Background and Objective: Generic benchmarks rarely capture specialized clinical reasoning that requires heterogeneous evidence and multistep inference. We developed a knowledge graph-constrained multi-agent workflow, named MAC-TCM, for constructing benchmarks in traditional Chinese medicine, and instantiated it as ZBN-Bench for the Zhao Bingnan academic school of dermatology.Methods: A domain knowledge base was assembled from over 1000 curated clinical cases, 54 books, over 1000 full-text articles, and over 3000 indexed records, and encoded as a graph of 3300 nodes and 12,370 edges along the symptom-syndrome-disease-treatment-formula-herb chain. Candidate items were produced under graph-path and question-direction constraints, filtered by answer-reachability validation and four-dimensional soft scoring. An external sample of 300 items was independently rated by four domain experts, and fourteen large language models were evaluated under a prespecified zero-shot protocol.Results: Entity extraction against manual review reached 92.4% precision and 89.6% sensitivity (Cohen’s kappa 0.85). The automatic quality filter showed high specificity against low-quality candidates, reducing factual errors from 27.2% under unconstrained generation to 5.4%. The final benchmark contains 1500 items (750 single-choice, 450 multiple-choice, 300 chained-choice), with an expert consensus score of 3.802 out of 4.0 and inter-rater Fleiss’ kappa 0.78. Automatic scoring agreed closely with expert consensus (mean absolute error 0.035). The best-performing model reached 62.30% overall accuracy, with degradation on high-difficulty (26.54%) and chained-choice (18.62%) items.Conclusions: The workflow provides a traceable software pipeline and a reproducibility-oriented protocol for constructing specialty benchmarks in traditional Chinese medicine. ZBN-Bench offers a school-specific evaluation resource for model benchmarking and methodological research, although broader external validation remains necessary before downstream clinical interpretation.

To analyze voice signals and identify asthma patients using voice signal analysis and machine learning techniques, we collected clear, low-noise fixed-pattern voice signals from 50 asthma patients and 50 healthy controls to build an analysis database. The research conducted multi-dimensional voice signal analysis based on MATLAB and selected voice feature indicators with significant differences between asthma patients and healthy controls. After dimensionality reduction analysis on differential phonetic features, the processed features were incorporated into subsequent SVM and RF modeling and classification research. The study established over 400 voice feature indicators related to diagnosis, of which 20 indicators showed significant differences between asthma patients and healthy controls (P < 0.01). In the classification study, both the SVM and RF models achieved identical accuracy rates of 87% on the test set, with AUC values of 0.95 for SVM and 0.93 for RF. This demonstrates their comparable performance in terms of overall classification accuracy, while the disparity in AUC values suggests that the SVM model may achieve a better trade-off between sensitivity and specificity. Thus, this paper not only provides a new method for non-invasive early detection of asthma but also lays the foundation for further application and optimization of this method in real-world settings.

The objective of this study was to develop and evaluate a non-invasive method for distinguishing patients with chronic obstructive pulmonary disease (COPD) from those with respiratory tract infections (RTI) using voice signal analysis and machine learning. Fixed-pattern voice signals were collected from 25 COPD patients and 25 RTI patients (serving as the control/comparison group). Multi-dimensional voice feature analysis was performed to identify features significantly differentiating the two groups. Statistically significant features were selected and subjected to dimensionality reduction. Logistic Regression (LR) and Random Forest (RF) models were then trained and evaluated for classification performance in distinguishing COPD from RTI. Over 400 voice features were initially analyzed. Eighteen features showed highly significant differences between COPD and RTI patients (P <; 0.05). In the task of distinguishing COPD patients from RTI patients, the LR model achieved a test set area under the curve AUC of 0.95, significantly outperforming the RF model (AUC = 0.76). This study demonstrates the feasibility of using voice analysis and machine learning, particularly the LR model, as a promising non-invasive tool for differentiating COPD from RTI. It provides a foundation for the practical application and further optimization of this voice-based approach in clinical settings requiring differential diagnosis of respiratory conditions.

This paper briefly introduces the design and implementation of an intelligent traditional Chinese medicine Q A system based on traditional Chinese medicine knowledge graph and large language models. By integrating the traditional Chinese medicine knowledge graph and natural language processing technology, the system aims to provide efficient traditional Chinese medicine Q A services.

