Sir: The development of separate secure services for women (Hassell & Bartlett, Psychiatric Bulletin, September 2001, 25, 340-342) is essential. However, in my experience the campaign to develop such services is brushing aside the limitations of both our phenomenological understanding of a particular group of patients, and the evidence base for interventions for this group. These patients, who I anticipate will be instantly recognisable to clinicians, present with many features of borderline personality disorder and some “are subject to episodic descent into psychosis”. I do not want to revisit the failings of the personality disorder/mental illness dichotomy. However, even when our current pharmacopoeia is exhausted, many of these patients continue to manifest evidence of severe psychological vulnerabilities, which have been present from an early age, and for which longterm psychological interventions are recommended. Forensic psychiatrists have become increasingly cautious about admitting male patients who present with problems primarily attributable to their personality. Disproportionately more women than men with a primary diagnosis of personality disorder are admitted to secure services (Reference Coid, Kahtan and GaultCoid et al, 2000). Does this gender bias reflect a particular therapeutic optimism or medical paternalism?