Medical records are important resources in evaluation of residents' clinical perfomance. Medical redcords may be audited by the resident himself (self audit). Self audit is more convenient and effective than peer review in modifing a doctor's clinical behavior. The purpose of this study is to evaluate the validity and reliability of structured self audit protocol for Family Medicine residents. Medical records of the hypertensive patients who visited Seoul National University Hospital Family Practice Center from May to August of 1991 were selected randomly and audited by both the residents themselves and fellows. Validity was evaluated by item by item comparison of audits by residents and fellows using rates calculated by kappa statistics. Reliability was evaluated by item by item comparison of selected charts between three groups of paired residents. The agreement rate of response of the residents and fellows was relatively high (agreement rate=77.1%, kappa=0.5874) and that of groups of paired residents was also high (agreement rate=77.1%, kappa=0.5870). We concluded that this self audit protocol has both good reliability and validity to evaluate medical records of residents in Family Medicine.
