Evaluating the accuracy of ChatGPT in delivering patient instructions for medications: an exploratory case study.

BACKGROUND: The use of ChatGPT in healthcare is still in its early stages; however, it has the potential to become a cornerstone in modern healthcare systems. This study aims to assess the accuracy of output of ChatGPT compared with those of CareNotes in providing patient instructions for three medications: tirzepatide, citalopram, and apixaban. METHODS: An exploratory case study was conducted using a published questionnaire to evaluate ChatGPT-generated reports against patient instructions from CareNotes. The evaluation focused on the completeness and correctness of the reports, as well as their potential to cause harm or lead to poor medication adherence. The evaluation was conducted by four pharmacy experts and 33 PharmD interns. RESULTS: The evaluators indicated that the ChatGPT reports of tirzepatide, citalopram, and apixaban were correct but lacked completeness. Additionally, ChatGPT reports have the potential to cause harm and may negatively affect medication adherence. CONCLUSION: Although ChatGPT demonstrated promising results, particularly in terms of correctness, it cannot yet be considered a reliable standalone source of patient drug information.