Xiangxun Lai

Xiangxun Lai

Auto-generated profile, based on publicly available information

Xiangxun Lai

Xiamen University of Technology

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 May 2025 in JMIR Medical Informatics
Verified
Journal Article
Open access
XLXiangxun Lai
Xiangxun Lai
JCJiacheng Chen
Jiacheng Chen
YLYue Lai
Yue Lai
SHShengqi Huang
Shengqi Huang
YCYongdong Cai
Yongdong Cai
CHCaihua Huang
Caihua Huang
Background: Regular exercise recommendations (ERs) and physical activity (PA) are crucial for the prevention and management of chronic diseases. However, creating effective exercise programs demand substantial time and specialized expertise from both medical and sports professionals. Large language models (LLMs), such as ChatGPT, offer a promising solution by helping create personalized ERs. While LLMs show potential, their use in exercise planning remains in its early stages and requires further exploration. Objectives: This study aims to systematically review and classify the applications of LLMs in ERs and PA. It also seeks to identify existing gaps and provide insights into future research directions for optimizing LLM integration in personalized health interventions. Methods: A scoping review methodology was used to identify studies related to LLM applications in ERs and PA. Literature searches were conducted in Web of Science, PubMed, IEEE, and arXiv for English language papers published up to March 21, 2024. Keywords included LLMs, chatbots, ERs, PA, fitness plan, and related terms. Two independent reviewers (XL and CH) screened and selected studies based on predefined inclusion criteria. Thematic analysis was used to synthesize findings, which were presented narratively. Results: An initial search identified 598 papers, of which 1.8% (11/598) of studies were included after screening and applying selection criteria. Of these, ChatGPT-based models were used in 55% (6/11) of the studies. In addition, 73% (8/11) of the studies used expert evaluations and user feedback to assess model usability, and 45% (5/11) of the studies used experimental designs to evaluate LLM interventions in ERs and PA. Key findings indicated that LLMs can generate tailored ERs, save time in clinical practice, and enhance safety by incorporating patient-specific data. They also increased engagement and supported behavior change. This made PA guidance more accessible, especially in remote or underserved communities. Conclusions: This review highlights the promising applications of LLMs in ERs and PA but emphasizes that they remain a supplement to human expertise. Expert validation is essential to ensure safety and mitigate risks. Future research should prioritize pilot testing, clinician training programs, and large-scale clinical trials to enhance feasibility, transparency, and ethical integration.

17 June 2025 in Journal of Medical Internet Research
Verified
Journal Article
Open access
XLXiangxun Lai
Xiangxun Lai
YLYue Lai
Yue Lai
JCJiacheng Chen
Jiacheng Chen
KPKaijaing Pan
Kaijaing Pan
QGQi Gao
Qi Gao
CHCaihua Huang
Caihua Huang
Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Large language model (LLM)-based AI coaches show promise for personalized exercise and health interventions. Their complex capabilities (communication, planning, movement analysis, monitoring) necessitate rigorous, multidimensional evaluation, but standardized frameworks are lacking. Objective: This scoping review systematically maps current evaluation strategies for LLM-based AI coaches in exercise and health, identifies strengths and limitations, and proposes future directions for robust, standardized validation. Methods: Following PRISMA-ScR guidelines, we systematically searched six databases using keywords for LLMs, exercise/health coaching, and evaluation. Studies describing LLM-based coaching systems with reported performance evaluation methods were included. Data on models, applications, evaluation strategies, and outcomes were charted. Results: Seventeen studies published between March 2023 and March 2025 met the inclusion criteria. Most utilized proprietary models (e.g., GPT-4), while some used open-source or custom models. Six studies incorporated multimodal inputs (video, sensor data). Evaluation strategies were highly heterogeneous, including quantitative metrics (Accuracy, F1, MAE), empirical methods (user studies, expert comparisons), and expert/user-centered feedback (expert scores [Kappa ≈ 0.79–0.82], user surveys [MITI, SASSI]). However, evaluations often lacked real-world testing, longitudinal assessment, and standardized benchmarks. Conclusions: Evaluating LLM-based exercise and health coaches requires multifaceted strategies: quantitative metrics for objective tasks, empirical validation for user interaction, and expert assessment for personalization and safety. Current evaluations are fragmented, lacking standardization, ecological validity, and longitudinal assessment. Future progress demands robust, multidimensional frameworks emphasizing real-world validation, integrating RAG for accuracy, and developing specialized, efficient multimodal models or agents for reliable and scalable AI coaching.

