TCMLLM-PR:evaluation of large language models for prescription recommendation in traditional Chinese medicine

Objective To develop and evaluate a fine-tuned large language model(LLM)for traditional Chinese medicine(TCM)prescription recommendation named TCMLLM-PR.Methods First,we constructed an instruction-tuning dataset containing 68654 samples(ap-proximately 10 million tokens)by integrating data from eight sources,including four TCM textbooks,Pharmacopoeia of the People’s Republic of China 2020(CHP),Chinese Medicine Clinical Cases(CMCC),and hospital clinical records covering lung disease,liver disease,stroke,diabetes,and splenic-stomach disease.Then,we trained TCMLLM-PR using Chat-GLM-6B with P-Tuning v2 technology.The evaluation consisted of three aspects:(i)compari-son with traditional prescription recommendation models(PTM,TCMPR,and PresRecST);(ii)comparison with TCM-specific LLMs(ShenNong,Huatuo,and HuatuoGPT)and general-domain ChatGPT;(iii)assessment of model migration capability across different disease datasets.We employed precision,recall,and F1 score as evaluation metrics.Results The experiments showed that TCMLLM-PR significantly outperformed baseline models on TCM textbooks and CHP datasets,with F1@10 improvements of 31.80%and 59.48%,respectively.In cross-dataset validation,the model performed best when migrating from TCM textbooks to liver disease dataset,achieving an F1@10 of 0.1551.Analysis of real-world cases demonstrated that TCMLLM-PR's prescription recommendations most closely matched actual doctors’prescriptions.Conclusion This study integrated LLMs into TCM prescription recommendations,leverag-ing a tailored instruction-tuning dataset and developing TCMLLM-PR.This study will pub-licly release the best model parameters of TCMLLM-PR to promote the development of the decision-making process in TCM practices(https://github.com/2020MEAI/TCMLLM).

Integrated data mining and network pharmacology to discover a novel traditional Chinese medicine prescription against diabetic retinopathy and reveal its mechanism

Background:Diabetic retinopathy(DR)is currently the leading cause of blindness in elderly individuals with diabetes.Traditional Chinese medicine(TCM)prescriptions have shown remarkable effectiveness for treating DR.This study aimed to screen a novel TCM prescription against DR from patents and elucidate its medication rule and molecular mechanism using data mining,network pharmacology,molecular docking and molecular dynamics(MD)simulation.Method:TCM prescriptions for treating DR was collected from patents and a novel TCM prescription was identified using data mining.Subsequently,the mechanism of the novel TCM prescription against DR was explored by constructing a network of core TCMs-core active ingredients-core targets-core pathways.Finally,molecular docking and MD simulation were employed to validate the findings from network pharmacology.Result:The TCMs of the collected prescriptions primarily possessed bitter and cold properties with heat-clearing and supplementing effects,attributed to the liver,lung and kidney channels.Notably,a novel TCM prescription for treating DR was identified,composed of Lycii Fructus,Chrysanthemi Flos,Astragali Radix and Angelicae Sinensis Radix.Twenty core active ingredients and ten core targets of the novel TCM prescription for treating DR were screened.Moreover,the novel TCM prescription played a crucial role for treating DR by inhibiting inflammatory response,oxidative stress,retinal pigment epithelium cell apoptosis and retinal neovascularization through various pathways,such as the AGE-RAGE signaling pathway in diabetic complications and the MAPK signaling pathway.Finally,molecular docking and MD simulation demonstrated that almost all core active ingredients exhibited satisfactory binding energies to core targets.Conclusions:This study identified a novel TCM prescription and unveiled its multi-component,multi-target and multi-pathway characteristics for treating DR.These findings provide a scientific basis and novel insights into the development of drugs for DR prevention and treatment.

