International Circulation: Even if it is electronic， how do you know that the data accurate without sending someone to check?
　　《国际循环》:即使数据电子化了，如何不派人去核实就能确定数据的准确性呢？
　　Dr. Harrington: That is the one of the misconceptions. That these data monitoring missions are to protect against rampant fraud and mistakes. I would submit that the biggest mistakes made in the medical record is not transcription to the case report form， but from the doctor writing data into the medical record.  One of the underpinnings of the large sample size is that the numbers will wash out the mistakes. There are methods to look for evidence of fraud: statistical outlier analyses， distribution analyses， etc.
　　Robert Harrington：这就是错误概念之一，即数据检查的任务是防止欺诈行为和错误。我认为病历方面最多见的错误不是将病历向病例报告表誊写时的错误，而是医生填写病历时的错误。大规模临床试验的基础之一就是研究的规模会稀释这些错误。有其他方法来寻找欺诈的证据：统计异常值分析、分布分析等。
　　International Circulation: What was the previous function of these data monitoring persons?
　　《国际循环》:这些数据监察人员的过去的作用是什么？
　　Dr. Harrington: They are accomplishing very little. They occasionally find a mismatch between the medical record and the case report form. Usually， they only find things that do not matter. Part of the reason why they go is because it has become a business for some to send them. There is this creation of whole industry around this. Imagine this. In a world electronic health records， you are in the hospital and you are an investigator. You have enrolled 50 patients for a trial of acute MI. All of their data is in the electronic record. Now imagine， a message comes from the principal investigator for the study. The message says that “I have the consent of these 50 patients in your system and here are the 100 details I need to know about each of them in your medical record.” The message is allowed into the electronic health record， plucks out the relevant information， and deposits it into a trial database. Complete efficiency. You have eliminated the middle person， eliminated the risk of transcription mistakes， and you have done it instantaneously. You have saved time， work， and effort. You can also combine that data with the other thousands of patients in the trial and run analyses for outliers.
　　Robert Harrington：他们完成的工作很少。偶尔会发现病历和病例报告表的不符之处。通常，他们发现的问题都不算什么问题。原因之一是他们去监察只是因为派他们监察是一项业务。这是整个行业所制造的。想象一下，在一个世界电子健康档案中，你作为一个研究者，在医院里，你为一个急性心肌梗死的试验入选了50例患者，所有的数据都在电子档案中。现在想象一下，来自主要研究者的信息说“我有你系统中50例患者的知情同意书，我需要知道关于你病历中的每个患者的100个详细信息。”信息输入电子健康档案，导出相关资料，并存入试验数据库，这样效率非常高。这也消除了中间人，消除了抄写错误，并且可以即刻完成，省时省力。同样还可以将那些数据和这个试验中其他几千名患者的数据进行对比，并进行异常值分析。
　　International Circulation: Dr. James stated in a previous interview that one can only use this data for simpler questions. Could you explain this?
　　《国际循环》:James博士在过去的一个采访中表示只能在较简单问卷上使用这个数据库。您能解释这个吗？
　　Dr. Harrington: He is right. In Swedish study， they can only ask a few follow up questions. They know if you died， they know if you were rehospitalized， and they might know some of the medications you were taking， so you have to confine your question to what you are collecting. In the SAFE-PCI trial， on radial vs. femoral access in women， in addition to plucking out data， we have asked a few more questions. This requires more work. When you randomize people to different parts of the study， you can ask a separate set of questions. And while this costs more money， it will not be much more. You are not creating a new platform， you are just adding on to what is already known.
　　Robert Harrington：他是对的。在瑞典研究中，他们只进行了几个随访问题，以了解患者是否死亡，是否再次入院，可能还了解患者正在服用的药物，因此，需要根据需要搜集的信息对问题进行限制。SAFE-PCI试验在女性中就股动脉和桡动脉入路对比，在这个试验中，除了获得数据之外，我们还多问了几个问题。这需要更多的工作。当你将患者随机到研究的不同组的时候，你可以问不同的问题。这虽然多花钱，但不会太多。你并不是在创造一个新平台，而只是在已知的基础上增加。