Grading a developmental continuum--elegy on the rise and fall of the endometrial biopsy

Abstract

Due to the original study design the frequency distribution of the slide categories sent to each of the three expert pathologists was essentially bimodal. The intraclass correlation coefficient for the 1-14 continuous categories reveals excellent agreement (0.82), whereas the kappa statistic for the dichotomized outcomes (>2 day lag) is poor (0.54). The discrepancy between the intraclass correlation coefficient and kappa values seems to be due to the bimodal distribution of the slide ratings. Some impartial mechanism was needed to make an unbiased selection of the slides used for endometrial dating. Imprecision in endometrial dating is due primarily to the limitations of the instrument and not the observer. Maturation stage of the endometrium is not critical for predicting a successful pregnancy.