Article type
Year
Abstract
Introduction/Objective: To assess the agreement between appropriateness criteria for the use of colonoscopy developed according to the RAND-UCLA method and evidence from published studies.
Methods: In 1994, a national multidisciplinary panel of 9 experts used the RAND-UCLA method to develop criteria for the appropriateness of performing colonoscopy. Based on a comprehensive review of the literature, a list of 402 possible clinical indications for colonoscopy was developed. The panellists were provided with the literature review as a basis for evaluating the degree of appropriateness of each indication. These evaluations were consolidated into 3 categories: appropriate, uncertain and inappropriate. A similar categorisation was made for indications retrieved from 18 studies (8 RCTs, 1 uncontrolled trial, 9 observational studies) summarised in the literature review. Proportion of agreement and kappa coefficients were calculated for both theoretical indications and those encountered in 577 primary care patients referred for colonoscopy.
Results: The 8 RCTs covered only 22 / 402 indications. A 68% agreement was found between appropriateness criteria and criteria derived from RCTs' (kappa=0.33). A similar agreement (67% agreement, kappa=0 41) was found for the 63 indications which could be compared with the 18 relevant studies. The agreement between appropriateness criteria and those derived from RCTs' was higher when based on the 577 real cases: 12 theoretical and 51 encountered indications; 75% agreement, kappa=0 47.
Discussion: Appropriateness of only a few indications for colonoscopy could be based on evidence found in the published literature, which underscores the limitation of an approach based on 'evidence based appropriateness criteria' only. However, the agreement between appropriateness criteria developed with the RAND-UCLA method and evidence from RCTs, and other studies, was moderate. This lack of agreement could indicate that panellists paid insufficient attention to published evidence when rating appropriateness of indications. On the other hand, additional elements could have influenced their judgements, such as a perceived low effectiveness in spite of proven efficacy. Further improvements in the development of appropriateness criteria and practice guidelines derived from those should better incorporate published evidence and complementary judgements from multi-disciplinary expert panels.
Methods: In 1994, a national multidisciplinary panel of 9 experts used the RAND-UCLA method to develop criteria for the appropriateness of performing colonoscopy. Based on a comprehensive review of the literature, a list of 402 possible clinical indications for colonoscopy was developed. The panellists were provided with the literature review as a basis for evaluating the degree of appropriateness of each indication. These evaluations were consolidated into 3 categories: appropriate, uncertain and inappropriate. A similar categorisation was made for indications retrieved from 18 studies (8 RCTs, 1 uncontrolled trial, 9 observational studies) summarised in the literature review. Proportion of agreement and kappa coefficients were calculated for both theoretical indications and those encountered in 577 primary care patients referred for colonoscopy.
Results: The 8 RCTs covered only 22 / 402 indications. A 68% agreement was found between appropriateness criteria and criteria derived from RCTs' (kappa=0.33). A similar agreement (67% agreement, kappa=0 41) was found for the 63 indications which could be compared with the 18 relevant studies. The agreement between appropriateness criteria and those derived from RCTs' was higher when based on the 577 real cases: 12 theoretical and 51 encountered indications; 75% agreement, kappa=0 47.
Discussion: Appropriateness of only a few indications for colonoscopy could be based on evidence found in the published literature, which underscores the limitation of an approach based on 'evidence based appropriateness criteria' only. However, the agreement between appropriateness criteria developed with the RAND-UCLA method and evidence from RCTs, and other studies, was moderate. This lack of agreement could indicate that panellists paid insufficient attention to published evidence when rating appropriateness of indications. On the other hand, additional elements could have influenced their judgements, such as a perceived low effectiveness in spite of proven efficacy. Further improvements in the development of appropriateness criteria and practice guidelines derived from those should better incorporate published evidence and complementary judgements from multi-disciplinary expert panels.