Clinical efficacy and safety of TCM prescriptions combined with nucleoside(acid)analogues in treating chronic hepatitis B:a meta-analysis

Objective There are many clinical reports on traditional Chinese medicine(TCM)combined with nucleoside(acid)analogues(NAs)for the treatment of chronic hepatitis B(CHB),but its efficacy and safety are not completely clear.This meta-analysis aims to evaluate the clinical efficacy and safety thus providing evidence for clinical applications.Methods We searched Chinese databases the China National Knowledge Infrastructure(CNKI),Wanfang Data,and China Science and Technology Journal Database(VIP),as well as English databases Pub Med and Cochrane Library,from time of establishment to April 14,2021.Literature quality was evaluated according to the bias risk assessment criteria of Cochrane Collaboration network.Rev Man 5.3 and Stata 12.0 software were used to perform this research.Results A total of 23 articles,3282 patients,and 25 TCM prescriptions were included in this study.NAs plus TCM remarkably improved the clinical total effective rate[Odds ratio(OR)=3.92,P<0.00001],TCM syndrome score(Mean difference=-3.73,P<0.00001),hepatitis B virus(HBV)DNA negative conversion rate(OR=1.49,P=0.0001),hepatitis Be antigen(HBe Ag)negative conversion rate(OR=2.03,P<0.00001),alanine aminotransferase levels[Std mean difference(SMD)=-0.95,P<0.00001],and aspartate aminotransferase levels(SMD=-0.70,P=0.0004).Adverse reaction rates did not increase in the combined treatment group(OR=0.97,P=0.84).A comprehensive analysis of the 25 TCM prescriptions suggested that the combination of spleen-strengthening prescriptions with NAs showed better effects than other prescriptions.Conclusion TCM in combination with NAs,demonstrated better clinical efficacy against CHB than NAs alone.In addition,the combination of spleen-strengthening prescriptions and NAs was identified as the best therapeutic strategy.However,more randomized controlled trials of high quality are needed to provide more reliable clinical basis for the application of TCM.

Research on the Progress of Treating Gastroesophageal Reflux Disease with Modified Classic Traditional Chinese Medicine Prescription

Gastroesophageal reflux disease (GERD) is a high-incidence digestive system disease. Western medicine mainly uses drugs such as proton pump inhibitors to inhibit gastric acid secretion, but some patients are accompanied by symptoms such as non-acid reflux and gas reflux, which cannot effectively treat the disease. It is necessary to actively explore other treatment schemes. Traditional Chinese medicine (TCM) has a long history of research on gastroesophageal reflux disease, which emphasizes the treatment based on syndrome differentiation as a whole. Through the treatment of various and multi-component TCM prescriptions, the patient’s body condition can be adjusted, and the treatment effect on gastroesophageal reflux disease is reliable, which has obvious therapeutic advantages. To further clarify the treatment of gastroesophageal reflux disease, this study reviewed and analyzed the research progress of the treatment of liver disease with modified prescriptions, and the report is as follows.

China Association of Chinese Medicine Paper notice for seminar on clinical application of TCM classic prescriptions and herbs in the treatment of emotional diseases

Hosted by ChinaAssociation of Chinese Medicine,and organized by World Journal of Integrated traditional and western Medicine,seminar on clinical application of TCM classic prescriptions and herbs in the treatment of emotional diseaseis planned to be held in Yichang,Hubei province from June 22nd to 24th.

Rules of Ren Shunping's prescription for trearing Spleen Deficiency and Qi Stagnation Syndrome of Functional Dyspepsia Based on Data Mining

