Academic Orthopaedic Surgeons Prefer Peer-Reviewed Guidelines Over Artificial Intelligence and Social Media for...

Background and objective Patients increasingly seek recommendations for improved health and wellness, as well as medical advice, through social media and large language models (LLMs), yet the quality of these sources varies considerably. Effective clinical communication requires an understanding of how patients engage with online medical content, as these sources may shape expectations even before a clinical encounter. However, to date, no study has assessed how orthopaedic surgeons themselves use these same resources. Hence, this study aimed to evaluate orthopaedic surgeons' self-reported use of LLMs, social media, and peer-reviewed clinical resources for patient care and personal medical care, along with the perceived understandability and helpfulness of each source. Methods In this cross-sectional study, an electronic survey was administered to a convenience sample of all orthopaedic surgeons at a single academic institution. Respondents indicated whether they had ever used LLMs (e.g., ChatGPT), social media platforms (e.g., TikTok, Instagram), or peer-reviewed clinical resources (e.g., guidelines from the North American Spine Society (NASS)) for patient care or their own medical care. Those who reported use were asked whether each source was easy to understand and helpful. Proportions are reported with 95% Wilson confidence intervals (CIs). McNemar's exact test was used to compare paired utilization rates across sources and contexts, while Fisher's exact test was employed to compare comprehensibility between distinct user subgroups. All analyses were exploratory given the pilot design. Results Sixteen orthopaedic surgeons completed the survey. Peer-reviewed clinical guidelines were used by 15 surgeons (93.8%; 95% CI: 71.7%-98.9%) for patient care and 14 (87.5%; 95% CI: 64.0%-96.5%) for personal care, significantly more frequently than LLMs or social media for both contexts (all p ≤ 0.016). LLMs were used by five surgeons (31.2%; 95% CI: 14.2%-55.6%) for patient care and seven (43.8%; 95% CI: 23.1%-66.8%) for personal care. LLM use for personal care was significantly more frequent than social media use (p = 0.031). Social media use was rare, reported by two surgeons (12.5%) for patient care and one (6.2%) for personal care. Every surgeon who used an LLM in either context rated it as helpful, and the majority found it easy to understand. Comprehensibility ratings did not differ significantly between LLM users and guideline users for either patient care (80.0% vs. 66.7%; p = 1.000) or personal care (85.7% vs. 64.3%; p = 0.613). Concordance analysis showed that 14 of 16 surgeons used guidelines in both contexts, whereas eight had never used an LLM and 14 had never used social media in either setting. Conclusions Academic orthopaedic surgeons demonstrate a strong and statistically significant preference for peer-reviewed clinical guidelines over LLMs and social media for both personal and patient medical care. LLMs occupy an intermediate position in surgeons' information-seeking preferences and were rated as helpful by all users, suggesting cautious but growing adoption. Social media use for medical purposes was rare among surgeons, contrasting sharply with published estimates exceeding 70% among the general public. This disconnect, with patients seeking orthopaedic information on platforms where expert content is scarce, represents a clinically meaningful gap.