Utilizing Artificial Intelligence and Chat Generative Pretrained Transformer to Answer Questions About Clinical Scenarios in Neuroanesthesiology

OBJECTIVE

We tested the ability of chat generative pretrained transformer (ChatGPT), an artificial intelligence chatbot, to answer questions relevant to scenarios covered in 3 clinical guidelines, published by the Society for Neuroscience in Anesthesiology and Critical Care (SNACC), which has published management guidelines: endovascular treatment of stroke, perioperative stroke (Stroke), and care of patients undergoing complex spine surgery (Spine).

METHODS

Four neuroanesthesiologists independently assessed whether ChatGPT could apply 52 high-quality recommendations (HQRs) included in the 3 SNACC guidelines. HQRs were deemed present in the ChatGPT responses if noted by at least 3 of the 4 reviewers. Reviewers also identified incorrect references, potentially harmful recommendations, and whether ChatGPT cited the SNACC guidelines.

RESULTS

The overall reviewer agreement for the presence of HQRs in the ChatGPT answers ranged from 0% to 100%. Only 4 of 52 (8%) HQRs were deemed present by at least 3 of the 4 reviewers after 5 generic questions, and 23 (44%) HQRs were deemed present after at least 1 additional targeted question. Potentially harmful recommendations were identified for each of the 3 clinical scenarios and ChatGPT failed to cite the SNACC guidelines.

CONCLUSIONS

The ChatGPT answers were open to human interpretation regarding whether the responses included the HQRs. Though targeted questions resulted in the inclusion of more HQRs than generic questions, fewer than 50% of HQRs were noted even after targeted questions. This suggests that ChatGPT should not currently be considered a reliable source of information for clinical decision-making. Future iterations of ChatGPT may refine algorithms to improve its reliability as a source of clinical information.

Differences in Perioperative Management of Patients Undergoing Complex Spine Surgery: A Global Perspective

BACKGROUND

The aim of this survey was to understand institutional spine surgery practices and their concordance with published best practices/recommendations.

METHODS

Using a global internet-based survey examining perioperative spine surgery practice, reported institutional spine pathway elements (n=139) were compared with the level of evidence published in guideline recommendations. The concordance of clinical practice with guidelines was categorized as poor (≤20%), fair (21%-40%), moderate (41%-60%), good (61%-80%), or very good (81%-100%).

RESULTS

Seventy-two of 409 (17.6%) institutional contacts started the survey, of which 31 (7.6%) completed the survey. Six (19.4%) of the completed surveys were from respondents in low/middle-income countries, and 25 (80.6%) were from respondents in high-income countries. Forty-one incomplete surveys were not included in the final analysis, as most were less than 40% complete. Five of 139 (3.6%) reported elements had very good concordance for the entire cohort; hospitals with spine surgery pathways reported 18 elements with very good concordance, whereas institutions without spine surgery pathways reported only 1 element with very good concordance. Reported spine pathways included between 7 and 47 separate pathway elements. There were 87 unique elements in the reviewed pathways. Only 3 of 87 (3.4%) elements with high-quality evidence demonstrated very good practice concordance.

CONCLUSIONS

This global survey-based study identified practice variation and low adoption rates of high-quality evidence in the care of patients undergoing complex spine surgery.