The task of information extraction is confronted with many difficulties in the field of traditional Chinese medicine. On top of that, existing methods, including traditional neural network and emerging large language model-assisted information extraction, show low accuracy and recall rates. In this paper, an instruction fine-tuning paradigm is proposed for traditional Chinese medicine text information extraction, which has achieved ideal results on data sets.

To provide proof of the evidence-based medicine and decision-making information for the clinical decision of functional gastrointestinal disorders(FGIDs), this study evaluated and compared the efficacy, safety, and economy of four oral Chinese patent medicines(CPMs) in the treatment of FGIDs using the method of rapid health technology assessment. The literature was systematically retrieved from CNKI, Wanfang, VIP, SinoMed, EMbase, PubMed, Cochrane Library and ClinicalTrials.gov from the establishment of the databases to May 1, 2022. Two evaluators screened out the literature, extracted data, evaluated the quality of the literature, and descriptively analyzed the results according to the prepared standard. Eventually, 16 studies were included, all of which was rando-mized controlled trial(RCT). The results showed that Renshen Jianpi Tablets, Renshen Jianpi Pills, Shenling Baizhu Granules, and Buzhong Yiqi Granules all had certain effects on the treatment of FGIDs. Renshen Jianpi Tablets treated FGIDs and persistent diarrhea. Shenling Baizhu Granules treated diarrhea with irritable bowel syndrome and FGIDs. Buzhong Yiqi Granules treated diarrhea with irritable bowel syndrome, FGIDs, and chronic diarrhea in children. Renshen Jianpi Pills treated chronic diarrhea. The four oral CPMs all have certain effects on the treatment of FGIDs and have specific advantages for specific patients. Compared with other CPMs, Renshen Jianpi Tablets have higher clinical universality. However, there are problems such as insufficient clinical research evidence, generally low quality of evidence, lack of comparative analysis among medicines, and lack of academic evaluation. More high-quality clinical research and the economic research should be carried out in the future, so as to provide more evidence for the evaluation of the four CPMs.

Objective To systematically analyze the clinical research evidence on traditional Chinese medicine（TCM） in the prevention and treatment of chronic atrophic gastritis（CAG）, summarize the efficacy, and provide guidance for future research. Methods With the help of the big data platform of evidence-based literature of traditional Chinese medicine which covers multiple dominant digestive diseases of TCM, the clinical researches, systematic reviews, guidelines and consensuses on the prevention and treatment of CAG by TCM published between the year 2000 and 2020 were searched for. The publication time, research type, sample size and intervention duration, the distribution of TCM syndromes and constitution in the observational studies, the category and role of prevention and treatment TCM programs（prevention, treatment, rehabilitation/secondary prevention, uncertainty）, clinical evaluation indicators and other information of the included articles were statistically analyzed, and the distribution of evidence was presented in the form of both graphics and text. Results A total of 796 articles were included, involving 775 clinical studies, 19 systematic reviews and 2 expert consensus. Total number of annual publications was slowly moving upward, with highest volume published in 2015 and 2019, both of which were 51. The clinical researches were mainly interventional studies（714 articles）, including 518 randomized controlled trials and 196 non-randomized controlled trials; besides, there were 61 observational studies. The sample size in 46. 50%（332/714） of the interventional studies was 60～100 cases, and the duration of the intervention in 47. 48%（339/714） of the intervention studies was 3～6 months. The intervention explored in the included studies was mainly Chinese herbal medicine. The focused outcome indicators were the total effective rate, endoscopic appearance, histopathology, TCM syndrome scores, clinical symptom improvement, and Hp negative conversion rate, while less attention was paid to the quality of life and psychological evaluation. The methodological quality of systematic reviews was generally low, with the AMSTAR scale scores mainly at 4 to 8 points, and 16 systematic reviews（84. 21%） concluded that "evidence shows potential efficacy".Conclusion TCM has certain advantages in the treatment of CAG, but most evidence have deficiencies in the rationality of research design, the uniformity of diagnostic criteria, and the scientificity of efficacy evaluation. In the future, it is necessary to strengthen the standardization of clinical diagnosis and treatment, and carry out rigorous large sample size, multi-center, randomized controlled trials to improve the evidence of TCM in the prevention and treatment of CAG.

