【摘要】：在臨床試驗(yàn)中,通常會(huì)收集很多不同的數(shù)據(jù)模塊,比如不良事件,疾病歷史,人口統(tǒng)計(jì)學(xué)等,當(dāng)數(shù)據(jù)模塊中的錄入數(shù)據(jù)與數(shù)據(jù)標(biāo)準(zhǔn)有違背時(shí),會(huì)產(chǎn)生相應(yīng)的數(shù)據(jù)差異。目前大部分?jǐn)?shù)據(jù)差異是通過(guò)電子化邏輯檢查產(chǎn)生,其有助于及時(shí)發(fā)現(xiàn)和處理問(wèn)題數(shù)據(jù),但是電子化邏輯檢查表仍有一定的不足之處,如重復(fù)邏輯檢查,錯(cuò)誤邏輯檢查等.目的分析病例報(bào)告表中數(shù)據(jù)差異率比較高的數(shù)據(jù)模塊,探討優(yōu)化邏輯檢查程序。在保證數(shù)據(jù)一致性,完整性,準(zhǔn)確性的前提下,減少不必要的電子化邏輯檢查程序,提高數(shù)據(jù)管理員的工作效率。方法1.收集來(lái)自某大型跨國(guó)企業(yè)數(shù)據(jù)庫(kù)中,完成年份在2011年到2013年間的113個(gè)臨床試驗(yàn)數(shù)據(jù)。調(diào)查每個(gè)臨床試驗(yàn)的基本情況：試驗(yàn)編號(hào),臨床期數(shù),完成年份,治療領(lǐng)域,病人總數(shù),數(shù)據(jù)點(diǎn)總數(shù)。2.比較不同臨床期數(shù),不同完成年份和不同治療領(lǐng)域下,各臨床試驗(yàn)數(shù)據(jù)模塊差異率的高低。3.分析數(shù)據(jù)差異率較高的模塊上的邏輯檢查程序和數(shù)據(jù)差異,通過(guò)以下三個(gè)方面：系統(tǒng)優(yōu)化,程序簡(jiǎn)化以及電子化檢查和人工檢查的平衡,探討改良電子化邏輯檢查表的方法。4.針對(duì)每個(gè)數(shù)據(jù)模塊提出的優(yōu)化方法,抽取項(xiàng)目人數(shù)適中,數(shù)據(jù)模塊差異率有代表性的臨床試驗(yàn),對(duì)其進(jìn)行回顧性驗(yàn)證,分析應(yīng)用以上優(yōu)化方案后,所減少的邏輯檢查數(shù)目,以及進(jìn)而減少的不必要數(shù)據(jù)差異數(shù)目,驗(yàn)證其可行性。結(jié)果：1.比較發(fā)現(xiàn)不同臨床試驗(yàn)完成年份,不同試驗(yàn)期數(shù)和不同治療領(lǐng)域下,差異率高的數(shù)據(jù)模塊主要都集中在不良事件數(shù)據(jù)模塊,合并用藥數(shù)據(jù)模塊,實(shí)驗(yàn)室數(shù)據(jù)模塊,服藥記錄數(shù)據(jù)模塊。其中不良事件數(shù)據(jù)模塊的差異率范圍是4%-15%,合并用藥數(shù)據(jù)模塊的差異率范圍是7.5%-27.6%,服藥記錄數(shù)據(jù)模塊的差異率范圍是1.3%-18.3%,實(shí)驗(yàn)室數(shù)據(jù)模塊的差異率范圍是0.6%-23.7%。2.對(duì)于這四個(gè)數(shù)據(jù)模塊上設(shè)置的電子化邏輯檢查程序,建議通過(guò)以下三個(gè)方面對(duì)其進(jìn)行改善：①對(duì)病例報(bào)告表上設(shè)置的關(guān)聯(lián)問(wèn)題,敘述性文本和缺失數(shù)據(jù),進(jìn)行系統(tǒng)優(yōu)化,在數(shù)據(jù)錄入階段進(jìn)行控制,減少數(shù)據(jù)錄入時(shí)可能發(fā)生的錯(cuò)誤,從而減少數(shù)據(jù)清理時(shí)所需的邏輯檢查②對(duì)于同一數(shù)據(jù)點(diǎn)上設(shè)置的多條邏輯檢查,通過(guò)程序簡(jiǎn)化,去除邏輯重復(fù)或相近的檢查。③數(shù)據(jù)庫(kù)中設(shè)置的邏輯檢查并不是越多越好,實(shí)際項(xiàng)目中,需考慮電子化檢查和人工檢查的平衡,對(duì)于實(shí)驗(yàn)室數(shù)據(jù)模塊,建議將原本的電子化檢查改為人工檢查(Lab Review Tool)。3.對(duì)于不良事件模塊,發(fā)現(xiàn)臨床試驗(yàn)A通過(guò)上述優(yōu)化方法可以減少22.45%的邏輯檢查程序：對(duì)于合并用藥模塊,發(fā)現(xiàn)臨床試驗(yàn)B通過(guò)上述優(yōu)化方法可以減少26.3%的邏輯檢查程序：對(duì)于服藥記錄模塊,發(fā)現(xiàn)臨床試驗(yàn)C通過(guò)上述優(yōu)化方法可以減少21.1%的邏輯檢查程序；對(duì)于實(shí)驗(yàn)室記錄模塊,發(fā)現(xiàn)臨床試驗(yàn)D通過(guò)上述方法可以減少16.7%的邏輯檢查程序。結(jié)論：1.臨床數(shù)據(jù)差異率高的模塊主要集中在不良事件模塊,合并用藥模塊,服藥記錄模塊和實(shí)驗(yàn)室數(shù)據(jù)模塊,使今后臨床試驗(yàn)數(shù)據(jù)管理計(jì)劃的制定更有側(cè)重性。2.優(yōu)化后的電子化邏輯檢查表,既減少了不必要的電子化邏輯檢查程序,又在確保數(shù)據(jù)一致性,準(zhǔn)確性,完整性的基礎(chǔ)上減輕了數(shù)據(jù)管理人員的工作負(fù)擔(dān),為將來(lái)臨床數(shù)據(jù)管理中電子化檢查的標(biāo)準(zhǔn)化提供有價(jià)值的實(shí)驗(yàn)依據(jù)和參考。
[Abstract]:In clinical trials, many different data modules are usually collected, such as adverse events, disease history, demography, and so on. When the input data in the data module are contrary to the data standard, the corresponding data differences are generated. Most of the data differences are generated by electronic logic check, which helps to find and place the data in time. But the electronic logic check table still has some shortcomings, such as repeated logic check, error logic check and so on. Aim to analyze the data module with high data difference rate in the case report table, and discuss the optimization logic check program. In order to reduce the unnecessary electricity on the premise of ensuring the consistency, integrity and accuracy of the data. Methods 1. to improve the efficiency of data administrators. Method 1. collected from a large multinational enterprise database to complete 113 clinical trial data from 2011 to 2013. Investigate the basic situation of each clinical trial: test number, clinical period, completion year, treatment