CDISC編碼標(biāo)準(zhǔn)和21CFR Part11在醫(yī)學(xué)研究系統(tǒng)中的應(yīng)用

1、周周 洪洪 Zhou H數(shù)據(jù)采集和處理系統(tǒng)提供商 Data acquisition and processing system provider CDISC編碼標(biāo)準(zhǔn)和21CFR Part11在醫(yī)學(xué)研究系統(tǒng)中的應(yīng)用The application of CDISC and 21CFR Part11 in clinical research標(biāo)準(zhǔn)讓數(shù)據(jù)互相認(rèn)識Standardized data can communicate seamlessly全球PM2.5最高的地區(qū)在北非和中國的華北、華東、華中全部。世界衛(wèi)生組織(WHO)認(rèn)為，PM2.5小于10是安全值，中國的這些地區(qū)全部高于50接近80。（南方

2、都市報）PM2.5監(jiān)測網(wǎng) http:/舉例：PM2.5事件-美國大使館公布PM2.5-政府認(rèn)可并同樣公布臨床研究數(shù)據(jù)是否需要標(biāo)準(zhǔn)化？DOES CLINICAL DATA NEED STANDARDIZATION ?新藥研發(fā)到上市的路線圖Pathway from Compound to New DrugR&D (研發(fā))Preclinical (臨床前期)PH I (一期臨床)PH II (二期臨床)PH III (三期臨床)Submission ( 申報審批)Year (年)Every step generates dataEvery step needs data management

3、Process(流程)藥物上市后的研究Post-marketing clinical reserachEvery step generates dataEvery step needs data management不良反應(yīng)監(jiān)測AE結(jié)果評價研究outcome市場策略 Strategy回顧和前瞻Retro & Pro新藥研究和上市后研究的區(qū)別pre & post marketing researchEvery step generates dataEvery step needs data management采集Collection運輸Transportation收樣Recei

4、pt分離Process入庫Storage分析Assays轉(zhuǎn)運Shipment出庫Retrieving新藥研究Trial有生物標(biāo)本的采集 Sample collectionEvery step generates dataEvery step needs data management上市后研究 Post maketing clinical research采集Collection存儲數(shù)據(jù)Storage分析Assays一般沒有生物標(biāo)本的采集Without sample collection臨床數(shù)據(jù)采集的目的Objective of clinical research datap 數(shù)據(jù)標(biāo)準(zhǔn)化的目的

5、 Objective 藥物臨床研究 Clinical trial 提交FDA評審 FDA data submission 與國際同行交換數(shù)據(jù) Data exchange eg: STS EACTS臨床數(shù)據(jù)的應(yīng)用流程Operational Process for Clinical Data EDC電子病歷EMR源數(shù)據(jù)Source DataRemote backup統(tǒng)計師Statistics心電圖ECG實驗室Labs研究者InvestigatorDatabaseEDC 和臨床數(shù)據(jù)的現(xiàn)狀Current Status for EDC and Clinical DataEDC是目前最流行的數(shù)據(jù)采集方法E

6、lectronic Data Capture (EDC) is the commonly used method to collect clinical data有很多EDC系統(tǒng)，每個的數(shù)據(jù)結(jié)構(gòu)都不一樣There are so many EDC vendors now, each of them has different database structure, setup, different variable naming convention and labeling and data attributes 有的界面和CRF一致some have the data entry inter

7、face similar to the CRFs 有的不一致， 完全是列表式 Some have a listing-form of data entry screen EDCRemote backupDatabaseEDCEDC 和臨床數(shù)據(jù)的現(xiàn)狀Current Status for EDC and Clinical Data提供給統(tǒng)計師的數(shù)據(jù)各不一樣Data delivered to statistician are very different! 數(shù)據(jù)呈遞給FDA也各不一樣Data submitted to FDA are very different! 由于這種非標(biāo)準(zhǔn)化情況，F(xiàn)DA很難橫

