K statistic as a measure of quality control in cervicovaginal cytology

Quality assessment schemes are widespread in most branches of pathology but are uncommon in the more subjective areas of histopathology and cytology. Researchers in many fields have become increasingly aware of the observer as an important source of measurement error. The validity of any method of reporting evidence of an abnormal process in cellular material is based on the degree of correlation with the actual disease process as it exists in the tissue and its reproducibility. Correlations can be tested in retrospective studies in which diagnoses based on cellular evidence are matched against the disease process present in biopsy specimens. Correlations can also be tested by examination of a set of unknown cellular preparations obtained in the presence of proven disease. While reproducibility is indirectly related to correlation, it is meant to imply satisfactory utilization of the method by other groups of cytotechnologists and cytopathologists. While cytopathology will continue to play an important role as a screening technique for the detection of cancer of the uterine cervix, its usefulness in the study of the early manifestations of the disease process is yet to be realized on a universal basis.