A comparison of large language model-generated and published perioperative neurocognitive disorder recommendations: a cross-sectional web-based analysis
Bibliographic record
Abstract
BACKGROUND: Perioperative neurocognitive disorders (PNDs) are common complications after surgery and anaesthesia, particularly in older adults, leading to increased morbidity, mortality, and healthcare costs. Therefore, major medical societies have developed recommendations for the prevention and treatment of PNDs. Our study evaluated the reliability of large language models, specifically ChatGPT-4 and Gemini, in generating recommendations for PND management and comparing them with published guidelines. METHODS: We conducted an online cross-sectional web-based analysis over 48 h in June 2024. Artificial intelligence (AI)-generated recommendations were produced in six different locations across five countries (Switzerland, Belgium, Turkey, Canada, and the East and West Coasts of the USA). The English prompt 'a table of a bundle of care for perioperative neurocognitive disorders' was entered into ChatGPT-4 and Gemini, generating tables evaluated by independent reviewers. The primary outcomes were the Total Disagreement Score (TDS) and Quality Assessment of Medical Artificial Intelligence (QAMAI), which compared AI-generated recommendations with published guidelines. RESULTS: The study generated 14 tables, with TDS and QAMAI scores showing similar results for ChatGPT-4 and Gemini (2 [1-3] vs 2 [2-3], P=0.636 and 4 [4-4] vs 4 [3-4], P=0.424, respectively). AI-generated recommendations aligned well with published guidelines, with the highest alignment observed in ChatGPT-4-generated recommendations. No complete agreement with guidelines was achieved, and lack of cited sources was a noted limitation. CONCLUSIONS: Large language models can generate perioperative neurocognitive disorder recommendations that align closely with published guidelines. However, further validation and integration of clinician feedback are required before clinical application.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.074 | 0.280 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.009 | 0.008 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".