Medical cases are an important mean to record the clinical treatment process for TCM doctors. With the help of big data processing technology and other means, digging and using of medical cases information is of great significance for the inheritance of TCM. TCM Cases Cloud integrates voice recognition, optical character recognition (OCR), four-diagnosis instrument of TCM and other intelligent information collection technologies, as well as abundant algorithms such as association analysis, Bayes, hierarchical clustering, etc. to dig the laws of syndrome differentiation, selection of prescriptions, medication, acupoint, etc. of famous TCM doctors in medical records. This article briefly introduced the platform from five aspects: the excavation of the experience of famous TCM doctors in modern times, the summary of the academic thoughts of ancient medical cases, the analysis of the law of diagnosis and treatment of specialized diseases, the regularity of a single drug and drug pairs, and the exploration of the compatibility of TCM prescriptions.

该研究旨在探讨药食同源组方联合益生菌对环磷酰胺(CTX)诱导免疫低下小鼠免疫功能的调节作用.除空白对照组用生理盐水灌胃外,其余各组小鼠腹腔注射环磷酰胺(CTX)100 mg/kg(按小鼠体质量计),连续腹腔注射4 d.将建模成功体重相近的小鼠随机分为6组:正常对照组、免疫抑制组、药食同源组方Ⅲ号、药食同源改良组(Ⅳ号)、益生菌组(Ⅴ号)和Ⅳ和Ⅴ联用组(Ⅵ号).连续灌胃给药14 d,观察Ⅲ、Ⅳ、Ⅴ和Ⅵ对小鼠胸腺指数、脾脏指数、吞噬能力、脾淋巴细胞增殖、NK细胞活性以及血清中IgG、IgA和IgM生成水平的影响.结果显示:Ⅲ、Ⅳ、Ⅴ和Ⅵ组均能显著提高模型小鼠胸腺指数(1.46、1.67、1.50、1.92 g/kg)、脾脏指数(2.32、2.54、2.52、2.82 g/kg)、吞噬能力(0.43、0.45、0.44、0.62)、脾淋巴细胞增殖作用(1.39、1.45、1.42、1.43)、NK细胞活性(0.35、0.43、0.36、0.71)(P<0.05);Ⅳ、Ⅴ和Ⅵ均能促使IgG(224.21、163.21、253.35 mg/L)、IgA(34.86、22.09、41.17 mg/L)和IgM(84.01、78.23、100.28 mg/L)生成水平均明显提高(P<0.05).研究结果表明,药食同源改良组方与益生菌联用组对环磷酰胺致免疫低下小鼠的免疫能力具有显著增强作用.

目的 从安全性、有效性、经济性、创新性、适宜性、可及性6个维度对人参健脾片治疗功能性胃肠病(FGIDs)进行综合临床评价,为国家基本药物遴选提供依据.方法 采用名义群体法、访谈法和调查问卷法等多种方式构建临床综合评价指标体系,通过文献检索、问卷调查、企业资料搜集等方法获取人参健脾片及3个同类药物(参苓白术颗粒、补中益气颗粒、人参健脾丸)临床证据.采用多准则决策分析模型对药物的临床价值进行综合评估,采用层次分析法计算准则层、指标层、备选方案权重.采用等频离散化原则对评价结果进行分级.结果 安全性证据表明,人参健脾片不良反应主要表现为轻微腹胀和头晕,无严重不良反应,基于现有研究,认为安全性证据较充分,风险较可控,安全性评为A级;临床研究表明,人参健脾片治疗FGIDs在改善腹泻症状方面较布拉氏酵母菌效果好,在改善餐后饱胀不适感、腹部疼痛和慢性腹泻症状体征方面较马来酸曲美布汀片效果好,有效性评为A级;人参健脾片与参苓白术颗粒、补中益气颗粒、人参健脾丸相比,日用药费最低,经济性评为A级;人参健脾片能改善FGIDs的多种临床症状,并且在药品制备工艺方面有多项专利,创新性强,评为A级;在药品的临床使用和患者依从性等方面,人参健脾片问卷调查得分较高,适宜性评为A级;药品药材供应可持续,价格低廉,疗程费用低,销售范围广,购买方便,可及性评为A级.结论 人参健脾片治疗FGIDs的临床价值评A类,建议该药可按程序转化为基本药物目录用药.

