Prevalence and associated factors of ChatGPT use in medical education and clinical practice

Abstract

Background: Healthcare education has witnessed significant transformations, with artificial intelligence (AI) playing an increasing role in medical imaging, diagnostics, and decision-making. ChatGPT, an AI-powered language model, is gaining popularity among medical students and physicians. However, excessive reliance on such tools may impact critical thinking and reduce dependence on professional judgment. Our objectives are to determine the prevalence of ChatGPT use among medical students and senior doctors in educational and clinical settings. To explore factors associated with its use, including its perceived impact on academic performance, decision-making, and information reliability. Methods: This cross sectional study includes four medical institutions in Peshawar, Khyber Pakhtunkhwa, Pakistan: Northwest School of Medicine, Pak International Medical College, Rehman Medical College, and Khyber Girl Medical College, and two dental colleges: Sardar Begum Dental College and Rehman College of Dentistry. It included 370 participants (176 males and 194 females). The participants were selected using a non-probability sampling technique. Results: ChatGPT use was widespread among students and doctors. However, there was no significant association between its use and academic performance (p>0.05). Fifth-year students showed a positive association with using ChatGPT for self-diagnosis (p<0.05). Fourth-year students reported it as a reliable source of information (p<0.01), while second-year students strongly favored it as a reputable source (p<0.001). Practitioners generally disagreed with these views. Conclusion: ChatGPT is increasingly used in medical education and practice. While it offers support in learning and decision-making, overdependence may hinder critical thinking. It should be viewed as a complementary tool, not a replacement, for expert medical judgment.

Ähnliche Arbeiten