Fulin Pu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Fulin Pu has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Xiaoying Wei
Nursing Department, The Third Affiliated Hospital, Sun Yat-Sen University, Guangzhou, Guangdong, China
2 shared publications - Yanling Chen
Department of Anesthesiology, Hospital of Stomatology, Guanghua School of Stomatology, Sun Yat-Sen University, 56 Lingyuan West Road, Guangzhou, Guangdong, 510055, China, 86 02083871483
2 shared publications - Haifen Liao
School of Nursing, Sun Yat-sen University, Guangzhou, China
1 shared publication - Jishuang Hong
Department of Anesthesiology, Hospital of Stomatology, Guanghua School of Stomatology, Sun Yat-sen University, Guangzhou, Guangdong, 510060, People’s Republic of China
1 shared publication - Jishuang Hong
Department of Anesthesiology, Hospital of Stomatology, Guanghua School of Stomatology, Sun Yat-Sen University, 56 Lingyuan West Road, Guangzhou, Guangdong, 510055, China, 86 02083871483
1 shared publication - Lijuan Zhang
State Key Laboratory of Oncology in South China, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, 651 Dongfeng East Road, Guangzhou, 510060, China, 86 02087343852
1 shared publication
Latest Publications
- Top Cited PapersOpen Access
- 18 June 2026
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 14 (1) , e93054
Evaluation of Five Large Language Models for Parental Education in Pediatric Anesthesia: Reliability and Readability Study
Background: Although large language models (LLMs) show potential for patient education, their accuracy, usability, and comprehensibility lack validation in high-risk pediatric anesthesia. Rigorous evaluation is therefore essential prior to widespread clinical use in perioperative parental anesthesia education. Objective: This study aims to evaluate the accuracy, reliability, and readability of responses generated by 5 LLMs to parental inquiries regarding pediatric anesthesia, and to assess their suitability for clinical use in perioperative caregiver education. Methods: Two expert anesthesiologists identified 33 parental questions on pediatric anesthesia by screening authoritative resources and Google Trends. On December 14, 2025, these questions were submitted to 5 LLMs (DeepSeek-V3.2, ChatGPT-5, Gemini 2.5 Flash, Copilot, and Perplexity) via official web interfaces with default settings and zero-shot prompting, with each query in a separate conversation. Responses were standardized for blinded assessment. Two pediatric anesthesiologists with ≥10 years of clinical experience independently evaluated accuracy and reliability using the 4-point Likert accuracy scale, DISCERN, Ensuring Quality Information for Patients (EQIP), Journal of the American Medical Association ( JAMA ) benchmark, and Global Quality Score (GQS). After text preprocessing, readability was evaluated using 6 algorithms (Automated Readability Index [ARI], Flesch Reading Ease Score [FRES], Gunning Fog Index [GFI], Flesch-Kincaid Grade Level [FKGL], Coleman-Liau Index [CL], and the Simple Measure of Gobbledygook [SMOG]) via an online calculator. Interrater reliability was analyzed using the intraclass correlation coefficient (ICC); differences across models were assessed with the Kruskal-Wallis H test; and deviations from the sixth-grade benchmark were evaluated using 1-sample Wilcoxon signed-rank tests ( P <.05 considered significant). Results: All 5 LLMs demonstrated high clinical accuracy (>90%; P =.12), with Gemini reaching 100%. Nevertheless, safety risks and content hallucinations were still observed. Excluding Gemini and Copilot, the remaining 3 models (ChatGPT, DeepSeek, and Perplexity) each produced unsafe content in 3.03% (n=1) of the 33 queries. Hallucinations were detected in all models except Gemini, with DeepSeek and Perplexity showing the highest hallucination rate (3/33, 9.09%). Furthermore, Perplexity showed superior reliability on DISCERN (median 41; P <.05), yet no model achieved a “good” rating. Gemini achieved the highest EQIP (median 66.67%; P <.05) despite lower GQS (median 3). Transparency was universally poor ( JAMA median ≤1), with DeepSeek and ChatGPT showing a “floor effect.” ChatGPT had superior readability, but all models exceeded the recommended 6-grade complexity level. Conclusions: In this study, 5 LLMs generally provided clinically accurate information when responding to parental questions about pediatric anesthesia. However, limitations were also identified, including hallucinated content, safety-related deficiencies, limited source transparency, and readability levels exceeding recommended standards. Therefore, LLM-generated information should be interpreted with caution and should not replace clinician guidance.
