Artificial intelligence in patient care: evaluating artificial intelligence's accuracy and accessibility in...

Artificial Intelligence (AI) is rapidly advancing and profoundly influencing healthcare, offering the potential to revolutionize access to medical information. As medical misinformation proliferates and online searches for health-related advice increase, there is an escalating need for dependable patient information. This study evaluates the effectiveness of an AI chatbot in delivering information for blepharoplasty candidates.

Numerous frequently asked questions on blepharoplasty sourced from the ASPS website were asked to ChatGPT. The responses were then rigorously cross-referenced with relevant scholarly literature and meticulously reviewed by the research team to determine their accuracy. Additionally, the questions were classified into three categories, and the responses were evaluated using Flesch-Kincaid readability metrics, along with ANOVA and trend analysis tests.

Despite minor variations, ChatGPT's responses to blepharoplasty FAQs largely aligned with current literature. Overall readability analysis showed a Flesch Reading Ease score of 31.48, indicating high school complexity with a Flesch-Kincaid Grade Level at 10.92, and 20.97% use of passive voice. The ANOVA results showed that there were no significant differences in readability between the categories (p-values: Flesch Reading Ease = 0.816, Flesch-Kincaid = 0.616, Passive Sentences = 0.115). Trend analysis also showed that the level of response complexity stayed the same across the questions.

ChatGPT is an evolving tool that holds potential for patients in accessing and comprehending medical information. While it can offer accurate insights, it's important to recognize that its answers might not consistently reach complete accuracy or be suitable for patients with varying educational levels.