Université Grenoble Alpes, CNRS, Grenoble INP, TIMC-IMAG, 38000, Grenoble, France.
CIC 1406, Université Grenoble Alpes, 38000, Grenoble, France.
Int J Comput Assist Radiol Surg. 2018 Jan;13(1):95-103. doi: 10.1007/s11548-017-1645-y. Epub 2017 Aug 20.
Evaluation of surgical technical abilities is a major issue in minimally invasive surgery. Devices such as training benches offer specific scores to evaluate surgeons but cannot transfer in the operating room (OR). A contrario, several scores measure performance in the OR, but have not been evaluated on training benches. Our aim was to demonstrate that the GOALS score, which can effectively grade in the OR the abilities involved in laparoscopy, can be used for evaluation on a laparoscopic testbench (MISTELS). This could lead to training systems that can identify more precisely the skills that have been acquired or must still be worked on.
32 volunteers (surgeons, residents and medical students) performed the 5 tasks of the MISTELS training bench and were simultaneously video-recorded. Their performance was evaluated with the MISTELS score and with the GOALS score based on the review of the recording by two experienced, blinded laparoscopic surgeons. The concurrent validity of the GOALS score was assessed using Pearson and Spearman correlation coefficients with the MISTELS score. The construct validity of the GOALS score was assessed with k-means clustering and accuracy rates. Lastly, abilities explored by each MISTELS task were identified with multiple linear regression.
GOALS and MISTELS scores are strongly correlated (Pearson correlation coefficient = 0.85 and Spearman correlation coefficient = 0.82 for the overall score). The GOALS score proves to be valid for construction for the tasks of the training bench, with a better accuracy rate between groups of level after k-means clustering, when compared to the original MISTELS score (accuracy rates, respectively, 0.75 and 0.56).
GOALS score is well suited for the evaluation of the performance of surgeons of different levels during the completion of the tasks of the MISTELS training bench.
微创手术中，手术技术能力的评估是一个主要问题。训练平台等设备可以提供特定的分数来评估外科医生，但无法在手术室（OR）中转移。相反，有几个评分标准可以衡量 OR 中的表现，但尚未在训练平台上进行评估。我们的目的是证明 GOALS 评分可以有效地对腹腔镜手术中涉及的能力进行分级，也可以用于腹腔镜测试平台（MISTELS）的评估。这可能会导致培训系统能够更准确地识别已经获得或仍需改进的技能。
32 名志愿者（外科医生、住院医师和医学生）完成了 MISTELS 训练平台的 5 项任务，并同时进行视频记录。他们的表现由 MISTELS 评分和基于两名经验丰富、盲法腹腔镜外科医生对记录进行评估的 GOALS 评分进行评估。GOALS 评分的并行效度使用 Pearson 和 Spearman 相关系数与 MISTELS 评分进行评估。GOALS 评分的结构效度使用 K 均值聚类和准确率进行评估。最后，使用多元线性回归确定每个 MISTELS 任务探索的能力。
GOALS 和 MISTELS 评分具有很强的相关性（整体评分的 Pearson 相关系数为 0.85，Spearman 相关系数为 0.82）。GOALS 评分对于构建训练平台的任务是有效的，与原始 MISTELS 评分相比，在 K 均值聚类后组间具有更好的准确率（准确率分别为 0.75 和 0.56）。
GOALS 评分非常适合评估不同水平的外科医生在完成 MISTELS 训练平台任务时的表现。
