Endometrial biopsy dating. Interobserver variation and its impact on clinical practice.

Abstract

To determine the degree of interobserver variation and the level of correlation in endometrial biopsy (EB) interpretation, five pathologists interpreted 30 EBs randomly selected from all EBs performed according to standard criteria. An EB was considered out of phase (OOP) if there was greater than a two-day lag in histologic development relative to the onset of the next menstrual period. Interobserver variation among the five pathologists was 0.85 and 0.82 days for one- and two-day dating ranges, respectively. Coefficients of correlation for histologic interpretation ranged from 0.839 to 0.934 (P < .0001). Seventy percent to 74% of the time all five pathologists agreed that a given EB was in phase or OOP. The low interobserver variation and high correlation between observers further validates the EB as a diagnostic test. However, up to 30% of the time a clinical management decision may be altered depending upon who interprets the EB. Thus, even with low interobserver variation, the clinical consequence of small changes in EB interpretation may be significant.