药物不良事件信息资源整合与数据挖掘研究
|
摘要
目前，药物不良事件的发生日益成为一个严重的公共卫生问题。药物上市前虽然经过严格的不良事件实验研究，但仍不能够发现所有的潜在不良事件。在20世纪60年代“反应停”（thalidomide）事件之后，许多国家引入了药物警戒（phamacovigilance）系统对上市药品进行监测。美国药品与食品管理局（Food andDrug Administration，FDA）的药物不良事件报告系统（Adverse Event ReportingSystem，AERS）数据库主要用于发现那些在临床试验阶段由于出现频次低而没有被识别出的罕见严重不良事件，或者新的药品不良事件，即安全性信号。如果在AERS中发现药物潜在的安全问题，FDA将进行流行病学研究以进一步评价该不良事件，确定药物与不良事件之间的因果关系。基于对药物不良事件的安全评价，FDA可能采取一系列的法规调整以提高产品安全及保障公众健康，如更新药品说明信息，限制使用药品，向公众介绍新的安全相关信息，或在少数情况下，从市场上撤销该药品。
当前，关于不良事件数据挖掘的多数研究都集中于利用小部分数据进行不良事件的数据挖掘，避免对大规模数据的利用和研究；对药物不良事件进行药物的作用机制、药代动力学及生理作用等方面的深度挖掘的研究，以及对某类药物的不良事件进行比较性数据挖掘的研究、进行药物的作用机制、药代动力学及生理作用等方面的深度挖掘的研究，以及对某类药物的不良事件进行比较性数据挖掘的研究、AERS与其他数据源的集成难以实现难以实现，而这类大规模、深层次的挖掘在揭示不同药物类别的不良事件特征、药物不良事件的原因以及基因相关性方面都具有重要意义，是药物不良事件监测乃至临床用药安全研究的重要方向。缺乏对药物不良事件相关数据资源的知识整合严重限制了上述研究。
药物不良事件信息资源的知识整合既是有效利用海量医疗信息资源的现实需求，也是提高药物不良事件数据挖掘效率所需认真研究并必须解决的关键问题。近年来药物领域本体的发展虽然为资源整合研究提供了实现契机，然而由于药物领域本体的复杂性、数据缺乏规范化以及领域本体映射的技术难题，药物不良事件领域数据在知识集成与深度聚合方面始终未能求得理想的解决方案，药物不良事件的数据挖掘也因此未能扩展到对大规模数据的利用和分析。
领域本体可以提供相关知识决策和推理支持，促进大规模药物安全信号的检测和药物不良事件的深度挖掘。本研究利用生物医学领域本体将AERS相关信息资源有机整合起来,实现知识集成、信息聚合、与其他医疗数据资源之间的互操作、丰富了药物不良事件数据挖掘的资源并促进对药物安全信号的检测。
本研究的主要内容包括：
（1）提出药物领域本体映射与聚合模型
实现本体映射以及对药物信息的分类与聚合将为药物相关知识决策和推理支持提供前提条件，同时也是构建领域知识库的重要基础，对于进一步针对药物的用机制、药代动力学及生理作用等方面的深度数据挖掘具有重要意义。由于领域本体自身结构的复杂性和领域本体之间的异构性，药物领域本体映射方法成为实现本体映射的难点之一。本研究提出药物领域本体映射与聚合模型模型，并以该模型为指导，对药物领域本体RxNorm与NDF-RT（美国国家药物文件—参考术语）进行映射实例研究，提出了RxNorm与NDF-RT两个领域本体之间映射及信息分类与聚合的一种新方法。研究结果证明该模型不仅具有可行性，也显示出其对多本体能够充分复用的实践价值；该模型也将在语义层面上进一步深化信息资源的知识组织方法，促进数字资源语义体系的构建。模型的不足之处在于，模型的使用是以现有本体为基础的，因此现有本体中的概念关系以及分类聚合信息的不足将将最终影响本体映射分类聚合的效果。另外，领域本体的其他特性也可能是改善知识组织方法的因素，因此，未来研究中应对领域本体进行更全面的调研，抽取有效的共有特征，促进模型的完善。
（2）基于RxNorm的AERS药名规范化初步研究
调查AERS药名被RxNorm的收录情况，是探索如何充分发挥RxNorm在AERS数据挖掘中作用的第一步，也是至关重要的一步。
本研究计算2004年到2010年AERS中全部药物名称与RxNorm精确匹配的比例，并与UMLS进行比较分析。结果显示了RxNorm和UMLS对AERS中唯一药物名称精确匹配的整体收录范围分别为13,565(4.8%)个与21,272(7.5%)个。2011AA版UMLS集成了160个源词汇表， UMLS对AERS的药名覆盖分别来自包括RxNorm在内的各种来源词表，其中RxNorm映射的数量排列第一。然后手工分析了频次大于1000的200个未被映射的高频AERS药物名称及分析388个随机选择的频次小于1000的低频药物名称，调查了某些药名未被映射的原因。尽管在AERS中，数据来源广泛且存在录入错误，但是高频词仍然能够显示出特定领域的词汇使用习惯。我们的研究将为RxNorm本体的完善提供依据。本章的研究也对下章研究中选择自然语言处理工具MedEx（以RxNorm为基础）提供了依据。
（3）构建数据挖掘知识整合库（AERS-DM）
在AERS药名规范化进行调查研究基础上，选择利用自然语言处理工具MedEx对AERS中药名进行规范化，并对其自然语言处理效果进行评价。在药物领域本体映射与聚合模型模型的基础上，使用贪婪算法将AERS中的药名聚合到RxNorm和NDF-RT中的药物分类信息。对于药物不良事件，通过映射方法将其映射到MedDRA中的PT和SOC代码进行聚合。最终建立开源的药物——不良事件数据挖掘知识整合库(AERS-DM)（网址：http://informatics.mayo.edu/adepedia/index.php/Download），最后通过实例研究，证实了AERS-DM数据集的挖掘效果。
AERS-DM中的信息集成了药物及不良事件知识库。AERS-DM具有规范化代码和聚合的功能，可以为AERS药物安全信号的挖掘以及相关数据挖掘领域提供更多的资源。该数据集包含两个表。一个表存储药物及不良事件的规范化信息，另一个表存储药物和不良事件的聚合信息。共有37,029,228对药物及不良事件记录。AERS中的药名被规范为14,490个RxNorm药名（由RxNorm代码表示），其中10,221个规范化的药名可以归到NDF-RT类别中，占71%。对于AERS-DM中的不良事件，共有14,740个MedDRA中的PT术语被聚合到MedDRA的SOC代码，占MedDRA中所有PT术语的76%。AERS-DM中，RxNorm代码表示的药名与MedDRA中的PT唯一对，即规范化后的药名与不良事件的唯一配对，共有4,639,613，将不良事件按组织器官聚合后，药物与不良事件组织器官的配对共205,725对。
（4）AERS-DM数据挖掘知识整合库的数据挖掘实证研究
AERS-DM是一个规范化和聚合的数据挖掘知识整合库，优势在于药物数据的规范化，以及药物数据和不良事件数据的分类聚合，这些分类聚合知识全部来自于生物医学本体中所含有的知识结构。传统的利用AERS进行的不良事件检测研究大多仅针对少量药物，进行大规模数据挖掘的研究数量较少。在本研究中，我们利用常用抗癌药物成分信息对药物作用机制、生理作用、治疗意向的药物聚类与药物不良事件的聚类，以及年龄与性别的药物不良事件差别进行了大规模的系统分析，进一步证实了AERS-DM的语义挖掘潜力。
传统的不良事件检测依赖比例失衡测度，主要是量化出药物-不良事件关联的“始料未及”的程度，并试图克服自发报告系统中不良事件缺乏疾病发生率背景信息的缺点。在此研究中我们提出了一种新的不良事件检测方法，在这种方法中，通过将AERS数据与电子病历数据连接起来，从而获得不良事件的疾病发生率信息，并实现大规模药物不良事件之间的比较研究。本研究证实了AERS-DM作为AERS的一个高级版本，是一个可用于数据挖掘的丰富资源。
本文的创新点包括：
（1）理论创新
提出药物领域本体映射与聚合模型。由于本体开发的局限性，当前领域本体各有特点，因此本研究提出的药物领域本体映射与聚合模型，充分利用不同本体的特点，通过本体映射，将某一本体的分类信息与其他本体的内容形成互补，实现某一领域多个本体的分类聚合功能，从而节约本体开发成本，实现本体充分复用。
（2）方法创新
（i）在药物领域本体映射与聚合模型的基础上，开发出一套系统的分类聚合算法，实现利用NDF-RT与RxNorm对AERS数据库中的药物进行分类聚合。方法创新体现在两方面：①利用RxNorm中的丰富关系来推理出可以映射到NDF-RT本体并能进一步进行药物分类的术语。②同时利用临床药物名和通用药物名来找到NDF-RT的多轴分类，以此避免单独使用通用药物名进行映射可能漏掉的分类。与现有的其他方法相比，此方法适用于更加复杂的情况。
（ii）利用自然语言处理工具与生物医学本体对AERS大规模数据进行规范化和信息聚合，使药物不良事件的大规模信号检测成为可能。在此基础上，实现了一种新的不良事件检测方法，通过将AERS数据与电子病历数据连接起来，获得不良事件的疾病发生率信息，实现大规模药物不良事件之间的比较研究。
At present, drug-related adverse events is becoming a serious public healthproblem. Though rigorous experimental studies are conducted before the drugs are putinto markets, not all potential adverse reactions can be found. After the "Thalidomide"tragedies after the1960s, many countries introduced pharmacovigilance system forthe monitoring of marketed drugs. United States drug and Food Administration (Foodand Drug Administration,FDA) drug adverse event reporting systems (Adverse EventReporting System,AERS) is mainly used to find rare serious adverse events that werenot identified in clinical trials due to low frequencies or new adverse drug events,namely, safety signals. If potential security problems of drugs are found in the AERS,FDA will conduct an epidemiological study to further evaluate the adverse event todetermine the causal relationship between the drug and adverse events. Based onsafety assessment of drug adverse events, FDA may take a series of regulations toimprove product safety for the protection of public health, such as updating druginstruction information, making restrictions on the use of drugs, giving newsecurity-related information to the public, or in a few cases, withdrawing the drugfrom the market.
