From Paperwork to Productivity - Harnessing the Power of Large Language Models for Plastic Surgery Administrative Tasks
Nikita O. Shulzhenko*, Edward S. Lee
Division of Plastic and Reconstructive Surgery, Rutgers-NJMS, Newark, NJ
Administrative burden can be a significant source of stress and burnout. Large language models (LLMs) offer a promising solution for streamlining administrative tasks to reduce overall workload without compromising patient care. This proof-of-concept study explored the potential of tailoring LLMs to streamline the completion of administrative tasks in plastic surgery.
Custom prompts were recursively developed using a LLM to generate common yet generic medical documentation for typical plastic surgery patients, such as those requiring breast reduction, panniculectomy, or chest masculinization. These documents were used to iteratively develop algorithms to produce common time-consuming administrative deliverables. Completely deidentified patient documentation was then subject to these algorithms and the output verified for accuracy, validity, and rapidity compared to traditional means.
The LLM algorithm resulted in suitably automating common tedious administrative tasks. From hypothetical and deidentified medical records, the LLM was rapidly able to report accurate ICD-10 and CPT codes, levels of service, as well as produce patient-specific documentation for prior authorizations, disability claims, and letters of medical necessity. These outputs were reviewed by the authors and found to be generally accurate with minimal revision.
The results of this proof-of-concept study demonstrate the potential of LLMs to alleviate the administrative workload of plastic surgeons. LLM processing was accurate and efficient, resulting in a significant reduction in the time and effort required to complete administrative tasks. Further exploration into this technology is warranted. With discerning review, the use of these deliverables should result in no compromise in patient care but may alleviate the psychosocial and opportunity costs of growing administrative burdens and translate into more efficient and efficacious healthcare delivery.
Back to 2023 Abstracts