- 18 June 2026
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 14 (1) , e93054
Top Cited PapersOpen AccessEvaluation of Five Large Language Models for Parental Education in Pediatric Anesthesia: Reliability and Readability Study
Background: Although large language models (LLMs) show potential for patient education, their accuracy, usability, and comprehensibility lack validation in high-risk pediatric anesthesia. Rigorous evaluation is therefore essential prior to widespread clinical use in perioperative parental anesthesia education. Objective: This study aims to evaluate the accuracy, reliability, and readability of responses generated by 5 LLMs to parental inquiries regarding pediatric anesthesia, and to assess their suitability for clinical use in perioperative caregiver education. Methods: Two expert anesthesiologists identified 33 parental questions on pediatric anesthesia by screening authoritative resources and Google Trends. On December 14, 2025, these questions were submitted to 5 LLMs (DeepSeek-V3.2, ChatGPT-5, Gemini 2.5 Flash, Copilot, and Perplexity) via official web interfaces with default settings and zero-shot prompting, with each query in a separate conversation. Responses were standardized for blinded assessment. Two pediatric anesthesiologists with ≥10 years of clinical experience independently evaluated accuracy and reliability using the 4-point Likert accuracy scale, DISCERN, Ensuring Quality Information for Patients (EQIP), Journal of the American Medical Association ( JAMA ) benchmark, and Global Quality Score (GQS). After text preprocessing, readability was evaluated using 6 algorithms (Automated Readability Index [ARI], Flesch Reading Ease Score [FRES], Gunning Fog Index [GFI], Flesch-Kincaid Grade Level [FKGL], Coleman-Liau Index [CL], and the Simple Measure of Gobbledygook [SMOG]) via an online calculator. Interrater reliability was analyzed using the intraclass correlation coefficient (ICC); differences across models were assessed with the Kruskal-Wallis H test; and deviations from the sixth-grade benchmark were evaluated using 1-sample Wilcoxon signed-rank tests ( P <.05 considered significant). Results: All 5 LLMs demonstrated high clinical accuracy (>90%; P =.12), with Gemini reaching 100%. Nevertheless, safety risks and content hallucinations were still observed. Excluding Gemini and Copilot, the remaining 3 models (ChatGPT, DeepSeek, and Perplexity) each produced unsafe content in 3.03% (n=1) of the 33 queries. Hallucinations were detected in all models except Gemini, with DeepSeek and Perplexity showing the highest hallucination rate (3/33, 9.09%). Furthermore, Perplexity showed superior reliability on DISCERN (median 41; P <.05), yet no model achieved a “good” rating. Gemini achieved the highest EQIP (median 66.67%; P <.05) despite lower GQS (median 3). Transparency was universally poor ( JAMA median ≤1), with DeepSeek and ChatGPT showing a “floor effect.” ChatGPT had superior readability, but all models exceeded the recommended 6-grade complexity level. Conclusions: In this study, 5 LLMs generally provided clinically accurate information when responding to parental questions about pediatric anesthesia. However, limitations were also identified, including hallucinated content, safety-related deficiencies, limited source transparency, and readability levels exceeding recommended standards. Therefore, LLM-generated information should be interpreted with caution and should not replace clinician guidance.
- Open Access
- 1 January 2026
- Published by Taylor & Francis in Journal of Multidisciplinary Healthcare
- Vol. ume 19, 1-17
Quality and Reliability of Deep Venous Thrombosis-Related Videos on TikTok and Bilibili
Deep venous thrombosis (DVT) is a prevalent, life-threatening condition with inadequate public awareness. Social media platforms are significant sources of health information. However, they often suffer from variable content quality. This study systematically evaluated DVT-related videos on TikTok and Bilibili, which represent the dominant short-video and medium-to-long video platforms in China. The top 150 DVT-related videos from each platform, TikTok and Bilibili, sorted by default, were retrieved and screened. The basic characteristics of the included videos, as well as the user engagement metrics, were recorded. The quality of the included videos was assessed using the Global Quality Scale (GQS), the modified DISCERN (mDISCERN) score, and the Journal of the American Medical Association (JAMA) benchmark criteria. TikTok videos had higher engagement (all P < 0.001) and shorter duration (median 88s vs 233.5s, P < 0.001). Quality-wise, except for a good GQS rating for Bilibili, quality scores across both platforms were generally moderate or poor. Specifically, Bilibili scored higher on GQS (median 4 vs 3, P < 0.001), while TikTok performed better on mDISCERN and JAMA (both P < 0.001). Furthermore, user engagement was negatively correlated with GQS scores but positively correlated with both mDISCERN and JAMA scores, revealing a potential mismatch between popularity and professional quality. Despite high user engagement, DVT videos on both platforms demonstrate deficient informational quality and reliability, underscoring a significant gap in accessible public health education. Therefore, multi-stakeholder collaboration is imperative to enhance content standards and facilitate effective health dissemination.