Most research on ADE data mining focused on small-scale data, avoiding theresearch with large-scale data, unable to deeply mine ADE in terms of the mechanismof action, pharmocodynamics and physiological effect of drugs. However, suchlarge-scale and deep-level data mining are of great significance for revealing differentfeatures of ADE among different drug categories, the etiology of ADE, as well asgenetic aspects, and is the important direction for monitoring ADE and clinical drugsafety studies. Lacking knowledge integration of drug-related adverse eventsresources greatly limits the above mentioned studies.
The knowledge integration of drug adverse events is not only the real demand forefficiently using a large amount of medical information resources, but also the keyissues which should be seriously studied and solved for promoting the efficiency ofdata mining of drug adverse events. In recent years though the development of drug ontology provided realization chance of information resource integration, but stillfailed to achieve the ideal solution for the knowledge integration and deepaggregation of data in drug adverse events due to the complex of drug ontologies, thelack of normalization of data and the technical difficulties of ontology mapping.Therefore, data mining of adverse drug events fail to expand to the utilization andanalysis of large scale data.
Domain ontologies can provide knowledge for decision making and reasoningsupport, promoting large-scale drug safety signal detection and deep mining of ADE.This study used biomedical ontologies to integrate AERS-related informationresources, realizing knowledge integration, information aggregation, andinteroperability with other medical data resources, enriching resources for ADEmining and promoting drug safety signal detection.The main contents of this study include:
（1）Proposing a theoretical model for mapping between multiple domainontologies
Realization of ontology mapping as well as drug classification and aggregationwill not only provide preconditions for drug-related knowledge decision-making andreasoning support, but also be the important foundation for building the knowledgebase in the field, bearing important significance the deep mining in terms ofmechanisms of action, pharmacokinetic and physiological effect of drugs. Due to thecomplexity of the domain ontology itself and between heterogeneous domainontologies, mapping methods for domain ontologies become one of difficulties inontology mapping. This study proposed a theoretical model for mapping betweenmultiple domain ontologies and aggregation. Guided by this model, a mappingexample between RxNorm and NDF-RT (The National Drug File-ReferenceTerminology) was conducted with a new approach for mapping, and classification andaggregation for drug information in RxNorm based on the classification mechanismprovided by NDF-RT was realized.
Research results show that the model is not only feasible, but also with practicalvalue in terms of fully reusing multiple ontologies; the theoretical models will alsofurther deepen knowledge organization method of information resources at thesemantic level, and promote the construction of digital resource systems. Theinadequacies of the model include that the empirical use of theoretical models is based on existing ontologies, deficiencies in concepts and classifications may influence theresults of the classification and aggregation from ontology mapping.In addition, othercharacteristics of domain ontology may also be the factors for improving knowledgeorganization methods, hence, the future research should conduct more comprehensiveresearch on domain ontologies, extract more effective common features and promotethe perfection of the model.
（2）Evaluation of RxNorm for Covering Drug Names inAERS
The investigation of AERS drug names covered by RxNorm is the first step tofully explore the way that RxNorm exerts effect in AERS data mining, and a crucialstep.
Using the AERS “DRUG” data from the first quarter of2004through the end of2010, we calculated the coverage of AERS unique drug names and all drugoccurrences by RxNorm and UMLS with data mining techniques. Results showed thecoverage of AERS unique drug names by RxNorm and UMLS is respectively13,565(4.8%) and21,272(7.5%). Then we manually analyzed200AERS drug namesuncovered by RxNorm with frequency of more than1000and388samples withfrequency of less than1000to investigat the reasons of non-coverage and proposedsome ways for enhancing RxNorm. Although different sources including health careprofessionals and consumers as mentioned above contribute to the collection of AERSand their drug name entries may vary greatly even including typos, high-frequencydrug frequencies can still reflect clinical usage habit in specific domain. This studyprovides the foundation for improving RxNorm, also for choosing the naturallanguage processing tool MedEx (based on Rxnorm).（3）BuildingAERS-DM
On the basis of the AERS drug name normalization investigation, the druginformation in the AERS is normalized to RxNorm, a standard terminology source formedication, using a natural language processing (NLP) medication extraction tool,MedEx. Drug class information is then obtained from the National DrugFile-Reference Terminology (NDF-RT) using a greedy algorithm, with the theoreticalmodel for mapping between multiple domain ontologies and aggregation. Adverseevents are aggregated through mapping with the Preferred Term (PT) and SystemOrgan Class (SOC) codes of MedDRA. Finally our study yields an aggregated knowledge-enhanced AERS data mining set (AERS-DM). Case studies wereperformed to demonstrate the usefulness of our approaches.
We have built an open-source Drug-ADE knowledge resource that is normalizedand aggregated using standard biomedical ontologies. The data resource could providemore perspectives to mine the AERS for ADE detection and be used by the datamining research community. Two tables are formed: one stores the normalizedDrug-ADE information and the other stores the aggregated information of Drug-ADE.The data in the two tables can be connected through the RxNorm codes. In total, theAERS-DM contains37,029,228Drug-ADE records. Seventy-one percent(10,221/14,490) of normalized drug concepts in the AERS were classified to9classesin NDF-RT. The number of unique pairs is4,639,613between RxNorm concepts andMedDRA PT codes and205,725between RxNorm concepts and SOC codes afterADE aggregation.（4）Empirical Study on Data Mining inAERS-DM
AERS-DM is a normalized and aggregated data set for data mining in AERS,with the advantage of normalization and aggregation data for drug and ADE classes,which all come from the knowledge structure asserted in biomedical ontologies.Traditionally ADE detection studies with the AERS were carried out for only a smallnumber of drugs, and few studies were focused on large-scale mining[1]. In this studywe demonstrated the semantic mining potential in AERS-DM by using theinformation on popular cancer drug ingredients to conduct systematic analysis of drugclusters in terms of mechanism of action, physiologic effect, treatment intention andADEs, as well as ADE differences in terms of age and sex.
Traditional ADE detection rely on the use of disproportionality measuresattempting to quantify the degree of “unexpectedness” of a drug-ADE association,and trying to overcome the disadvantage of lacking incidence information of ADEs inspontaneous reports including AERS. In this study we demonstrated a novel ADEdetection method where the incidence information of ADEs could be obtained throughconnecting AERS data with EHRs, realizing the comparative research on ADE oflarge-scale drugs. As an advanced version of AERS, AERS-DM may serve as anintriguingly substantial resource for data mining as shown in this study. Innovations in this study include:
(1) Theoretical innovation
The theoretical model for mapping between multiple domain ontologies wasproposed. Currently each domain ontology has distinct features due to limitations inontology development. For example, some domain ontology provides classificationand aggregation information, and some with no such information, is complementaryto others in the coverage and contents. The theoretical model for mapping betweenmultiple domain ontologies proposed in this study fully uses different features ofdomain ontologies to realize the classification and aggregation function throughontology mapping, thus saving ontology development cost and realizing ontologyreuse.
(2) Method innovation
(i) Based on the theoretical model for mapping between multiple domainontologies, a systematic algorithm was developed realizing classification andaggregation of drugs with NDF-RT for RxNorm codes normalized in AERS. Themethod innovation was shown in two aspects:①The rich semantic connectionswithin RxNorm were fully utilized to infer the related concepts that can be used fordrug classification in NDF-RT.②both clinical drug names and generic drug nameswere used to find multi-axis classifications in NDF-RT, thus avoiding to missclassifications by using generic drug names only. Compared with other existingmethods, this method is suitable for more complex situations.
(ii) Natural language processing methods and biomedical ontology were used forlarge-scale data normalization and information aggregation in AERS, making massivesignal detection of adverse drug events possible. Based on that, a novel ADE detectionmethod was proposed, where the incidence information of ADEs could be obtainedthrough connecting AERS data with EHRs, realizing the comparative research onADE of large-scale drugs.