23 November 2025 in International Journal of Medical Informatics
Verified
Journal Article
Open access
XLXiangxun Lai
Xiangxun Lai
YLYue Lai
Yue Lai
JCJiacheng Chen
Jiacheng Chen
SHShengqi Huang
Shengqi Huang
QGQi Gao
Qi Gao
CHCaihua Huang
Caihua Huang
The quality assessment of personalized exercise prescriptions is currently hampered by the subjectivity and inefficiency of traditional rating scales. Artificial intelligence (AI) presents a transformative opportunity for objective, scalable evaluation. This pilot study aimed to develop an AI-assisted evaluation framework and assess the feasibility, reliability, and efficiency of its core component, the Adaptive Precision Boolean Rubric (Adaptive-PBR). Based on ACSM guidelines, we developed a 50-item Precision Boolean Rubric (PBR) and a 10-item Likert scale. To maximize ecological validity, we utilized GPT-4 (via ChatGPT Plus Web Interface) with a dual-run consistency protocol to generate case-specific, 20-item Adaptive-PBRs. Twelve experts evaluated five diverse clinical cases (yielding 180 rating points) using all three instruments under randomized conditions. The Adaptive-PBR demonstrated excellent inter-rater reliability (ICC = 0.83), significantly outperforming the Likert scale (ICC = 0.65) and matching the full PBR (ICC = 0.82). Quantitatively, it achieved high scoring precision (Median 0.78, IQR 0.70-0.85) while reducing evaluation time by approximately 63 % (mean 7.1 vs. 19.5 min) compared to the full PBR. Crucially, the Adaptive-PBR mitigated the subjective variability and experience-related bias observed with the Likert scale. The AI-assisted Adaptive-PBR establishes a feasible, reliable, and efficient evaluation standard. By combining granular criteria with AI-driven adaptability, it offers a robust alternative to subjective scales, with immediate potential as a quality assurance tool in clinical training and practice.

28 January 2026 in Research Square
Verified
Preprint
Open access
PDF file
XLXiangxun Lai
Xiangxun Lai
YLYue Lai
Yue Lai
JHJiayue He
Jiayue He
CHCaihua Huang
Caihua Huang
Background Telerehabilitation is a key component of the global digital health transformation, offering a potent solution to health inequities, such as China's maldistribution of rehabilitation resources. However, its widespread adoption depends on professional acceptance, which is not well understood in the Chinese context. Objective This study aimed to investigate the utility, perceptions, contributing factors, and barriers regarding telerehabilitation among registered Chinese rehabilitation professionals. Methods A national cross-sectional web-based survey was conducted (n = 228) among registered rehabilitation therapists and physicians using a snowball sampling technique. The validated questionnaire assessed demographics, prior experience, perceived utility, preferred modalities, barriers, and attitudes. Data were analyzed using descriptive statistics, Chi-square tests, and Spearman's rank correlation. Results A significant "awareness-to-practice" gap was identified: while 67% were aware of telerehabilitation, only 23% had clinical experience with it. A majority (57%) held a neutral attitude towards its development. The primary perceived barrier was the constraint on "hands-on" techniques (75.4%). "Patient value" (68.4%), "policy" (55.3%), and "physiotherapist's expertise" (52.2%) were key implementation factors. Significant negative correlations were found between gender and attitudes (r=-.148, p = .025) and between working city-tier and perceived appropriate fees (r=-.211, p = .001). Conclusion Telerehabilitation in China is at a pivotal juncture, characterized by professional hesitancy rather than opposition. This "awareness-to-practice" gap is driven by a lack of standardized training, clear policy, and robust reimbursement structures. Addressing these structural, educational, and financial bottlenecks is imperative to unlock telerehabilitation's potential for improving healthcare accessibility and equity in China.

