Audit and summative assessment: a completed audit cycle

Objective To develop and test a system for delivering a completed audit cyc le for summative assessment of general practitioner registrars in the Unite d Kingdom. Design A trainer-based questionnaire on criteria for a completed audit cycl e, followed by two marking exercises of audit projects submitted by general practice registrars. Setting Training practices in the West of Scotland between 1997 and 1998. S ubjects Trainers and registrars in the above practices. Results 116 (89%) agreed that two collections of data were an essential or desirable part of an audit project. All 57 registrars who started in August 1997 successfully completed an audit cycle, seven (12%) after resubmission . Using two rather than three independent assessors to screen che projects, the marking instrument was shown to have a sensitivity of 95% (95% confide nce interval (CI) +/-3.9%) and a specificity of 77% (95% CI +/-7.5%). All a ssessors found the new system easier to mark and 47 registrars (87%) found completing an audit cycle as or easier than expected. Conclusion Evidence from the pilot project has shown that a general practic e registrar's ability to review and critically analyse a piece of his/her w ork, with appropriate management of any necessary change, can be tested fea sibly by means of a completed audit cycle within the registrar year. The pr ocess retains adequate levels of sensitivity and specificity and requires f ewer assessors for marking the projects.