8、向的比較不同廠家，同類藥的療效和安全性Because of non-standard and non-exchangeable, there was no way for FDA to evaluate drug applications cross different sponsors EDCRemote backupDatabaseEDC樣本數(shù)據(jù)示例SampleGrace研究項目暴露出的問題The Problemsp 大部分臨床醫(yī)學(xué)研究 Most clinical trials 沒有采用數(shù)據(jù)標(biāo)準(zhǔn)化 dont employ a standard for data exchange 沒有使用標(biāo)

9、準(zhǔn)化的工具和技術(shù) dont use standardized analytic tools or techniques p 后果Result 其結(jié)果是造成了臨床數(shù)據(jù)分析中效率和系統(tǒng)化的困難，特別是對多個臨床試驗Analyzing clinical trial data efficiently and systematically is difficult and time consuming, especially across many trials 不如這樣一個簡單的問題：有多少女性參與了GRACE的臨床試驗？e.g. How many women participate in clini

10、cal trials with GRACE? 如何自動統(tǒng)計女性數(shù)量How do you automate counting women? 面對這樣的一個簡單的問題，我們往往會感到無能為力 Facing on such a simple question, we often feel lack of power 沒有標(biāo)準(zhǔn)的數(shù)據(jù)集名 No standard file names 怎么可能自動地用計算機找出包含要找數(shù)據(jù)的數(shù)據(jù)集How does a computer find the file that contains the data ? 沒有標(biāo)準(zhǔn)的變量名 No standard variable

11、 names 怎么可能自動地用計算機找出包含性別的數(shù)據(jù)名How does a computer find the column that contains sex information 沒有標(biāo)準(zhǔn)的數(shù)據(jù)編碼 No standard terminology 怎么可能自動地用計算機告知1還是2代表女性How does a computer know which code represents which sex ? 臨床數(shù)據(jù)標(biāo)準(zhǔn)化非常關(guān)鍵和重要！ CLINICAL DATA STANDARDIZATION IS VERY IMPORTANT AND CRITICAL! 數(shù)據(jù)的可交換性Data Ex

12、changeability 定義Definition 數(shù)據(jù)的可交換性是它具有的屬性，它使得數(shù)據(jù)可以以各種方式傳送和使用而不會失去它的內(nèi)容和意義Data exchangeability is the attributes of data that enable the data to be communicated and used in many ways without loss or alter the information content or meaning assigned to the data. 只有標(biāo)準(zhǔn)化的數(shù)據(jù)才具有可交換性O(shè)nly standardized data are

13、 exchangeableCDISC是什么? CDISC:臨床數(shù)據(jù)交換標(biāo)準(zhǔn)協(xié)會Clinical Data Interexchange Standards Consortium, started with small group of volunteers in 1997 是一個全球的、開放的、多學(xué)科的、非營利的組織，創(chuàng)建了支持臨床研究數(shù)據(jù)和元數(shù)據(jù)采集、交換、提交和存檔的標(biāo)準(zhǔn)It is a global, open, multidisciplinary, non-profit organization that has established standards to support the a

14、cquisition, exchange, submission and archive of clinical research data and metadata. 主要的模型 There are two major models in CDISC: SDTM:針對臨床數(shù)據(jù) Standard Data Tabulation Model for clinical trial data ADaM:針對分析數(shù)據(jù)Analysis Data Model 其他正在建立的標(biāo)準(zhǔn)Other still-developing standard CDASH:針對CRF表 Clinical Data Acquis

15、ition Standards Harmonization for CRF standardsSDTM的基本模型SDTMs fundamental model for operational dataObservation觀察Generic structure（衍伸結(jié)構(gòu)）Unique identifiersTopic variable or parameterTiming Variables Qualifiers.Interventions干預(yù)Findings發(fā)現(xiàn)Events事件General classes （一般分類）SubjectCMEXEGIELBPEAEDSSDTM Domains