field, total number of patients, data points general .2. compared the number of different clinical periods, different years and different fields of treatment, the difference rate of each clinical trial data module,.3. analysis of the logic check program and data difference on the module with high difference rate of data, through the following three aspects: system optimization, program simplification, and the balance of electronic inspection and manual inspection. The method of improving the electronic logic Checklist (.4.), the optimization method proposed by each data module to extract the moderate number of project number and the representative clinical trial of the data module difference rate is reviewed, and the number of the reduced logic checks and the unnecessary data difference after the application of the above optimization are analyzed. Results: 1. the results are as follows: the data modules with high difference rate are mainly concentrated in the bad event data module, the combined drug data module, the laboratory data module, the drug recording data module, and the adverse event data module. The range of the difference rate is 4%-15%. The range of the difference rate of the combined drug data module is 7.5%-27.6%. The difference rate range of the drug recording data module is 1.3%-18.3%. The difference rate range of the laboratory data module is the electronic logic check program set on the four data modules by 0.6%-23.7%.2.. It is suggested to use the following three aspects. To improve: (1) the correlation problem set on the case report table, the narrative text and the missing data, the system optimization, the control in the data entry stage, the reduction of the possible errors when the data entry is reduced, thus reducing the logical check needed in the data cleaning. The program is simplified to remove logical repetition or close inspection. 3. The logical check set in the database is not the more the better. In the actual project, the balance between the electronic inspection and the manual inspection should be considered. For the laboratory data module, it is suggested to change the original electronic check into the Lab Review Tool.3. for the bad event module, It is found that the clinical trial A can reduce 22.45% of the logic check program through the above optimization: for the combined medication module, it is found that the clinical trial B can reduce the logic check program by the above optimization method: for the drug recording module, it is found that the clinical trial C can reduce the logical inspection process by 21.1% through the above optimization method. In the laboratory record module, it is found that the clinical trial D can reduce 16.7% of the logic check program through the above methods. Conclusion: 1. the modules with high clinical data difference rate are mainly concentrated in the adverse event module, the combined drug module, the medicine recording module and the laboratory data module, so that the future clinical trial data management plan is formulated. More emphasis is placed on the electronic logic checklist after.2. optimization, which reduces the unnecessary electronic logic check program, and reduces the work burden of data managers on the basis of ensuring data consistency, accuracy and integrity, and provides valuable experimental basis for standardization of electronic inspection in future clinical data management. Reference resources.
【學(xué)位授予單位】：復(fù)旦大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP311.13;R95

