Investigator sites accept the need for technology like electronic data capture (EDC). But they remain puzzled about why some problematic EDC solutions are selected by clinical trial sponsor firms that seem indifferent to the operational problems created by such systems.
“Sponsors and CROs really need to think about this,” says Christine Pierre, CEO of the 16-year-old site network RXTrials, and founder of Site Solutions Summit. “Vendors shouldn't be chosen based on who had the biggest, glossiest presentation, but rather, what is their product like for the end user? Randomization goes back to the volunteer who 's waiting and waiting and waiting, and maybe even being sent away. Do we really want new trial subjects' first entree to research to go like this?” Sites have multiple concerns about EDC:
There's no standardization around how data are recorded. Yes, the Clinical Data Interchange Standards Consortium (CDISC) has been developing standards. Some of those standards should simplify the design and navigation of electronic case report forms. But sponsors tend to have idiosyncratic ways of structuring the data they need even within those parameters. And since the FDA never required the usage of CDISC standards, the sponsor community interprets them as suggestions.
“Everyone is so averse to being told to how to collect their data, there is fairly strong resistance,” says Patricia Larrabee, founder and CEO of Rochester Clinical Research. Staffers at the site level muddle through what each sponsor has decided for each trial, however odd it may be in medical terms.
There are so many EDC systems, each of which must be mastered at a site. EDC systems fall into two categories: those developed in home-grown fashion by the sponsor or contract research organization (CRO), and those developed by vendors like Oracle. The home-grown systems are fading from the industry, but they're still used. Either type can present problems if a sponsor has made its own tweaks over the years. Most sponsors can't resist doing so. Every customization can introduce bugs and dependencies that may take years to fix.
If the quirks aren't fully discussed in training—and training tends to be highly generic—then the quirks lead to queries. If a site's query rate is high, its reputation is affected. So sites feel the limitations of EDC systems are affecting them.
Larrabee remembers a situation in which a sponsor used a major EDC platform. Each screen had a tiny X at the bottom. No one knew what the X connoted; it was not discussed in training. In the end, the X generated a whopping 300 queries. A monitor had to come to the sites to sort it out. “But typically, the monitor hasn't seen the system, either,” says Larrabee. Why? Because...
EDC systems usually aren't ready when the trial starts. “Sometimes we'll have 15 patients already in the study before the system is operational,” says Larrabee. “Half the time the system isn't ready when a trial starts, and we don't even know what it's going to look like.” That state of affairs, needless to say, is a function of the sponsor world's historic penchant for endlessly revising trial protocols at the expense of every other element of the process. With thousands of trials starting and stopping all the time, it's unlikely EDC systems will ever be ready when sites are.
On top of that, there are often password delays. It can take a few weeks for a sponsor or CRO to generate a password for site staff. And if the site needs a password in a hurry—say, the person who's handling data input is out sick—there's no getting a quick password for a temp or substitute worker.
Extra work and the training penalize the site. Pierre says reimbursement for sites has remained flat over the last decade, but the cost of doing business has increased. A lot of that has to do with EDC.
Most new studies require about eight hours of EDC training. And the site is generally not compensated for that time. “Even when you get a job at McDonald's and they train you, you are on the clock. But sites are expected to do this on their own time,” says Pierre. To add insult to injury, the training may not be relevant to the exact system being used or the study being conducted. Even so, the sites must sit through it.
Thus some sites must hire extra staff simply to deal with EDC. Larrabee notes that her site has had to add four employees to handle data entry since the advent of EDC, and it's getting ready to add a fifth. Most industries generally use technology to reduce personnel costs; the pharmaceutical trade is a counter-example, backing a costly technology that requires more people.
Why have EDC systems been developed in a way that's so unfriendly for the people who use them the most? There is rhetoric around how important sites are to sponsors. But in the end it is a power play, according to Pierre. “Sites have traditionally been viewed as the nonentity in the clinical research food chain,” she says. “Up until very recently, the site perspective was inconsequential to the industry.”
Fortunately, that's changing. Pierre says more sponsors and CROs are looking at the dismal numbers. Only a quarter of sites stay in research after four years. Almost 65 percent of investigators do one study and nothing more. So sponsors realize it's time to partner with sites, not alienate them.
More and more, Pierre says she's seeing sponsors ask sites what they might like to see in an EDC system, and making the effort to validate server strength before a high-volume study. At some startup meetings, sponsors provide screen shots so sites can have a preview of their EDC experience. Other sponsors have launched site relationship management efforts to strengthen ties with sites.
And sites, at least large ones, are getting better about standing up for themselves. Some are asking ahead of time how much paper will be used and budgeting for EDC accordingly. In contracting discussions, some sites are asking for what they need instead of just taking whatever the sponsors offer. But that process—sites becoming empowered—is slow.
“I think there has been significant change in the culture of how sponsors and sites interact. It's very new, though, and unfortunately it hasn't yet filtered down to EDC,” says Pierre. Adds Weerasinghe: “The concept of EDC is great, but there are a lot of improvements necessary for it to work at the site level. It's just not there yet.”
—by Suz Redfearn
Editor's note: The first part of this series can be found here.d9A2t49mkex