16、（SDTM域）(dataset structures)The patient/subject focused information model of the clinical reality (general classes of observations on subjects: interventions, findings, events). This model has been developed by CDISC/SDS team.Interventions干預(yù)Events事件ConMedsExposureAEMedHistDispositionFindings結(jié)果ECGPhys

17、ExamLabsVitalsDemogOther其他Subj Char*Substance UseIncl Excl*RELATESSUPPQUALStudy SumStudy DesignQS*, MB*CommentsCP*, DV*CDISC SDTM 域 DomainsCDISC SDTM 域注解 Domains comments分類分類解釋解釋EX試驗用藥記錄CM伴隨用藥記錄和其它物理治療SU 物質(zhì)使用（煙、酒、咖啡等）干預(yù)干預(yù)Interventions事件事件EventsAE不良反應(yīng)事件DS試驗完成、中止、脫落報告MH病史（試驗藥物相關(guān)疾病病史及其它病史）結(jié)果結(jié)果Findings分

18、類分類解釋解釋EG心電圖IE入選、排除標(biāo)準(zhǔn)LB實驗室檢查QS調(diào)查問卷（HAMILTON DEPRESSION SCLAE, SF36, ADAS.）PE體格檢查SC個體特征（比如患者姓名、眼睛顏色、生育狀態(tài)、過敏狀態(tài)）VS生命體征注： HAMILTON DEPRESSION SCLAE 指 漢密頓抑郁量表 SF36 簡明健康狀況問卷 ADAS 阿爾茨海默病評分其他其他OtherDM受試者人口學(xué)資料CO非異常描述、其它注釋FDA對數(shù)據(jù)標(biāo)準(zhǔn)化的態(tài)度FDAs Stands on Clinical Data StandardizationFDA藥品審評中心于2004年7月就接納CDISC CDER a

19、dopted CDISC SDTM 7/04今年二月， FDA發(fā)布了關(guān)于數(shù)據(jù)標(biāo)準(zhǔn)化的指導(dǎo)原則草案FDA released a draft guidance in Feb. 2012 on study data standardizationFDA鼓勵申辦方采用標(biāo)準(zhǔn)化數(shù)據(jù)，越早越好FDA is encouraging manufacturers to adopt these new data submission standards early in the product development cycle FDA認(rèn)可CDISC作為標(biāo)準(zhǔn)之一 CDISC is one of standard F

20、DA endorsed越來越多的藥廠已開始采用CDISC標(biāo)準(zhǔn)The more and more companies are submitting data FDA in CDISC standard再過2-3年，CDISC可能成為FDA的正式標(biāo)準(zhǔn)It could be an official requirement of CDISC standard by FDA in 2-3 yearsFDA申報標(biāo)準(zhǔn)數(shù)據(jù)包Submission Data Package FDA申報標(biāo)準(zhǔn)數(shù)據(jù)包以下The submission data package to FDA usually include the fol

21、lowing SDTM datasets in SAS transport format (.xpt) SDTM annotated CRFs Define.xml for SDTM ADaM (in .xpt) Define.xml for ADaM Review guide 審閱指導(dǎo)申報數(shù)據(jù)進入FDA的流程 Flow for Submission Data within FDAJanus System審辦方SponsorSDTMADaM統(tǒng)計評審員FDA Statistical Reviewer醫(yī)學(xué)評審員FDA Medical Reviewer電子申報接口Electronic Submiss

22、ion Gateway (ESG)常用創(chuàng)建SDTM和ADaM流程Commonly Used Procedure for Creating SDTM and ADaMOperational Database (Raw)Clinical Data Repository (CDR)SDTMADaM幾乎是通過SAS在數(shù)據(jù)庫外的后期制作Created via SAS轉(zhuǎn)換為SDTM和ADaM存在的問題Issues in Mapping SDTM and ADaM from Raw目前的SDTM轉(zhuǎn)換是單向的，沒法溯源It is an one-way mapping (i.e. RawSDTM) withou

