Scholarlink

Purpose <p><?tk 3?>To develop and evaluate a scoring algorithm that generates a resident-level intraoperative composite entrustability score from local EPA microassessment data that accounts for procedural difficulty, procedure repetition, and rater stringency/variability.</p> Methods <p><?tk 3?>We analyzed 949 intraoperative EPA microassessments collected in 2024–2025 from a large academic general surgery residency program, including 57 residents (PGY1-5), 72 attending surgeons, and 14 intraoperative EPA categories. A multilevel model with random intercepts for residents, EPAs, and attendings and a fixed effect for procedural repetition generated resident composite entrustability scores, which were evaluated for distributional properties, correlation with raw means, known groups validity by training level, and sources of measurement imprecision.</p> Results <p><?tk 3?>Estimated composite entrustability scores were normally distributed (mean 2.67, range 1.29–3.77) and highly correlated with raw averages (<i>r</i> = 0.98), indicating refinement of the scoring. Senior residents demonstrated significantly higher scores than junior residents (3.32 vs. 2.33, <i>p</i> &lt; 0.001), supporting known groups validity. Procedural repetition was positively associated with entrustability (β = 0.028, <i>p</i> &lt; 0.001). Variability in procedure difficulties and attending rater stringency was observed.</p> Conclusions <p><?tk 3?>Multilevel modeling offers a practical and psychometrically sound approach for synthesizing intraoperative EPA microassessments into resident-level entrustability estimates, effectively separating trainee performance from variation due to procedural difficulty, rater stringency, and learning over time. This scoring algorithm supports more accurate and rigorous use of local EPA data for programmatic decision-making.</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

Developing a scoring algorithm for intraoperative entrustability among general surgery residents using local EPA data

  • 期刊论文

  • 发表日期：

    2026年7月14日

查看全文
  • Dandan Chen,
  • Sophia Mckinley,
  • Jonah Thomas,
  • Emily Witt,
  • Roy Phitayakorn,
  • Jonathan Greer,
  • Joy Moses,
  • Douglas Smink

摘要

Purpose

To develop and evaluate a scoring algorithm that generates a resident-level intraoperative composite entrustability score from local EPA microassessment data that accounts for procedural difficulty, procedure repetition, and rater stringency/variability.

Methods

We analyzed 949 intraoperative EPA microassessments collected in 2024–2025 from a large academic general surgery residency program, including 57 residents (PGY1-5), 72 attending surgeons, and 14 intraoperative EPA categories. A multilevel model with random intercepts for residents, EPAs, and attendings and a fixed effect for procedural repetition generated resident composite entrustability scores, which were evaluated for distributional properties, correlation with raw means, known groups validity by training level, and sources of measurement imprecision.

Results

Estimated composite entrustability scores were normally distributed (mean 2.67, range 1.29–3.77) and highly correlated with raw averages (r = 0.98), indicating refinement of the scoring. Senior residents demonstrated significantly higher scores than junior residents (3.32 vs. 2.33, p < 0.001), supporting known groups validity. Procedural repetition was positively associated with entrustability (β = 0.028, p < 0.001). Variability in procedure difficulties and attending rater stringency was observed.

Conclusions

Multilevel modeling offers a practical and psychometrically sound approach for synthesizing intraoperative EPA microassessments into resident-level entrustability estimates, effectively separating trainee performance from variation due to procedural difficulty, rater stringency, and learning over time. This scoring algorithm supports more accurate and rigorous use of local EPA data for programmatic decision-making.