Roundtable: Are We There Yet? How About Now? Now?

Electronic data capture (EDC) is an emerging paradigm for gathering information in clinical trials. Ask anyone who has sorted through stacks of accumulated paper at the end of a study, and they'll say EDC is the wave of the future. But while many companies are on board with the technological benefits, enthusiasm wanes when it comes to actual implementation. Even some of EDC's biggest champions admit to its obstacles: "It's a disruptive technology that doesn't give immediate returns," says James Tiede, vice president of integrated data services, global clinical operations at Johnson & Johnson.

Gary Tyson, vice president, clinical development practice at Campbell Alliance

Tiede is one of several industry executives who gathered to discuss companies' obstacles for full-blown EDC implementation—and also to examine the technology's benefits. Hosted by Pharmaceutical Executive and management consultancy Campbell Alliance, the executive roundtable brought together a group of early EDC adopters, all of whom had offered thought leadership to the industry and their company in pushing forward the technology.

The conversation was moderated by Pharm Exec's editor-in-chief Patrick Clinton.

Along with J&J's Tiede, Clinton was joined by Kelly Vaillant, global head, clinical data management at Amgen (attending via speaker phone); Steven Olsen, associate director, global development and informatics at Bristol-Myers Squibb; Gary Tyson, vice president, clinical development practice at Campbell Alliance; Ted Chin, vice president, electronic data management at GlaxoSmithKline; and Paula Brown Stafford, executive vice president, global data management at Quintiles. The following is an edited transcript of the session.

Ted Chin, vice president for electronic data management at GlaxoSmithKline

CLINTON: Most companies know the benefits of EDC—the real-time access to data, faster study completion, and cost savings are clear. Still, some leaders are hesitant to fully integrate EDC into their organizations. Why should they integrate EDC into a clinical trials process that has worked for them for years?

TYSON: One of the messages we tell senior management is that EDC has a lot of tremendous benefits, but you have to be patient if you want to see them. You're not going to see it in year one. It's just a classic case of people needing to be aware that with any kind of major change, you have a ramp-up curve before you see any benefit.

TIEDE: I agree. It's a disruptive technology that doesn't give immediate return. But through the repeated application and deployment of integrating EDC into how we do our business, that's where you get the return. That takes more than just a year.

Paula Brown Stafford, executive vice president, global data management at Quintiles

OLSEN: If you look at the conversation at EDC conferences four years ago, everyone was saying, "Let's try it, let's pilot it." Now you go to conferences and people are saying, "I'm in this stage of an implementation," "We're at 90 percent already," or "We're in the year where we're moving from 25 to 75 percent." Companies are now in that stage where they've committed to it, and we're somewhere along a continuum of getting past the pilot and into whatever is full mutation for an organization. There's a growing understanding through our peer community and a sense of community among the pharma companies.

TYSON: There's a willingness to question fundamental assumptions that people have embraced since they were junior-level staff in the industry, but it's hard to do because change is hard. One of the key challenges is the willingness to look at fundamental assumptions about roles, specifically whether they are roles between the sponsor and the site. For example, in many people's minds, there's a very clear distinction between what monitors do and what data managers do. But in an EDC world, that becomes a very gray area—it becomes an arbitrary decision that has to get made. It's not dictated.

If you go with the assumption of what monitors have always done and what data managers have always done, will EDC work? Sure, but it won't work optimally, so it becomes a classic silo-breaking assumption that leaders really need to revisit. When we talk about the challenges in the industry, I think letting go of models that have worked well for new technology is hard to do, but it's critical.

Steven Olsen, associate director, global development and informatics at Bristol-Myers Squibb

CHIN: Within, at least GSK, we've made the decision to really drive EDC into our business and we've seen significant benefits: We've increased our data management capacity greater than 50 percent; we've seen queries and cycle times drop; and we've seen last patient/last visits database. Our battle lines now are drawn on integrating our clinical data, meaning all of our clinical data, not just the eCRF pieces, labs, X-rays, PET scans, and pharmacogenomic data. Query turnaround time is very fast and costs have come down. It's truly beginning to drive a very different way of looking at the data and making inferences of that data.