23、t traceability. E.g. age was calculated variable. From the mapped SDTM, it didnt tell how age was calculated太多的數(shù)據(jù)庫外的手工程序，非常容易出錯! Too much outside database programming，easy to make mistakes in the conversion processes增加了大量的質(zhì)控和驗證工作Increased a lot of validation and QC work不能保證申報數(shù)據(jù)的質(zhì)量和溯源性Lack of insuran

24、ce of quality and traceability for submitted data package沒有有用的軟件或工具No powerful tools for automatic SDTM mapping理想的步驟The Should-be ProcedureOperational Database (Raw)Clinical Data Repository (CDR)SDTMADaM無系統(tǒng)外轉(zhuǎn)換No outside-system mapping目前還沒有這樣的系統(tǒng)No system available yetShould be xml format, not SASSAS

25、datasets are two dimensional, i.e. it is “flat” Xml is multi-dimensional, i.e. it is stereoFDA is thinking to request for xml format in the future部分的解決方案Partial Solutions 對一個新study，數(shù)據(jù)標(biāo)準(zhǔn)化應(yīng)從數(shù)據(jù)庫的設(shè)計開始For new study, data standardization can be started with the database design 利用CDASH的標(biāo)準(zhǔn)表建庫 Use CDASH stan

26、dard forms 用SDTM給變量命名和標(biāo)識，但不是所有的變量都能這樣做Use SDTM to name and label variables (not all variables can be directly named in SDTM, e.g. date field 用SDTM域來分類 Group data using SDTM domain 即使這樣，一個完全的SDTM標(biāo)準(zhǔn)庫也會面臨挑戰(zhàn)Even you setup database in CDASH, it is still not possible to have a100% SDTM database. However,

27、it is a SDTM-like database 對以前的歷史遺留研究，用SAS轉(zhuǎn)換仍然是唯一的辦法For legacy studies, mapping through SAS is still the only wayCDISC 總結(jié)SummaryCDISC是用于數(shù)據(jù)申報的一個趨勢CDISC standard is the current trend for regulatory submission數(shù)據(jù)標(biāo)準(zhǔn)化對申辦方和FDA都非常重要和關(guān)鍵It is very important and critical for both sponsors and FDA to have a qua

28、lity and efficient regulatory review但是，我們還面臨很多挑戰(zhàn)However, there are a lot of challenges in 對CDISC指導(dǎo)原則的理解understanding the CDISC guidance 如何具體實施 Implementing the modules 需要研發(fā)有效的可直接轉(zhuǎn)換和溯源的軟件或工具Developing some powerful tools or software for direct mapping from ODM to SDTM and being able to trace back the

29、 source 如何保證電子數(shù)據(jù)的有效和安全?ENSURES SECURITY AND DATA ADMINISTRATION ARE EFFECTIVELY MANAGED21CFR Part11 是什么?l 21 CFR Part 11 是聯(lián)邦法規(guī)21章第11款 主要規(guī)定內(nèi)容涉及電子記錄和電子簽名 l 食品、醫(yī)藥制造行業(yè)多遵照此標(biāo)準(zhǔn)，此法規(guī)確保了電子數(shù)據(jù)的有效性和可靠性 l 遵照此標(biāo)準(zhǔn)而保留的數(shù)據(jù)才可以作為通過檢驗或者今后追溯的有效數(shù)據(jù)來源l 現(xiàn)在很多大型自動化控制系統(tǒng)已經(jīng)符合此標(biāo)準(zhǔn)規(guī)定21CFR Part11的重點Keyponits電子記錄 E-records 封閉系統(tǒng)的管理開放系統(tǒng)的管理簽名的顯示簽名/記錄的連接電子簽名 E-signatures簽名的唯一性和證據(jù)電子簽名的構(gòu)成和管理識別代碼和密碼的管理我們D