腹腔镜直肠手术经验可缩短机器人直肠切除术学习曲线:累积和分分析
本研究用累积和(CUSUM)图分析两位有腹腔镜直肠手术经验的外科医生从腹腔镜转向机器人全直肠系膜切除术(TME)的学习曲线。回顾性分析2006-2015年384例(294例腹腔镜,90例机器人)TME,按1:1倾向评分匹配后,以匹配的腹腔镜病例定义质量指标标准。结果显示一位医生无明显学习过程,另一位医生在约15例后所有质量指标达标。
要点
- 有经验的腹腔镜直肠外科医生转机器人TME学习曲线短,约15例后质量指标达标
- 匹配的腹腔镜病例可作为机器人手术学习曲线的个体化质控标准
- 并非所有有经验医生都需要明显学习过程,个体差异存在
核心结论
- 对于有经验的腹腔镜结直肠外科医生,机器人TME的正式学习过程可能较短,可达到与传统腹腔镜相似的表现水平。
研究设计与样本
- 回顾性病例系列分析,2006-2015年共384例(294例腹腔镜,90例机器人)直肠癌TME,来自两位外科医生。
- 每位医生的连续机器人TME病例按年龄、BMI、ASA分级、T分期和肿瘤位置高度与腹腔镜TME病例进行1:1倾向评分匹配。
- 以匹配的腹腔镜病例定义质量指标标准:手术时间、R分期、淋巴结清扫数目、住院时间和主要并发症(Clavien-Dindo 3-5级)。
主要发现
- 匹配后所有协变量的绝对标准化均数差均≤0.25。
- 外科医生A的CUSUM曲线与腹腔镜标准相比未见明显学习过程。
- 外科医生B在手术时间、淋巴结清扫数目和主要并发症方面显示学习过程,CUSUM曲线初始向上拐点;15例后所有质量指标基本达标。
关键数值与推荐
- 连续指标偏离均值超过四分之一标准差,或二分类指标超过观察风险,定义为脱靶,CUSUM曲线出现向上拐点。
- 外科医生B在15例机器人TME后所有质量指标基本达标。
临床怎么用
- 有丰富腹腔镜直肠手术经验的外科医生可较快安全过渡到机器人TME,学习曲线可能较短。
- 可利用匹配的腹腔镜病例作为个体化质控标准,通过CUSUM监测机器人手术质量指标。
局限性
- 回顾性设计,仅纳入两位外科医生,样本量有限,结论外推需谨慎。
- 学习曲线定义基于特定质量指标和统计方法,可能受病例选择和匹配质量影响。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: The learning curve for robotic colorectal surgery is ill-defined. This study aimed to investigate the learning curve of experienced laparoscopic rectal surgeons when starting with robotic total mesorectal excision (TME) using cumulative sum (CUSUM) charts. METHODS: This retrospective case series analysed patients who underwent curative and elective laparoscopic or robotic TMEs for rectal cancer performed by two surgeons. The first consecutive robotic TME cases of each surgeon were 1:1 propensity score matched to their laparoscopic TME cases using age, body mass index, American Society of Anesthesiologists grade, T stage (AJCC) and tumour location height. The matched laparoscopic cases defined individual standards for the quality indicators: operating time, R stage, lymph node harvest, length of hospital stay and major complications (Clavien-Dindo grade 3-5). Deviation of more than a quarter of a standard deviation from the mean for the continuous indicators, or exceeding the observed risk for the binary indicators was defined as off-target with an upward inflection in the CUSUM curve. RESULTS: From 2006 to 2015, 384 (294 laparoscopic; 90 robotic) TMEs met the inclusion criteria. Surgeon A performed 206 (70.1%) of the laparoscopic and 43 (47.8%) of the robotic cases. Surgeon B performed 88 (29.9%) of the laparoscopic and 47 (52.2%) of the robotic cases. After matching, no covariate exhibited an absolute standardised mean difference >0.25. For surgeon A, the CUSUM curves showed no apparent learning process compared to his laparoscopic standards. For surgeon B, a learning process for operation time, lymph node harvest and major complications was demonstrated by an initial upward inflection of the CUSUM curves; after 15 cases, all quality indicators were generally on target. CONCLUSIONS: For experienced laparoscopic colorectal surgeons, the formal learning process for robotic TME may be short to reach a similar performance level as obtained in conventional laparoscopy.
引用
- Odermatt M, Ahmed J, Panteleimonitis S, et al. Prior experience in laparoscopic rectal surgery can minimise the learning curve for robotic rectal resections: a cumulative sum analysis. Surg Endosc. 2017. PMID: 28271267. doi:10.1007/s00464-017-5453-9
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。