Scholarlink

<p>Soft tissue surgical robots are leader-follower systems, with a learning curve for the surgeons. This is a key consideration when initiating a robotic service. This timely review synthesises the evidence. A systematic search captured studies in robotic colorectal surgery learning curves using cumulative summation (CUSUM) methodology. 39 studies were included (5684 patients). Most studies were single-surgeon (67%), single-centre (92%), using da Vinci robots (97%), with 30–662 patients. CUSUM was time-based (total operating time, console time), or ‘risk adjusted’ (RA-CUSUM) to cumulate surgical ‘success’ and ‘failure’. Most learning curves had 1 or 2 inflection points, implying 2 (49%) or 3 phases (46%), respectively. The first inflection occurred after 3-110 cases; the second inflection, if present, after a further 9–93 cases (12–137 cases overall). There was heterogeneity in prior surgical experience, patient characteristics, and case-mix. Some studies investigated learning curves for procedural components (docking time, operative steps). Across studies, study size correlated with cases to first (p &lt; 0.001) and second inflection (p &lt; 0.05), with line-of-best-fit slopes 0.08 and 0.13, respectively. Learning curves are highly variable in morphology and cases to inflection. Patient, surgeon, training, and hospital factors likely affect this. A novel finding is the significant correlation between study size and location of CUSUM inflection. This highlights a key methodological artefact rather than a purely skill-related phenomenon. Caution should therefore be exercised when interpreting inflection points as thresholds of proficiency or safety. This artefact and the heterogeneity in learning contexts limits generalisability and use in robotic service planning.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Robotic colorectal surgery: making sense of the learning curve

  • 期刊论文

  • 发表日期：

    2026年3月3日

查看全文
  • Atanu Pal,
  • Malaka Jayawardene

摘要

Soft tissue surgical robots are leader-follower systems, with a learning curve for the surgeons. This is a key consideration when initiating a robotic service. This timely review synthesises the evidence. A systematic search captured studies in robotic colorectal surgery learning curves using cumulative summation (CUSUM) methodology. 39 studies were included (5684 patients). Most studies were single-surgeon (67%), single-centre (92%), using da Vinci robots (97%), with 30–662 patients. CUSUM was time-based (total operating time, console time), or ‘risk adjusted’ (RA-CUSUM) to cumulate surgical ‘success’ and ‘failure’. Most learning curves had 1 or 2 inflection points, implying 2 (49%) or 3 phases (46%), respectively. The first inflection occurred after 3-110 cases; the second inflection, if present, after a further 9–93 cases (12–137 cases overall). There was heterogeneity in prior surgical experience, patient characteristics, and case-mix. Some studies investigated learning curves for procedural components (docking time, operative steps). Across studies, study size correlated with cases to first (p < 0.001) and second inflection (p < 0.05), with line-of-best-fit slopes 0.08 and 0.13, respectively. Learning curves are highly variable in morphology and cases to inflection. Patient, surgeon, training, and hospital factors likely affect this. A novel finding is the significant correlation between study size and location of CUSUM inflection. This highlights a key methodological artefact rather than a purely skill-related phenomenon. Caution should therefore be exercised when interpreting inflection points as thresholds of proficiency or safety. This artefact and the heterogeneity in learning contexts limits generalisability and use in robotic service planning.