AI's deep dive into complex pediatric inguinal hernia issues: a challenge to traditional guidelines?

This study utilized ChatGPT, an artificial intelligence program based on large language models, to explore controversial issues in pediatric inguinal hernia surgery and compare its responses with the guidelines of the European Association of Pediatric Surgeons (EUPSA).

Six contentious issues raised by EUPSA were submitted to ChatGPT 4.0 for analysis, for which two independent responses were generated for each issue. These generated answers were subsequently compared with systematic reviews and guidelines. To ensure content accuracy and reliability, a content analysis was conducted, and expert evaluations were solicited for validation. Content analysis evaluated the consistency or discrepancy between ChatGPT 4.0's responses and the guidelines. An expert scoring method assess the quality, reliability, and applicability of responses. The TF-IDF model tested the stability and consistency of the two responses.

The responses generated by ChatGPT 4.0 were mostly consistent with the guidelines. However, some differences and contradictions were noted. The average quality score was 3.33, reliability score was 2.75, and applicability score was 3.46 (out of 5). The average similarity between the two responses was 0.72 (out of 1), Content analysis and expert ratings yielded consistent conclusions, enhancing the credibility of our research.

ChatGPT can provide valuable responses to clinical questions, but it has limitations and requires further improvement. It is recommended to combine ChatGPT with other reliable data sources to improve clinical practice and decision-making.