Yang Xu,* Qiankun Liu,* Jiaxue Pang, Chunlu Zeng, Xiaoqing Ma, Pengyao Li, Li Ma, Juju Huang, Hui Xie

College of Nursing, Bengbu Medical University, Bengbu, Anhui, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Hui Xie, Bengbu Medical University, Donghai Avenue, Longzihu District, Bengbu City, Anhui Province, People’s Republic of China, Email hui2122@sina.com

Purpose: Exercise is a vital adjunct therapy for patients with hypertension comorbidities. However, medical personnel and patients face significant obstacles in implementing exercise prescription recommendations. AI has been developed as a beneficial tool in the healthcare field. The performance of intelligent tools such as ChatGPT 4.0 and Intelligent Health Promotion Systems (IHPS) in issuing exercise prescriptions for patients with hypertension comorbidities remains to be verified.
Patients and Methods: After collecting patient information through IHPS hardware and questionnaire systems, the data were input into the software terminals of ChatGPT 4.0 and IHPS according to the five stages of the Transtheoretical Model, resulting in exercise prescriptions. Subsequently, experts from various fields scored the accuracy, comprehensiveness, and applicability of each prescription, along with providing professional recommendations based on their expertise. By comparing the performance of both systems, their capability to serve this specific group was evaluated.
Results: In most cases, ChatGPT scored significantly higher than IHPS in terms of accuracy, comprehensiveness, and applicability. However, when patients exhibited certain functional movement disorders, GPT’s exercise prescriptions involved higher health risks, whereas the more conservative approach of IHPS was advantageous.
Conclusion: The path of generating exercise prescriptions using artificial intelligence, whether via ChatGPT or IHPS, cannot achieve a completely satisfactory state.But can serve as a supplementary tool for professionals issuing exercise prescriptions to patients with hypertension comorbidities, especially in alleviating the financial burden of consulting costs. Future research could further explore the performance of AI in issuing exercise prescriptions, harmonize it with physiological indicators and phased feedback, and develop an interactive user experience.

Keywords: artificial intelligence, exercise prescription, hypertension, intelligent health promotion systems, transtheoretical model