Zhi Yang

Zhi Yang

Auto-generated profile, based on publicly available information

Zhi Yang

School of Nursing and Health Sciences, Hong Kong Metropolitan University, Kowloon, China

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

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

Representative Publications

View All Publications
1 December 2018 in Nurse Education Today
Verified
Journal Article
Closed access
ZYZhi Yang
Zhi Yang
YZYing Zhou
Ying Zhou
WCWai-Yee Chung
Wai-Yee Chung
QTQiubi Tang
Qiubi Tang
LJLian Jiang
Lian Jiang
TWThomas K.S. Wong
Thomas K.S. Wong

10 January 2025 in JMIR Medical Informatics
Verified
Journal Article
Open access
Shiben ZHUShiben ZHU
Shiben ZHU
WHWanqin Hu
Wanqin Hu
ZYZhi Yang
Zhi Yang
JYJiani Yan
Jiani Yan
FZFang Zhang
Fang Zhang
Background: Large language models (LLMs) have been proposed as valuable tools in medical education and practice. The Chinese National Nursing Licensing Examination (CNNLE) presents unique challenges for LLMs due to its requirement for both deep domain–specific nursing knowledge and the ability to make complex clinical decisions, which differentiates it from more general medical examinations. However, their potential application in the CNNLE remains unexplored. Objective: This study aims to evaluates the accuracy of 7 LLMs including GPT-3.5, GPT-4.0, GPT-4o, Copilot, ERNIE Bot-3.5, SPARK, and Qwen-2.5 on the CNNLE, focusing on their ability to handle domain-specific nursing knowledge and clinical decision-making. We also explore whether combining their outputs using machine learning techniques can improve their overall accuracy. Methods: This retrospective cross-sectional study analyzed all 1200 multiple-choice questions from the CNNLE conducted between 2019 and 2023. Seven LLMs were evaluated on these multiple-choice questions, and 9 machine learning models, including Logistic Regression, Support Vector Machine, Multilayer Perceptron, k-nearest neighbors, Random Forest, LightGBM, AdaBoost, XGBoost, and CatBoost, were used to optimize overall performance through ensemble techniques. Results: Qwen-2.5 achieved the highest overall accuracy of 88.9%, followed by GPT-4o (80.7%), ERNIE Bot-3.5 (78.1%), GPT-4.0 (70.3%), SPARK (65.0%), and GPT-3.5 (49.5%). Qwen-2.5 demonstrated superior accuracy in the Practical Skills section compared with the Professional Practice section across most years. It also performed well in brief clinical case summaries and questions involving shared clinical scenarios. When the outputs of the 7 LLMs were combined using 9 machine learning models, XGBoost yielded the best performance, increasing accuracy to 90.8%. XGBoost also achieved an area under the curve of 0.961, sensitivity of 0.905, specificity of 0.978, F1-score of 0.901, positive predictive value of 0.901, and negative predictive value of 0.977. Conclusions: This study is the first to evaluate the performance of 7 LLMs on the CNNLE and that the integration of models via machine learning significantly boosted accuracy, reaching 90.8%. These findings demonstrate the transformative potential of LLMs in revolutionizing health care education and call for further research to refine their capabilities and expand their impact on examination preparation and professional training.

12 August 2024 in International Journal of Nursing Sciences
Verified
Journal Article
Open access
ZYZhi Yang
Zhi Yang
MTMimi Mun Yee Tse
Mimi Mun Yee Tse
HHHuiting Huang
Huiting Huang
HFHaiyun Fang
Haiyun Fang
JCJoanne Wai Yee Chung
Joanne Wai Yee Chung
TSThomas Kwok Shing Shing
Thomas Kwok Shing Shing
This study aimed to assess the feasibility of an online compassion training program for nursing students and preliminarily investigate its effects on mindfulness, self-compassion, and stress reduction. This study employed a randomized controlled trial design. Second-year students from a nursing college in Guangzhou, China, were recruited as research participants in August 2023. The intervention group participated in an 8-week online compassion training program via the WeChat platform, comprising three stages: mindfulness (weeks 1–2), self-compassion (weeks 3–5), and compassion for others (weeks 6–8). Each stage included four activities: psychoeducation, mindfulness practice, weekly diary, and emotional support. Program feasibility was assessed through recruitment and retention rates, program engagement, and participant acceptability. Program effectiveness was measured with the Mindful Attention Awareness Scale, Self-Compassion Scale - Short Form, and Perceived Stress Scale. A total of 28 students completed the study (13 in the intervention group, 15 in the control group). The recruitment rate was 36.46%, with a high retention rate of 93.3%. Participants demonstrated high engagement: 69.2% accessed learning materials every 1–2 days, 93.3% practiced mindfulness at least weekly, with an average of 4.69 diary entries submitted per person and 23.30 WeChat interactions with instructors. Regarding acceptability, all participants expressed satisfaction with the program, with 92.4% finding it “very helpful” or “extremely helpful.” In terms of intervention effects, the intervention group showed a significant increase in mindfulness levels from pre-intervention (51.54 ± 10.93) to post-intervention (62.46 ± 13.58) (P < 0.05), while no significant change was observed in the control group. Although there were no statistically significant differences between the two groups in post-intervention self-compassion and perceived stress levels, the intervention group showed positive trends: self-compassion levels increased (35.85 ± 8.60 vs. 40.85 ± 5.54), and perceived stress levels slightly decreased (44.77 ± 8.65 vs. 42.00 ± 5.77). This pilot study demonstrated the feasibility of an online compassion training program for nursing students and suggested its potential effectiveness in enhancing mindfulness, self-compassion, and stress reduction. Despite limitations such as small sample size and lack of long-term follow-up, preliminary evidence indicates promising prospects for integrating such training into nursing education. Further research is warranted to confirm these findings and assess the sustained impact of this approach on nursing education and practice.

