Variation in the assessment of adequacy in cervical smears

Cervical smear adequacy: cellularity references were found to increase both interobserver agreement and unsatisfactory rate

OBJECTIVES

To determine the degree of interobserver variation in the assessment of conventional cervical smear adequacy as defined by The Bethesda System (TBS) 2001, and to determine the effect of using reference images of known squamous cellularity when performing squamous adequacy assessments.

METHODS

Experimental pre-test/post-test design utilizing 70 conventionally prepared cervical smears. Sample smears containing scant squamous cellularity were independently rated on two occasions by six cytotechnologists. Time 1 was without the use of reference images, and Time 2 was aided by cellularity reference images. The kappa statistic was used to compare rater agreement.

RESULTS

The level of agreement increased from an average kappa of 0.26 (SD 0.10) for Time 1, to an average kappa of 0.40 (SD 0.15) for Time 2. The difference in mean kappa values at the two assessments was statistically significant (t = 3.71; P = 0.002). Unanimous agreement among the raters was observed for 15 samples (21.42%) at Time 1 (only one of which was classified as unsatisfactory) and 21 samples (30.00%) at Time 2 (12 of which were classified as unsatisfactory).

CONCLUSION

Interobserver agreement increased after cellularity reference images were implemented. Using TBS 2001 squamous adequacy criteria and images of known squamous cellularity as references resulted in a decreased number of smears reported as satisfactory.