欢迎来到《四川大学学报(医学版)》 2025年5月6日 星期二

本体在糖尿病临床决策支持系统中的应用

周祎灵, 石清阳, 陈向阳, 李舍予, 沈百荣

周祎灵, 石清阳, 陈向阳, 等. 本体在糖尿病临床决策支持系统中的应用[J]. 四川大学学报(医学版), 2023, 54(1): 208-216. DOI: 10.12182/20220860201
引用本文: 周祎灵, 石清阳, 陈向阳, 等. 本体在糖尿病临床决策支持系统中的应用[J]. 四川大学学报(医学版), 2023, 54(1): 208-216. DOI: 10.12182/20220860201
ZHOU Yi-ling, SHI Qing-yang, CHEN Xiang-yang, et al. Ontologies Applied in Clinical Decision Support Systems for Diabetes[J]. Journal of Sichuan University (Medical Sciences), 2023, 54(1): 208-216. DOI: 10.12182/20220860201
Citation: ZHOU Yi-ling, SHI Qing-yang, CHEN Xiang-yang, et al. Ontologies Applied in Clinical Decision Support Systems for Diabetes[J]. Journal of Sichuan University (Medical Sciences), 2023, 54(1): 208-216. DOI: 10.12182/20220860201

本体在糖尿病临床决策支持系统中的应用

基金项目: 四川省科技厅项目(No. 2022YFH0114)资助
详细信息
    通讯作者:

    李舍予: E-mail:lisheyu@gmail.com

Ontologies Applied in Clinical Decision Support Systems for Diabetes

More Information
  • 摘要: 基于电子健康系统的临床决策支持系统(clinical decision support system, CDSS)可以辅助基层医生进行临床决策,提高诊疗效率。其中利用本体构建CDSS的医学知识库和患者数据使CDSS的推理过程和决策结果具有可解释性。本文就糖尿病治疗领域的相关本体及基于本体的CDSS的进展与挑战进行综述。首先,阐明我国糖尿病诊疗的现状与挑战为亟需提高医疗服务效率与质量;在介绍本体的相关知识后,概述使用本体构建CDSS的框架、方法及特点;然后综述国内外糖尿病治疗领域的本体及基于本体的CDSS的案例,总结其构建方法及特点;最后提出该领域的展望:将循证医学与本体相融合,构建可信的临床推荐体系是目前CDSS的发展重点。

     

    Abstract: A clinical decision support system (CDSS) integrated with electronic health records helps physicians at the grassroots make patient-appropriate and evidence-based treatment decisions and improves the efficiency of diagnosis and treatment. Furthermore, using ontologies to build up the medical knowledge base and patient data for CDSS enhances the automation and transparency of the reasoning process of CDSS and helps generate interpretable and accurate treatment recommendations. Herein, we reviewed the relevant ontologies in the field of diabetes treatment and the progress and challenges concerning ontology-based CDSSs. Firstly, we elaborated on the current status and challenges of diabetes treatment in China, highlighting the urgent need to improve the efficiency and quality of medical services. Then, we presented background information about ontologies and gave an overview of the framework, methodology, and features of using ontologies to construct CDSS. After that, we reviewed the ontologies and instances of ontology-based CDSS in the field of diabetes treatment in China and abroad and summarized their construction methods and features. Last but not the least, we discussed the future prospects of the field, suggesting that integrating evidence-based medicine with ontologies to build a reliable clinical recommendation system should be the current focus of CDSS development.

     

  • 我国糖尿病患者已逾亿人[14]。糖尿病的管理存在终身性和复杂性,其负担已严重影响了国民健康乃至经济发展[1, 5]。一方面,糖尿病需要长期/终身的药物治疗,这意味着巨大的直接医疗支出,而其并发症发生的年轻化则严重影响劳动力的社会贡献能力[1, 45]。另一方面,现有的糖尿病管理模式要求社区全科医生、糖尿病专科医生及相关健康专业人士密切协作[3, 67],在医药成本的基础上增加了大量的人力资源和管理成本。特别是我国目前在医疗资源、地理交通、经济发展等诸多领域存在一定地域差异,基层卫生服务水平和人员能力在不同地区间也存在一定差异[5]。在社会经济和医疗资源发展相对后进的地区,长期有效的糖尿病管理尤为困难[1, 45]。而近年来不断增加的糖尿病人口并未给我国基层医疗体系的发展留出更多的缓冲时间和空间[5]

    在这样的背景下,我国涌现出不少以大型三级医院医疗资源下沉为特色的行业医疗联合体[6]。这种模式一定程度上缓解了糖尿病给基层医疗服务带来的压力,但受到单个医疗从业者服务效率的局限,人力资源不足仍是联合体扩大和服务水平提升的掣肘[6, 8]。因此,如何有效提高医疗服务效率成为当下我国医疗领域亟需解决的问题。

    近年来信息技术、人工智能、大数据和循证医学的发展,给医疗服务效率的提升带来了希望[9]。其中,基于电子健康系统(electronic health record, EHR)的临床决策支持系统(clinical decision support system, CDSS)是解决这一核心问题的重要解决方案之一[1011]。实现CDSS的技术路线很多,其中基于本体知识库的CDSS无疑是国内外公认方法最为严谨、最具扩展性的方案之一[1215]。近年来,国内外CDSS类产品不断涌现,其中不乏针对基层及住院糖尿病的工程[1418],但使用本体作为知识库来源的糖尿病CDSS在我国相对有限,其概念对我国临床医生来说相对陌生。本文就该领域的研究进展及面临的问题进行综述。

    本体一词来源于本体论哲学,在信息科学中,意指一种对知识的表征,由一个领域内的一组概念以及它们之间的关系构成[1920]。它是表示领域背景知识的方式之一,能够被计算机理解、操作与推理[12]。本体首先明确知识范围,在给定的知识范围下,确定该范围涉及的概念及属性(如特点、参数、特性),统一概念和属性的语义,确保对信息的共同理解。然后内置概念、属性各自之间的关系与公理,形成层级或简单的网状关系,提供一个概念到另一个概念的导航,实现本体内部的自动推理[2021]。本体分为领域本体、上层本体和两者构成的混合本体。领域本体是针对某个特定的领域知识。上层本体是描述所有领域本体之中普遍的共同概念及关系的本体,目的在于使不同领域间能进行融合与互相操作,比如基本形式化本体(basic formal ontology, BFO)是生物医学领域应用最广泛的上层本体,在全球有超过300个本体项目使用BFO作为上层本体[2223]

    本体的编辑需要本体语言,后者的本质是一种形式化语言,能编码领域知识,也能描述推理规则。尽管本体语言很多,历史上包括KIF、OKBC、SHOE、XOL、OIL、DAML+OIL、XML、RDF、RDFs等[24]。但目前最为通用的本体语言是2009年被万维网联盟(W3C)正式推荐的OWL2(web ontology language-2)[25]。这是一种基于描述逻辑的本体语言OWL(Web Ontology Language)的拓展和修订版本,具有良好的表达能力、推断能力和计算属性[25]。用户可以使用本体编辑器对本体语言进行编辑,目前常用的本体编辑器包括Protégé、Web Protégé、TopBraid Composer、Neon工具箱、OilED、FOAF编辑器、WebOnto、OntoEdit、WebODE等[20, 26]。这些开源编辑器均支持OWL2格式语言的编辑。本体借助本体语言和编辑器可编辑和存储知识中的概念及其简单的关系,而知识中派生复杂的规则可以借助SWRL(semantic web rule language)编辑,SWRL规则可引入OWL本体。本体和SWRL规则将文本型的、隐性的知识结构化、透明化、显性表达,一起实现复杂知识的自动推理[1112]

