Colposcopists' agreement on cervical biopsy site

To determine the inter-observer agreement among colposcopists on the most abnormal area of the cervix from which a biopsy would be obtained and whether any attributes predict agreement.

Fifty cervigrams were reviewed and 72 colposcopists from five countries indicated the site to biopsy and whether an ECC should be obtained. Prior to the study, six Canadian colposcopists met to achieve consensus on the most diseased area for biopsy. Consensus was also reached on whether an ECC was indicated. For each cervigram, percent agreement was determined between each study colposcopist and the consensus. Data were analyzed to determine the attributes associated with the consensus response.

The percent overall agreement of the colposcopists with the consensus diagnoses had a mean of 0.70 (95% CI, 0.65-0.75). The use of ECC was most common in Canada (15% of cases). The following factors were assessed by multivariate analysis to determine their influence on individual agreement with the consensus recommendation for the site to biopsy: country, duration of practice (less than or greater than 1 year), professional group (nurse, family doctor, pathologist, gynecologist, gynecologic oncologist), expert status (recognized national/international expert vs colposcopist), and gender. No factor was significantly associated.

This international study was feasible and the level of inter-observer agreement among colposcopists on the location of the most severe lesions in cervical images is good.