ChatGPT: Is This Patient Education Tool for Urological Malignancies Readable for the General Population?

Abstract

Background With widespread adoption of technological advancements in everyday life, patients are now increasingly able and willing to obtain information about their health conditions, treatment options, and indeed expected outcomes via the convenience of any device than can access the worldwide web. This introduces another aspect of patient care in the provision of healthcare for the modern doctor. ChatGPT is the first of an increasing number of self learning programs that have been released recently which may revolutionize and impact healthcare delivery. Methods The aim of this study is to obtain an objective measure of the readability of information provided on ChatGPT when compared with current validated patient information sheets provided by government health institutions in Western Australia. The same structured questions were input into the program for three major urological malignancies (urothelial, renal, and prostate), with the response generated evaluated with a validated readability scoring system – Flesch-Kincaid reading ease score. The same scoring system was then applied to current patient information sheets in circulation from Cancer Council Australia and UpToDate. Results Findings in this study looking at ease of readability of information provided on ChatGPT as compared to other government bodies and healthcare institutions confirm that they are non-inferior and may be a useful tool or adjunct to the traditional clinic based consultations. Ease of use of the information generated from ChatGPT was increased further when the question was modified to target an audience of 16 years of age, the average level of education attained by an Australian. Discussion Future research can be done to look into incorporating the use of similar technologies to increase efficiency in the healthcare system and reduce healthcare costs.

Comparing patient education tools for chronic pain medications: Artificial intelligence chatbot versus traditional patient information leaflets

“…To mitigate these risks, ongoing monitoring, robust validation mechanisms, and human oversight are necessary to ensure the delivery of accurate responses while acknowledging the potential of chatbots to offer accessible and timely information. [ 12 ] Thia and Saluja[ 13 ] found ChatGPT comparable to traditional materials in readability but suggested using specific prompts to improve ease of use for certain audiences. [ 13 ]…”

“…[ 12 ] Thia and Saluja[ 13 ] found ChatGPT comparable to traditional materials in readability but suggested using specific prompts to improve ease of use for certain audiences. [ 13 ]…”