Kaijaing Pan

Dr. Kaijaing Pan


Research Focus & Metrics

Academic Career Length

2Years (2025-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.

15 June 2026 in Healthcare
Verified
Journal Article
Open access
PDF file
KPKaijaing Pan
Kaijaing Pan
CHCaihua Huang
Caihua Huang
XLXinyu Lin
Xinyu Lin
SHShengqi Huang
Shengqi Huang
Background: Large language models and other generative artificial intelligence (AI) tools are increasingly being embedded in digital healthcare services, including mobile health applications, telerehabilitation, remote monitoring, and hybrid care pathways. In this review, digital healthcare refers to technology-mediated healthcare services in which digital platforms, mobile applications, wearables, remote communication, and AI-enabled interfaces support health assessment, self-management, rehabilitation, clinical decision support, or service delivery. When AI-generated exercise guidance moves from general education to individualized recommendations about dose, progression, contraindications, or rehabilitation, it may become directly actionable and safety-relevant. Objectives: This review aimed to clarify when AI-generated exercise guidance in digital healthcare may warrant safety-relevant governance attention and to outline implementation considerations for explainability, human oversight, and service-level governance. It addresses a gap in the literature: general AI-governance and exercise-prescription discussions rarely specify how point-of-use explanations, review thresholds, and escalation safeguards can be organized for directly actionable AI exercise guidance. Methods: We conducted a governance-oriented narrative review of peer-reviewed literature and representative regulatory or guidance documents. This review was not designed as a systematic review, scoping review, or exhaustive evidence map; transparent source mapping was used to support conceptual synthesis. Searches and source mapping focused on generative AI, large language models, explainable AI, clinical decision support, digital health, mobile health, exercise prescription, rehabilitation, trust, automation bias, and human oversight. Sources were included when they informed the safety, explainability, governance, or real-world implementation of patient-facing AI-generated exercise guidance. Extracted material was grouped by evidentiary role and synthesized through framework synthesis and governance mapping to distinguish literature-supported observations, author interpretation, and proposed implementation tools. Results: The included sources were first organized into five thematic groups: digital exercise delivery and exercise-prescription evidence; explainability, trust, and automation bias literature; professional responsibility, ethics, and patient disclosure literature; regulatory and policy documents; and digital literacy, patient/clinician attitudes, and equity literature. The synthesis then proceeded from safety relevance to explanation needs, human oversight and escalation needs, and selected regulatory and policy signals before translating these strands into conceptual and implementation-oriented outputs rather than empirically validated instruments. AI-generated exercise guidance was most safety-relevant in scenarios involving individualized dose, progression, contraindication-sensitive action, or rehabilitation strategy. Across the included sources, generic transparency alone was not sufficient to support reviewable use; relevant explanation elements included evidence sources, risk warnings, reasoning paths, and reasonable alternatives. Oversight considerations varied with embodied risk, clinical ambiguity, user vulnerability, and likelihood of direct enactment. Implementation considerations linked interface design, clinical review, escalation, auditability, and post-deployment monitoring. Conclusions: AI-generated exercise guidance in digital healthcare may warrant governance attention as a patient-safety and accountability issue when it influences actionable exercise decisions. The proposed framework offers a conceptual basis for designing more reviewable and accountable mobile and remote exercise-support services. Future work can validate these outputs in patient-facing services, clinician review workflows, usability studies, implementation pilots, and safety evaluations.

1 July 2026 in Risk Management and Healthcare Policy
Verified
Journal Article
Open access
PDF file
KPKaijaing Pan
Kaijaing Pan
XLXinyu Lin
Xinyu Lin
SHShengqi Huang
Shengqi Huang
CHCaihua Huang
Caihua Huang
LLM-based exercise and health guidance creates a distinctive risk-management challenge: seemingly modest changes in product claims, target users, data inputs, personalization, automation, human oversight, or updates may move a tool from general wellness support toward higher-risk medical use. In exercise prescription and rehabilitation, an unsafe recommendation can affect physical load, recognition of warning symptoms, and timely referral. We conducted a structured narrative review and doctrinal/comparative legal analysis, anchored in China’s National Medical Products Administration (NMPA) framework and informed by the European Union Medical Device Regulation (MDR), the EU Artificial Intelligence Act, and US Food and Drug Administration (FDA) and International Medical Device Regulators Forum (IMDRF) materials. Peer-reviewed literature was primarily searched for 2019– 2026, with foundational regulatory, legal, and technical guidance included where directly relevant. We propose a six-dimensional trigger matrix covering intended use and claims, user context, depth of personalization, data and sensor sources, automation, human oversight and closed-loop control, and upgrade and change pathways. The framework includes anchored Green/Yellow/Red coding rules, non-compensatory aggregation rules, a structured governance checklist, and a lifecycle pathway for evidence generation, risk management, and change control. In a preliminary application exercise, three independent raters applied the coding rules to seven standardized hypothetical scenarios and achieved complete agreement on all dimension-level and overall designations (Fleiss’ kappa = 1.00). This small exercise supports initial reproducibility of the rubric but does not establish legal classification accuracy, clinical validity, or real-world effectiveness. The proposed framework is intended to support earlier risk identification, evidence planning, procurement review, and dialogue among developers, healthcare institutions, and regulators; it does not replace product-specific legal analysis or regulatory determination.

17 June 2025 in JMIR Preprints
Verified
Preprint
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
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.