Key points are not available for this paper at this time.
Background: Artificial Intelligence (AI) is being integrated into healthcare, with the anticipation of improved efficiency. Nevertheless, there exists a significant gap in understanding its appropriate utilization and potential risks. Objectives: This survey is designed to capture the insights of current rheumatology fellows in the United States (US) regarding their comprehension of this subject. Methods: The survey was created utilizing Qualtrics and was shared via email to all rheumatology program directors and coordinators in the US to be forwarded to their current fellows for completion. The survey remained accessible from October to December 2023. Results: The survey garnered responses from 90 rheumatology fellows, with a demographic breakdown revealing 55% females, 38% males, and 2% chose not to disclose their gender. Insights into the academic journey of the fellows revealed 55% in their first year, 34% in their second year, and 2% pursuing a third-year research track. In terms of age, the distribution included 11% fellows in the 26-29 age bracket, 66% in the 30-35 range, 9% in the 36-39 group, and 12% above 40 years. Exploring their awareness of artificial intelligence (AI), 61% fellows affirmed their knowledge, while 5% disagreed, and 26% remained neutral. The exploration of AI tools, including CHATGPT and Open Evidence, was reported by 60% fellows, whereas 32% denied such engagement. Fellows mentioned that they used the AI for gathering personal information, literature review, composing letters to address insurance concerns, etc. A majority (60%) expressed a positive inclination towards incorporating AI education into their training programs, while 10% expressed disagreement, and 17% remained neutral (Figure 1). Confidence in AI's potential to complement or enhance rheumatology practice exhibited variation, with 44% fellows expressing positivity, 8% expressing negativity, and 38% remaining neutral, suggesting potential knowledge gaps (Figure 2). Specific AI-related skills considered beneficial encompassed expediting patient notes (77%), streamlining billing and administrative tasks (66%), and aiding prior authorizations. Some fellows envisioned potential applications of AI in evidence-based medicine (40%) and imaging analysis (38%). Concerns about AI replacing rheumatologists were predominantly dismissed by 74% fellows, while 10% were neutral, and only 3% concurred. Concerns about AI implementation included errors in charting/transcription (60%), over-reliance on AI recommendations (54%), AI generating false information (49%), and a lack of understanding of AI algorithms (47%). Notably, 50% fellows raised concerns about accountability in case of misjudgment, 34% were worried about diminished face-to-face patient discussions if AI took a primary role, and 51% expressed data privacy concerns. Despite these challenges, 63% fellows maintained optimism about integrating AI into their future practices, while 23% remained neutral, and only 2% disagreed. Conclusion: This distinctive survey of rheumatology fellows significantly adds to the broader discourse on AI in healthcare. A thorough literature review has uncovered studies that align with and support our findings, encompassing aspects such as variable awareness levels among healthcare professionals concerning AI, the imperative need for AI training in medical schools, the expectations for AI to enhance healthcare tasks, and concerns surrounding AI-generated errors, over-reliance, and data privacy. Ongoing research efforts will be instrumental in furthering our comprehension of the expansive role of AI in shaping the future landscape of rheumatology and healthcare. REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests: None declared.
Purohit et al. (Sat,) studied this question.