Division of Gastroenterology, St. Michael's Hospital, University of Toronto, Toronto, Ontario, Canada.
Division of Gastroenterology, St. Michael's Hospital, University of Toronto, Toronto, Ontario, Canada; Department of Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada.
Gastrointest Endosc. 2018 Mar;87(3):766-775. doi: 10.1016/j.gie.2017.08.020. Epub 2017 Aug 30.
Colonoscopy performance is typically assessed by a supervisor in the clinical setting. There are limitations of this approach, however, because it allows for rater bias and increases supervisor workload demand during the procedure. Video-based assessment of recorded procedures has been proposed as a complementary means by which to assess colonoscopy performance. This study sought to investigate the reliability, validity, and feasibility of video-based assessments of competence in performing colonoscopy compared with live assessment.
Novice (<50 previous colonoscopies), intermediate (50-500), and experienced (>1000) endoscopists from 5 hospitals participated. Two views of each colonoscopy were videotaped: an endoscopic (intraluminal) view and a recording of the endoscopist's hand movements. Recorded procedures were independently assessed by 2 blinded experts using the Gastrointestinal Endoscopy Competency Assessment Tool (GiECAT), a validated procedure-specific assessment tool comprising a global rating scale (GRS) and checklist (CL). Live ratings were conducted by a non-blinded expert endoscopist. Outcomes included agreement between live and blinded video-based ratings of clinical colonoscopies, intra-rater reliability, inter-rater reliability and discriminative validity of video-based assessments, and perceived ease of assessment.
Forty endoscopists participated (20 novices, 10 intermediates, and 10 experienced). There was good agreement between the live and video-based ratings (total, intra-class correlation [ICC] = 0.847; GRS, ICC = 0.868; CL, ICC = 0.749). Intra-rater reliability was excellent (total, ICC = 0.99; GRS, ICC = 0.99; CL, ICC = 0.98). Inter-rater reliability between the 2 blinded video-based raters was high (total, ICC = 0.91; GRS, ICC = 0.918; CL, ICC = 0.862). GiECAT total, GRS, and CL scores differed significantly among novice, intermediate, and experienced endoscopists (P < .001). Video-based assessments were perceived as "fairly easy," although live assessments were rated as significantly easier (P < .001).
Video-based assessments of colonoscopy procedures using the GiECAT have strong evidence of reliability and validity. In addition, assessments using videos were feasible, although live assessments were easier.
结肠镜检查的性能通常由临床环境中的主管进行评估。然而，这种方法存在局限性，因为它允许评分者偏见，并增加了主管在手术过程中的工作量需求。已经提出了基于视频的记录程序评估作为评估结肠镜检查性能的补充手段。本研究旨在调查基于视频的评估与现场评估相比，对结肠镜检查能力的可靠性、有效性和可行性。
来自 5 家医院的新手（<50 次结肠镜检查）、中级（50-500 次）和经验丰富（>1000 次）内镜医师参与了研究。每个结肠镜检查都拍摄了两个视图：内镜（腔内）视图和记录内镜医师手部运动的录像。两名盲法专家使用胃肠道内镜检查能力评估工具（GiECAT）独立评估记录的程序，该工具是一种经过验证的特定于程序的评估工具，包括全球评分量表（GRS）和检查表（CL）。现场评分由非盲法专家内镜医师进行。结果包括临床结肠镜检查的现场和基于视频的评分之间的一致性、内部评分者的可靠性、外部评分者的可靠性和基于视频的评估的区分效度，以及评估的便利性。
共有 40 名内镜医师参与了研究（20 名新手、10 名中级和 10 名经验丰富的内镜医师）。现场和基于视频的评分之间具有良好的一致性（总评分，组内相关系数 [ICC] = 0.847；GRS，ICC = 0.868；CL，ICC = 0.749）。内部评分者的可靠性非常好（总评分，ICC = 0.99；GRS，ICC = 0.99；CL，ICC = 0.98）。两名盲法视频评分者之间的外部评分者的可靠性较高（总评分，ICC = 0.91；GRS，ICC = 0.918；CL，ICC = 0.862）。GiECAT 总评分、GRS 和 CL 评分在新手、中级和经验丰富的内镜医师之间差异显著（P <.001）。尽管现场评估被评为更容易，但基于视频的评估被认为“相当容易”（P <.001）。
使用 GiECAT 对结肠镜检查程序进行基于视频的评估具有可靠和有效的证据。此外，视频评估是可行的，尽管现场评估更容易。
