Proficiency testing in cytology laboratories in Ontario, Canada: a decade of experience. III. A precision study of consistency and reproducibility in cytology reporting.

The Ontario Laboratory Proficiency Testing Program's testing model and the results of early surveys of cytology laboratories have been previously described. To test consistency and accuracy, the same slides from one survey were recirculated to the same laboratory in the next survey. There was no statistical difference in either the accuracy or consistency in reporting by the various categories of personnel. Analysis of the data by category of diagnosis showed the reporting of "no abnormal cells" and "benign atypia" to be slightly more accurate and "moderate dysplasia" and "severe dysplasia" to be slightly less accurate than the reporting in other categories. Consistency in reporting appeared to be slightly less accurate in severe dysplasia than in the other categories. The analysis, which includes the results from all licensed cytology laboratories in Ontario, provides evidence that, in the reporting of cervical cytologic material, it is feasible to differentiate among the various categories within the spectrum of squamous epithelial abnormalities with a general level of accuracy and consistency that, although satisfactory, is not yet ideal.