Objective:To explore the medication rules of Professor Ren Shunping treating spleen deficiency and qi stagnation syndrome of functional dyspepsia,so as to provide reference for clinical prevention and development of new drugs.Methods:The 144 Chinese herbal formulas considered as active prescribed by professor Ren Shunping for the first time treating the patients with spleen deficiency and qi stagnation syndrome of functional dyspepsia were collected,entered into TCM inheritance auxiliary platform(V2.5),and then mined by some data mining methods of the software,such as the association rules and unsupervised entropy hierarchical clustering.Results:This study involved 144 Chinese herbal formulas,including a total of 56 Chinese herbal medicines.Of these Chinese herbal medicines,the herbs of cold and warm nature,and pungent,bitter and sweet in flavors were used in a quite high frequency,all of which act on spleen,stomach,lungs,large intestine,liver and gallbladder meridians.One core Chinese herbal formula was excavated,composed of patchouli,perilla stem,radix bupleuri,scutellaria,ginger pinellia,codonopsispilosula,magnolia bark,tangerine peel,large-headed atractylodes,immature bitter orange,wood incense,fructus amomi,inula britannica chinensis,ginger,medicated leaven,fried rice and fried malt based on the association rules.And a total of 17 groups of core used drugs and 2 groups of new Chinese herbal formulas undiscovered were mined based on unsupervised entropy hierarchical clustering.Conclusion:In treating spleen deficiency and qi stagnation syndrome of functional dyspepsia,professor Ren is good at treating spleen deficiency and qi stagnation syndrome of functional dyspepsia from the perspective of"tong jiang",using the"five-in-one"strategy of strengthening the spleen,evacuating the liver,coming down qi from the stomach,dispersing lung,and bowel-reflexing to treat patients as a whole.And the herbal medicines were light in dose,lovely and mild in nature for the formulars.At the same time,he paid attention to the combination of syndromes and symptoms,and flexible arbitration.

Analysis on the Pharmacists Intervention Results of the Problems from 2000 Prescriptions of Chinese Herbal Pieces

OBJECTIVE: To interfere with the improper prescriptions of Chinese herbal pieces issued by the outpatient doctors so as to reduce the doctor's prescription error rate and ensure the safety of medication for patients. METHODS:Based on Prescription Administrative Policy, Management Standard of Chinese Herbal Pieces in Hospital, and Beijing Chinese Herbal Pieces Dispensing Regulation, 2000 prescriptions were collected and analyzed according to the different intervention, and solutions were put forward for a variety of reasons. RESULTS: In 2000 prescriptions, there were 932 prescriptions containing toxic drug overdose without double signature, 577 prescriptions of dose error, 282 prescriptions of ingredients error, 85 prescriptions with wrong route of administration, 73 prescriptions of herb incompatibility without double signature, and 51 prescriptions with repeat herbs, accounting for 46.5%, 28.85%, 14.1%, 4.25%, 3.65% and2.55% of the total number of prescriptions respectively. After the intervention of pharmacists, 2000 prescriptions were all successfully intervened to protect the safety of medication. CONCLUSIONS: TCM pharmacist should double check prescriptions with errors issued by doctors, so as to reduce the prescription error rate, and ensure the safety of medication and rehabilitation of patients, and then improve values of pharmacists.

Analysis on the Medication Law of Drug Pair Acorus Tatarinowii-Polygala Tenuifolia Based on the Dictionary of Traditional Chinese Medicine Prescriptions

Objective:Using data mining method to dig and sort out the prescriptions with Acorus tatarinowiiPolygala tenuifolia drug pair,and summarize the medication characteristics and compatibility of their prescriptions after preliminary screening.Methods:By searching the Dictionary of Traditional Chinese Medicine Prescriptions,a standardized database of prescriptions was established,and the properties,tastes and meridian tropism of prescriptions were classified,and the indications of prescriptions and core combinations of traditional Chinese medicines were analyzed.Results:178 prescriptions were collected,of which 210 were related.Most of the drugs were warm in nature and sweet in taste,and mainly return to the heart and kidney meridians.Under the same confidence and different support,the core drug combination for treating brain diseases was Acorus tatarinowiiPolygala tenuifolia-Ginseng,and the core drug combination for treating asthenia was Acorus tatarinowii-Poria cocos-Rehmannia glutinosa-Polygala tenuifolia.Conclusion:The compatibility characteristics of the prescriptions containing Acorus Tatarinowii and Polygala tenuifolia in the Dictionary of Traditional Chinese Medicine Prescriptions are remarkable,which provides reference for scientific guidance of clinical rational drug use and basic research of prescriptions containing Acorus tatarinowii and Polygala tenuifolia.