当前，关于不良事件数据挖掘的多数研究都集中于利用小部分数据进行不良事件的数据挖掘，避免对大规模数据的利用和研究；对药物不良事件进行药物的作用机制、药代动力学及生理作用等方面的深度挖掘的研究，以及对某类药物的不良事件进行比较性数据挖掘的研究、进行药物的作用机制、药代动力学及生理作用等方面的深度挖掘的研究，以及对某类药物的不良事件进行比较性数据挖掘的研究、AERS与其他数据源的集成难以实现难以实现，而这类大规模、深层次的挖掘在揭示不同药物类别的不良事件特征、药物不良事件的原因以及基因相关性方面都具有重要意义，是药物不良事件监测乃至临床用药安全研究的重要方向。缺乏对药物不良事件相关数据资源的知识整合严重限制了上述研究。
药物不良事件信息资源的知识整合既是有效利用海量医疗信息资源的现实需求，也是提高药物不良事件数据挖掘效率所需认真研究并必须解决的关键问题。近年来药物领域本体的发展虽然为资源整合研究提供了实现契机，然而由于药物领域本体的复杂性、数据缺乏规范化以及领域本体映射的技术难题，药物不良事件领域数据在知识集成与深度聚合方面始终未能求得理想的解决方案，药物不良事件的数据挖掘也因此未能扩展到对大规模数据的利用和分析。
领域本体可以提供相关知识决策和推理支持，促进大规模药物安全信号的检测和药物不良事件的深度挖掘。本研究利用生物医学领域本体将AERS相关信息资源有机整合起来,实现知识集成、信息聚合、与其他医疗数据资源之间的互操作、丰富了药物不良事件数据挖掘的资源并促进对药物安全信号的检测。
本研究的主要内容包括：
（1）提出药物领域本体映射与聚合模型
实现本体映射以及对药物信息的分类与聚合将为药物相关知识决策和推理支持提供前提条件，同时也是构建领域知识库的重要基础，对于进一步针对药物的用机制、药代动力学及生理作用等方面的深度数据挖掘具有重要意义。由于领域本体自身结构的复杂性和领域本体之间的异构性，药物领域本体映射方法成为实现本体映射的难点之一。本研究提出药物领域本体映射与聚合模型模型，并以该模型为指导，对药物领域本体RxNorm与NDF-RT（美国国家药物文件—参考术语）进行映射实例研究，提出了RxNorm与NDF-RT两个领域本体之间映射及信息分类与聚合的一种新方法。研究结果证明该模型不仅具有可行性，也显示出其对多本体能够充分复用的实践价值；该模型也将在语义层面上进一步深化信息资源的知识组织方法，促进数字资源语义体系的构建。模型的不足之处在于，模型的使用是以现有本体为基础的，因此现有本体中的概念关系以及分类聚合信息的不足将将最终影响本体映射分类聚合的效果。另外，领域本体的其他特性也可能是改善知识组织方法的因素，因此，未来研究中应对领域本体进行更全面的调研，抽取有效的共有特征，促进模型的完善。
（2）基于RxNorm的AERS药名规范化初步研究
调查AERS药名被RxNorm的收录情况，是探索如何充分发挥RxNorm在AERS数据挖掘中作用的第一步，也是至关重要的一步。
本研究计算2004年到2010年AERS中全部药物名称与RxNorm精确匹配的比例，并与UMLS进行比较分析。结果显示了RxNorm和UMLS对AERS中唯一药物名称精确匹配的整体收录范围分别为13,565(4.8%)个与21,272(7.5%)个。2011AA版UMLS集成了160个源词汇表， UMLS对AERS的药名覆盖分别来自包括RxNorm在内的各种来源词表，其中RxNorm映射的数量排列第一。然后手工分析了频次大于1000的200个未被映射的高频AERS药物名称及分析388个随机选择的频次小于1000的低频药物名称，调查了某些药名未被映射的原因。尽管在AERS中，数据来源广泛且存在录入错误，但是高频词仍然能够显示出特定领域的词汇使用习惯。我们的研究将为RxNorm本体的完善提供依据。本章的研究也对下章研究中选择自然语言处理工具MedEx（以RxNorm为基础）提供了依据。
（3）构建数据挖掘知识整合库（AERS-DM）
在AERS药名规范化进行调查研究基础上，选择利用自然语言处理工具MedEx对AERS中药名进行规范化，并对其自然语言处理效果进行评价。在药物领域本体映射与聚合模型模型的基础上，使用贪婪算法将AERS中的药名聚合到RxNorm和NDF-RT中的药物分类信息。对于药物不良事件，通过映射方法将其映射到MedDRA中的PT和SOC代码进行聚合。最终建立开源的药物——不良事件数据挖掘知识整合库(AERS-DM)（网址：http://informatics.mayo.edu/adepedia/index.php/Download），最后通过实例研究，证实了AERS-DM数据集的挖掘效果。
AERS-DM中的信息集成了药物及不良事件知识库。AERS-DM具有规范化代码和聚合的功能，可以为AERS药物安全信号的挖掘以及相关数据挖掘领域提供更多的资源。该数据集包含两个表。一个表存储药物及不良事件的规范化信息，另一个表存储药物和不良事件的聚合信息。共有37,029,228对药物及不良事件记录。AERS中的药名被规范为14,490个RxNorm药名（由RxNorm代码表示），其中10,221个规范化的药名可以归到NDF-RT类别中，占71%。对于AERS-DM中的不良事件，共有14,740个MedDRA中的PT术语被聚合到MedDRA的SOC代码，占MedDRA中所有PT术语的76%。AERS-DM中，RxNorm代码表示的药名与MedDRA中的PT唯一对，即规范化后的药名与不良事件的唯一配对，共有4,639,613，将不良事件按组织器官聚合后，药物与不良事件组织器官的配对共205,725对。
（4）AERS-DM数据挖掘知识整合库的数据挖掘实证研究
AERS-DM是一个规范化和聚合的数据挖掘知识整合库，优势在于药物数据的规范化，以及药物数据和不良事件数据的分类聚合，这些分类聚合知识全部来自于生物医学本体中所含有的知识结构。传统的利用AERS进行的不良事件检测研究大多仅针对少量药物，进行大规模数据挖掘的研究数量较少。在本研究中，我们利用常用抗癌药物成分信息对药物作用机制、生理作用、治疗意向的药物聚类与药物不良事件的聚类，以及年龄与性别的药物不良事件差别进行了大规模的系统分析，进一步证实了AERS-DM的语义挖掘潜力。
传统的不良事件检测依赖比例失衡测度，主要是量化出药物-不良事件关联的“始料未及”的程度，并试图克服自发报告系统中不良事件缺乏疾病发生率背景信息的缺点。在此研究中我们提出了一种新的不良事件检测方法，在这种方法中，通过将AERS数据与电子病历数据连接起来，从而获得不良事件的疾病发生率信息，并实现大规模药物不良事件之间的比较研究。本研究证实了AERS-DM作为AERS的一个高级版本，是一个可用于数据挖掘的丰富资源。
本文的创新点包括：
（1）理论创新
提出药物领域本体映射与聚合模型。由于本体开发的局限性，当前领域本体各有特点，因此本研究提出的药物领域本体映射与聚合模型，充分利用不同本体的特点，通过本体映射，将某一本体的分类信息与其他本体的内容形成互补，实现某一领域多个本体的分类聚合功能，从而节约本体开发成本，实现本体充分复用。
（2）方法创新
（i）在药物领域本体映射与聚合模型的基础上，开发出一套系统的分类聚合算法，实现利用NDF-RT与RxNorm对AERS数据库中的药物进行分类聚合。方法创新体现在两方面：①利用RxNorm中的丰富关系来推理出可以映射到NDF-RT本体并能进一步进行药物分类的术语。②同时利用临床药物名和通用药物名来找到NDF-RT的多轴分类，以此避免单独使用通用药物名进行映射可能漏掉的分类。与现有的其他方法相比，此方法适用于更加复杂的情况。
（ii）利用自然语言处理工具与生物医学本体对AERS大规模数据进行规范化和信息聚合，使药物不良事件的大规模信号检测成为可能。在此基础上，实现了一种新的不良事件检测方法，通过将AERS数据与电子病历数据连接起来，获得不良事件的疾病发生率信息，实现大规模药物不良事件之间的比较研究。
At present, drug-related adverse events is becoming a serious public healthproblem. Though rigorous experimental studies are conducted before the drugs are putinto markets, not all potential adverse reactions can be found. After the "Thalidomide"tragedies after the1960s, many countries introduced pharmacovigilance system forthe monitoring of marketed drugs. United States drug and Food Administration (Foodand Drug Administration,FDA) drug adverse event reporting systems (Adverse EventReporting System,AERS) is mainly used to find rare serious adverse events that werenot identified in clinical trials due to low frequencies or new adverse drug events,namely, safety signals. If potential security problems of drugs are found in the AERS,FDA will conduct an epidemiological study to further evaluate the adverse event todetermine the causal relationship between the drug and adverse events. Based onsafety assessment of drug adverse events, FDA may take a series of regulations toimprove product safety for the protection of public health, such as updating druginstruction information, making restrictions on the use of drugs, giving newsecurity-related information to the public, or in a few cases, withdrawing the drugfrom the market.
