AI-Based Health Symptom Checking, Self-Medication, Health Anxiety, and Physical Self-Efficacy Among College Students

Abstract

College students increasingly use generative artificial intelligence (AI) to interpret health symptoms. However, little is known about how this behavior is associated with self-medication, health anxiety, and physical self-efficacy in Philippine higher education. This study examined these relationships using a cross-sectional correlational design. A researcher-developed 10-item AI Health Symptom Checking Scale was pilot-tested with 30 undergraduate students and then administered online, together with established measures of self-medication beliefs, health anxiety, and physical self-efficacy, to 291 students recruited through purposive-convenience sampling. The main survey was completed in one day in May 2026. Internal consistency was assessed using Cronbach's alpha and McDonald's omega, and relationships were examined using Pearson correlations. The new scale demonstrated good internal consistency (α = .877; ω = .881). Its four factors had statistically significant, generally small positive associations with selected self-medication dimensions (r = .159–.253, p < .05) and small-to-moderate positive associations with health anxiety (r = .155–.301, p < .05). Physical self-efficacy was not significantly associated with AI symptom checking or self-medication, but its two dimensions were negatively associated with health anxiety (r = −.140 to −.175, p < .05). Greater engagement with AI for symptom checking was therefore associated with, but not shown to cause, stronger self-medication beliefs and health anxiety. Universities should strengthen digital health literacy and encourage professional verification of AI-generated health information.