【相似文獻(xiàn)】

相關(guān)會(huì)議論文 前4條

1 徐宗昌;黃書(shū)峰;孫寒冰;;面向故障診斷的IETM故障數(shù)據(jù)模塊研究[A];Proceedings of 14th Chinese Conference on System Simulation Technology & Application(CCSSTA’2012)[C];2012年

2 彭菲菲;孫忠林;魏永山;;網(wǎng)絡(luò)數(shù)據(jù)資源動(dòng)態(tài)聚合中Web數(shù)據(jù)模塊技術(shù)的研究與應(yīng)用[A];IT服務(wù)促進(jìn)企業(yè)信息化——第十一屆中國(guó)Java技術(shù)及應(yīng)用交流大會(huì)文集[C];2008年

3 高秀麗;孫忠林;魏永山;;基于Web的網(wǎng)絡(luò)數(shù)據(jù)資源動(dòng)態(tài)關(guān)聯(lián)方法研究與應(yīng)用[A];IT服務(wù)促進(jìn)企業(yè)信息化——第十一屆中國(guó)Java技術(shù)及應(yīng)用交流大會(huì)文集[C];2008年

4 徐宗昌;張光明;李博;;基于IETM的裝備交互式訓(xùn)練研究[A];Proceedings of 14th Chinese Conference on System Simulation Technology & Application(CCSSTA’2012)[C];2012年

相關(guān)重要報(bào)紙文章 前4條

1 劉義玲;捷訊推出新款無(wú)線數(shù)據(jù)模塊[N];通信產(chǎn)業(yè)報(bào);2001年

2 林琪;組件技術(shù)在Delphi中的實(shí)現(xiàn)[N];中國(guó)計(jì)算機(jī)報(bào);2001年

3 周艷會(huì)邋作者單位系總后后勤科學(xué)研究所;利用Delphi讀寫(xiě)Excel信息[N];計(jì)算機(jī)世界;2007年

4 彭;捷訊SpringPort連接無(wú)線有線[N];中國(guó)計(jì)算機(jī)報(bào);2001年

相關(guān)碩士學(xué)位論文 前10條

1 袁柳柳;通過(guò)臨床數(shù)據(jù)差異分析優(yōu)化電子化邏輯檢查表[D];復(fù)旦大學(xué);2014年

2 阮先軍;油港設(shè)備智能交互式電子技術(shù)手冊(cè)研究[D];武漢理工大學(xué);2013年

3 于春慧;交互式電子技術(shù)手冊(cè)系統(tǒng)的研究與設(shè)計(jì)[D];南京郵電大學(xué);2013年

4 李宗亮;基于S1000D標(biāo)準(zhǔn)的IETM系統(tǒng)技術(shù)研究[D];廣東工業(yè)大學(xué);2006年

5 吳星佐;基于日志挖掘的IETM數(shù)據(jù)模塊推薦模型研究[D];哈爾濱工程大學(xué);2013年

6 韓美善;多模數(shù)據(jù)卡PC側(cè)軟件數(shù)據(jù)模塊的研究與實(shí)現(xiàn)[D];西安電子科技大學(xué);2012年

7 向芬芬;IETM技術(shù)在便攜式工卡系統(tǒng)中的應(yīng)用研究[D];南京航空航天大學(xué);2012年

8 楊天賀;基于S1000D規(guī)范的IETM的研究與應(yīng)用[D];南京航空航天大學(xué);2011年

9 于強(qiáng)強(qiáng);電能質(zhì)量在線監(jiān)測(cè)系統(tǒng)的設(shè)計(jì)與數(shù)據(jù)模塊的實(shí)現(xiàn)[D];山東大學(xué);2008年

10 任興旺;CALS體系下基于多色集合的IETM研究[D];廣東工業(yè)大學(xué);2007年

，

本文編號(hào)：2167969