Most research on ADE data mining focused on small-scale data, avoiding theresearch with large-scale data, unable to deeply mine ADE in terms of the mechanismof action, pharmocodynamics and physiological effect of drugs. However, suchlarge-scale and deep-level data mining are of great significance for revealing differentfeatures of ADE among different drug categories, the etiology of ADE, as well asgenetic aspects, and is the important direction for monitoring ADE and clinical drugsafety studies. Lacking knowledge integration of drug-related adverse eventsresources greatly limits the above mentioned studies.
The knowledge integration of drug adverse events is not only the real demand forefficiently using a large amount of medical information resources, but also the keyissues which should be seriously studied and solved for promoting the efficiency ofdata mining of drug adverse events. In recent years though the development of drug ontology provided realization chance of information resource integration, but stillfailed to achieve the ideal solution for the knowledge integration and deepaggregation of data in drug adverse events due to the complex of drug ontologies, thelack of normalization of data and the technical difficulties of ontology mapping.Therefore, data mining of adverse drug events fail to expand to the utilization andanalysis of large scale data.
Domain ontologies can provide knowledge for decision making and reasoningsupport, promoting large-scale drug safety signal detection and deep mining of ADE.This study used biomedical ontologies to integrate AERS-related informationresources, realizing knowledge integration, information aggregation, andinteroperability with other medical data resources, enriching resources for ADEmining and promoting drug safety signal detection.The main contents of this study include:
（1）Proposing a theoretical model for mapping between multiple domainontologies
Realization of ontology mapping as well as drug classification and aggregationwill not only provide preconditions for drug-related knowledge decision-making andreasoning support, but also be the important foundation for building the knowledgebase in the field, bearing important significance the deep mining in terms ofmechanisms of action, pharmacokinetic and physiological effect of drugs. Due to thecomplexity of the domain ontology itself and between heterogeneous domainontologies, mapping methods for domain ontologies become one of difficulties inontology mapping. This study proposed a theoretical model for mapping betweenmultiple domain ontologies and aggregation. Guided by this model, a mappingexample between RxNorm and NDF-RT (The National Drug File-ReferenceTerminology) was conducted with a new approach for mapping, and classification andaggregation for drug information in RxNorm based on the classification mechanismprovided by NDF-RT was realized.
Research results show that the model is not only feasible, but also with practicalvalue in terms of fully reusing multiple ontologies; the theoretical models will alsofurther deepen knowledge organization method of information resources at thesemantic level, and promote the construction of digital resource systems. Theinadequacies of the model include that the empirical use of theoretical models is based on existing ontologies, deficiencies in concepts and classifications may influence theresults of the classification and aggregation from ontology mapping.In addition, othercharacteristics of domain ontology may also be the factors for improving knowledgeorganization methods, hence, the future research should conduct more comprehensiveresearch on domain ontologies, extract more effective common features and promotethe perfection of the model.
（2）Evaluation of RxNorm for Covering Drug Names inAERS
The investigation of AERS drug names covered by RxNorm is the first step tofully explore the way that RxNorm exerts effect in AERS data mining, and a crucialstep.
Using the AERS “DRUG” data from the first quarter of2004through the end of2010, we calculated the coverage of AERS unique drug names and all drugoccurrences by RxNorm and UMLS with data mining techniques. Results showed thecoverage of AERS unique drug names by RxNorm and UMLS is respectively13,565(4.8%) and21,272(7.5%). Then we manually analyzed200AERS drug namesuncovered by RxNorm with frequency of more than1000and388samples withfrequency of less than1000to investigat the reasons of non-coverage and proposedsome ways for enhancing RxNorm. Although different sources including health careprofessionals and consumers as mentioned above contribute to the collection of AERSand their drug name entries may vary greatly even including typos, high-frequencydrug frequencies can still reflect clinical usage habit in specific domain. This studyprovides the foundation for improving RxNorm, also for choosing the naturallanguage processing tool MedEx (based on Rxnorm).（3）BuildingAERS-DM
On the basis of the AERS drug name normalization investigation, the druginformation in the AERS is normalized to RxNorm, a standard terminology source formedication, using a natural language processing (NLP) medication extraction tool,MedEx. Drug class information is then obtained from the National DrugFile-Reference Terminology (NDF-RT) using a greedy algorithm, with the theoreticalmodel for mapping between multiple domain ontologies and aggregation. Adverseevents are aggregated through mapping with the Preferred Term (PT) and SystemOrgan Class (SOC) codes of MedDRA. Finally our study yields an aggregated knowledge-enhanced AERS data mining set (AERS-DM). Case studies wereperformed to demonstrate the usefulness of our approaches.
We have built an open-source Drug-ADE knowledge resource that is normalizedand aggregated using standard biomedical ontologies. The data resource could providemore perspectives to mine the AERS for ADE detection and be used by the datamining research community. Two tables are formed: one stores the normalizedDrug-ADE information and the other stores the aggregated information of Drug-ADE.The data in the two tables can be connected through the RxNorm codes. In total, theAERS-DM contains37,029,228Drug-ADE records. Seventy-one percent(10,221/14,490) of normalized drug concepts in the AERS were classified to9classesin NDF-RT. The number of unique pairs is4,639,613between RxNorm concepts andMedDRA PT codes and205,725between RxNorm concepts and SOC codes afterADE aggregation.（4）Empirical Study on Data Mining inAERS-DM
AERS-DM is a normalized and aggregated data set for data mining in AERS,with the advantage of normalization and aggregation data for drug and ADE classes,which all come from the knowledge structure asserted in biomedical ontologies.Traditionally ADE detection studies with the AERS were carried out for only a smallnumber of drugs, and few studies were focused on large-scale mining[1]. In this studywe demonstrated the semantic mining potential in AERS-DM by using theinformation on popular cancer drug ingredients to conduct systematic analysis of drugclusters in terms of mechanism of action, physiologic effect, treatment intention andADEs, as well as ADE differences in terms of age and sex.
Traditional ADE detection rely on the use of disproportionality measuresattempting to quantify the degree of “unexpectedness” of a drug-ADE association,and trying to overcome the disadvantage of lacking incidence information of ADEs inspontaneous reports including AERS. In this study we demonstrated a novel ADEdetection method where the incidence information of ADEs could be obtained throughconnecting AERS data with EHRs, realizing the comparative research on ADE oflarge-scale drugs. As an advanced version of AERS, AERS-DM may serve as anintriguingly substantial resource for data mining as shown in this study. Innovations in this study include:
(1) Theoretical innovation
The theoretical model for mapping between multiple domain ontologies wasproposed. Currently each domain ontology has distinct features due to limitations inontology development. For example, some domain ontology provides classificationand aggregation information, and some with no such information, is complementaryto others in the coverage and contents. The theoretical model for mapping betweenmultiple domain ontologies proposed in this study fully uses different features ofdomain ontologies to realize the classification and aggregation function throughontology mapping, thus saving ontology development cost and realizing ontologyreuse.
(2) Method innovation
(i) Based on the theoretical model for mapping between multiple domainontologies, a systematic algorithm was developed realizing classification andaggregation of drugs with NDF-RT for RxNorm codes normalized in AERS. Themethod innovation was shown in two aspects:①The rich semantic connectionswithin RxNorm were fully utilized to infer the related concepts that can be used fordrug classification in NDF-RT.②both clinical drug names and generic drug nameswere used to find multi-axis classifications in NDF-RT, thus avoiding to missclassifications by using generic drug names only. Compared with other existingmethods, this method is suitable for more complex situations.
(ii) Natural language processing methods and biomedical ontology were used forlarge-scale data normalization and information aggregation in AERS, making massivesignal detection of adverse drug events possible. Based on that, a novel ADE detectionmethod was proposed, where the incidence information of ADEs could be obtainedthrough connecting AERS data with EHRs, realizing the comparative research onADE of large-scale drugs.
引文
[1] Almenoff JS PE, Gibbs TG, DuMouchel W, Evans SJ, Yuen N.: Novel statisticaltools for monitoring the safety of marketed drugs. Clin Pharmacol Ther2007,82(2):157-166.
[2]孔雅慧:全球药物不良反应监测系统概述.药物流行病学杂志2011,20(7):364-368.
[3]叶小飞:基于自发呈报系统与循证医学的药品不良反应信号挖掘.第二军医大学;2011.
[4] Handler SM, Altman RL, Perera S, Hanlon JT, Studenski SA, Bost JE, Saul MI,Fridsma DB: A systematic review of the performance characteristics of clinicalevent monitor signals used to detect adverse drug events in the hospital setting.Journal of the American Medical Informatics Association2007,14(4):451-458.
[5] FDA Adverse Event Reporting System (FAERS)(formerly AERS)
[http://www.fda.gov/cder/aers/default.htm]
[6] Hochberg AM, Reisinger SJ, Pearson P, O’Hara DJ, Hall K: Using data miningto predict safety actions from FDA adverse event reporting system data. DrugInformation Journal2007,41(5):633-642.
[7] Rawlins M: Spontaneous reporting of adverse drug reactions. I: the data. Britishjournal of clinical pharmacology1988,26(1):1.
