Tracking AI/LLM Literacy and Knowledge in Urology Outpatients (TALK-U)

Overview

This study is a one-time, questionnaire-based survey of adults attending a urology outpatient clinic. The investigators will examine whether patients who use AI chat tools based on large language models (LLMs), such as ChatGPT, before seeking care differ from non-users in (1) anxiety, (2) satisfaction with the clinic visit, and (3) awareness/understanding of their health condition. Participants who provide written informed consent will complete a brief survey during their clinic visit (about 5-7 minutes). The survey includes a standardized anxiety scale (GAD-7), visual analog scales (0-10) for anxiety before and after the visit, satisfaction, and disease awareness, as well as questions about AI/LLM knowledge and use. Treating clinicians will not have access to participants' survey responses, and participation will not affect clinical care.

The TALK-U study is a cross-sectional, observational survey conducted in the urology outpatient clinic of Dr. Abdurrahman Yurtaslan Ankara Oncology Training and Research Hospital. The primary exposure is self-reported AI/LLM knowledge and use for health information (e.g., whether participants have used ChatGPT or similar tools, frequency, awareness of AI, and purpose of use). The primary outcome is anxiety measured by GAD-7 at the time of the clinic encounter. Secondary outcomes include (i) anxiety assessed on a 0-10 visual analog scale before and after the visit, (ii) outpatient visit satisfaction (0-10 VAS), and (iii) perceived disease awareness/understanding (0-10 VAS). Eligible adult first-time urology outpatients who can complete the survey independently and provide informed consent will be invited during a 2-month recruitment period. Analyses will compare outcomes between AI/LLM users and non-users and explore associations with age group (\<40, 40-60, \>60), education level, and urology subspecialty category. Data will be analyzed using appropriate parametric/nonparametric tests depending on distribution, with p \< 0.05 considered statistically significant.

Conditions