This project investigates how ChatGPT can improve the readability of patient education materials on interventional radiology (IR) procedures using original texts from the Cardiovascular and Interventional Radiological Society of Europe.
It involves calculating various readability scores and simplifying the text to a fifth-grade level, resulting in significant improvements in readability metrics but a decrease in the credibility scores.
The study concludes that while ChatGPT effectively makes the material easier to understand, it compromises the reliability of the original text, indicating a need for human oversight in the simplification process.