological therapy and management. Rees comments on the need for a “robust evidence base for its effectiveness”, in this respect. It should be noted that the development of the model over the past 25 years has involved both conceptual developments and numerous small scale studies of both process and outcome. These are recorded in three books and over 50 papers published in peer-reviewed journals and a number of further papers are due to appear in a special section of the June issue of the British Journal of Medical Psychology. Much of this work has been focused on borderline personality disorder. Bowing to current definitions of ‘robust’, and despite considerable ethical and design problems, we have now embarked on a large scale randomised controlled trial of 24-session CAT in this category of patients. Despite favourable referees' reports and a completed feasibility study, this inexpensive study of a group of currently neglected patients who have a very low spontaneous recovery rate, a high suicide rate and are high consumers of resources has failed to attract research and development funding.