    在生物医学领域,开放生物及生物医学本体(open biological and biomedical ontology, OBO)是本体开发交流的重要平台社区,为生物医学领域本体的开发制定一系列开发原则[2729]。由美国国立卫生院(National Institutes of Health, NIH)资助的国家生物医学本体(national center for biomedical ontology, NCBO)中心建立了目前生物医学领域最全面的本体存储库,并提供了全面开放获取的生物医学本体搜索引擎(BioPortal,网址https://bioportal.bioontology.org/[3033]。Ontobee、EMBL-EBI Ontology Lookup Service、Pubmed、Google Scholar和Semantic Scholar等均是生物医学领域本体检索的途径。

    本体可以通过整合和拓展现有本体来构建,从而节省资源[3436]。在整合已有本体时有多种方式,比如自上而下的方式,先根据目的确定该领域涉及的核心概念及模块,构建本体的大框架;然后根据每个框架选择并整合符合要求的本体或者本体中的一部分;再确定概念之间的关系,评估是否符合本体的要求;最后评估整合的模型[35]。也可以通过本体网络模型整合现有本体构建新本体,即通过映射对齐各个本体重叠的部分;对于各个本体不重叠的部分,则通过创建一个本体模型,以加入新信息拓展的方式将不重叠的部分进行整合[21]

    CDSS是一种将医疗观察与医疗知识结合,充分运用可供利用的、合适的计算机技术,针对半结构化或非结构化医学问题,通过人机交互的方式,提高临床诊疗效率的系统[37]。工作流程核心是在正确的时间,向正确的人,采用正确的临床策略,通过正确的渠道,提供正确的信息[38]。CDSS应用场景较多,包括辅助继续教育、药物成本控制、预警系统、疾病诊断、疾病治疗、临床管理和药物监控等[37]。尽管CDSS可应用于糖尿病领域的上述任何场景,但本文重点探讨糖尿病治疗的CDSS。这类CDSS的有效实施取决于患者资料、医学知识库、推理方法、信息技术与架构、与临床工作流程的整合实施、及人机交互[39]

    患者数据作为CDSS的前端数据,其质量是CDSS生成准确推理结论的前提[40]。患者数据可来源于患者或者医生等其他医务工作者的直接输入,也可来自于EHR、移动设备和传感器等,其中随着EHR在医院的广泛应用,EHR成为患者信息的良好来源,不仅提供较完整的患者信息,包括人口学信息、实验室检查、药物处方、生命体征、疾病诊断、手术信息等,而且提供的信息带有时间维度,能不断更新[40]。大数据的兴起为CDSS有效推理提供了更多的输入信息,同时也提出了挑战,最直接的挑战便是患者信息繁多复杂、非结构化、且不同类型的数据在内容和形式上各异,导致数据的准确性和可用性较低。想要提高数据的可用性,首先需要对患者数据结构化、标准化,并且明确数据在特定语境下的准确语义。自然语言处理(natural language processing, NLP)等信息技术可对非结构化的数据进行提取,形成半结构化或者结构化的数据。词汇表、分类法、同义词表、主题映射、逻辑模型和本体等方法可以标准化半结构化或者结构化的数据[39]

    本体相比于其他标准化的方法,有其独特的特点。本体内置的语义关系使数据不在单一的词汇中被理解,而是在特定语境下被识别,极大改善患者数据的质量和可用性;并且以标准化和网状或者层级关系的方式表示,将患者数据连接起来形成可供计算机理解的个人医疗信息,内置的逻辑可供自动推理、生成相应的结论。

    患者数据的本体可以通过选择并集成已有可用的、较完善的、有效清晰的医学本体,然后在集成的本体上以更改与拓展的方式进行搭建。现有许多高质量医学本体可供选择,比如上层本体可以选择BFO、普通医学本体(ontology for general medical science, OGMS);疾病、症状及表型的模块可以选择人类表型本体、人类疾病本体和症状本体;测量单位模块可以选择测量单位本体;时间维度模块可以选择SWRL时间本体;药物模块可以选择解剖学治疗学及化学分类系统(anatomical therapeutic chemical, ATC)编码和药物本体;实验室检查模块可以选择观测指标标识符逻辑命名与编码系统(the logical observation identifiers names and codes system, LOINC)等[3536]

    医学知识库的构建决定了CDSS的临床可解释性和患者安全性。可靠的医学知识库构建无疑是高质量CDSS的保障。医学知识库是基于知识的智能体,需要具备完备性、有效性、一致性和高度结构化[39]。目前知识库的构建往往采用可扩展标记语言(extensible makeup language, XML)、带有或不带有概率的决策树和本体等[41]。这些知识库的知识来源可以是已发表的临床实践指南(clinical practice guidelines, CPGs)、医学文献及教科书、专家临床经验,也可以是通过人工智能(包括机器学习或深度学习网络等模型)获取的数据驱动的“黑箱”知识[4243]。前者通过系统性的文献检索,对相关领域文献和信息进行系统性采集、整理和归纳,总结出其中的内在逻辑,并将其结构化和形式化,进而使其可以由计算机处理。这些工作均需人工完成,维护成本高昂,且存在信息更新可能不及时等时效性问题。而后者采用“黑箱”算法进行关联分析,可以更高效地得到更多的知识、信息,但其结果的可解释性和外推性往往需要进一步人工核查,采用的模型方法对结果潜在影响较大,且依赖于数据的可用性及质量。考虑到CDSS可能直接影响医疗决策及对应的患者预后,其所依据的医学知识库对制作过程的透明化和推荐结果的可解释性具有苛刻要求,利用基于人工总结的文献规则仍是CDSS医学知识库最重要的知识来源。

    本体和SWRL规则可以以结构化、计算机可理解的方式显性地表达和存储叙述性的CPGs及决策过程。基于本体的医学知识库可分为两部分:领域内医学知识本体和规则库。领域内医学知识本体的构建可主要分为三步:首先根据EHR、该疾病相关的CPGs、医学文献、专家经验及教科书明确该疾病治疗领域的组成要素,确定背景知识本体中的核心概念及组成模块;再确定每个模块使用的术语集;最后可以通过任务网络模型,比如GLIF、EON、SAGE和Asbru,将CPGs及决策过程结构化,形成资源描述框架(resource description framework, RDF)格式的网状模型,即概念之间的关系和公理,并内置于领域内知识本体中[41]。规则库的构建可以通过SWRL等语言结构化该疾病领域中复杂及衍生的规则。领域内知识本体与SWRL规则可使用同一组术语集,使两者融入更加容易。

    由构建过程可知,基于本体的医学知识库具有以下三个显著的优势:其一,本体可梳理并整合某疾病治疗领域的相关概念、概念间的网状或者层级关系、临床路径与决策过程,为相关概念提供明确的定义,发现领域中的空白区域;其二,其本身具有丰富的语义表达性和推理性,可以尽可能充分地表达CPGs的内容,并且本体对内容的呈现方式使人和计算机均能理解;其三,本体对领域知识的表达是显性可见的,且可根据其元数据进行溯源,使后续的推理具有透明性,使决策结果具有可解释性。