[8] Szarfman A, Machado SG, ONeill RT: Use of Screening Algorithms andComputer Systems to Efficiently Signal Higher-Than-Expected Combinationsof Drugs and Events in the US FDAs Spontaneous Reports Database. DrugSafety2002,25(6):381-392.
[9] Lindquist M, Stahl M, Bate A, Edwards IR, Meyboom RHB: A retrospectiveevaluation of a data mining approach to aid finding new adverse drug reactionsignals in the WHO international database. Drug Safety2000,23(6):533-542.
[10] Almenoff J, Tonning JM, Gould AL, Szarfman A, Hauben M, Ouellet-HellstromR, Ball R, Hornbuckle K, Walsh L, Yee C: Perspectives on the use of datamining in pharmacovigilance. Drug Safety2005,28(11):981-1007.
[11] Finney D: Systemic signalling of adverse reactions to drugs. Methods ofinformation in medicine1974,13(1):1.
[12] Bate A, Evans S: Quantitative signal detection using spontaneous ADRreporting. Pharmacoepidemiology and drug safety2009,18(6):427-436.
[13]尚鹏辉，詹思延:数据挖掘在药物不良事件信号检出和分析中的应用.中国药品应用与监测2009,6(3):187-190.
[14] Szarfman A, Tonning JM, Levine JG, Doraiswamy PM: Atypical antipsychoticsand pituitary tumors: a pharmacovigilance study. Pharmacotherapy: The Journalof Human Pharmacology and Drug Therapy2006,26(6):748-758.
[15] Van Puijenbroek EP, Egberts ACG, Meyboom RHB, Leufkens HGM: Signallingpossible drug–drug interactions in a spontaneous reporting system: delay ofwithdrawal bleeding during concomitant use of oral contraceptives anditraconazole. British journal of clinical pharmacology1999,47(6):689-693.
[16] Thakrar BT, Grundschober SB, Doessegger L: Detecting signals of drug–druginteractions in a spontaneous reports database. British journal of clinicalpharmacology2007,64(4):489-495.
[17] Harpaz R, Chase H, Friedman C: Mining multi-item drug adverse effectassociations in spontaneous reporting systems. BMC bioinformatics2010,11(Suppl9):S7.
[18] Tatonetti NP, Fernald GH, Altman RB: A novel signal detection algorithm foridentifying hidden drug-drug interactions in adverse event reports. Journal ofthe American Medical Informatics Association2012,19(1):79-85.
[19] Poluzzi E, Raschi E, Motola D, Moretti U, De Ponti F: Antimicrobials and therisk of torsades de pointes: the contribution from data mining of the US FDAAdverse Event Reporting System. Drug Safety2010,33(4):303-314.
[20] Ali AK: Pharmacovigilance analysis of adverse event reports for aliskirenhemifumarate, a first-in-class direct renin inhibitor. Therapeutics and ClinicalRisk Management2011,7:337.
[21] Chen Y, Guo JJ, Healy DP, Lin X, Patel NC: Risk of hepatotoxicity associatedwith the use of telithromycin: a signal detection using data mining algorithms.The Annals of pharmacotherapy2008,42(12):1791-1796.
[22] Kadoyama K, Kuwahara A, Yamamori M: Hypersensitivity reactions toanticancer agents: Data mining of the public version of the FDA adverse eventreporting system, AERS. Journal of experimental&clinical cancer research: CR2011,30(1):93.
[23] Sakaeda T, Kadoyama K, Okuno Y: Adverse event profiles of platinum agents:Data mining of the public version of the FDA adverse event reporting system,AERS, and reproducibility of clinical observations. International journal ofmedical sciences2011,8(6):487-491.
[24] Sakaeda T, Kadoyama K, Yabuuchi H, Niijima S, Seki K, Shiraishi Y, Okuno Y:Platinum agent-induced hypersensitivity reactions: Data mining of the publicversion of the FDA adverse event reporting system, AERS. International journalof medical sciences2011,8(4):332.
[25] Davis S, King B, Raine JM: Spontaneous reporting–UK. Pharmacovigilance2002:199-215.
[26] Wysowski DK, Swartz L: Adverse drug event surveillance and drugwithdrawals in the United States,1969-2002: the importance of reportingsuspected reactions. Archives of internal medicine2005,165(12):1363.
[27] Moore TJ, Psaty BM, Furberg CD: Time to act on drug safety. JAMA: thejournal of the American Medical Association1998,279(19):1571-1573.
[28] Cruzan S: FDA Talk Paper: FDA proposes to Withdraw Seldane Approval.1997.
[29]钱轶峰，罗宝章，叶小飞等:检测联合用药不良事件信号的数据挖掘方法.中国卫生统计2010,27(1):31-34.
[30] Lee F, Harris W, Cheng H, Shenoi J, Zhao S, Wang J, Champion T, Izard J, GoreJL, Porter M: Pathologic Response Rates of Gemcitabine/Cisplatin VS.Methotrexate/Vinblastine/Adriamycin/Cisplatin Neoadjuvant Chemotherapy forMuscle Invasive Urothelial Bladder Cancer.
[31] Ritchie MD, Denny JC, Crawford DC, Ramirez AH, Weiner JB, Pulley JM,Basford MA, Brown-Gentry K, Balser JR, Masys DR: Robust replication ofgenotype-phenotype associations across multiple diseases in an electronicmedical record. The American Journal of Human Genetics2010,86(4):560-572.
[32] Tatonetti N, Denny J, Murphy S, Fernald G, Krishnan G, Castro V, Yue P, Tsau P,Kohane I, Roden D: Detecting drug interactions from adverse-event reports:interaction between paroxetine and pravastatin increases blood glucose levels.Clinical Pharmacology&Therapeutics2011,90(1):133-142.
[33] Guideline ICHHT: Clinical safety data management: definitions and standardsfor expedited reporting.1994.
[34] Kohn LT, Corrigan J, Donaldson MS: To err is human: building a safer healthsystem. Volume6: Natl Academy Pr;2000.
[35][http://www.sda.gov.cn/WS01/CL0747/63731.html]
[36] What Is a Medication Error? Rockville M: National Coordinating Council forMedication Error Reporting and Prevention.1998.
[37] Nebeker JR, Barach P, Samore MH: Clarifying adverse drug events: a clinician'sguide to terminology, documentation, and reporting. Annals of internal medicine2004,140(10):795-801.
[38] VI CWG: Management of Safety Information from Clinical Trials. Geneva:WHO.
[39] Zhou L, Plasek JM, Mahoney LM, Chang F, DiMaggio D, Rocha RA: MappingPartners Master Drug Dictionary to RxNorm Using an NLP-based Approach.Journal of Biomedical Informatics2011, DOI: http://www.sciencedirect.com/science/article/pii/S1532046411001948.
[40] UMLS许可协议[http://www.nlm.nih.gov/research/umls/license.html]
[41] Peters L, Bodenreider O: Using the RxNorm web services API for qualityassurance purposes. In:2008. American Medical Informatics Association;2008:591-595.
[42] National Drug File–Reference Terminology (NDF-RT) Documentation
[http://evs.nci.nih.gov/ftp1/NDF-RT]
[43] Bodenreider O: Using SNOMED CT in combination with MedDRA forreporting signal detection and adverse drug reactions reporting. In:2009.American Medical Informatics Association;2009:45.
[44] Estoup JB: Gammes Stenographiques.4th edition edition. Paris;1916.
[45] Manning CD,曼宁, Schutze H,许策,春法,苑:统计自然语言处理基础:电子工业出版社;2005.
[46]王伟:信息计量学及其医学应用.第一版edition:人民卫生出版社;2009.
[47] Kumar SK, Harding J: Ontology mapping using description logic and bridgingaxioms. Computers in Industry2012.
[48] Wang R, Wang L, Liu L, Chen G, Wang Q: Combination of the ImprovedMethod for Ontology Mapping. Physics Procedia2012,25:2167-2172.
[49] Hartung M, Gro A, Rahm E: COnto–Diff: generation of complex evolutionmappings for life science ontologies. Journal of Biomedical Informatics2013,46(1):15-32.
[50] Bock J, Hettenhausen J: Discrete particle swarm optimisation for ontologyalignment. Information Sciences2012,192:152-173.
[51]米杨,曹锦丹:基于PROMPT的本体映射实例分析.情报学报2010(006):987-991.
[52]吕刚,郑诚,胡春玲:基于概念分类的多本体映射方法研究倡.计算机应用研究2011,28(9).
[53]王效岳,胡泽文,白如江: WordNet与SUMO本体之间的映射机制研究.现代图书情报技术2011,1:005.
[54] Palchuk MB, Klumpenaar M, Jatkar T, Zottola RJ, Adams WG, Abend AH:Enabling hierarchical view of RxNorm with NDF-RT drug classes. In:2010.American Medical Informatics Association;2010:577-581.
[55] Pathak J, Murphy SP, Willaert BN, Kremers HM, Yawn BP, Rocca WA, ChuteCG: Using RxNorm and NDF-RT to Classify Medication Data Extracted fromElectronic Health Records: Experiences from the Rochester EpidemiologyProject. In:2011. American Medical Informatics Association;2011:1089-1098.
[56] Pathak J, Chute CG: Analyzing categorical information in two publiclyavailable drug terminologies: RxNorm and NDF-RT. Journal of the AmericanMedical Informatics Association2010,17(4):432-439.