- 1 January 2026
- Published by Taylor & Francis in Journal of Multidisciplinary Healthcare
- Vol. ume 19, 1-17
Open AccessQuality and Reliability of Deep Venous Thrombosis-Related Videos on TikTok and Bilibili
Deep venous thrombosis (DVT) is a prevalent, life-threatening condition with inadequate public awareness. Social media platforms are significant sources of health information. However, they often suffer from variable content quality. This study systematically evaluated DVT-related videos on TikTok and Bilibili, which represent the dominant short-video and medium-to-long video platforms in China. The top 150 DVT-related videos from each platform, TikTok and Bilibili, sorted by default, were retrieved and screened. The basic characteristics of the included videos, as well as the user engagement metrics, were recorded. The quality of the included videos was assessed using the Global Quality Scale (GQS), the modified DISCERN (mDISCERN) score, and the Journal of the American Medical Association (JAMA) benchmark criteria. TikTok videos had higher engagement (all P < 0.001) and shorter duration (median 88s vs 233.5s, P < 0.001). Quality-wise, except for a good GQS rating for Bilibili, quality scores across both platforms were generally moderate or poor. Specifically, Bilibili scored higher on GQS (median 4 vs 3, P < 0.001), while TikTok performed better on mDISCERN and JAMA (both P < 0.001). Furthermore, user engagement was negatively correlated with GQS scores but positively correlated with both mDISCERN and JAMA scores, revealing a potential mismatch between popularity and professional quality. Despite high user engagement, DVT videos on both platforms demonstrate deficient informational quality and reliability, underscoring a significant gap in accessible public health education. Therefore, multi-stakeholder collaboration is imperative to enhance content standards and facilitate effective health dissemination.
- Open Access
- 1 January 2023
- Published by Wiley in European Journal of Cancer Care
- Vol. 2023 (1)
Factors Associated with Adherence to Complete Decongestion Therapy in Women with Breast Cancer‐Related Lymphedema
Background . Adherence to complete decongestion therapy (CDT) is of utmost importance for patients with breast cancer‐related lymphedema (BCRL). However, past studies have demonstrated that adherence to CDT in BCRL patients is less than satisfactory. Our study aimed to describe the level of adherence to CDT among BCRL patients and utilize the Health Belief Model to analyse the key factors that influence adherence. Methods . We employed a convenient sampling method to select 158 patients with BCRL who were in the maintenance stage of CDT in a grade A tertiary tumour hospital in Guangzhou. A questionnaire survey was used to investigate patients’ adherence to CDT, as well as their sociodemographic characteristics, medical history characteristics, disease knowledge, and health beliefs. Results . The CDT adherence score of BCRL patients was 96.46 ± 18.46 points, with an average score of 2.61 ± 0.50 points per item. The adherence rate was 65.18 ± 12.47%, indicating a moderate level of adherence. Factors that significantly influenced CDT adherence in BCRL patients included age under 45 years, having 0 or 1 child, experiencing 9 lymphedema symptoms, and having low levels of disease knowledge and health beliefs ( p < 0.05). Conclusions . The level of adherence to CDT in BCRL patients is moderate and requires improvement. To enhance adherence to CDT, medical professionals should focus on patients who are younger, have fewer children, exhibit fewer symptoms of lymphedema, possess limited disease knowledge, and have weak health beliefs.
- 1 January 2023
- Published by Wiley in European Journal of Cancer Care
- Vol. 2023 (1)
Open AccessFactors Associated with Adherence to Complete Decongestion Therapy in Women with Breast Cancer‐Related Lymphedema
Background . Adherence to complete decongestion therapy (CDT) is of utmost importance for patients with breast cancer‐related lymphedema (BCRL). However, past studies have demonstrated that adherence to CDT in BCRL patients is less than satisfactory. Our study aimed to describe the level of adherence to CDT among BCRL patients and utilize the Health Belief Model to analyse the key factors that influence adherence. Methods . We employed a convenient sampling method to select 158 patients with BCRL who were in the maintenance stage of CDT in a grade A tertiary tumour hospital in Guangzhou. A questionnaire survey was used to investigate patients’ adherence to CDT, as well as their sociodemographic characteristics, medical history characteristics, disease knowledge, and health beliefs. Results . The CDT adherence score of BCRL patients was 96.46 ± 18.46 points, with an average score of 2.61 ± 0.50 points per item. The adherence rate was 65.18 ± 12.47%, indicating a moderate level of adherence. Factors that significantly influenced CDT adherence in BCRL patients included age under 45 years, having 0 or 1 child, experiencing 9 lymphedema symptoms, and having low levels of disease knowledge and health beliefs ( p < 0.05). Conclusions . The level of adherence to CDT in BCRL patients is moderate and requires improvement. To enhance adherence to CDT, medical professionals should focus on patients who are younger, have fewer children, exhibit fewer symptoms of lymphedema, possess limited disease knowledge, and have weak health beliefs.