    构建一个良好的基于本体的医学知识库仍存在一些挑战。其一,临床医学知识本身具有高度的复杂性和不确定性,存在概念模糊和规则不明确,导致临床医学知识本体相应的概念存在模糊性,这可能会导致推理中途停止或者得到错误结论。但随着医学领域的发展,不确定的程度和知识盲区的范围将会逐渐缩小。其二,传统的临床实践指南有时难以满足高度的个性化治疗的需求。其三,医学知识结构化形成本体和规则的过程可能存在信息失真,原因在于不同构建者对模糊的知识的解读不同,对知识的定义、提取和结构化的标准不同。因此亟需一个结构化医学知识和处理模糊边界知识的指导标准,以保证本体和规则能准确且完善地表达医学信息。此外,结构化语言有时不能很好地处理知识间的细微差别,这也可能导致信息失真。因此构建规则和本体时需要在临床专家和方法学家的指导下对细节进行严格区分。最后医学知识本体和规则的创建、维护与更新均需要跨领域的专家团队共同投入大量的精力与时间,其中,专家团队成员包括领域内的临床专家、合成证据创建指南的专家、专业的知识库工程师、熟识临床知识规范编辑及编辑工具的编辑者等。

    在疾病治疗的临床决策过程中,需要基于患者病情(患者数据)结合临床知识(一定规则)进行推理。因此CDSS在实现过程中需要推理系统,借助本体内置的关系和SWRL等规则,对患者数据本体和医学知识本体的信息进行自动调取、连接,实现决策过程和决策行为,生成推理结论、给出治疗方案推荐。基于本体的推理方法常见的有Tableaux运算的方法、逻辑编程改写的方法、一阶查询重写的方法、产生式规则的方法等。其中,Tableaux运算适用于检查本体的可满足性等。支持这种推理方法的推理器有FaCT++、Racer、Pellet、HermiT[42]。一阶查询重写的方法可以通过SPARQL高效地结合不同数据格式的数据源。基于产生式规则的方法主要可概括为:先通过模式匹配的算法将每条规则中的条件与患者数据本体和医学知识本体进行匹配,当这条规则中的所有条件均被匹配时,该条规则就被触发;再通过一定的策略从被触发的多条规则中选择一条规则;最后按照规则中的执行序列运行。支持这种推理方法的引擎有Jena等[11]

    我们针对糖尿病治疗领域,检索2012年到2021年在OBO、NCBO BioPortal、Ontobee、EMBL-EBI Ontology Lookup Service、Pubmed、Google Scholar和Semantic Scholar等生物医学领域本体检索库中发表的本体及相关文章,共发现有14篇相关的文献[34-36, 4454],其中有4个基于OWL2的可用本体[36, 44, 46, 48],分别是中国糖尿病本体(Chinese Diabetes Ontology, CDO,2021)[44]、FHIR and SSN-based T1D Ontology(FASTO,2017)[46]、Diabetes Mellitus Treatment Ontology(DMTO,2015)[36]、BioMedBridges Diabetes Ontology(DIAB,2015)[48];5篇只讲述关于本体开发过程的文章[35, 4445, 4748];1篇描述了本体开发过程及其应用在语义问答中的文章[49];8篇描述本体开发过程及其应用在CDSS中的文章[34, 36, 46, 5054]表1)。

    表  1  糖尿病治疗相关的本体及其特点
    Table  1.  Characteristics of ontologies for diabetes treatment
    Name of ontology or developer (Released date/update date)Domain/
    objective
    Main componentsSource of knowledgeLanguage and toolMain terminologyTop-level ontology or standardsReused existing ontologyUsed in clinical practice
    Need patient medical recordsKnowledge baseRule
    CDO (2021/2022)[44] Knowledge base Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    SWRL NA Protégé
    OWL2
    SWRL
    SNOMED-CT BFO
    OGMS
    HPO
    DOID
    DRON
    No
    MADHUSANKA S (2020)[45]* Knowledge base Yes Clinical practice guideline OWL2 CPGs Protégé
    OWL2
    NA NA NA No
    BRAVO M (2020)[35]* Patient profile Yes Patient profile
    Exercise and diet
    BMI
    Medication
    Treatment plans
    Not applicable Other ontologies Protégé
    OWL2
    LOINC NA NDF-RT
    DOID
    SYMP
    No
    SHERIMON P C (2020)[50]* Treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    OWL2 NICE guidelines Protégé
    OWL2
    NA NA NA No
    OMDP (2019/NA)[34]* Prognosis
    Diagnosis
    Treatment
    Yes Gene
    Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Time
    Treatment plans
    Medication and interaction
    SWRL CPGs
    Domain expert
    Literature
    EHR
    Protégé
    OWL2
    SWRL
    RxNorm
    SNOMED CT
    LOINC
    CPT
    BFO
    OGMS
    DOID
    SYMP
    GO
    Diet ontology
    Physical activity ontology
    No
    FASTO (2018/2018)[46] Treatment for type 1 diabetes Yes Sensor data
    Patient value and preference
    Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Time
    Treatment plans
    Medication and interaction
    SWRL CPGs
    Domain expert
    Literature
    Official web sites
    EHR
    Protégé 5.1
    OWL2
    SWRL
    RxNorm, SNOMED CT, LOINC
    BFO
    SSN
    HL7 FHIR
    Vital-sign ontology
    DDO
    DMTO
    SWRL TO
    OAE
    No
    DMTO (2017/2017)[36] Treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Time
    Treatment plans
    Medication and interaction
    SWRL
    DL-safe
    CPGs
    Domain expert
    Literature
    Official web sites
    EHR
    Book
    Protégé 5.0
    SWRL
    OWL2
    SNOMED CT
    RxNorm
    NDF-RT
    BFO
    OGMS
    DDO
    DINTO
    RO
    SMASH
    PATO
    OntoFood
    SYMP
    TIME
    No
    DDO (before2015/
    NA)[47]*
    Diagnosis Yes Patient profile
    Exercise and diet
    Medication
    Basic knowledge of diabetes
    SWRL CPGs
    Domain expert
    Literature
    Official web sites
    EHR
    Book
    Protégé 5.0
    SWRL
    OWL2
    SNOMED CT
    RxNorm
    NDF-RT
    UMLS
    BFO
    OGMS
    DOID
    SYMP
    UO
    OGMD
    No
    CHEN R C (2016/NA)[51]* Determining the ranking of antidiabetic medications Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Medication and interaction
    Fuzzy
    Jena
    TOPSIS algorithm
    CPGs
    Domain expert
    EHR
    Protégé
    WebProtégé
    Jena
    OWL2
    NA NA NA No
    IRS-T2D (2016/NA)[52]* Treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    SWRL
    OWL2
    Domain expert
    Literature
    Protégé
    OWL
    SWRL
    NA NA NA No
    OntoDiabetic (2015/NA)[53]* Assessing risk factors and treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    DL-safe
    OWL2
    CPGs Protégé
    OWL2
    DL-safe
    NA NA Answered
    Questionnaire ontology
    Adaptive
    Questionnaire ontology
    No
    DMO (2014/NA)[49]* Patient register, automated system for EHR data Yes Patient profile
    Basic knowledge of diabetes
    Medication
    NA CPGs
    Domain expert
    Protégé 4.3 SNOMED CT NA NA No
    CHEN R C (2012/NA)[54]* Antidiabetic medications selection Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Medication and interaction
    SWRL CPGs Protégé
    SWRL
    OWL DL
    NA NA NA No
     CDO: Chinese diabetes ontology; BFO: basic formal ontology; OGMS: ontology for general medical science; HPO: human phenotype ontology; DOID: human disease ontology; DRON: the drug ontology; LOINC: the logical observation identifiers names and codes system; NDF-RT: national drug file-reference terminology; SYMP: the symptom ontology; CPT: current procedure terminology; GO: gene ontology; OMDP: ontology-based model for diabetic patients; HL7: health level 7; FHIR: fast healthcare interoperability resources; FASTO: FHIR and SSN-based T1D ontology; DMO: diabetes management ontology; SWRL: semantic web rule language; OWL: web ontology language; DL: description logic; SSN: semantic sensor network ontology; OAE: ontology of adverse events; DINTO: the drug-drug interactions ontology; PATO: phenotype and trait ontology; SMASH: semantic mining of activity, social, and health ontology; DMTO: diabetes mellitus treatment ontology; DDO: diabetes mellitus diagnosis ontology; UO: units of measurement ontology; OGMD: ontology of glucose metabolism disorder; UMLS: unified medical language system; BMI: body mass index; EHR: electronic health record; NICE: the national institute for health and care and excellence; CPGs: clinical practice guidelines. * The OWL of the ontology is available; NA: not available.
    下载: 导出CSV 
    | 显示表格