[57] Warnekar PP, Bouhaddou O, Parrish F, Do N, Kilbourne J, Brown SH, LincolnMJ: Use of RxNorm to exchange codified drug allergy information betweenDepartment of Veterans Affairs (VA) and Department of Defense (DoD). In:2007. American Medical Informatics Association;2007:781.
[58] Peters L, Kapusnik-Uner JE, Bodenreider O: Methods for managing variation inclinical drug names. In:2010. American Medical Informatics Association;2010:637.
[59] Li Y, Salmasian H, Harpaz R, Chase H, Friedman C: Determining the Reasonsfor Medication Prescriptions in the EHR using Knowledge and NaturalLanguage Processing. In:2011. American Medical Informatics Association;2011:768-776.
[60] U.S. Food and Drug Administration Drugs@FDA [http://www.a ccessdata.fda.gov/scripts/cder/drugsatfda/]
[61] Drugs.com [http://www.drugs.com/]
[62] Francis WN, Kucera H: Computational analysis of present-day AmericanEnglish. Providence, RI: Brown University Press;1967.
[63] Wang X, Hripcsak G, Markatou M, Friedman C: Active computerizedpharmacovigilance using natural language processing, statistics, and electronichealth records: a feasibility study. Journal of the American Medical InformaticsAssociation2009,16(3):328-337.
[64] Harpaz R, Haerian K, Chase HS, Friedman C: Statistical Mining of PotentialDrug Interaction Adverse Effects in FDA’s Spontaneous Reporting System. In:2010. American Medical Informatics Association;2010:281.
[65] Kennedy DL, Goldman SA, Lillie RB: Spontaneous reporting in the UnitedStates. Pharmacoepidemiology2000:149-174.
[66] Harpaz R, Perez H, Chase H, Rabadan R, Hripcsak G, Friedman C: Biclusteringof adverse drug events in the FDA's spontaneous reporting system. ClinicalPharmacology&Therapeutics2010,89(2):243-250.
[67] Moore TJ, Furberg CD, Glenmullen J, Maltsberger JT, Singh S: Suicidalbehavior and depression in smoking cessation treatments. PloS one2011,6(11):e27016.
[68] Vilar S, Harpaz R, Chase HS, Costanzi S, Rabadan R, Friedman C: Facilitatingadverse drug event detection in pharmacovigilance databases using molecularstructure similarity: application to rhabdomyolysis. Journal of the AmericanMedical Informatics Association2011,18(Suppl1):i73-i80.
[69] Liu Y, LePendu P, Iyer S, Shah NH: Using Temporal Patterns in MedicalRecords to Discern Adverse Drug Events from Indications. AMIA Proceedings2012.
[70] Xu H, Stenner SP, Doan S, Johnson KB, Waitman LR, Denny JC: MedEx: amedication information extraction system for clinical narratives. Journal of theAmerican Medical Informatics Association2010,17(1):19-24.
[71] Friedman C, Shagina L, Lussier Y, Hripcsak G: Automated encoding of clinicaldocuments based on natural language processing. Journal of the AmericanMedical Informatics Association2004,11(5):392-402.
[72] Moore TJ, Glenmullen J, Furberg CD: Prescription drugs associated withreports of violence towards others. PloS one2010,5(12):e15337.
[73] U.S. Food and Drug Administration Drugs@FDA [http://www.accessdata.fda.gov/scripts/cder/drugsatfda/]
[74] Lu Y, Cederbaum AI: Cisplatin-induced hepatotoxicity is enhanced by elevatedexpression of cytochrome P4502E1. Toxicological Sciences2006,89(2):515-523.
[75] Friedman C: Towards a comprehensive medical language processing system:methods and issues. In Proceedings of the AMIA annual fall symposium:1997.American Medical Informatics Association;1997:595-599.
[76] Shah NH, Bhatia N, Jonquet C, Rubin D, Chiang AP, Musen MA: Comparisonof concept recognizers for building the Open Biomedical Annotator. BMCbioinformatics2009,10(Suppl9):S14.
[77] Savova GK, Masanz JJ, Ogren PV, Zheng J, Sohn S, Kipper-Schuler KC, ChuteCG: Mayo clinical Text Analysis and Knowledge Extraction System (cTAKES):architecture, component evaluation and applications. Journal of the AmericanMedical Informatics Association2010,17(5):507-513.
[78] LePendu P, Iyer SV, Fairon C, Shah NH: Annotation analysis for testing drugsafety signals using unstructured clinical notes. Journal of BiomedicalSemantics2012,3(Suppl1):S5.
[79] Patrick J, Li M: High accuracy information extraction of medicationinformation from clinical notes:2009i2b2medication extraction challenge.Journal of the American Medical Informatics Association2010,17(5):524-527.
[80] Rosenbloom ST, Awad J, Speroff T, Elkin PL, Rothman R, Spickard III A,Peterson J, Bauer BA, Wahner-Roedler DL, Lee M: Adequacy of representationof the National Drug File Reference Terminology Physiologic Effects referencehierarchy for commonly prescribed medications. In AMIA Annual SymposiumProceedings:2003. American Medical Informatics Association;2003:569-578.
[81] Zhu Q, Jiang G, Chute CG: Profiling structured product labeling with NDF-RTand RxNorm. J Biomedical Semantics2012,3:16.
[82] MedDRA [http://en.wikipedia.org/wiki/MedDRA]
[83] Nadkarni PM, Darer JD: Determining correspondences between high-frequencyMedDRA concepts and SNOMED: a case study. BMC Medical Informatics andDecision Making2010,10(1):66.
[84] Mougin F, Dupuch M, Grabar N: Improving the mapping between MedDRAand SNOMED CT. Artificial Intelligence in Medicine2011:220-224.
[85] Semantic MEDLINE Prototype [http://skr3.nlm.nih.gov/SemMedDemo/index.jsp]
[86] Kilicoglu H, Fiszman M, Rodriguez A, Shin D, Ripple A, Rindflesch TC:Semantic MEDLINE: a web application for managing the results of PubMedSearches. In Proceedings of the third international symposium for semanticmining in biomedicine:2008;2008:69-76.
[87] Yancik R: Cancer burden in the aged. Cancer1997,80(7):1273-1283.
[88] Wang J, Huang Y: Pharmacogenomics of sex difference in chemotherapeutictoxicity. Current drug discovery technologies2007,4(1):59-68.
[89] Fattinger K, Roos M, Vergères P, Holenstein C, Kind B, Masche U, Stocker DN,Braunschweig S, Kullak‐Ublick GA, Galeazzi RL: Epidemiology of drugexposure and adverse drug reactions in two Swiss departments of internalmedicine. British journal of clinical pharmacology2000,49(2):158-167.
[90] Raman A, Lombardo J, Chandrasekhar R, Fakih M: Bevacizumab relatedadverse events among various age groups of elderly patients with advancedcolorectal cancer. J Clin Oncol2007,25(18S):14546.
[91] Sloan JA, Goldberg RM, Sargent DJ, Vargas-Chanes D, Nair S, Cha SS,Novotny PJ, Poon MA, O’Connell MJ, Loprinzi CL: Women experience greatertoxicity with fluorouracil-based chemotherapy for colorectal cancer. Journal ofclinical oncology2002,20(6):1491-1498.
[92] Singh S, Parulekar W, Murray N, Feld R, Evans WK, Tu D, Shepherd FA:Influence of sex on toxicity and treatment outcome in small-cell lung cancer.Journal of clinical oncology2005,23(4):850-856.
[93] MediLexicon's FDA Approved Drugs Listings [http://www.medilexicon.com/drugs-list/cancer.php]
[94] CIMMiner [http://discover.nci.nih.gov/cimminer.]
[95] Yoshida S: The results of gefitinib prospective investigation. Med Drug J2005,41:772-789.
[96] Murashige N, Tanimoto T, Oshima Y, Lim KH, Chang YH, Inoue A, KobayashiK, Nukiwa T: Interstitial lung disease and gefitinib. New England Journal ofMedicine2010,363(16):1578-1580.
[97] Aapro M, Macciocchi A, Gridelli C: Palonosetron improves prevention ofchemotherapy-induced nausea and vomiting in elderly patients. The journal ofsupportive oncology2005,3(5):369.
[98] Harpaz R, Haerian K, Chase HS, Friedman C: Statistical Mining of PotentialDrug Interaction Adverse Effects in FDA’s Spontaneous Reporting System. In:2010. American Medical Informatics Association:281.
[99] medilexicon [www.medilexicon.com]
[2]孔雅慧:全球药物不良反应监测系统概述.药物流行病学杂志2011,20(7):364-368.
[3]叶小飞:基于自发呈报系统与循证医学的药品不良反应信号挖掘.第二军医大学;2011.
[4] Handler SM, Altman RL, Perera S, Hanlon JT, Studenski SA, Bost JE, Saul MI,Fridsma DB: A systematic review of the performance characteristics of clinicalevent monitor signals used to detect adverse drug events in the hospital setting.Journal of the American Medical Informatics Association2007,14(4):451-458.
