Comparison of a Simplified Colonoscopy Bowel Preparation Scale With the Validated Boston Bowel Preparation Scale

Abstract

Introduction: A multi-society task force suggests adequate bowel cleansing permit detection of polyps >5mm. We validated a grading scale defining Adequate as enabling detection of polyps > 5mm AND no need to shorten the interval to the next colonoscopy. We aimed to assess intra- and inter-reader agreement between this 2 category scale and the Boston Bowel Preparation Scale (BBPS). Methods: 80 de-identified complete colonoscopies were videotaped. Readers were experienced board certified GI physicians; all attended a pre-study session to review representative videos. Grading was based on optimal cleansing on withdrawal. 8 readers reviewed 40 cases total; each video was read by 4 readers. Reader panels (comprised of a combination of 4 of the 8 readers assigned to each case) were randomized. Readers read assigned cases twice (T1, T2), 4 weeks apart. This study focused on 2 readers who read their 40 assigned cases a third time (T3) using the BBPS after completing on-line training. Agreement (%) and the kappa statistic quantified intra- and inter-reader agreement between readers. Results: For all 8 readers, there were 320 readings at T1 and T2. A rating of Adequate varied between readers for the 2-category scale (range 50%-98%; p = 0.001). Intra-reader agreement ranged from 63% to 98% (median=81%) and kappa from 0.25 to 0.88 (median 0.54). Overall inter-reader kappa= 0.35 for T1 and 0.26 for T2. For the 2 readers assessing T1, T2, and T3, there were 80 readings at each timepoint. An Adequate grade occurred significantly more with the BBPS (80-83%) than the 2-category scale (38-58%) (p-values range 0.001 to 0.004). Intra-reader agreement for the 2-category scale for T1 and T2 was k=0.25 (Reader 1) and k=0.42 (Reader 2). Intra-reader agreement between the 2-category scale and BBPS for Reader 1: T1/BBPS k=0.15, T2/BBPS k=0.34; Reader 2: T1/BBPS k=0.51, T2/BBPS k=0.17. For 18 cases common to both readers, adequacy ratings were also lower with the 2-category scale (44%-67%) compared to the BBPS (67%-83%) (p=NS). Inter-reader agreement for the 2 readers at T1 was k=0.14, T2 was k=0.44, and T3 was k=0.57. Conclusion: Bowel preparation is more likely to be rated Adequate using the BBPS compared to the 2-category scale. Intra- and inter-reader agreement for the 2-category scale was fair to moderate, and agreement between this scale and the BBPS varied greatly, ranging from slight to moderate. The 2-category scale may not be appropriate for clinical practice. Funding by Ferring Pharmaceuticals.