Evaluation of the safety, accuracy, and helpfulness of the GPT-4.0 Large Language Model in neurosurgery

Abstract

BackgroundAlthough prior work demonstrated the surprising accuracy of Large Language Models (LLMs) on neurosurgery board-style questions, their use in day-to-day clinical situations warrants further investigation. This study assessed GPT-4.0′s responses to common clinical questions across various subspecialties of neurosurgery. MethodsA panel of attending neurosurgeons formulated 35 general neurosurgical questions spanning neuro-oncology, spine, vascular, functional, pediatrics, and trauma. All questions were input into GPT-4.0 with a prespecified, standard prompt. Responses were evaluated by two attending neurosurgeons, each on a standardized scale for accuracy, safety, and helpfulness. Citations were indexed and evaluated against identifiable database references. ResultsGPT-4.0 responses were consistent with current medical guidelines and accounted for recent advances in the field 92.8 % and 78.6 % of the time respectively. Neurosurgeons reported GPT-4.0 responses providing unrealistic information or potentially risky information 14.3 % and 7.1 % of the time respectively. Assessed on 5-point scales, responses suggested that GPT-4.0 was clinically useful (4.0 ± 0.6), relevant (4.7 ± 0.3), and coherent (4.9 ± 0.2). The depth of clinical responses varied (3.7 ± 0.6), and “red flag” symptoms were missed 7.1 % of the time. Moreover, GPT-4.0 cited 86 references (2.46 citations per answer), of which only 50 % were deemed valid, and 77.1 % of responses contained at least one inappropriate citation. ConclusionCurrent general LLM technology can offer generally accurate, safe, and helpful neurosurgical information, but may not fully evaluate medical literature or recent field advances. Citation generation and usage remains unreliable. As this technology becomes more ubiquitous, clinicians will need to exercise caution when dealing with it in practice.