基于数据挖掘的古方治疗产后腹痛用药规律研究

目的基于数据挖掘技术,总结治疗产后腹痛古方的组方配伍及用药规律。方法整理收集《中华医方》妇科篇中治疗产后腹痛的方剂,进行筛选,建立数据库。运用数据挖掘软件进行频数统计、关联规则分析、聚类分析、因子分析、复杂网络构建等。结果使用的药物类别以补益药、活血化瘀药、温里药、清热药、理气药较为集中,使用高频次药物为当归、肉桂、川芎、白芍、桃仁等。四气以温、平居多,五味以辛、苦为主,归经多属肝、脾、肺、肾经。关联强度较高的二联规则17组,3联规则10组,肉桂-当归、川芎-当归、白芍-当归等为高频药对,川芎-肉桂-当归、延胡索-当归-肉桂、白芍-肉桂-当归等为高频角药组合。高频药物聚类分析及因子分析,获得8个中药聚类组合,12个公因子,复杂网络分析提示核心方剂为生化汤、少腹逐瘀汤、下瘀血汤。结论虚、寒、瘀是产后腹痛的病变特点,血虚不荣、血瘀阻滞是产后腹痛的主要病机。组方以益气补血药、温里散寒药、活血化瘀药为主,配伍清热药、理气药、止血药。数据挖掘所得到核心组合和方剂对临床应用及新药开发有重要的指导和借鉴意义。

基于数据挖掘的痛经用药规律研究

目的:基于数据挖掘技术分析《中医方剂大辞典》中治疗痛经方剂的用药规律。方法:根据纳入与排除标准筛选出《中医方剂大辞典》中治疗痛经的方剂,对其进行方药名称规范,运用Excel、IBM SPSS Modeler 18.0、IBM SPSS Statistics 26.0等进行频数统计、关联规则分析、聚类分析及因子分析。结果:共纳入151首治疗痛经的方剂,涉及中药191味;其中出现频次≥10的中药为当归、川芎、延胡索、白芍、香附、甘草等共36味,药效以补虚药、活血化瘀药、理气药为主,药性偏温、寒、平,药味则偏辛、苦、甘,多入肝、脾、心经。选取出现频次≥25的13味高频药物进行关联规则分析,得到“当归-川芎“”当归-香附-川芎“”当归-熟地黄-白芍-川芎”等常用药对21组,并通过提升关联规则的支持度得到“当归-川芎-白芍”核心组合。进行高频药物聚类分析得出4类药物组合,C1:白芍、熟地黄、川芎、当归;C2:牡丹皮、茯苓;C3:香附、甘草;C4:桃仁、红花,因子分析得到5个公因子。结论《:中医方剂大辞典》中治疗痛经以补血活血、化瘀散结、疏肝理气为主。

基于《中医方剂大辞典》的雷丸组方用药规律分析

目的 基于数据挖掘方法探究《中医方剂大辞典》中雷丸的组方用药规律。方法 以《中医方剂大辞典》为信息源,梳理含雷丸的方剂,构建雷丸方药数据库;采用文献分析和数据挖掘方法,探讨雷丸的增效减毒及其方剂主治病症类型、药物配伍情况;利用关联规则分析探寻其在常见病中的配伍特点;使用皮尔逊相关性分析发掘药-证的相关性,总结其用药规律。结果 含有雷丸方剂的剂型主要为丸剂和散剂,高频用药为槟榔、大黄、木香等,与雷丸配伍的主要为驱虫药、泻下药及理气药等,且味多为苦、辛,归胃经、脾经。其中“槟榔-雷丸”出现的药对配伍频数最高,其次是“大黄-雷丸”“木香-雷丸”;雷丸方剂的主治病症为虫积的频数最高,其次为疳证、积滞等。皮尔逊相关性分析结果表明,治疗虫积配伍槟榔、木香、使君子、芜荑、鹤虱等;治疗疳证配伍胡黄连、黄连、青皮等;治疗积滞配伍神曲、陈皮、牵牛子、三棱等。结论 雷丸与槟榔等配伍可加强杀虫消积的功效;与胡黄连、大黄等配伍可增加泻下攻积的作用,以上可能会引起轻微的胃肠道不良反应;与陈皮、神曲等配伍可保护胃黏膜,减轻胃肠不良反应。但雷丸毒性物质基础研究等未见报道,仍需深入研究,为雷丸临床应用提供更确切的基础。