研究中医体质辨识方法,中医四诊客观化方法,基于深度学习研究健康状态辨识和智能推荐的算法模型,参考中医养生知识库构建方案库,利用互联网实现多终端的健康服务,研发智能化中医健康管理云平台,通过微信小程序、移动平板和Web服务3种模式,提供具有中医特色的健康管理.

建立"三结合"的中药注册审评证据体系,中医理论证据评价体系的形成至关重要.现有的中医理论证据体系尚不成熟,对于古代文献的检索缺乏规范的检索步骤且证据的评价要素不够全面,为此对古代文献的检索环节进行规范,提出DSPMRE六要素以明确检索问题和需求,在证据评价方面创新性提出文献的完整性,将文献来源、文献的共识性共同提炼为证据评价要素,初步形成中医理论证据等级推荐意见,以此完善文献检索、分析评价、分级推荐3 个关键环节,构建"三结合"体系下支撑中药新药研发和注册审批的中医理论证据体系.

中医疾病术语标准是科研与临床工作的重要基石,为了更好地了解中医疾病术语标准的变化,本文对比研究了《中医临床诊疗术语疾病部分》(GB/T16751-1997)(下文简称 97 版)和《中医临床诊疗术语第 1 部分:疾病修订版(2020)》(下文简称 20版)两版国家标准后发现,20 版标准新增的类病描述使术语的分类更加详细准确,有利于数据的分科统计;新增骨伤科、妇产科、儿科等临床常见的西医疾病为中医病名,着力于中医专科疾病名的补充与细化;删除了"早孕"和"多胎"2 个生理性术语.但 20 版在概念词的选择中忽略了临床实用性,如将"痹证类病"作为概念词,"痹证"作为入口词;部分易混淆的中医病名的存在也影响了标准的易用性,如胃胀病、胃痞病、胃疡病、胃络痛、胃痛等.因此,在未来的中医疾病标准研制中,建议增加其相对应的西医病名,方便病历书写和临床数据规范处理;同时参考实用性、普及性、准确性选择疾病标准词,避免将常用词作为可选用词,而应将其作为概念词;减少易混淆中医病名的存在,降低术语使用错误率,促进标准的推广应用.

目的:构建"药-病-证-效-结局指标"的临床应用证据链模型,实现文献检索、证据筛选、结局指标的量化和提取,为脾胃病临床决策提供循证支持.方法:基于循证证据金字塔结构和中医辨证论治特点,将随机对照临床研究、名医经验等类型文献中的信息结构化,以药物疗效评价为核心,采用关联数据技术,对全部证据建立关联关系.结果:脾胃病循证证据链实现了文献自动检索、自动分类,文献质量自动评估,结局指标量化、提取及查询等功能,显著提高了循证检索系统的检索效率,帮助用户在指标层面上浏览中医证据,发现证据之间潜在的联系.结论:脾胃病循证证据链增强了中医领域证据资源的联通性,可以为用户提供实时、准确、个性化的知识服务.

目的 应用数据挖掘技术分析中成药治疗便秘的组方规律,为临床提供参考依据.方法 收集中医药知识服务平台中治疗便秘的中成药,应用Excel2019建立数据库,并统计中药使用频次,分析药性、药味、归经及功效分类,运用SPSS Statistics 21.0对频次≥15的中药进行聚类分析,运用SPSS Modeler 18.0对所有中药进行关联规则分析.结果 纳入中成药109种,涉及中药225味,使用频次>10的中药有22味,如大黄、黄芩、当归、牵牛子等;药性以寒、温、微寒、平为多,药味以苦、辛、甘为多,归经以脾、肝、胃经为主.关联规则分析显示,大黄-黄芩、大黄-陈皮支持度较高.聚类分为5类.结论 治疗便秘中成药的核心药物为大黄、黄芩、当归等;常用配伍有大黄-黄芩、大黄-陈皮等,治疗按便秘类型辨证论治,可运用清热通便、顺气通便、润肠通便、利湿通便等治法.