[5] FDA Adverse Event Reporting System (FAERS)(formerly AERS)
[http://www.fda.gov/cder/aers/default.htm]
[6] Hochberg AM, Reisinger SJ, Pearson P, O’Hara DJ, Hall K: Using data miningto predict safety actions from FDA adverse event reporting system data. DrugInformation Journal2007,41(5):633-642.
[7] Rawlins M: Spontaneous reporting of adverse drug reactions. I: the data. Britishjournal of clinical pharmacology1988,26(1):1.
[8] Szarfman A, Machado SG, ONeill RT: Use of Screening Algorithms andComputer Systems to Efficiently Signal Higher-Than-Expected Combinationsof Drugs and Events in the US FDAs Spontaneous Reports Database. DrugSafety2002,25(6):381-392.
[9] Lindquist M, Stahl M, Bate A, Edwards IR, Meyboom RHB: A retrospectiveevaluation of a data mining approach to aid finding new adverse drug reactionsignals in the WHO international database. Drug Safety2000,23(6):533-542.
[10] Almenoff J, Tonning JM, Gould AL, Szarfman A, Hauben M, Ouellet-HellstromR, Ball R, Hornbuckle K, Walsh L, Yee C: Perspectives on the use of datamining in pharmacovigilance. Drug Safety2005,28(11):981-1007.
[11] Finney D: Systemic signalling of adverse reactions to drugs. Methods ofinformation in medicine1974,13(1):1.
[12] Bate A, Evans S: Quantitative signal detection using spontaneous ADRreporting. Pharmacoepidemiology and drug safety2009,18(6):427-436.
[13]尚鹏辉，詹思延:数据挖掘在药物不良事件信号检出和分析中的应用.中国药品应用与监测2009,6(3):187-190.
[14] Szarfman A, Tonning JM, Levine JG, Doraiswamy PM: Atypical antipsychoticsand pituitary tumors: a pharmacovigilance study. Pharmacotherapy: The Journalof Human Pharmacology and Drug Therapy2006,26(6):748-758.
[15] Van Puijenbroek EP, Egberts ACG, Meyboom RHB, Leufkens HGM: Signallingpossible drug–drug interactions in a spontaneous reporting system: delay ofwithdrawal bleeding during concomitant use of oral contraceptives anditraconazole. British journal of clinical pharmacology1999,47(6):689-693.
[16] Thakrar BT, Grundschober SB, Doessegger L: Detecting signals of drug–druginteractions in a spontaneous reports database. British journal of clinicalpharmacology2007,64(4):489-495.
[17] Harpaz R, Chase H, Friedman C: Mining multi-item drug adverse effectassociations in spontaneous reporting systems. BMC bioinformatics2010,11(Suppl9):S7.
[18] Tatonetti NP, Fernald GH, Altman RB: A novel signal detection algorithm foridentifying hidden drug-drug interactions in adverse event reports. Journal ofthe American Medical Informatics Association2012,19(1):79-85.
[19] Poluzzi E, Raschi E, Motola D, Moretti U, De Ponti F: Antimicrobials and therisk of torsades de pointes: the contribution from data mining of the US FDAAdverse Event Reporting System. Drug Safety2010,33(4):303-314.
[20] Ali AK: Pharmacovigilance analysis of adverse event reports for aliskirenhemifumarate, a first-in-class direct renin inhibitor. Therapeutics and ClinicalRisk Management2011,7:337.
[21] Chen Y, Guo JJ, Healy DP, Lin X, Patel NC: Risk of hepatotoxicity associatedwith the use of telithromycin: a signal detection using data mining algorithms.The Annals of pharmacotherapy2008,42(12):1791-1796.
[22] Kadoyama K, Kuwahara A, Yamamori M: Hypersensitivity reactions toanticancer agents: Data mining of the public version of the FDA adverse eventreporting system, AERS. Journal of experimental&clinical cancer research: CR2011,30(1):93.
[23] Sakaeda T, Kadoyama K, Okuno Y: Adverse event profiles of platinum agents:Data mining of the public version of the FDA adverse event reporting system,AERS, and reproducibility of clinical observations. International journal ofmedical sciences2011,8(6):487-491.
[24] Sakaeda T, Kadoyama K, Yabuuchi H, Niijima S, Seki K, Shiraishi Y, Okuno Y:Platinum agent-induced hypersensitivity reactions: Data mining of the publicversion of the FDA adverse event reporting system, AERS. International journalof medical sciences2011,8(4):332.
[25] Davis S, King B, Raine JM: Spontaneous reporting–UK. Pharmacovigilance2002:199-215.
[26] Wysowski DK, Swartz L: Adverse drug event surveillance and drugwithdrawals in the United States,1969-2002: the importance of reportingsuspected reactions. Archives of internal medicine2005,165(12):1363.
[27] Moore TJ, Psaty BM, Furberg CD: Time to act on drug safety. JAMA: thejournal of the American Medical Association1998,279(19):1571-1573.
[28] Cruzan S: FDA Talk Paper: FDA proposes to Withdraw Seldane Approval.1997.
[29]钱轶峰，罗宝章，叶小飞等:检测联合用药不良事件信号的数据挖掘方法.中国卫生统计2010,27(1):31-34.
[30] Lee F, Harris W, Cheng H, Shenoi J, Zhao S, Wang J, Champion T, Izard J, GoreJL, Porter M: Pathologic Response Rates of Gemcitabine/Cisplatin VS.Methotrexate/Vinblastine/Adriamycin/Cisplatin Neoadjuvant Chemotherapy forMuscle Invasive Urothelial Bladder Cancer.
[31] Ritchie MD, Denny JC, Crawford DC, Ramirez AH, Weiner JB, Pulley JM,Basford MA, Brown-Gentry K, Balser JR, Masys DR: Robust replication ofgenotype-phenotype associations across multiple diseases in an electronicmedical record. The American Journal of Human Genetics2010,86(4):560-572.
[32] Tatonetti N, Denny J, Murphy S, Fernald G, Krishnan G, Castro V, Yue P, Tsau P,Kohane I, Roden D: Detecting drug interactions from adverse-event reports:interaction between paroxetine and pravastatin increases blood glucose levels.Clinical Pharmacology&Therapeutics2011,90(1):133-142.
[33] Guideline ICHHT: Clinical safety data management: definitions and standardsfor expedited reporting.1994.
[34] Kohn LT, Corrigan J, Donaldson MS: To err is human: building a safer healthsystem. Volume6: Natl Academy Pr;2000.
[35][http://www.sda.gov.cn/WS01/CL0747/63731.html]
[36] What Is a Medication Error? Rockville M: National Coordinating Council forMedication Error Reporting and Prevention.1998.
[37] Nebeker JR, Barach P, Samore MH: Clarifying adverse drug events: a clinician'sguide to terminology, documentation, and reporting. Annals of internal medicine2004,140(10):795-801.
[38] VI CWG: Management of Safety Information from Clinical Trials. Geneva:WHO.
[39] Zhou L, Plasek JM, Mahoney LM, Chang F, DiMaggio D, Rocha RA: MappingPartners Master Drug Dictionary to RxNorm Using an NLP-based Approach.Journal of Biomedical Informatics2011, DOI: http://www.sciencedirect.com/science/article/pii/S1532046411001948.
[40] UMLS许可协议[http://www.nlm.nih.gov/research/umls/license.html]
[41] Peters L, Bodenreider O: Using the RxNorm web services API for qualityassurance purposes. In:2008. American Medical Informatics Association;2008:591-595.
[42] National Drug File–Reference Terminology (NDF-RT) Documentation
[http://evs.nci.nih.gov/ftp1/NDF-RT]
[43] Bodenreider O: Using SNOMED CT in combination with MedDRA forreporting signal detection and adverse drug reactions reporting. In:2009.American Medical Informatics Association;2009:45.
[44] Estoup JB: Gammes Stenographiques.4th edition edition. Paris;1916.
[45] Manning CD,曼宁, Schutze H,许策,春法,苑:统计自然语言处理基础:电子工业出版社;2005.
[46]王伟:信息计量学及其医学应用.第一版edition:人民卫生出版社;2009.
[47] Kumar SK, Harding J: Ontology mapping using description logic and bridgingaxioms. Computers in Industry2012.
[48] Wang R, Wang L, Liu L, Chen G, Wang Q: Combination of the ImprovedMethod for Ontology Mapping. Physics Procedia2012,25:2167-2172.
[49] Hartung M, Gro A, Rahm E: COnto–Diff: generation of complex evolutionmappings for life science ontologies. Journal of Biomedical Informatics2013,46(1):15-32.
[50] Bock J, Hettenhausen J: Discrete particle swarm optimisation for ontologyalignment. Information Sciences2012,192:152-173.
[51]米杨,曹锦丹:基于PROMPT的本体映射实例分析.情报学报2010(006):987-991.
[52]吕刚,郑诚,胡春玲:基于概念分类的多本体映射方法研究倡.计算机应用研究2011,28(9).
[53]王效岳,胡泽文,白如江: WordNet与SUMO本体之间的映射机制研究.现代图书情报技术2011,1:005.