    在5个仅介绍了开发过程的本体中,CDO(2021)[44]是一个以中文描述糖尿病知识库的本体,目前创建了1 484类,其中关系较为简单,暂时还未有文献对该本的设计进行详细的描述。MADHUSANKA S(2020)[45]则基于糖尿病药物治疗的临床实践指南,提出了构建临床实践指南及决策过程的本体的新方法。BRAVO M(2020)[35]描述了一个基于医学知识和质量设计标准搭建的糖尿病患者信息本体。DIAB(2015)[48]是一个用于整合小鼠疾病模型和人群研究中2型糖尿病疾病表型的本体,为基础和临床研究的沟通铺设桥梁。糖尿病诊断本体(Diabetes mellitus diagnosis ontology, DDO,2015前)[47]详细描述了用于糖尿病诊断的本体,现已整合于DMTO[36]。另外糖尿病本体(Diabetes Mellitus Ontology, DMO,2014)[49]介绍了用本体统一各个EHR中糖尿病患者数据的早期构想,并且将该本体应用在语义搜索问答中,以期在EHR中识别2型糖尿病的表型(表1)。

    表2总结了文献报道的8个基于本体的糖尿病治疗CDSS及其开发情况[34, 36, 46, 5054]。这些CDSS在框架上有较大异质性,而其共性可以归纳为图1。根据信息采集输入端可以分为问答器、EHR、传感器和多模态融合。

    表  2  基于本体的糖尿病治疗的临床决策支持系统
    Table  2.  Ontology-based clinical decision support systems for diabetes treatment
    Author (Date)PurposeAssociated ontology and rulesUser interfaceUsersInputInference engineOutputTime dimensionIntegrate with EHREvaluation
    SHERIMON P C (2020)[50] Treatment Ontology based NICE guideline
    SWRLrules
    A graphical communication interface Patient
    Healthcare providers
    Systeminterface
    Java engine
    SPARQL
    Pellet
    Treatment plans based on risk score Present and historical Not capable Not
    CHEN L (2019)[34] Prognosis
    Diagnosis
    Treatment
    OMDP A graphical communication interface
    EHR
    Patient
    EHR
    Patient profile
    D2RQ
    Pellet Treatment plans (medication, diet, exercise, education) Present and historical Capable 766 patients with clinical profiles
    El-SAPPAGH S (2019)[46] Treatment for type 1 diabetes FASTO Remote sensor
    A graphical communication interface
    EHR
    Patient
    Sensor
    Multiple EHR
    Patient profile
    FHIR
    D2RQ
    SPARQL
    Pellet
    Treatment plans (medication, diet, exercise, education) Present and historical
    Real-time sensor
    Capable A patient with clinical profiles
    El-SAPPAGH S (2018)[36] Treatment DMTO EHR EHR Patient profile SPARQL/SQWRL
    DL queries
    Hermit/Pellet
    FaCT++
    Treatment plans (medication, diet, exercise, education) Present and historical Capable Not
    CHEN (2016)[51] Determining the ranking of antidiabetic medications Drug knowledge ontology
    Fuzzy rules
    Jenarules
    A graphical communication interface Doctor Systeminterface
    Java engine
    Fuzzy
    Jena
    TOPSISalgorithm
    The target for HbA1c The ranking of antidiabetic medications Present and historical Not capable 10 patients with clinical profiles
    IRS-T2D (2016)[52] Treatment Ontology for antidiabetic drugs
    Ontology for patient profile
    SWRL rules
    A graphical communication interface Doctor Systeminterface
    Ontology for patient profile
    JESS Treatment plans Present Not capable 30 patients with clinical profiles
    OntoDiabetic (2015)[53] Assessing risk factors and treatment Domain ontology
    Process ontology
    Patient ontology
    A graphical communication interface Patient
    Healthcare provider
    Systeminterface
    Java engine
    OWL modeler
    Pellet
    Hermit
    Treatment plans Present Not capable Data from community
    CHEN R C (2012)[54] Antidiabetic medications selection Ontology for antidiabetic drugs
    Ontology for patient profile
    SWRL rules
    A graphical communication interface Doctor Systeminterface
    Java engine
    Ontology for patient profile
    JESS
    Pellet
    Treatment plans for medications Present Not capable 20 patients with clinical profiles
     NICE: the national institute for health and care excellence; OMDP: ontology-based model for diabetic patients; EHR: electronic health record; HL7: health level 7; FHIR: fast healthcare interoperability resources; FASTO: FHIR and SSN-based T1D ontology; DMTO: diabetes mellitus treatment ontology; DDO: diabetes mellitus diagnosis ontology; SWRL: semantic web rule language; OWL: web ontology language; JESS: Java expert system shell.
    下载: 导出CSV 
    | 显示表格
    图  1  糖尿病治疗领域的基于本体的临床决策支持系统的共同框架
    Figure  1.  Common framework for ontology-based clinical decision support systems for diabetes treatment
    CPGs: clinical practice guidelines; EHR: electronic health record.

    基于问答的CDSS通过用户主动录入信息作为信息输入端。其录入者可以是患者、医生、其他医疗服务提供者及多用户,应用场景可以是医院问诊或者患者在家的自我管理。一般通过统一的问答窗口收集患者的相关信息,收集到的信息与制定好的问题回答本体结合,形成标准化和结构化的患者信息,然后在Java引擎中转换为推理器所需要的格式,再输入进推理器,推理器结合数据、知识库和推理规则并推理生成结论。基于问答窗口的CDSS采用预设的结构化或非结构化问题收集患者信息,其对患者信息的采集效率和准确性高度依赖问题的设置和用户回答问题的质量。结构化信息的录入简化了输入信息的处理难度,但可能增加用户使用的难度。特别是影响糖尿病患者的决策信息很多,除了基本人口信息之外,还有血糖监测、生活方式等信息,这些都给临床医生和患者录入信息增加了难度。非结构化信息,例如自然语言、检查单扫描件等,需要通过一定技术与回答本体对应进行数据的标准化和结构化。该输入简化并方便了用户使用,但增加了信息识别的难度。因此,由临床、医学信息、计算机科学和临床流行病专家共同组成的研究团队对问题及问答器的设置,对提升CDSS的使用效果及用户体验至关重要。在已发表的基于本体的CDSS中,CHEN R C(2012)[54]通过医生输入患者的信息,采集的患者信息种类较少。而OntoDiabetic(2016)[53]和IRS-T2D(2016)[52]则需要用户输入更多的数据信息。CHEN R C(2016)[51]在此基础上引入了信息的时间维度,大大丰富了输入信息,但也给用户使用的便捷性带来了一定挑战。这些CDSS的信息录入往往发生在医患交流过程中,更方便医生综合患者的实际情况、价值观与偏好。一些CDSS(CHEN R C,2016)[51]使用了模糊算法计算个性化的糖化血红蛋白控制目标,并采用TOPSIS算法考虑多维度因素,综合评价用药方案并给出推荐药物的排序。然而,这些CDSS输入端高度依赖医生和患者的充分沟通和信息录入的准确性。糖尿病患者包括血糖监测记录在内的医疗信息较多,仅依据用户录入方式很难全面采集患者信息。