8 April 2024 in JMIR Preprints
Verified
Preprint
Open access
XLXiangxun Lai
Xiangxun Lai
JCJiacheng Chen
Jiacheng Chen
SHShengqi Huang
Shengqi Huang
YCYongdong Cai
Yongdong Cai
ZSZhifeng Sun
Zhifeng Sun
CHCaihua Huang
Caihua Huang
BACKGROUND Despite a growing awareness among healthcare professionals regarding the benefits of exercise and physical activity, the integration of exercise recommendations into healthcare practices remains sub-optimal. The complexity of the recommendations, which necessitate interdisciplinary knowledge, poses challenges in formulating the structured exercise plans. The rapid development of Large Language Models (LLMs) and Chatbots, offers a potential solution by leveraging AI's capabilities to enhance the precision and personalization of fitness and health recommendations. OBJECTIVE This paper provides a systematic summary of existing research in the landscape of LLMs and Chatbot applications for exercise recommendations and physical activity, encompassing important topics, theoretical frameworks, methodologies, and potential avenues for further exploration. METHODS A comprehensive literature search was conducted across databases, including Web of Science, PubMed, IEEE, and arXiv. The focus was on studies that utilized LLMs or Chatbots in recommendations for exercise, physical activity or fitness. The search was limited to English-language articles, and the selection process involved two reviews and arbitration in case of disagreement among reviewers. RESULTS The review identified 598 articles initially, narrowing down to 11 reports after applying inclusion and exclusion criteria. The current research domain is remarkably active, playing a pivotal role in the field of exercise health, and studies predominantly focus on integrating LLMs and Chatbots with Just-In-Time Adaptive Interventions (JITAIs) to design personalized exercise programs and evaluate the quality of AI-generated exercise recommendations. The primary LLMs utilized in this field are predominantly from the ChatGPT series developed by OpenAI. CONCLUSIONS Notably, the review underscores the need for further advancements to match the expertise of exercise professionals and address individualized, evidence-based physical activity recommendations. While LLMs and Chatbots demonstrate promising utility in exercise recommendations and physical activity recommendations, there is a critical need for further development to enhance their accuracy, adaptability, and alignment with professional healthcare standards. The future direction emphasizes the fine-tuning of models, enhanced integration of wearable technology with AI and dynamic feedback loops in optimizing personalized health interventions.

12 April 2025 in JMIR Preprints
Verified
Preprint
Open access
XLXiangxun Lai
Xiangxun Lai
YLYue Lai
Yue Lai
QGQi Gao
Qi Gao
CHCaihua Huang
Caihua Huang
UNSTRUCTURED Artificial intelligence (AI) holds significant potential to transform exercise health through applications like personalized training, injury prevention, and targeted rehabilitation. However, the scarcity of high-quality, real-world data—due to issues like data fragmentation, inconsistent formats, privacy constraints, and confidentiality concerns—severely restricts AI development and deployment in this domain. Learning from advancements in healthcare AI, strategies such as federated learning for privacy-preserving collaboration, synthetic data generation to augment limited datasets, transfer learning to leverage existing models, and structured data-sharing protocols offer viable solutions. Furthermore, an innovative closed-loop data generation process is proposed: leveraging large language model APIs to create extensive synthetic datasets from initial high-quality inputs, followed by rigorous human expert validation and refinement through real-world feedback. This paper argues that overcoming the data bottleneck necessitates an integrated strategy combining these methods. Implementing such comprehensive measures will enhance AI model robustness and generalizability, ultimately accelerating the delivery of tangible benefits to athletes and health practitioners.