Interobserver and intraobserver reliability of the proposed International Continence Society, Society of Gynecologic Surgeons, and American Urogynecologic Society pelvic organ prolapse classification system

Abstract

Objective Our purpose was to determine the intraobserver and interobserver reliability of site-specific measurements and stages with the proposed international Continence Society, Society of Gynecologic Surgeons, and American Urogynecologic Society 1994 draft prolapse terminology document.Study design Women who completed informed consent procedures underwent pelvic examinations by two investigators, each blinded to the results of the other's examination. The reproducibility of the nine site-specific measurements and the summary stage and substage were analyzed with Spearman's correlation coefficient (rs) and Kendel tau B Correlation Coefficient (tau b), respectively. Similar analyses were performed on supine and upright examinations performed at two different times by one examiner.Results Experienced examiners averaged 2.05 minutes per examination and new examiners averaged 3.73 minutes. In the study of interobserver reliability, 48, subjects, mean age 61 +/- 14 years, parity 3 +/- 2, weight 74 +/- 31 kg, comprised the study population. Correlations for each of the nine measurements were substantial and highly significant (rs 0.817, 0.895, 0.522, 0.767, 0.746, 0.747, 0.913, 0.514, and 0.488, p = 0.0008 to < 0.0001). Staging and substaging were highly reproducible (tau b 0.702 and 0.652). In no subject did the stage vary by more than one; in 69% stages were identical. In the study of intraobserver reliability, for 25 subjects correlations for each of the nine measurements were equally strong (rs 0.780, 0.934, 0.765, 0.759, 0.859, 0.826, 0.812, 0.659, 0.431). Measurements from the upright examinations reflected greater prolapse. Staging and substaging were highly reproducible (tau b 0.712 and 0.712). In no subject did the stage vary by more than one; in 64% stages were identical. All stage discrepancies represented an increase in the upright position.Conclusions There is good reproducibility of measures with the proposed system. The data suggest that the reliability is independent of examiner experience. Patient position is likely important in maximizing the severity of the prolapse.