Background: The determination of proficiency in procedural skills has evoked keen interest, but meaningful guidelines are limited by the absence of pertinent clinical data. Colposcopic accuracy is defined as the clinical correlation between a colposcopic impression and a histologic report. The colposcopic accuracy of physicians in a university-based family practice residency colposcopy program was evaluated.
Methods: Demographic information, clinical findings, and laboratory results of patients evaluated by colposcopy were prospectively recorded. Colposcopic accuracy was calculated based on the agreement of the colposcopic impression with the histologic interpretation within one histologic grade.
Results: Colposcopic examinations were performed on 282 patients. Histologic evidence of premalignant cervical disease was identified in 115 patients as follows: mild dysplasia, 72; moderate dysplasia, 24; and severe dysplasia, 19. The colposcopic impression agreed within one histologic grade in 188 of 205 patients for a colposcopic accuracy rate of 91.7% (95% CI = 87.1% to 95.1%).
Conclusions: Given an acceptable latitude of clinical correlation between the colposcopic impression and histologic interpretation, the colposcopic accuracy of family physicians compared favorably with that reported by other colposcopists. The more common colposcopic errors were overestimation of low-grade disease and underestimation of high-grade disease. Colposcopic accuracy at an essential minimal proficiency level of 80% should form the basis for assessing specialized perceptual ability and therefore determining colposcopic competency.