Factors predictive of length of stay for women in medium secure settings.

Attempts to understand determinates of length of stay in secure settings has been given increased impetus by minimum standard setting and payment by results initiatives. Factors predictive of length of stay in the extant literature include, index offence, previous engagement in therapy, symptomatology, personality pathology and need. These factors were assessed for their predictive validity in a sample of 70 consecutive admissions to a women's medium secure service. Patients were divided into short- and long-stay groups A loglinear analysis of intake variables indicated that the highest order interaction (duration of stay × diagnosis × index offence/previous engagement in therapy) was significant. Analysis of dynamic variables indicated that longer stays were associated with total symptomatology, paranoid pathological personality scale scores and Camberwell Assessment of Need Forensic Version need items of psychotic symptoms and psychological distress. Short-stay patients showed higher levels of therapy engagement. Findings are confirmatory of some of the key variables to be considered when estimating length of stay and when planning treatment.