The meridian theory is the pioneer of clinical diagnosis and treatment of diseases in Traditional Chinese Medicine (TCM). From Shang Han Lun to Pi Wei Lun, the meridian theory has contributed important theoretical organization materials and clinical practice experience to the establishment of the diagnosis system of external and internal injuries. The acupoints contained in its clinical acupuncture and moxibustion record symptoms, and some laws summarized have been absorbed and used for reference. It shows the positive significance of its exploration in clinical diagnosis and treatment. A system of differentiation and treatment of external and internal injuries with acupuncture has not been formed, even though the meridian theory of TCM has a long history with many areas being explored, such as diseases, acupoints, acupuncture methods and stimulation amount. Therefore, this paper starts from the academic development history of meridians, reviews and analyzes the contribution and limitations of TCM acupuncture and moxibustion in the diagnosis and treatment of internal injury, in order to enlighten the current study and understanding of TCM.

中医临床用药禁忌种类繁多,审核业务复杂.针对上述难题,探索了中医临床用药禁忌规则构建方法与应用模式,研发了中医临床用药禁忌智能审核系统,系统功能由基础服务、规则库管理维护、数据资源管理维护与统计监测 4 部分构成,采用浏览器/服务器架构实现,该系统将中医临床用药禁忌规则从应用程序代码中分离出来,可让临床药师、医师对系统内中医临床用药禁忌规则进行灵活配置,具有良好的实用性与先进性.

高脂血症是临床的常见病,是心、脑等重要靶器官动脉粥样硬化的高危因素之一.中医治疗高脂血症理论基础丰富,治疗方式多样,疗效确切且安全性高.本文从单味中药、中药复方、针灸埋线、食疗、运动疗法方面总结中医治疗高脂血症的相关临床研究,为该病临床治疗和后续研究提供参考.

目的 利用古今医案云平台(V2.3.2)整理并挖掘名老中医治疗慢性胃炎的用药规律.方法 收集古今医案云平台(V2.3.2)名医医案库中75位名老中医及其书籍中的医案数据,以古今医案云平台(V2.3.2)进行用药频次统计分析、聚类分析、复杂网络分析.结果 共纳入医案423个,涉及中药269味,主要有陈皮、柴胡、茯苓、法半夏、白术、白芍、黄连、黄芩、甘草、党参、砂仁、木香;药性以温、平、微寒为主;药味多属辛、苦、甘;归经主归脾、胃、肺、肝经.高频中药系统聚类分析聚为4组,C1:柴胡、黄芩、法半夏、黄连;C2:陈皮、茯苓、党参、白术、甘草、白芍、砂仁、木香.C2可进一步分为C3与C4,C3:党参、白术、甘草、白芍、砂仁、木香;C4:陈皮、茯苓.复杂网络分析得到4组中医证候及其核心处方,痰热中阻证常用法半夏、枳实、黄连、瓜蒌、吴茱萸、海螵鞘、郁金、延胡索、柴胡、茯苓;肝胃不和证常用柴胡、黄芩、白芍、桂枝、木香、陈皮、茯苓;脾胃气虚证常用陈皮、白术、茯苓、党参、木香;寒热错杂证常用黄连、半夏、黄芩、党参、干姜.结论 名老中医治疗慢性胃炎以行气健脾、清热燥湿、疏肝降逆、滋补胃气为诊疗思路,从虚实、气血、寒热、脏腑4个方面进行辨证.

目的 运用数据挖掘方法分析治疗脂溢性脱发的内服中药复方专利的用药规律,为中药组方优化及临床新药研发提供参考.方法 检索国家知识产权局网站建库至2021年11月1日发布的治疗脂溢性脱发的中药复方专利,采用古今医案云平台(V2.4.3)分析高频中药(频数≥10)的类别、性味、归经,以及中药关联网络,进行层次聚类分析及复杂网络分析,采用Excel2010及R4.2.1软件进行高频药物因子分析.结果 纳入中药复方专利44项,涉及中药179味,其中使用频次最高者为制何首乌,用药以寒性、甘味药为主,归经以肝经为首;高频中药因子分析提取出5个公因子;关联规则及聚类分析分别得到药对组合10个和聚类方3个,核心药物为制何首乌、当归、茯苓、白鲜皮、墨旱莲、女贞子、桑椹、牡丹皮、生地黄、熟地黄、菟丝子、侧柏叶.结论 脂溢性脱发病因病机多与肝肾不足、脾失健运、血热风热等相关,中药复方专利治疗本病以清热解毒、滋补肝肾药为主.