Artificial Intelligence Generates Multilingual Cancer Patient Materials with Superior Readability and Equivalent Accuracy, Clarity, and Completeness to Expert-Written Texts
Los Angeles, CA - May 2025 –
A new study published in PLOS One spearheaded by researchers at the University of Southern California unveils the transformative capabilities of generative artificial intelligence (AI) in crafting patient education materials (PEMs) for urological cancer care. Utilizing the advanced GPT-4, this study demonstrates AI's ability to generate highly readable and accurate educational content aligned with current oncological guidelines.
In the rapidly evolving landscape of cancer treatment, patient education materials must be regularly updated to reflect the newest guidelines. These materials need to be both informative and easily understandable to empower patients and mitigate misinformation. Historically, PEMs have struggled with readability and timely updates, posing challenges for patients seeking clear and current health information.
The study employed a tri-phasic generative AI framework, leveraging GPT-4 to distill complex clinical guidelines into layperson-friendly summaries—a task completed in an astonishing average time of just over 50 seconds per document. The PEMs produced by GPT-4 showcased superior readability, requiring a lower education level to comprehend compared to traditional materials. This is significant as improving accessibility and understanding is crucial for patient engagement.
In a rigorous assessment involving 32 specialists from different team of the young academic urologists of the European Association of Urology, the AI-generated PEMs were evaluated against traditional counterparts for prostate, bladder, kidney, and testicular cancer information. They were found equal in accuracy, completeness, and clarity, dispelling concerns about the dependability of AI in handling sensitive medical data. Additionally, translations facilitated by AI scored high in grammatical correctness and clarity across multiple languages, broadening access to vital information globally.
The authors emphasized the importance of AI in not only enhancing the accessibility of medical information but also in reducing the workload involved in producing high-quality education materials. "Our AI model, as part of the BRIDGE-AI initiative, bridges a critical gap, offering timely, contextually accurate information in a manner tailored to diverse educational backgrounds and languages," Dr. Rodler stated, highlighting the AI's potential to significantly improve worldwide health literacy.
Despite its successes, the study acknowledges limitations, including AI's occasional inaccuracies and reliance on existing data. The need for human oversight remains imperative to ensure ethical and responsible information dissemination. Ongoing advancements in AI are expected to refine these capabilities, fostering a synergistic relationship between technology and human expertise in healthcare communication.
Rodler, S., Cei, F., Ganjavi, C., Checcucci, E., De Backer, P., Rivero Belenchon, I., Taratkin, M., Puliatti, S., Veccia, A., Piazza, P., Baekelandt, L., Kowalewski, K. F., Gómez Rivas, J., Fankhauser, C. D., Moschini, M., Gandaglia, G., Campi, R., De Castro Abreu, A., Russo, G. I., Cocci, A., … YAU Collaborators (2025). GPT-4 generates accurate and readable patient education materials aligned with current oncological guidelines: A randomized assessment. PloS one, 20(6), e0324175. https://doi.org/10.1371/journal.pone.0324175