There is a major issue of cost versus benefits, because this is a process where the real benefits are downstream. In a business that values immediate return on investment, what are some of the improvements that outweigh the initial cost of implementing EDC?

TIEDE: We've tracked this within Centocor [J&J's research and clinical development organization] and we literally saw a ten-fold reduction of queries. We saw a three-fold reduction in the time to resolve queries, and more than a 60-percent reduction of database lock. Over the time we were tracking things, we increased our capacity and estimated that we avoided hiring 28 people over four years.

James Tiede, vice president and head, integrated data services, global clinical operations at Johnson & Johnson.

OLSEN: You're deploying a new technology so you have to put a help desk in place, as well as provide different training for the help staff and investigator staff. At least in my organization, folks are saying, "We didn't realize it was going to cost us that much to do this type of thing." But for us at BMS, there are fewer discrepancies within the tool that we actually have to manage. This is a big benefit. The organization has become much more efficient, as well. The teams are now becoming 20 to 40 percent more efficient.

What stands in the way of EDC becoming the industry standard for all clinical trials?

TIEDE: A lot of it is a management issue. EDC is redefining how companies do business and a lot of us are deeply embedded in legacy systems. The existing processing systems are fine and everybody's comfortable with them, so changing it and trying to deal with that risk is a challenge.

CHIN: I agree. You need to have the senior management commitment of vision of where you want to take the business. At GSK, the hammer went down from up top, and it took that commitment and vision to be able to really drive it into our business.

VAILLANT: The other reason pharma really hasn't adopted EDC as quickly as other industries is that there's really no burning business need. Paper works fine and our business isn't going to fold because we need to transition.

BROWN STAFFORD: I do think one of the barriers to success has been the slow pickup in terms of the clinical monitors, because since they've been around, they've done the same job.

OLSEN: A challenge for us as an industry deals with the investigators: We all use the same ones in similar trials, so what technology each of us exposes to the investigator staff looks very different. They see one from GSK, one from BMS, one from Quintiles, and they wonder why the processes are different. They're the ultimate customers here, and it's a struggle in some cases.

Johnson & Johnson's James Tiede says getting everyone on board for EDC implementation is a challenge because the existing systems are acceptable.

How do investigators feel about EDC?

BROWN STAFFORD: For the most part, they've bought into EDC. Most of the investigators that we interface with say, once they've used an EDC product, they don't want to go back to paper. The issue now is consolidation in terms of the technology because the investigators are clear, at least from what we've seen, that they're ready for EDC.

TYSON: Four years ago when I was talking to people about EDC, the attitude I would sometimes hear was, "Well, maybe it's just a fad." Now when you talk to people about EDC, they say, "We're not there yet, but we're piloting." That's an important sea change driven by the investigators, since they've started to embrace the technology.

CHIN: It gets to the point of just trying to enhance the user experience. There's technology overload at some of the sites so there's a lot of anxiety out there. The question is: How do we make it better for them?

You have all implemented EDC, and we heard what challenges you've faced in the process. What are some things you did well, and what are the things you wished you did differently?

TIEDE: Implementing is tough work. It really requires a lot of focus, dedication, commitment from the top of the organization, and then a lot of follow-through. For us, it was a real cross-functional and multidisciplinary effort, where we spent over a year talking about the process and roles before we even began our first trial. We put the thought in it and it really turned out to our benefit. For example, when our first vendor went bankrupt, we were able to change vendors. We didn't miss a single milestone with regard to new studies, and I'm convinced it was because we had a focused definition of roles and responsibility from the beginning.

BROWN STAFFORD: We did the very same thing. Something else that helped was the person I put in charge of the cross-functional implementation team. I didn't want someone from data, from clinical, and I didn't want a project manager. I chose a biostatistician because I wanted an unbiased, third-party opinion within the company that could lead this implementation effort and not have the baggage of "this is the way we've always done it." It helped us be very successful in terms of setting our processes, training, pricing, and business development.