补肾健脾通络方治疗老年衰弱综合征临床疗效观察

目的观察补肾健脾通络方治疗老年衰弱综合征临床疗效及安全性。方法研究纳入2021年1月—2022年6月收治的老年衰弱综合征患者共计102例,以随机数字表法将患者分为对照组与观察组,每组51例,对照组患者采取临床常规治疗,观察组患者采取常规治疗结合补肾健脾通络方治疗。观察各组数据:治疗总有效率、治疗前后患者中医证候总积分变化及Fried表型衰弱量表(frailty phenotype,FP)评分变化及白介素-6(interleukin 6,IL-6)、超敏C反应蛋白(C-ReactiveProtein,hs-CRP)水平变化、握力变化及起立-行走计时测试(Timed Up and Go test,TUG)变化、5次坐立试验(five-times-sit-to-stand test,FTSST)变化、步速变化、体质指数(Body Mass Index,BMI)变化及四肢骨骼肌质量指数(appendicular skeletal muscle mass index,ASMI)变化、简易精神状态评分(Mini-mental State Examination,MMSE)变化、匹兹堡睡眠指数(Pittsburgh sleep quality index,PSQI)评分变化、日常生活能力(Activity of Daily Living,ADL)评分变化、微型营养评估(mini nutritionaal assessment,MNA)评分变化、患者治疗满意度、不良反应。结果观察组患者治疗总有效率比对照组高[96.08%(49/51)vs84.31%(43/51)],P<0.05;治疗前,各组患者中医证候总积分、FP评分、IL-6、hs-CRP水平、握力、TUG测试结果及FTSST测试结果、步速、BMI、ASMI、MMSE评分、PSQI评分、ADL评分、MNA评分比较,P>0.05,治疗后各组患者中医证候总积分、FP评分、IL-6、hs-CRP水平、握力、TUG测试结果及FTSST测试结果、步速、BMI、ASMI、MMSE评分、PSQI评分、ADL评分、MNA评分均改善,观察组患者治疗后中医证候总积分、FP评分、IL-6、hs-CRP水平、握力、TUG测试结果及FTSST测试结果、步速、BMI、ASMI、MMSE评分、PSQI评分、ADL评分、MNA评分均优于对照组患者,P<0.05;观察组患者治疗满意度高于对照组[98.04%(50/51)vs84.31%(43/51)],P<0.05;观察组与对照组不良反应率均较低,且可自行缓解,P>0.05。结论补肾健脾通络方治疗老年衰弱综合征临床疗效显著,能缓解衰弱综合征患者的临床症状,抑制患者机体炎性反应,增加肌肉质量,改善睡眠,提高生活质量,且无不良反应,较为安全可靠,值得应用。

基于网络药理学的中药组方治疗骨质疏松症的用药规律研究

目的 分析国家专利中药组方治疗骨质疏松症的用药规律并进行网络药理学分析。方法 在国家知识产权局网站中检索治疗慢骨质疏松的中药组方,规范药名后进行中药频数分析、性味归经分析,组方规律分析,药物靶点频数分析,k-means降维PPI分析、GO-BP生物学功能富集分析和KEGGPathway分析。结果 共纳入112项专利中药组方,涉及156味中药,其中淫羊藿的使用频次最高。药性以温性、平性、寒常见;五味多为甘、苦、辛;归经多归属肝经、肾经、脾经。关联规则分析提示常用的药对组合为牡蛎—淫羊藿,三七—当归,黄芪—淫羊藿等。核心靶点主要是AR、ESR1、ESR2、NR1I2、SHBG、VDR、NR3C1,GABRA家族和GABRG家族等。k-mean聚类分析PPI互作网络可降维成4类,与临床分型密切相关。GO和KEGG-Pathway分析表明中药复方主要通过神经活性配体-受体相互作用,氯化物跨膜转运,类固醇激素生物合成,碳水化合物的消化和多种物质成瘾的信号通路,通过消化吸收,神经信号和递质传递,类固醇的代谢,氯离子的转运等功能发挥防治骨质疏松症的作用。结论 结果表明中药治疗原发性骨质疏松症有规律可循,为中药治疗骨质疏松提供了理论依据,为中药的临床研究和基础研究提供了新的思路。

