We compared the diagnoses from intraoperative frozen-section consultation with the final diagnosis using permanent tissue sections from 34 hospitals throughout North America. Participating pathologists who provided data from their practices volunteered for a pilot study of Q-Probes, a modular quality assurance program of the College of American Pathologists, Skokie, III. During 4 weeks, 186 pathologists evaluated 1952 frozen-section cases and deferred diagnoses to the final permanent sections in 77 cases (3.9%). Concordance between frozen-section and the final histologic diagnoses was 96.5%; whereas the mean and median concordance rates for participating institutions were 96.8% and 97.4%, respectively. Of 67 discordant diagnoses, 29 occurred from sampling of nonrepresentative tissue specimens, and 29 from diagnostic misinterpretations. The pathologists indicated that the diagnostic discordance had little or no effect on patient care in 94% of cases. We conclude that the North American pathologists studied interpret frozen sections with exceedingly high accuracy, approximating that reported from teaching hospitals.