[54] Palchuk MB, Klumpenaar M, Jatkar T, Zottola RJ, Adams WG, Abend AH:Enabling hierarchical view of RxNorm with NDF-RT drug classes. In:2010.American Medical Informatics Association;2010:577-581.
[55] Pathak J, Murphy SP, Willaert BN, Kremers HM, Yawn BP, Rocca WA, ChuteCG: Using RxNorm and NDF-RT to Classify Medication Data Extracted fromElectronic Health Records: Experiences from the Rochester EpidemiologyProject. In:2011. American Medical Informatics Association;2011:1089-1098.
[56] Pathak J, Chute CG: Analyzing categorical information in two publiclyavailable drug terminologies: RxNorm and NDF-RT. Journal of the AmericanMedical Informatics Association2010,17(4):432-439.
[57] Warnekar PP, Bouhaddou O, Parrish F, Do N, Kilbourne J, Brown SH, LincolnMJ: Use of RxNorm to exchange codified drug allergy information betweenDepartment of Veterans Affairs (VA) and Department of Defense (DoD). In:2007. American Medical Informatics Association;2007:781.
[58] Peters L, Kapusnik-Uner JE, Bodenreider O: Methods for managing variation inclinical drug names. In:2010. American Medical Informatics Association;2010:637.
[59] Li Y, Salmasian H, Harpaz R, Chase H, Friedman C: Determining the Reasonsfor Medication Prescriptions in the EHR using Knowledge and NaturalLanguage Processing. In:2011. American Medical Informatics Association;2011:768-776.
[60] U.S. Food and Drug Administration Drugs@FDA [http://www.a ccessdata.fda.gov/scripts/cder/drugsatfda/]
[61] Drugs.com [http://www.drugs.com/]
[62] Francis WN, Kucera H: Computational analysis of present-day AmericanEnglish. Providence, RI: Brown University Press;1967.
[63] Wang X, Hripcsak G, Markatou M, Friedman C: Active computerizedpharmacovigilance using natural language processing, statistics, and electronichealth records: a feasibility study. Journal of the American Medical InformaticsAssociation2009,16(3):328-337.
[64] Harpaz R, Haerian K, Chase HS, Friedman C: Statistical Mining of PotentialDrug Interaction Adverse Effects in FDA’s Spontaneous Reporting System. In:2010. American Medical Informatics Association;2010:281.
[65] Kennedy DL, Goldman SA, Lillie RB: Spontaneous reporting in the UnitedStates. Pharmacoepidemiology2000:149-174.
[66] Harpaz R, Perez H, Chase H, Rabadan R, Hripcsak G, Friedman C: Biclusteringof adverse drug events in the FDA's spontaneous reporting system. ClinicalPharmacology&Therapeutics2010,89(2):243-250.
[67] Moore TJ, Furberg CD, Glenmullen J, Maltsberger JT, Singh S: Suicidalbehavior and depression in smoking cessation treatments. PloS one2011,6(11):e27016.
[68] Vilar S, Harpaz R, Chase HS, Costanzi S, Rabadan R, Friedman C: Facilitatingadverse drug event detection in pharmacovigilance databases using molecularstructure similarity: application to rhabdomyolysis. Journal of the AmericanMedical Informatics Association2011,18(Suppl1):i73-i80.
[69] Liu Y, LePendu P, Iyer S, Shah NH: Using Temporal Patterns in MedicalRecords to Discern Adverse Drug Events from Indications. AMIA Proceedings2012.
[70] Xu H, Stenner SP, Doan S, Johnson KB, Waitman LR, Denny JC: MedEx: amedication information extraction system for clinical narratives. Journal of theAmerican Medical Informatics Association2010,17(1):19-24.
[71] Friedman C, Shagina L, Lussier Y, Hripcsak G: Automated encoding of clinicaldocuments based on natural language processing. Journal of the AmericanMedical Informatics Association2004,11(5):392-402.
[72] Moore TJ, Glenmullen J, Furberg CD: Prescription drugs associated withreports of violence towards others. PloS one2010,5(12):e15337.
[73] U.S. Food and Drug Administration Drugs@FDA [http://www.accessdata.fda.gov/scripts/cder/drugsatfda/]
[74] Lu Y, Cederbaum AI: Cisplatin-induced hepatotoxicity is enhanced by elevatedexpression of cytochrome P4502E1. Toxicological Sciences2006,89(2):515-523.
[75] Friedman C: Towards a comprehensive medical language processing system:methods and issues. In Proceedings of the AMIA annual fall symposium:1997.American Medical Informatics Association;1997:595-599.
[76] Shah NH, Bhatia N, Jonquet C, Rubin D, Chiang AP, Musen MA: Comparisonof concept recognizers for building the Open Biomedical Annotator. BMCbioinformatics2009,10(Suppl9):S14.
[77] Savova GK, Masanz JJ, Ogren PV, Zheng J, Sohn S, Kipper-Schuler KC, ChuteCG: Mayo clinical Text Analysis and Knowledge Extraction System (cTAKES):architecture, component evaluation and applications. Journal of the AmericanMedical Informatics Association2010,17(5):507-513.
[78] LePendu P, Iyer SV, Fairon C, Shah NH: Annotation analysis for testing drugsafety signals using unstructured clinical notes. Journal of BiomedicalSemantics2012,3(Suppl1):S5.
[79] Patrick J, Li M: High accuracy information extraction of medicationinformation from clinical notes:2009i2b2medication extraction challenge.Journal of the American Medical Informatics Association2010,17(5):524-527.
[80] Rosenbloom ST, Awad J, Speroff T, Elkin PL, Rothman R, Spickard III A,Peterson J, Bauer BA, Wahner-Roedler DL, Lee M: Adequacy of representationof the National Drug File Reference Terminology Physiologic Effects referencehierarchy for commonly prescribed medications. In AMIA Annual SymposiumProceedings:2003. American Medical Informatics Association;2003:569-578.
[81] Zhu Q, Jiang G, Chute CG: Profiling structured product labeling with NDF-RTand RxNorm. J Biomedical Semantics2012,3:16.
[82] MedDRA [http://en.wikipedia.org/wiki/MedDRA]
[83] Nadkarni PM, Darer JD: Determining correspondences between high-frequencyMedDRA concepts and SNOMED: a case study. BMC Medical Informatics andDecision Making2010,10(1):66.
[84] Mougin F, Dupuch M, Grabar N: Improving the mapping between MedDRAand SNOMED CT. Artificial Intelligence in Medicine2011:220-224.
[85] Semantic MEDLINE Prototype [http://skr3.nlm.nih.gov/SemMedDemo/index.jsp]
[86] Kilicoglu H, Fiszman M, Rodriguez A, Shin D, Ripple A, Rindflesch TC:Semantic MEDLINE: a web application for managing the results of PubMedSearches. In Proceedings of the third international symposium for semanticmining in biomedicine:2008;2008:69-76.
[87] Yancik R: Cancer burden in the aged. Cancer1997,80(7):1273-1283.
[88] Wang J, Huang Y: Pharmacogenomics of sex difference in chemotherapeutictoxicity. Current drug discovery technologies2007,4(1):59-68.
[89] Fattinger K, Roos M, Vergères P, Holenstein C, Kind B, Masche U, Stocker DN,Braunschweig S, Kullak‐Ublick GA, Galeazzi RL: Epidemiology of drugexposure and adverse drug reactions in two Swiss departments of internalmedicine. British journal of clinical pharmacology2000,49(2):158-167.
[90] Raman A, Lombardo J, Chandrasekhar R, Fakih M: Bevacizumab relatedadverse events among various age groups of elderly patients with advancedcolorectal cancer. J Clin Oncol2007,25(18S):14546.
[91] Sloan JA, Goldberg RM, Sargent DJ, Vargas-Chanes D, Nair S, Cha SS,Novotny PJ, Poon MA, O’Connell MJ, Loprinzi CL: Women experience greatertoxicity with fluorouracil-based chemotherapy for colorectal cancer. Journal ofclinical oncology2002,20(6):1491-1498.
[92] Singh S, Parulekar W, Murray N, Feld R, Evans WK, Tu D, Shepherd FA:Influence of sex on toxicity and treatment outcome in small-cell lung cancer.Journal of clinical oncology2005,23(4):850-856.
[93] MediLexicon's FDA Approved Drugs Listings [http://www.medilexicon.com/drugs-list/cancer.php]
[94] CIMMiner [http://discover.nci.nih.gov/cimminer.]
[95] Yoshida S: The results of gefitinib prospective investigation. Med Drug J2005,41:772-789.
[96] Murashige N, Tanimoto T, Oshima Y, Lim KH, Chang YH, Inoue A, KobayashiK, Nukiwa T: Interstitial lung disease and gefitinib. New England Journal ofMedicine2010,363(16):1578-1580.
[97] Aapro M, Macciocchi A, Gridelli C: Palonosetron improves prevention ofchemotherapy-induced nausea and vomiting in elderly patients. The journal ofsupportive oncology2005,3(5):369.
[98] Harpaz R, Haerian K, Chase HS, Friedman C: Statistical Mining of PotentialDrug Interaction Adverse Effects in FDA’s Spontaneous Reporting System. In:2010. American Medical Informatics Association:281.
[99] medilexicon [www.medilexicon.com]