Anticipating 600,000 Queries
March 17, 2008
Hold on to your desk. A major electronic data capture (EDC) vendor recently announced a revision to its technology that is ... entirely internet based. This was similar to announcing the development of a stone wheel that is ... round. It’s fine to brag about a round wheel with a press release. But don’t be emotionally bruised if Wired magazine leaves your new wheel off its cover next month.
On the other end of the innovation spectrum, Medidata Solutions has aimed higher. The New York City company has had a few serious hiccups with its service delivery. But the strength of its technology has allowed it to vault ahead of a long list of competitors in a highly competitive space.
Like its once-imposing round-wheel EDC rival, Medidata has updated its platform. This update will matter. Two new Medidata features in particular are likely to to raise the bar for other EDC suppliers. A few paragraphs down, we’ll discuss an ingenious way Medidata is letting data managers examine the consequences of a mid-study amendment before it is made live.
The clinical trial technology vendor is also touting something that, we suspect, sounds easy to deliver but was probably anything but. The training for its system is now online and fully integrated with its larger EDC system. There’s a single sign-on for everything. This is something that will make the life of the average clinical site significantly simpler.
Medidata’s chief technology officer, Glen de Vries, says it’s now impossible to access the live portions of a Medidata study without having completed the necessary training. Even the feedback form to evaluate the training is part of the same system. “If you have to log into somewhere else to get your training, that doesn’t make sense,” says de Vries.
So in the new version of its EDC platform, Medidata can address what it calls a major annoyance. The system can remember that a particular investigator has completed a specific training module and only show uncompleted courses, perhaps for protocol #9817AHP.
Medidata hopes happy investigators will sing the praises of its Rave EDC brand to their contacts on the sponsor side of the equation. “What you’re trying to do is make sure that the first system they walk in to and turn on can be Rave,” he says. “That’s how we think about how we will try to be the new hub for clinical development.”
“If you’ve got an investigator who’s working with you on thirty studies, you don’t need that investigator and that coordinator to be trained on Rave thirty times,” explains de Vries, adding that a sponsor can establish simple rules about which roles need to take which courses, and how frequently. The system, de Vries says, “will acknowledge they have taken the study. They won’t have to take it again. The relevant users will be presented with the relevant content, pass the test, and be back in the system. You are guaranteed at any moment that all your users have done all the relevant training.”
Medidata listened to sponsors who were struggling to keep track of who had been trained on which system for which trial. It decided to add more self-administered accounts. The inspiration was Amazon.com, where individual users can easily reset basic data about themselves. In Medidata’s world, easier self-administration means clinical research coordinators can use the system to view or print an official certificate showing that they have been trained on a particular study.
Having one system, not two, for training and administration should minimize the busywork of transferring credentials from one system to another, or locking individuals out of System A because they have not taken a tutorial in System B. “The actual net effect of how much easier it is to manage these users is dramatic,” says de Vries. “There is no need to reconcile those two systems. You can, within one system, create accounts for all your users and in one place decide what types of users need what types of training.”
Medidata is trying to use internet standards in elearning. The company is using a variety of Adobe tools, including Flash, but also adheres to an XML standard called SCORM, which stands for Sharable Content Object Reference Model. That should facilitate sponsors who want to generate their own training, though of course Medidata has standard courses for its own technology.
Sponsors will be able to upload SCORM-compliant or Adobe content and be good to go. For trainees, as we saw in a demo, it’s about as complicated as starting and stopping a YouTube video. “We put a lot of control in the users’ hands in terms of how they wanted to pace the course,” says Medidata’s VP of program strategy, Daniel Mudgett, who joined the company after working at Amgen.
The system has quizzes built in to test a user’s knowledge. It randomly poses different questions to minimize the ability to cheat. (Sponsors can determine how many correct answers are needed to earn a passing grade.) It’s easier on the eyes and more engaging than a slide deck. “The point is to make it interesting and keep you engaged,” says Mudgett. “Watching Powerpoints is really dry.”
Medidata chief medical officer Ed Ikeguchi, a physician, says sponsors are increasingly aware of the burden of taking investigators away from billable and productive activities, even for a day or two. The industry, he says, wants an easier way. The novelty of going to a hotel to be trained (again) wears off quickly, Ikeguchi says, especially for the most seasoned clinicians.
“After you’re ramped up and busy and have a lot of studies, it’s quite demanding on your time to get away,” says Ikeguchi. “The ability to handle something like this online is a wonderful possibility.”
In the same update of its EDC system, Medidata is offering something it calls publish in place (PIP). The firm calls PIP a faster way to explore the ramifications of mid-study changes. It lets data managers simulate the myriad effects of a study amendment.
PIP is an outgrowth of old Rave functionality called amendment manager, which facilitates study amendments to live trials.With nearly every study in the industry having at least one amendment, it’s high time that there’s a less painful method to see the effects of such changes on an EDC system.
PIP uses test data that are split off the live data for a particular trial. “We basically let you say, here is all your test data, here’s your new metadata, and this is what your data looks like in that context,” de Vries says.
Typically, de Vries says, many sponsors go through significant contortions and nervous waiting to amend their studies. Sponsors may make the changes in a trial live, only to discover something that was hard to anticipate. Perhaps there were so many dependencies or calculations within a set of linked case report forms (CRF) or edit checks that it was impossible to forecast what might happen after a change on CRF #32. Maybe there was something that everyone forgot about on CRF #7 that had untoward effects on CRF #32.
“They can take real world data examples and say, ‘I’ve changed this cascade of variables; now what does it look like?’ ” de Vries says. Such results are a click or two away, he says, not the result of hours or days of testing or loading data.
‘Houston, We Have ...’
Perhaps the study amendment looks great. Or maybe a few apparently innocuous changes in logic generated 600,000 new queries.
“That happens to people,” says de Vries. “It’s transformational in nature to test amendments before you roll them out. If I have a good set of test data, I can see what the final net effect will be.”
Medidata’s system could have interesting effects on reducing or shifting mid-study user-acceptance testing (UAT). It may also minimize the number of unnecessary apologies, late nights, tense phone calls or other melodrama that some sites and sponsors must be put through.
De Vries says the firm wants the industry’s data managers to be able to focus on data management, not just whether an EDC system is functioning well. “The data manager can concentrate on the changes—are we accurately reflecting the amendment we want in the study? Instead of ‘what do I have to do to validate that stuff?’”d9A2t49mkex