    在EHR或电子病历系统(electronic medical record, EMR)建立完善的医疗机构,糖尿病的大量就诊信息都有详细的记录和标准化的储存。基于EHR的CDSS是通过EHR与CDSS输入端对接,直接导入患者在医院的健康数据,包括人口学信息、生命体征、合并症及并发症、实验室检查、药物处方、放射学报告、手术记录等。El-SAPPAGH等先后设计了DDO(2015年前)[47],DMTO(2015)[36]和FASTO(2018)[46]三项糖尿病领域的本体,均可对接EHR数据。DDO用于糖尿病的诊断,DMTO是目前最完善的针对2型糖尿病治疗的本体,FASTO用于糖尿病患者特别是1型糖尿病患者的胰岛素实时管理。其中FASTO是在DMTO和DDO的基础上扩展而来。它最大的更新在于加入胰岛素剂量调整的规则,其他领域知识与规则也得到了不同程度的优化和扩充。此外,FASTO规定了患者数据接口的标准模式为FHIR标准(为不同系统之间的通信提供协议与标准),可以支持移动健康应用的开发,支持云环境上的临床决策支持系统,支持分布式EHR、移动设备和无线区域网络之间的互操作性。但其目前只针对1型糖尿病,其临床应用大为受限。此外,它与DMTO一样欠缺不同临床场景的治疗方案、含有多种必要的饮食营养素的饮食方案和用户友好的计量单位。OMDP(2019)[34]相较于FASTO和DMTO,增加了基因的知识和糖尿病的分型,整合了预防、筛查和治疗。但其目前并未公布患者数据接口标准化的模式以及由原始EHR数据转换为推理引擎可用数据的细节。

    基于传感器的CDSS通过输入端与传感器对接,收集患者数据。传感器可以提供患者实时的监测数据,包括基于可穿戴设备实时监控下的生命体征、血糖和血糖的波动轨迹。文献表明血糖波动对糖尿病的发病和预后均有着重要影响,因此连接传感器的数据将有助于CDSS得出有效的治疗方案。目前,语义传感器网络本体(semantic sensor network ontology, SSN)是W3C推荐的用于描述传感器的观测值、过程、特征、样本、观测属性以及执行器的本体[55]。Health level 7 FHIR标准也可为患者传感器数据、EHR数据和知识本体提供转换的标准,使其能互相理解共同操作[46, 56]。FASTO介绍了基于传感器的CDSS构建的方法和细节[46]。但由于可穿戴设备使用的限制,这些CDSS在我国的临床应用尚有待发展。

    多模态融合的CDSS的输入端可以同时通过多种途径收集患者信息,包括EHR、传感器、手机、问答窗口等。EHR能提供医院内患者的相关健康信息,传感器可以提供患者实时的血糖和生命体征数据,问答窗口可以设置问题补充EHR无法提供的信息,比如:饮食、价值观与偏好等。三者结合可以为决策提供最全面的患者信息。OMDP[34]和FASTO[46]均为多模态融合的CDSS,允许EMR数据和问答窗口输入信息,FASTO[46]允许接入传感器数据。

    目前公认的CDSS本体与推理均基于临床指南、医学文献或者专家临床经验,给出的治疗方案具有高度的可解释性。但这些CDSS均未在现实医疗系统中使用,因此缺乏关于它们在现实应用中处理大量患者数据时存储数据、逻辑推理、给出推断的能力与表现的评估报告。

    基于本体的CDSS可以利用本体高度的灵活性和适应性,将指南推荐意见结构化,并与患者数据对接,通过推理系统得出临床可解释的推荐意见。在我国糖尿病大爆发的背景下,这类CDSS将有助于推进诊疗效率和提升医疗公平性。然而,不同于其他学科,糖尿病临床研究结论存在高度不确定性,且患者价值观与偏好存在高度异质性,这进一步导致CDSS很难给出单一准确的建议。因此,如何将循证医学与本体相融合,构建可信的临床推荐体系则是目前CDSS发展的重点。目前,BMJ期刊与(Making GRADE the irresistible choice, MAGIC)证据生态基金会合作建立的系列临床实践指南——BMJ快速推荐提供了一个解决问题的范例[57]。这些临床实践指南使用了高度透明的循证医学证据和决策模式,使其自身可以快速顺利地被结构化并形成关系本体,进而通过推理系统形成个体化推荐。同时,透明的循证医学证据也可以借助交互式工具辅助医患共同决策,在最大程度上降低医患共同决策的成本。这些先进的方法值得我国学习,但我国更需要研发具有自主知识产权的先进糖尿病决策支持工具。

    *    *    *

    利益冲突 所有作者均声明不存在利益冲突

  • 图  1   糖尿病治疗领域的基于本体的临床决策支持系统的共同框架

    Figure  1.   Common framework for ontology-based clinical decision support systems for diabetes treatment

    CPGs: clinical practice guidelines; EHR: electronic health record.