基于数据挖掘的中医药治疗颈性眩晕证治方药规律研究

目的基于数据挖掘方法探讨中医药治疗颈性眩晕的证治方药规律。方法检索中国知识资源总库(CNKI)、中国学术期刊数据库(万方数据)、中文科技期刊数据库(VIP)2013年1月1日-2023年1月1日中医药治疗颈性眩晕研究文献,运用WPS Office 11.1.0.14036软件统计方剂、疗法、证型、药物及药物性味归经的频次、频率;运用SPSS Statistics 25.0、SPSS Modeler 18.0软件对方药进行聚类分析和关联规则分析。结果纳入研究文献155篇,包含中医证型24种,方剂162首,中药119味。证型以痰湿中阻证、气血亏虚证为主,常用方剂类别为祛痰剂、补益剂,其中半夏白术天麻汤、益气聪明汤使用频次最高;常用药物类别为补虚药、解表药、活血化瘀药、平肝息风药、化痰药;药性以温性为主,药味以甘味为主,归经以脾经为主。聚类分析得出4个聚类方,关联规则分析得出11组药物组合。结论颈性眩晕主要证型为痰湿中阻证、气血亏虚证,治以化痰祛湿、益气升阳,方剂多选用半夏白术天麻汤、益气聪明汤;用药多性温味甘,以白术、半夏、天麻、葛根、茯苓为主。

基于数据挖掘探讨中医治疗鼻衄用药规律

目的 基于中医传承辅助平台2.5软件挖掘中医治疗鼻衄(鼻出血)的用药规律,探讨新方组合。方法 检索中国知识资源总库(CNKI)、中国学术期刊数据库(万方数据)、中文科技期刊数据库(VIP)、中国生物医学文献数据库(CBM)建库至2022年12月1日中医治疗鼻衄的相关文献,录入中医传承辅助平台并建立处方数据库,运用频次统计、关联规则和复杂网络等方法分析处方药物使用频率、常用药对、组合规则,并探索新方。结果 筛选出有效方剂221首,涉及中药265味,其中使用频次≥20的常用中药33味,前3位分别为生地黄、黄芩、炙甘草。药性以寒、温为主,药味以苦、甘为主,归经主入肝经、肺经、心经、脾经、胃经。常用药物组合46个,核心组合36个,发现新方16首。生地黄为治疗鼻衄的核心药物,生地黄-黄芩、生地黄-当归、生地黄-黄芩-炙甘草、生地黄-牡丹皮-黄芩为常用药物组合。结论 中医治疗鼻衄秉清热凉血止血大法,佐以化瘀补血、补气摄血、滋阴降火,重视脏腑辨证精准治衄,并多用蜂蜜、炙甘草等甘味药取缓急、补中添益之功用,部分药物如黄芩、大蓟、栀子等用炒焦、炒炭以加强止血效力。

中医药治疗肺癌的用药规律

目的:探究中医药治疗肺癌的用药规律。方法:检索《肿瘤良方大全》《肿瘤方剂大辞典》《卫生部药品标准中药成方制剂》以及《国家药品监督管理局总局国家药品标准(修订)颁布件》中中医治疗肺癌的处方,将数据录入Excel进行数据处理,使用Lantern 5.0、SPSS Modeler 18.0及SPSS Statistics 25.0对肺癌方剂频次统计、系统聚类分析及关联规则分析。结果:共纳入650个处方,572种药,6 120次。药性以寒、温、平为主,药味以苦、甘、辛为主,归经使用最多为胃、脾、肾、肝、肺。使用频次最多的药物为黄芪、白花蛇舌草、甘草、茯苓、沙参、麦冬等。高频药物以清热解毒药和益气养阴药及活血化瘀药主;关联规则分析得到药对关联16条,三联24条,因子内得到13个公因子,系统聚类分析得出7个关联紧密的药组。隐结构得到14个隐变量,每个隐变量有2个隐类,共28个隐类。结论:补气养阴药和清热抗癌解毒、活血化瘀药、健脾利湿药、清热化痰治则,为中医药治疗肺癌提供借鉴,有待进一步结合临床及实验给予验证。

