Scholarlink

Background <p>Conversational artificial intelligence (AI) tools, such as large language models and chat-based systems, are increasingly used by clinicians and trainees. However, empirical evidence on how anesthesiology residents integrate conversational AI into perioperative workflows, especially within the operating room, is lacking. This exploratory single-center observational study aimed to characterize prevalence, patterns, and determinants of conversational AI use in the operating room.</p> Methods <p>We conducted a cross-sectional, anonymous online survey of anesthesiology and intensive care medicine residents in a single training program. The instrument was developed through iterative refinement and addressed knowledge, training, general AI usage habits, perioperative use, supervision, attitudes, safety perceptions, and data protection behaviors. Reporting followed STROBE and CHERRIES guidelines. The primary outcome was self-reported use of conversational AI in the operating room within the previous 3 months. Bivariate analyses and logistic regression examined associations with demographic, behavioral, and attitudinal factors.</p> Results <p>Among 101 responses, 86 residents met inclusion criteria. The mean age was 31.8 (SD 3.9) years; 58.1% were female. Conversational AI use in the operating room was reported by 58.1% (95% CI 47.6–68.0%). Higher general frequency of AI use was independently associated with operating room use (adjusted OR 3.10; 95% CI 1.83–5.92). Training seniority was also associated with operating room use (adjusted OR 1.97 per additional residency year; 95% CI 1.27–3.33). Female gender was associated with higher odds of operating room use (adjusted OR 3.97; 95% CI 1.17–15.28). Attitudinal variables showed strong univariable associations with operating room conversational AI use, but these associations were attenuated after multivariable adjustment. Cronbach’s alpha for the attitudinal scale was 0.79 (95% CI 0.71–0.85).</p> Conclusions <p>In this exploratory single-center study, conversational AI use in the operating room was reported by a substantial proportion of anesthesiology residents (~ 60%) and was associated with general AI usage patterns and residency seniority. Formal training, supervision, and institutional policies appeared limited. These findings highlight the need for competency-based educational initiatives and clear governance to support safe and effective integration of conversational AI into perioperative care.</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

Conversational artificial intelligence in the operating room: an exploratory cross-sectional single-center survey among anesthesiology residents

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月22日

查看全文
  • Dario Gaetano,
  • Viola Lomonaco,
  • Salvatore Ferraioli,
  • Vincenzo Pota,
  • Umberto Colella,
  • Maria Caterina Pace,
  • Pasquale Sansone

摘要

Background

Conversational artificial intelligence (AI) tools, such as large language models and chat-based systems, are increasingly used by clinicians and trainees. However, empirical evidence on how anesthesiology residents integrate conversational AI into perioperative workflows, especially within the operating room, is lacking. This exploratory single-center observational study aimed to characterize prevalence, patterns, and determinants of conversational AI use in the operating room.

Methods

We conducted a cross-sectional, anonymous online survey of anesthesiology and intensive care medicine residents in a single training program. The instrument was developed through iterative refinement and addressed knowledge, training, general AI usage habits, perioperative use, supervision, attitudes, safety perceptions, and data protection behaviors. Reporting followed STROBE and CHERRIES guidelines. The primary outcome was self-reported use of conversational AI in the operating room within the previous 3 months. Bivariate analyses and logistic regression examined associations with demographic, behavioral, and attitudinal factors.

Results

Among 101 responses, 86 residents met inclusion criteria. The mean age was 31.8 (SD 3.9) years; 58.1% were female. Conversational AI use in the operating room was reported by 58.1% (95% CI 47.6–68.0%). Higher general frequency of AI use was independently associated with operating room use (adjusted OR 3.10; 95% CI 1.83–5.92). Training seniority was also associated with operating room use (adjusted OR 1.97 per additional residency year; 95% CI 1.27–3.33). Female gender was associated with higher odds of operating room use (adjusted OR 3.97; 95% CI 1.17–15.28). Attitudinal variables showed strong univariable associations with operating room conversational AI use, but these associations were attenuated after multivariable adjustment. Cronbach’s alpha for the attitudinal scale was 0.79 (95% CI 0.71–0.85).

Conclusions

In this exploratory single-center study, conversational AI use in the operating room was reported by a substantial proportion of anesthesiology residents (~ 60%) and was associated with general AI usage patterns and residency seniority. Formal training, supervision, and institutional policies appeared limited. These findings highlight the need for competency-based educational initiatives and clear governance to support safe and effective integration of conversational AI into perioperative care.