OBJECTIVE: The present study examined the test-retest reliability of team consensus best-estimate diagnoses of axis I and II disorders. METHOD: As part of a series of family studies of outpatients with depressive and personality disorders, best-estimate diagnoses of relatives were derived in team diagnostic conferences held regularly over 4 years. Diagnoses were based on all available information, including direct interviews, family history data, and treatment records, and explicit guidelines were developed to resolve discrepancies between data sources. To evaluate the reliability of the team best-estimate diagnoses, 92 relatives were independently rediagnosed after a 2-year interval. RESULTS: The reliability of both axis I and II disorders was good to excellent. The results were similar for cases in which diagnoses were based on direct interviews plus informant data and cases in which diagnoses were based on informant data alone. CONCLUSIONS: These data indicate that the team consensus best-estimate diagnostic method can be applied consistently, even over an interval of several years.Abstract Teaser