    表  1   糖尿病治疗相关的本体及其特点

    Table  1   Characteristics of ontologies for diabetes treatment

    Name of ontology or developer (Released date/update date)Domain/
    objective
    Main componentsSource of knowledgeLanguage and toolMain terminologyTop-level ontology or standardsReused existing ontologyUsed in clinical practice
    Need patient medical recordsKnowledge baseRule
    CDO (2021/2022)[44] Knowledge base Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    SWRL NA Protégé
    OWL2
    SWRL
    SNOMED-CT BFO
    OGMS
    HPO
    DOID
    DRON
    No
    MADHUSANKA S (2020)[45]* Knowledge base Yes Clinical practice guideline OWL2 CPGs Protégé
    OWL2
    NA NA NA No
    BRAVO M (2020)[35]* Patient profile Yes Patient profile
    Exercise and diet
    BMI
    Medication
    Treatment plans
    Not applicable Other ontologies Protégé
    OWL2
    LOINC NA NDF-RT
    DOID
    SYMP
    No
    SHERIMON P C (2020)[50]* Treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    OWL2 NICE guidelines Protégé
    OWL2
    NA NA NA No
    OMDP (2019/NA)[34]* Prognosis
    Diagnosis
    Treatment
    Yes Gene
    Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Time
    Treatment plans
    Medication and interaction
    SWRL CPGs
    Domain expert
    Literature
    EHR
    Protégé
    OWL2
    SWRL
    RxNorm
    SNOMED CT
    LOINC
    CPT
    BFO
    OGMS
    DOID
    SYMP
    GO
    Diet ontology
    Physical activity ontology
    No
    FASTO (2018/2018)[46] Treatment for type 1 diabetes Yes Sensor data
    Patient value and preference
    Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Time
    Treatment plans
    Medication and interaction
    SWRL CPGs
    Domain expert
    Literature
    Official web sites
    EHR
    Protégé 5.1
    OWL2
    SWRL
    RxNorm, SNOMED CT, LOINC
    BFO
    SSN
    HL7 FHIR
    Vital-sign ontology
    DDO
    DMTO
    SWRL TO
    OAE
    No
    DMTO (2017/2017)[36] Treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Time
    Treatment plans
    Medication and interaction
    SWRL
    DL-safe
    CPGs
    Domain expert
    Literature
    Official web sites
    EHR
    Book
    Protégé 5.0
    SWRL
    OWL2
    SNOMED CT
    RxNorm
    NDF-RT
    BFO
    OGMS
    DDO
    DINTO
    RO
    SMASH
    PATO
    OntoFood
    SYMP
    TIME
    No
    DDO (before2015/
    NA)[47]*
    Diagnosis Yes Patient profile
    Exercise and diet
    Medication
    Basic knowledge of diabetes
    SWRL CPGs
    Domain expert
    Literature
    Official web sites
    EHR
    Book
    Protégé 5.0
    SWRL
    OWL2
    SNOMED CT
    RxNorm
    NDF-RT
    UMLS
    BFO
    OGMS
    DOID
    SYMP
    UO
    OGMD
    No
    CHEN R C (2016/NA)[51]* Determining the ranking of antidiabetic medications Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Medication and interaction
    Fuzzy
    Jena
    TOPSIS algorithm
    CPGs
    Domain expert
    EHR
    Protégé
    WebProtégé
    Jena
    OWL2
    NA NA NA No
    IRS-T2D (2016/NA)[52]* Treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    SWRL
    OWL2
    Domain expert
    Literature
    Protégé
    OWL
    SWRL
    NA NA NA No
    OntoDiabetic (2015/NA)[53]* Assessing risk factors and treatment Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Treatment plans
    Medication and interaction
    DL-safe
    OWL2
    CPGs Protégé
    OWL2
    DL-safe
    NA NA Answered
    Questionnaire ontology
    Adaptive
    Questionnaire ontology
    No
    DMO (2014/NA)[49]* Patient register, automated system for EHR data Yes Patient profile
    Basic knowledge of diabetes
    Medication
    NA CPGs
    Domain expert
    Protégé 4.3 SNOMED CT NA NA No
    CHEN R C (2012/NA)[54]* Antidiabetic medications selection Yes Patient profile
    Exercise and diet
    Basic knowledge of diabetes
    Medication and interaction
    SWRL CPGs Protégé
    SWRL
    OWL DL
    NA NA NA No
     CDO: Chinese diabetes ontology; BFO: basic formal ontology; OGMS: ontology for general medical science; HPO: human phenotype ontology; DOID: human disease ontology; DRON: the drug ontology; LOINC: the logical observation identifiers names and codes system; NDF-RT: national drug file-reference terminology; SYMP: the symptom ontology; CPT: current procedure terminology; GO: gene ontology; OMDP: ontology-based model for diabetic patients; HL7: health level 7; FHIR: fast healthcare interoperability resources; FASTO: FHIR and SSN-based T1D ontology; DMO: diabetes management ontology; SWRL: semantic web rule language; OWL: web ontology language; DL: description logic; SSN: semantic sensor network ontology; OAE: ontology of adverse events; DINTO: the drug-drug interactions ontology; PATO: phenotype and trait ontology; SMASH: semantic mining of activity, social, and health ontology; DMTO: diabetes mellitus treatment ontology; DDO: diabetes mellitus diagnosis ontology; UO: units of measurement ontology; OGMD: ontology of glucose metabolism disorder; UMLS: unified medical language system; BMI: body mass index; EHR: electronic health record; NICE: the national institute for health and care and excellence; CPGs: clinical practice guidelines. * The OWL of the ontology is available; NA: not available.
    下载: 导出CSV

    表  2   基于本体的糖尿病治疗的临床决策支持系统

    Table  2   Ontology-based clinical decision support systems for diabetes treatment

    Author (Date)PurposeAssociated ontology and rulesUser interfaceUsersInputInference engineOutputTime dimensionIntegrate with EHREvaluation
    SHERIMON P C (2020)[50] Treatment Ontology based NICE guideline
    SWRLrules
    A graphical communication interface Patient
    Healthcare providers
    Systeminterface
    Java engine
    SPARQL
    Pellet
    Treatment plans based on risk score Present and historical Not capable Not
    CHEN L (2019)[34] Prognosis
    Diagnosis
    Treatment
    OMDP A graphical communication interface
    EHR
    Patient
    EHR
    Patient profile
    D2RQ
    Pellet Treatment plans (medication, diet, exercise, education) Present and historical Capable 766 patients with clinical profiles
    El-SAPPAGH S (2019)[46] Treatment for type 1 diabetes FASTO Remote sensor
    A graphical communication interface
    EHR
    Patient
    Sensor
    Multiple EHR
    Patient profile
    FHIR
    D2RQ
    SPARQL
    Pellet
    Treatment plans (medication, diet, exercise, education) Present and historical
    Real-time sensor
    Capable A patient with clinical profiles
    El-SAPPAGH S (2018)[36] Treatment DMTO EHR EHR Patient profile SPARQL/SQWRL
    DL queries
    Hermit/Pellet
    FaCT++
    Treatment plans (medication, diet, exercise, education) Present and historical Capable Not
    CHEN (2016)[51] Determining the ranking of antidiabetic medications Drug knowledge ontology
    Fuzzy rules
    Jenarules
    A graphical communication interface Doctor Systeminterface
    Java engine
    Fuzzy
    Jena
    TOPSISalgorithm
    The target for HbA1c The ranking of antidiabetic medications Present and historical Not capable 10 patients with clinical profiles
    IRS-T2D (2016)[52] Treatment Ontology for antidiabetic drugs
    Ontology for patient profile
    SWRL rules
    A graphical communication interface Doctor Systeminterface
    Ontology for patient profile
    JESS Treatment plans Present Not capable 30 patients with clinical profiles
    OntoDiabetic (2015)[53] Assessing risk factors and treatment Domain ontology
    Process ontology
    Patient ontology
    A graphical communication interface Patient
    Healthcare provider
    Systeminterface
    Java engine
    OWL modeler
    Pellet
    Hermit
    Treatment plans Present Not capable Data from community
    CHEN R C (2012)[54] Antidiabetic medications selection Ontology for antidiabetic drugs
    Ontology for patient profile
    SWRL rules
    A graphical communication interface Doctor Systeminterface
    Java engine
    Ontology for patient profile
    JESS
    Pellet
    Treatment plans for medications Present Not capable 20 patients with clinical profiles
     NICE: the national institute for health and care excellence; OMDP: ontology-based model for diabetic patients; EHR: electronic health record; HL7: health level 7; FHIR: fast healthcare interoperability resources; FASTO: FHIR and SSN-based T1D ontology; DMTO: diabetes mellitus treatment ontology; DDO: diabetes mellitus diagnosis ontology; SWRL: semantic web rule language; OWL: web ontology language; JESS: Java expert system shell.
    下载: 导出CSV
  • [1]

    JIA W. Diabetes: A challenge for China in the 21st century. Lancet Diabetes Endocrinol,2014,2(4): e6–e7. DOI: 10.1016/S2213-8587(14)70027-0

    [2] 侯清涛, 李芸, 李舍予, 等. 全球糖尿病疾病负担现状. 中国糖尿病杂志,2016,24(1): 92–96. DOI: 10.3969/j.issn.1006-6187.2016.01.023
    [3]