强化用药管理联合精益管理对中药饮片处方合理性及用药安全性的影响

目的:分析强化用药管理联合精益管理对中药饮片处方合理性及用药安全性的影响。方法:抽取2018年7月—2019年11月期间我院中药房的中药饮片处方共240份,其中,2018年7月—12月服用中药饮片治疗的患者120例实施传统管理,纳入实施前,2019年6月—11月服用中药饮片治疗的患者120例实施强化用药管理联合精益管理,并纳入实施后。评价及对比实施前后处方用药不良情况、服药不良反应发生率、处方调剂质量、患者治疗满意度。结果:实施后的处方用药不良情况发生率低于实施前(P<0.05);实施后患者的服药不良反应率低于实施前患者(P<0.05);实施后的处方调剂差错率低于实施前(P<0.05);实施后患者的治疗满意度高于实施前(P<0.05)。结论:对中药饮片处方实施强化用药管理联合精益管理,可确保处方用药合理性,提高处方调剂质量,减少服药后不良反应,患者治疗满意度高,具有借鉴应用价值。

基于大语言预训练模型的中医个性化处方推荐研究

目的针对中医个性化处方推荐问题,研究自动化处方推荐任务,为中医临床辅助决策提供参考。方法基于大语言预训练文本生成模型设计一种中医个性化处方推荐算法。将中医处方推荐任务转化为端到端(seq2seq)的文本生成任务,即将临床症状描述文本通过模型转化为处方文本,以实现处方推荐任务的需求,并利用基于大语言预训练的BART(Bidirectional and Auto-Regressive Transformers)模型的预训练参数来提升模型对通用语义信息的理解,通过对训练集处方内中药排序提升模型的处方推荐性能。结果实验证明通过大语言预训练模型以及端到端的文本生成架构可有效提升模型的生成性能,同时对处方内中药依次排序可以获取更高准确率,并且通过中药的排列获取更多值得参考的有价值信息。中医个性化处方推荐模型在处方排序后分别在前5、10、15味生成的处方分别取得了58.60、53.79和49.67的准确率。结论中医个性化处方推荐模型取得了更优的处方推荐效果,表明其可为中医临床治疗疾病进行参考,达到辅助临床决策支持的效果。

干燥综合征合并肺间质病变住院患者中医证候分布及处方规律研究

目的采用数据挖掘方法分析干燥综合征(SS)合并肺间质病变(ILD)住院患者的中医证候分布及处方配伍规律。方法收集中国中医科学院广安门医院2010年1月-2021年12月符合筛选标准的SS-ILD住院患者临床资料,对中医证型与基本信息、临床表现、舌脉象、化验检查、中药处方进行分析。结果共纳入患者87例,不同证型患者性别、年龄、发病年龄、病程、吸烟史比较差异均无统计学意义(P>0.05),燥湿互结证患者饮酒史概率最高(P<0.05)。不同证型患者口干、关节疼痛等主诉及关节肿胀、胸闷、咳嗽等刻下症比较差异有统计学意义(P<0.05),ILD特殊体征及舌脉差异无统计学意义(P>0.05)。阴虚内热证患者尿素氮水平最高(P<0.05),其他化验检查结果比较差异无统计学意义(P>0.05)。共纳入方剂79首,涉及中药217味,高频用药以健脾益气、滋阴清热、养血活血为主,多性平、味甘,主归肺经、肝经、脾经,常用药对符合虚实夹杂病机,核心处方为八珍汤合养阴清肺汤加减。结论SS-ILD中医证候虚实夹杂,以气阴两虚为本,痰浊瘀血为标,病位在肺、肝、脾,临床治疗以益气养阴和中、兼化痰祛瘀通络为原则。

经方治疗子宫内膜异位症的研究进展

通过对《金匮要略》《伤寒论》中的经典方剂温经汤、当归芍药散、当归四逆汤、小柴胡汤、桂枝茯苓丸、抵挡汤、金匮肾气丸等治疗不同证型子宫内膜异位症(endometriosis,EMs)的文献进行梳理,阐释其治疗EMs的机理,为临床辨证不同类型EMs提供参考。