[MUSIC] Elements of the trial preparation phase are also important to ensuring the quality of the trial. Many of these have been discussed earlier in these courses, and we'll review them just quickly here. For example, data collection instruments play a really important role in ensuring the quality of a clinical trial. As with a study protocol, being clear, logical and organized can help prevent misunderstandings in how data are collected and/or reported. Some of the elements discussed earlier in the module on data collection instruments include visual simplicity, consistency within and across different data collection instruments. Whenever possible, clear communication, right on the instrument, of the formats for expected data, coding schemes used, and even occasionally explanations of how a particular data element is to be collected. Putting these right on the data collection instrument, instead of having them in a separate location that might be cross-referenced, makes it more likely that they'll be remembered, understood, and actually employed. Generally speaking, it has been my experience in trials that when confronted with a decision between making a single long data collection instrument or perhaps splitting that data collection instrument into multiple shorter instruments. Often the multiple shorter instruments will be easier to employ and less likely to lead to mistakes. Even if it seems inefficient because some of those instruments might have to repeat some information that was contained on other elements of it. So when given a choice, I will often prefer a few shorter instruments as opposed to one single longer instrument. Data management systems is yet another area in which lots of quality assurance processes can be employed and can be critical for maintaining the overall quality of a trial. Just as with data collection instruments, clear, logical, organized collection of the data in that system is likely to help prevent misunderstanding and promote better quality of the data. Many data management systems include built-in quality assurance methods. And several examples discussed in the data management module included things like data acceptability. An example would be range checks for numeric entries. Data consistency, comparing items one against another, example might be that a participant's age and his or her date of birth match each other for the date of a visit. And certain kinds of data logic checks, such as skip patterns or branching logic. So that the system automatically will check to make sure that there is not an entry for the number of pregnancies for male participants, or if there is an entry, that it is zero. Another important quality assurance procedure that can greatly improve overall quality is to attempt whenever possible to do extensive pilot testing of everything that will be used in a trial. This includes data collection instruments, which we would encourage study design staff to test extensively internally and then to share them with the clinical sites and the staff that will be actually employing them on participants. Perhaps testing them on each other or testing them on nonparticipant patients that might be similar to the population of the study that might be enrolled prior to the study opening. Lots of feedback, multiple iterations of designing the data collection instruments is common and often will prevent numerous revisions from having to be made once the trial opens. Which can be a considerable burden for the study and lead to greater possibility of mistakes and misunderstandings. Similarly for the data management system, extensive internal testing by the people developing that system. Followed whenever possible by external testing by the staff that will actually be using the system, but prior to it beginning to be used for the study. And again, feedback, iteration, making adjustments where appropriate to prevent the system from being used in ways that are not appropriate or end up with mistakes or problems in the subsequent data analysis. An exactly analogous approach for things like specimen management systems, image management system, if a study was collecting, say, X-rays or CT scan data. Testing all of those systems prior to them being used can greatly improve the quality of the end result of that system and therefore of the study. For example, many image management systems are going to include a deidentification phase. Since often the images collected clinically include specific identifiers that probably should not be collected and transmitted as part of the research study. Those systems, when tested, often can identify problems that could be addressed and fixed before the study begins. And that can be very helpful for improving the efficiency of the study and maintaining the quality. Clinical procedures should be pilot tested. If a particular kind of specimen is going to be collected in an unusual way in the study, that is a good opportunity to test it before the study begins to avoid problems or misunderstandings as to how to collect those kinds of specimens or perform those clinical procedures. Same thing for manuals and documents. Internal review, extensive internal review when possible with things like manuals of operations, just like for the data collection instruments. And then sharing them with the clinical site staff that will be using them and giving them an opportunity to provide feedback about things that were unclear or about which they were uncertain. So an important goal early in the trial preparation phase is to test everything possible. Another important element during the trial preparation phase is to consider how all study staff will be trained. And training includes many, many features in parts of a clinical trial, including understanding of the protocol and the study design. In practice, we often in our group include an assessment phase. In which anyone participating in the trial has to be able to pass a quiz that we administer online through a website, there are other ways to do that, to make sure that they have at least a basic understanding of the study. We also provide practice materials, elements of the protocol, perhaps packaged in a way to make it easier for reference. We also require study staff to demonstrate the ability to execute certain parts of the study protocol. So for example, staff that will be using the data management systems are expected to do practice data entry, demonstrate that they understand how the system works and will be able to do accurate data entry. We often employ a similar thing for specific clinical procedures so that if clinical staff are going to have to perform a certain kind of procedure, we might ask them to practice that on another person if that would be appropriate and safe. An example would be a noninvasive breathing test. We might ask them to perform that breathing test on someone else or more than once. And we would review the results of that testing and make sure that they had followed the correct protocol for obtaining the test data. There are important questions to consider when it comes to training of clinical site staff in a clinical trial. One of the important questions is whether some of the training or all of the training can be conducted remotely versus in-person training. It's been my experience that in-person training is more effective, but perhaps less efficient with regard to time and resources. And often in a trial, we will combine both of them. We'll have in-person training, we'll record the in-person sessions, make those available for staff that are unable to attend in person, and supplement the in-person training with additional remote resources. Even when using a remote mode of training for staff, there's also a question to be answered whether that needs to be live training versus recorded trainings. Sometimes there's a good opportunity for give and take for questions and live answers during a live session that is difficult to replicate in a recorded session. And so often we will provide recorded sessions, making it possible for staff to learn on their own schedule, but then ask them to participate in at least a few live sessions that supplement those recorded sessions. Another thing that we have found helps to really promote quality of training processes is to avoid assuming too much information on behalf of the participants. Being very basic with what you explain might seem inefficient and even potentially boring to study staff. But it does help prevent people who actually did not have the prior knowledge or background information, who often are hesitant to express that they didn't understand something. And make sure that everyone has the basic knowledge necessary to continue on in the study. Related to training is a process that many trials employ, often referred to as certification. Certification is a process to ensure and then document some minimum standards for understanding of the study or the study procedures. Many studies will employ different standards for different roles within a certain study so that people that are performing data entry will have to demonstrate proficiency with the data management system. But for example, a study physician not expected to do data entry would not have to demonstrate that level of proficiency. And instead might have to demonstrate proficiency, say, via a quiz or an assessment that they understand some element of the study that will be performed only by study physicians. Whenever possible, the certification standards as adapted for each of the roles in a study should be applied universally. It's okay, we think, to allow someone to complete certification based on demonstrating that same level of proficiency, for example, in another study or another element of the study. And we refer to this as grandfathering. So if a person has demonstrated proficiency with a particular kind of clinical test in a previous trial, then it's okay to certify them for a current trial without having them demonstrate that proficiency again. But we do try to avoid making exceptions for people who believe that they are too busy to comply with the certification procedures, or they find it difficult to imagine that someone wouldn't already know that they have those proficiencies. And we suggest administering certification and monitoring and managing it centrally whenever possible. In our group, we use an online certification tracking system that is available for view by all the staff at all of the sites so that there is complete transparency about who is and is not certified for a particular role. [MUSIC]