    XU Y, WANG L, HE J, et al. Prevalence and control of diabetes in Chinese adults. JAMA,2013,310(9): 948–959. DOI: 10.1001/jama.2013.168118

    [4]

    MAO W, YIP C M W, CHEN W. Complications of diabetes in China: health system and economic implications. BMC Public Health,2019,19(1): 269. DOI: 10.1186/s12889-019-6569-8

    [5]

    CHAN J C, ZHANG Y, NING G. Diabetes in China: A societal solution for a personal challenge. Lancet Diabetes Endocrinol,2014,2(12): 969–979. DOI: 10.1016/S2213-8587(14)70144-5

    [6] 安康, 苏巧俐, 徐娇, 等. 紧密型医联体内社区卫生服务中心全科医生医疗能力培训教学需求分析. 中国全科医学,2020,23(22): 2825–2830. DOI: 10.12114/j.issn.1007-9572.2020.00.387
    [7] 赵梅村, 蒋凤, 张婷, 等. 国外糖尿病专科医疗服务体系对完善我国糖尿病专科医联体建设的启示. 中华糖尿病杂志,2020,12(4): 263–267. DOI: 10.3760/cma.j.cn115791-20190922-00346
    [8] 蔡淳, 贾伟平. 中国糖尿病的社区化管理. 中国科学: 生命科学,2018,48(8): 820–826. DOI: 10.1360/N052018-00048
    [9]

    BAUMFELD-ANDRE E, CARRINGTON N, SIAMI F S, et al. The current landscape and emerging applications for real-world data in diagnostics and clinical decision support and its impact on regulatory decision making. Clin Pharmacol Ther,2022,112(6): 1172–1182. DOI: 10.1002/cpt.2565

    [10]

    OSHEROFF J A, TEICH J M, LEVICK D, et al. Improving outcomes with clinical decision support: An implementer’s guide. 2nd ed. Boca Raton: HIMSS Publishing, 2012. https://doi.org/10.4324/9780367806125.

    [11]

    GREENES R A. Clinical decision support: The road to broad adoption. 2nd ed. London, UK; Waltham, MA, USA: Elsevier/Academic Press, 2014.

    [12]

    NOY N F, MCGUINNESS D L. Ontology development 101: a guide to creating your first ontology. Technical Report KSL-01-05 and SMI-2001-0880, Stanford Knowledge Systems Laboratory and Stanford Medical Informatics, 2001.[2022-03-19]. http://ksl.stanford.edu/people/dlm/papers/ontology-tutorial-noy-mcguinness-abstract.html.

    [13]

    CONWAY N, ADAMSON K A, CUNNINGHAM S G, et al. Decision support for diabetes in Scotland: Implementation and evaluation of a clinical decision support system. J Diabetes Sci Technol,2018,12(2): 381–388. DOI: 10.1177/1932296817729489

    [14] 王小贤, 夏晨曦, 张芳芳, 等. 本体在糖尿病领域知识表示与语义推理研究和应用综述. 医学信息学杂志,2017,38(7): 56–61. DOI: 10.3969/j.issn.1673-6036.2017.07.014
    [15] 高星, 王岩, 秦盼盼, 等. 基于本体的糖尿病知识库构建. 中华医学图书情报杂志,2018,27(10): 8–13. DOI: 10.3969/j.issn.1671-3982.2018.10.002
    [16] 陈桂芬, 汪江, 杨志刚. 基于本体的规则推理和案例推理结合的糖尿病诊疗专家系统研究. 长春大学学报(自然科学版),2016,26(3): 19–25. DOI: 10.3969/j.issn.1009-3907.2016.03.004
    [17] 张莉, 王玉廷. 基于Protege的糖尿病本体构建. 科学咨询,2020(11): 9–11.
    [18] 刘智锋, 夏晨曦, 黄梨, 等. 糖尿病领域本体构建与语义推理实现. 中华医学图书情报杂志,2017,26(9): 7–11. DOI: 10.3969/j.issn.1671-3982.2017.09.002
    [19]

    CHATTERJEE A, PRINZ A, GERDES M, et al. An automatic ontology-based approach to support logical representation of observable and measurable data for healthy diet management: Proof-of-concept study. J Med Internet Res,2021,23(4): e24656. DOI: 10.2196/24656

    [20]

    DISSANAYAKE P I, COLICCHIO T K, CIMINO J J. Using clinical reasoning ontologies to make smarter clinical decision support systems: A systematic review and data synthesis. J Am Med Inform Assoc,2020,27(1): 159–174. DOI: 10.1093/jamia/ocz169

    [21]

    REYES-PEÑA C, TOVAR M, BRAVO M, et al. An ontology network for diabetes mellitus in Mexico. J Biomed Semantics,2021,12(1): 19. DOI: 10.1186/s13326-021-00252-2

    [22]

    SPEAR A D, CEUSTERS W, SMITH B. Functions in basic formal ontology. Appl Ontol,2016,11: 103–128. DOI: 10.3233/AO-160164

    [23]

    ARP R, SMITH B. Function, role, and disposition in basic formal ontology. Nat Prec,2008: 1–1. DOI: 10.1038/npre.2008.1941.1

    [24]

    CORCHO O, GÓMEZ-PÉREZ A. A roadmap to ontology specification languages//International conference on knowledge engineering and knowledge management. Berlin, Heidelberg: Springer, 2000: 80-96.

    [25]

    World Wide Web Consortium. OWL 2 web ontology language document overview. 2012. [2022-03-19]. https://www.w3.org/2007/OWL/draft/ED-owl2-overview-20120808.

    [26]

    SLIMANI T. Ontology development: A comparing study on tools, languages and formalisms. Indian J Sci Technol,2015,8(24): 1–12. DOI: 10.17485/ijst/2015/v8i34/54249

    [27]

    About the OBO Foundry(2022)[2021–11–23]. http://www.obofoundry.org/about-OBO-Foundry.html.

    [28]

    SMITH B, ASHBURNER M, ROSSE C, et al. The OBO Foundry: Coordinated evolution of ontologies to support biomedical data integration. Nat Biotechnol,2007,25(11): 1251. DOI: 10.1038/NBT1346

    [29]

    JACKSON R, MATENTZOGLU N, OVERTON JA, et al. OBO Foundry in 2021: Operationalizing open data principles to evaluate ontologies. Database (Oxford),2021,2021: baab069. DOI: 10.1093/database/baab069

    [30]

    RUBIN D L, LEWIS S E, MUNGALL C J, et al. National center for biomedical ontology: Advancing biomedicine through structured organization of scientific knowledge. OMICS,2006,10(2): 185–198. DOI: 10.1089/omi.2006.10.185

    [31]

    MUSEN M A, NOY N F, SHAH N H, et al. The National center for biomedical ontology. J Am Med Inform Assoc,2012,19(2): 190–195. DOI: 10.1136/amiajnl-2011-000523

    [32]

    MARTÍNEZ-ROMERO M, JONQUET C, O'CONNOR M J, et al. NCBO Ontology Recommender 2.0: An enhanced approach for biomedical ontology recommendation. J Biomed Semantics,2017,8(1): 21. DOI: 10.1186/s13326-017-0128-y

    [33]

    WHETZEL P L, NOY N F, SHAH N H, et al. BioPortal: enhanced functionality via new Web services from the National Center for Biomedical Ontology to access and use ontologies in software applications. Nucleic Acids Res,2011,39(Web Server issue): W541–W545. DOI: 10.1093/nar/gkr469

    [34]

    CHEN L, LU D, ZHU M, et al. OMDP: An ontology-based model for diagnosis and treatment of diabetes patients in remote healthcare systems. Int J Distrib SensNetw,2019,15(5): 1550147719847112.

