Nurse Town: An LLM-Powered Simulation Game for Nursing Education
Yuqi Hu, Qiwen Xiong, Lina Yi, Ilmi Yoon
Abstract
The shortage of nurses remains a persistent challenge in the U.S. healthcare system, largely driven by a scarcity of clinical training opportunities for nursing students. This underscores the urgent need for innovative approaches in nursing education. While serious games have shown effectiveness, they still face limitations, particularly in providing diverse experiences and overcoming resource constraints. Leveraging advances in artificial intelligence (AI), Nurse Town is a novel digital simulation game designed to enhance critical nursing competencies, including clinical decision-making, effective communication, and empathy. Using Large Language Models (LLMs), we developed virtual patient avatars capable of simulating medical scenarios and engaging in natural, context-aware conversations. The game also features an LLM-based automated performance assessment component. Compared to traditional training methods, Nurse Town significantly reduces the demand for educational resources while ensuring consistent training quality. We present a minimum viable product (MVP) featuring one virtual patient with hypertension. Future development plans include expanding scenario diversity, enhancing avatar realism, and integrating Virtual Reality (VR).