11 September 2023 in Geriatric Nursing
Verified
Journal Article
Closed access
ZYZhi Yang
Zhi Yang
MTMimi Mun Yee Tse
Mimi Mun Yee Tse
JCJoanne Wai Yee Chung
Joanne Wai Yee Chung
DCDoris Yin Kei Chong
Doris Yin Kei Chong
TWThomas Kwok Shing Wong
Thomas Kwok Shing Wong

27 June 2024 in JMIR Preprints
Verified
Preprint
Open access
Shiben ZHUShiben ZHU
Shiben ZHU
WHWanqin Hu
Wanqin Hu
ZYZhi Yang
Zhi Yang
JYJiani Yan
Jiani Yan
FZFang Zhang
Fang Zhang
BACKGROUND Large language models (LLMs) have been proposed as valuable tools in medical education and practice. The Chinese National Nursing Licensing Examination (CNNLE) presents unique challenges for LLMs due to its requirement for both deep domain–specific nursing knowledge and the ability to make complex clinical decisions, which differentiates it from more general medical examinations. However, their potential application in the CNNLE remains unexplored. OBJECTIVE This study aims to evaluates the accuracy of 7 LLMs including GPT-3.5, GPT-4.0, GPT-4o, Copilot, ERNIE Bot-3.5, SPARK, and Qwen-2.5 on the CNNLE, focusing on their ability to handle domain-specific nursing knowledge and clinical decision-making. We also explore whether combining their outputs using machine learning techniques can improve their overall accuracy. METHODS This retrospective cross-sectional study analyzed all 1200 multiple-choice questions from the CNNLE conducted between 2019 and 2023. Seven LLMs were evaluated on these multiple-choice questions, and 9 machine learning models, including Logistic Regression, Support Vector Machine, Multilayer Perceptron, k-nearest neighbors, Random Forest, LightGBM, AdaBoost, XGBoost, and CatBoost, were used to optimize overall performance through ensemble techniques. RESULTS Qwen-2.5 achieved the highest overall accuracy of 88.9%, followed by GPT-4o (80.7%), ERNIE Bot-3.5 (78.1%), GPT-4.0 (70.3%), SPARK (65.0%), and GPT-3.5 (49.5%). Qwen-2.5 demonstrated superior accuracy in the Practical Skills section compared with the Professional Practice section across most years. It also performed well in brief clinical case summaries and questions involving shared clinical scenarios. When the outputs of the 7 LLMs were combined using 9 machine learning models, XGBoost yielded the best performance, increasing accuracy to 90.8%. XGBoost also achieved an area under the curve of 0.961, sensitivity of 0.905, specificity of 0.978, F1-score of 0.901, positive predictive value of 0.901, and negative predictive value of 0.977. CONCLUSIONS This study is the first to evaluate the performance of 7 LLMs on the CNNLE and that the integration of models via machine learning significantly boosted accuracy, reaching 90.8%. These findings demonstrate the transformative potential of LLMs in revolutionizing health care education and call for further research to refine their capabilities and expand their impact on examination preparation and professional training.

21 September 2026 in BMC Nursing
Verified
Journal Article
Open access
PDF file
ZYZhi Yang
Zhi Yang
MTMimi Mun Yee Tse
Mimi Mun Yee Tse
HHHuiting Huang
Huiting Huang
HFHaiyun Fang
Haiyun Fang
YWYuan Wang
Yuan Wang
TWThomas Kwok Shing Wong
Thomas Kwok Shing Wong
Nursing students experience elevated levels of stress and mental health challenges, highlighting the need for accessible and scalable interventions to supp