    [35]

    BRAVO M, GONZÁLEZ D, ORTIZ J A R, et al. Management of diabetic patient profiles using ontologies. Contaduríay administración,2020,65(5): 12. DOI: 10.1177/1550147719847112

    [36]

    El-SAPPAGH S, KWAK D, ALI F, et al. DMTO: A realistic ontology for standard diabetes mellitus treatment. J Biomed Semantics,2018,9(1): 8. DOI: 10.1186/s13326-018-0176-y

    [37]

    SUTTON R T, PINCOCK D, BAUMGART D C, et al. An overview of clinical decision support systems: Benefits, risks, and strategies for success. NPJ Digit Med,2020,3: 17. DOI: 10.1038/s41746-020-0221-y

    [38]

    BERNER E S. Clinical decision support systems. New York: Springer Science+ Business Media, LLC, 2007.

    [39]

    MIDDLETON B, SITTIG D F, WRIGHT A. Clinical decision support: A 25 year retrospective and a 25 year vision. Yearb Med Inform,2016,Suppl 1(Suppl 1): S103–S116. DOI: 10.15265/IYS-2016-s034

    [40]

    ADEL E, El-SAPPAGH S, BARAKAT S, et al. Distributed electronic health record based on semantic interoperability using fuzzy ontology: A survey. Int J Comput Appl,2018,40(4): 223–241. DOI: 10.1080/1206212X.2017.1418237

    [41]

    PELEG M. Computer-interpretable clinical guidelines: A methodological review. J Biomed Inform,2013,46(4): 744–763. DOI: 10.1016/j.jbi.2013.06.009

    [42]

    SHOAIP N, El-SAPPAGH S, BARAKAT S, et al. Reasoning methodologies in clinical decision support systems: A literature review//DEY N, ASHOUR A S, FONG S J, et al. U-Healthcare Monitoring Systems: Volume One: Design and Applications. New York: Acad Press, 2019: 61-87. doi: 10.1016/B978-0-12-815370-3.00004-9.

    [43]

    El-RASHIDY N, El-SAPPAGH S, ISLAM S M, et al. Mobile health in remote patient monitoring for chronic diseases: Principles, trends, and challenges. Diagnostics,2021,11(4): 607. DOI: 10.3390/diagnostics11040607

    [44]

    Chinese Diabetes Mellitus Ontology.(2022-04-19)[2022–05–23]. https://bioportal.bioontology.org/ontologies/CDO.

    [45]

    MADHUSANKA S, WALISADEERA A, DANTANARAYANA G, et al. An ontological clinical decision support system based on clinical guidelines for diabetes patients in Sri Lanka. Healthcare (Basel),2020,8(4): 573. DOI: 10.3390/healthcare8040573

    [46]

    EL-SAPPAGH S, ALI F, HENDAWI A, et al. A mobile health monitoring-and-treatment system based on integration of the SSN sensor ontology and the HL7 FHIR standard. BMC Med Inform Decis Mak,2019,19(1): 97. DOI: 10.1186/s12911-019-0806-z

    [47]

    El-SAPPAGH S, ALI F. DDO: A diabetes mellitus diagnosis ontology. Appl Inform,2016,3(1): 5. DOI: 10.1186/s40535-016-0021-2

    [48]

    BioMedBridges Diabetes Ontology. (2015-11-27)[2021–11–23]. https://bioportal.bioontology.org/ontologies/DIAB.

    [49]

    RAHIMI A, LIAW S T, TAGGART J, et al. Validating an ontology-based algorithm to identify patients with type 2 diabetes mellitus in electronic health records. Int J Med Inform,2014,83(10): 768–778. DOI: 10.1016/j.ijmedinf.2014.06.002

    [50]

    SHERIMON P C, SHERIMON V. Design and implementation of ontology based risk assessment model in diabetes mellitus. ARIV Int J Technol,2020,1(2): 19–28.

    [51]

    CHEN R C, JIANG H Q, HUANG C Y, et al. Clinical decision support system for diabetes based on ontology reasoning and TOPSIS analysis. J Healthc Eng,2017,2017: 4307508. DOI: 10.1155/2017/4307508

    [52]

    MAHMOUD N, ELBEH H. IRS-T2D: Individualize recommendation system for type2 diabetes medication based on ontology and SWRL. INFOS,2016: 203–209. DOI: 10.1145/2908446.2908495

    [53]

    SHERIMON P C, KRISHNAN R. OntoDiabetic: an ontology-based clinical decision support system for diabetic patients. Arab J Sci Eng,2016,41(3): 1145–1160. DOI: 10.1007/s13369-015-1959-4

    [54]

    CHEN R C, HUANG Y H, BAU C T, et al. A recommendation system based on domain ontology and SWRL for anti-diabetic drugs selection. Expert Syst Appl,2012,39(4): 3995–4006. DOI: 10.1016/j.eswa.2011.09.061

    [55]

    Semantic Sensor Network Ontology. (2017-12-08)[2021–11–23]. https://www.w3.org/TR/vocab-ssn/.

    [56]

    Overview - FHIR v4.0. 1[2021–11–23]. http://www.hl7.org/fhir/overview.html.

    [57]

    SIEMIENIUK R A, AGORITSAS T, MACDONALD H, et al. Introduction to BMJ rapid recommendations. BMJ,2016,354: i5191. DOI: 10.1136/bmj.i5191

  • 期刊类型引用(4)

    1. 王敏,胡兆,徐晓巍,郑思,李姣,姚焰. 融合知识驱动和数据驱动的混合决策模型构建:以室性心动过速病因诊断为例. 协和医学杂志. 2025(02): 454-461 . 百度学术
    2. 张山,陆观,刘璐,王慧莹,吴瑛. 基于PADIS指南的知识图谱构建研究. 医学信息学杂志. 2024(12): 56-61 . 百度学术
    3. 陈向阳,李舍予. 非内分泌病房住院糖尿病患者的血糖管理. 中国全科医学. 2023(15): 1799-1803 . 百度学术
    4. 陆秋娴,肖冠坤,李毓萍,贾莉莎,江丽,彭天宇,崔佳,卜凡,郭云,李鸣. 低血糖负荷食物交换份法结合移动饮食管理对2型糖尿病患者血糖改善的效果探索. 现代预防医学. 2023(10): 1784-1789 . 百度学术

    其他类型引用(1)

cc

开放获取 本文遵循知识共享署名—非商业性使用4.0国际许可协议(CC BY-NC 4.0),允许第三方对本刊发表的论文自由共享(即在任何媒介以任何形式复制、发行原文)、演绎(即修改、转换或以原文为基础进行创作),必须给出适当的署名,提供指向本文许可协议的链接,同时标明是否对原文作了修改;不得将本文用于商业目的。CC BY-NC 4.0许可协议详情请访问 https://creativecommons.org/licenses/by-nc/4.0

图(1)  /  表(2)
计量
  • 文章访问数:  972
  • HTML全文浏览量:  406
  • PDF下载量:  102
  • 被引次数: 5
出版历程
  • 收稿日期:  2022-04-13
  • 修回日期:  2022-12-05
  • 网络出版日期:  2023-01-16
  • 发布日期:  2023-01-19

目录

/

返回文章
返回