[Excerpt from authors]Abstract: The objective was to explore the relationship between staff related variables and patient outcome in day treatment programmes for patients with personality disorders. The importance of staff size, skill mix and treatment intensity (hours of treatment per week) was examined, in addition to location-specific effects. Multi-centre data routinely collected under non-experimental conditions from nine units, all members of a cooperative network in Norway, were analysed using a multilevel analysis. The data set consisted of treatment unit characteristics for the period 1993-2005, constituting an unbalanced panel of 71 units, together with information from 1574 patients who completed day treatment according to the plan. Patient outcome was measured by change in Global Assessment of Functioning Scale (GAF). Twelve per cent of variation in patient outcome was attributed to the treatment unit level. Staff size and treatment intensity influenced outcome to a minor extent, while an increased proportion of nurses or other college-educated personnel was associated with improved patient outcome. A positive location-specific effect was found in one unit attached to a university. Potential cost savings seem to be apparent with respect to staff size and, to some extent, skill mix.