Objective To examine variation in reported quality and working conditions of pre-registration house officer (PRHO) posts. To use multilevel modelling to determine how much of the variation was due to the effects of consultant firms, hospitals, trusts and deaneries, as well as variation at the level of the individual doctor. Design Questionnaire survey of national sample of PRHOs at the end of the pre-registration year. Multi-level modelling was carried out using MLwiN. Setting PRHOs working in UK approved medical and surgical posts in NHS hospitals from August 1996 to January 1999. Participants One thousand, four-hundred and thirty-five PRHOs who had previously been part of an extended cohort study of medical student selection and training. They reported on a total of 4926 posts, on 2721 identifiable consultant firms, in 336 hospitals, in 264 trusts and in 17 deaneries. Main outcome measures PRHOs' perceptions of the overall quality of all PRHO posts they had worked in, and working conditions in the current post. Results Twenty percent of PRHO posts were described as excellent and 34% as very good, through to 6% reported as not very good, poor or bad. The overall rating of the post showed highly significant variation at the level of hospital/trust, and consultant firm. Other aspects of posts, such as working conditions showed variation at the level of deanery, trust, hospital and firms, with different patterns for the different measures. Assessments of stress in PRHOs (GHQ, burnout, response to uncertainty), and a desire to leave medicine, showed no variation at the level of deanery, trust, hospital or firm. Conclusion Some aspects of a PRHO post, in particular overall rated quality, but also many aspects of working conditions, show substantial variation at the level of hospital., trust and consultant firm, suggesting that differences reflect local variation in working practices and treatment of PRHOs, with the possibility of change and improvement. The lack of any variation at the level of deanery, trust, hospital or firm in stress of PRHOs suggests that these responses are idiosyncratic, individual responses by doctors themselves, rather than a general characteristic of posts, hospitals, trusts or deaneries.
