BACKGROUND: To assess the phenomenon of second medical opinion in general internal medicine we performed a retrospective audit and descriptive analysis in a teaching hospital. METHODS: Referrals to the out-patient clinic of the University Hospital Groningen were screened using the hit 'second opinion' in the automated search program of the electronic hospital information system (ZIS). This involved the screening of the records of all patients who had visited the clinic in the period between March 1998 and July 2000. In this period 117 female and 84 male patients had visited the out-patient clinic for a second medical opinion. The case records of all these patients were analyzed for patient characteristics, chief complaint, investigations performed, referring doctor, reason(s) for referral and, if applicable, established diagnosis. Subsequently, we determined which investigations had been performed when formulating a second opinion as well as the number of times the second opinion was beneficial in terms of new diagnoses or of changes in the management plan. RESULTS: The majority (85.5%) of the patients was referred for a diagnostic problem. The chief complaints were fatigue (34.3%) and abdominal pain (26.8%). Investigations were frequently repeated at the time of the consult. A new diagnosis was established in about 10% of cases. Meanwhile, in cases of second opinions regarding further management, beneficial advice was provided in 70% of cases. CONCLUSION: A second opinion in internal medicine is most often asked for diagnostic purposes in patients with functional complaints. Despite the repetition of investigations performed, often in only a limited number of patients, a relevant new diagnosis can be established.