WVU vision researchers explore using AI to help with patient education materials

WVU vision researchers explore using AI to help with patient education materials

Through the help of AI, complex, patient-focused ophthalmic medical education materials could become more reader-friendly thanks to researchers in the West Virginia University Department of Ophthalmology and Visual Sciences.

The project assesses the readability of educational brochures created by the American Society of Ophthalmic Plastic and Reconstructive Surgery (ASOPRS).

“Given that ASOPRS is comprised entirely of physicians, I began to wonder if the materials they distribute are sometimes too advanced and less readable for patients,” John Nguyen, M.D., professor of ophthalmology and director of the Ophthalmic Plastic and Reconstructive Surgery Program, said. “I wanted to explore new potential avenues for delivering complex information about ophthalmic plastic and reconstructive surgery to patients in plain language.”

Dr. Nguyen and a team comprising of residents, fellows and medical students prompted two AI models, ChatGPT 4.0 and Google Bard, to produce written educational materials on 16 common ophthalmic plastic and reconstructive surgery topics. The goal was to create these materials with and without a sixth-grade reading-level prompt modifier. Those materials were then analyzed using seven key readability metrics and compared to the equivalent ASOPRS materials.

Through this process, Nguyen and his team determined that when prompted to produce materials at a sixth-grade reading level, both ChatGPT 4.0 and Google Bard were able to significantly improve their readability scores when compared to the corresponding ASOPRS brochures. Nguyen added that they found ChatGPT to be the most successful of the AI models, as it was able to produce the highest readability scores the most consistently.

While the initial research phase is complete, the team will further assess the accuracy and completeness of the AI-produced reading materials and eventually share them with real patients to further evaluate their readability. The research team consisted of, Nguyen; Diane Wang, M.D., fellow;  Alen Eid, M.D., resident; and Stephen Chen, medical student.

Dr. Wang, who primarily worked on data analysis throughout this study, said that as a fellow still early on in her medical career, she is excited to see how implementations of AI will be applied to the medical field in the future.

“As clinicians, sometimes we forget that the complex information we discuss daily is far from the norm of what typical patients encounter in their day-to-day life,” she said. “This makes it our responsibility to play a key role in directing our patients to comprehensible, accurate and accessible information online to help them make more informed decisions about their health.”

To learn more about the research opportunities available at the Department of Ophthalmology and Visual Sciences, visit medicine.hsc.wvu.edu/eye/research