Paula Brown Stafford, from Quintiles, says once investigators use an EDC product, they don't want to go back to paper.

CHIN: GSK took a very aggressive approach to driving EDC into the business, with senior management well behind us. They gave the team a very aggressive target, which was to have at least 80 percent of our new clinical trials [using EDC] by the end of 2005. And here it was, mid-2004—we had 18 months to look at the entire organization. But that vision and commitment led to a very focused purpose and we drove this thing unbelievably hard. We did manage to deliver, not only 80 percent, but 90 percent of our studies by the end of 2005 that were across Phase I to Phase IV.

TYSON: In the end, aggressive deadlines focus resources, and that's why we're in favor of them. And like anything else in life, if you have an aggressive deadline, you can pull yourself up in the priority list and you can get the resources that you need to get it done. Without deadlines for organizational adoption, you get into a pilot-forever mode and you never see any benefit from EDC because no one really learns to work differently.

TIEDE: I didn't use the word pilot when we rolled out because to me, pilot doesn't give the sense of commitment. For us, it was deployment. I wanted to tell the organization that, in fact, we were making a change and we were not going to go back if the first one didn't work. This was the first step of a continuum.

You would think the biggest danger in the implementation is that you would go halfway—you replace the paper, but wouldn't fix or rebuild processes. How do you keep that from happening?

TIEDE: I can't tell you the number of times we've talked to organizations that have gotten the ball within ten yards of the goal, but for whatever reason couldn't get it over the goal line because they lost their focus on the training. It requires as much energy and as much focus as the "cool" re-engineering stuff, and it's where some organizations really struggle.

CHIN: Training really has evolved from an old world, where you put a factory worker on an assembly line and you train them to do these things. But these are tacit knowledge workers, so what we actually have to do is re-skill. Training is the onus of the company and re-skilling is our onus. These are the skills that allow you to be able to survive and really thrive in this 21st century. We also have to train investigators in ePRO [electronic patient reported outcomes] and all the other technology out there. [Investigators] are the part and parcel of driving this EDC continuum.

OLSEN: I have a specific example: We have a large EDC study that has high visibility in the organization. We had an investigator complain to us that our EDC tool wasn't as good as somebody else's and we weren't ready for ePRO yet. This guy had an e-tool that he had to learn and be trained on, yet he still had to use his black pen and a piece of paper. We felt a little bit schizophrenic about how we presented ourselves to the investigator. Just one of our studies requires six to nine hours of training time, so we're really putting a burden on our investigators—you can see why training them properly is important. They're the ultimate user.

TIEDE: One of the things we need to look at is, once we've done our first and second trial, who do we need to go back and certify? We're trying to build an internal database of those who have been through the experience, so we don't have to give them all the same training every time we do a trial, especially on the technology side.

VAILLANT: The success or failure of EDC is not going to be dictated by the technology. It's going to be dictated by the support infrastructure you put behind the technology, including training. I think CRAs [clinical research associates] really shift their focus from being data cleaners and data checkers to being relationship managers. That's really where their time is better spent. We're changing from an organization of viewers to an organization of managers. We're no longer doing the work. We need to develop a skill set to manage the work more effectively, which means having more project management skills.

You're talking about redefining roles. Can you talk a bit more about this idea of roles and how they can change? What about the people who have to deal with remote locations, like the monitors?

TYSON: Monitors can have 100 percent visibility of the data, so when they show up, everything's ready to roll—we see what a significant savings that is. That's a huge power to give to the monitors, and it's totally different from the past when they had to do the best they could with what they had. Now their job is an ongoing site relationship and data management role. The day-to-day interaction is a whole new role for them.

Tyson, says managers must be patient if they want to see the benefits of EDC.

The needs of the medical monitors, however, aren't the same in every company. It depends on the therapeutic area and specialty. So it points out a new core capability that clinical IT needs to have. This is a challenge because not all organizations have clinical IT, and not all IT organizations specialize in clinical trials, yet it's what clinical needs.

BROWN STAFFORD: What we are seeing is that the monitors still are going to the sites almost as frequently because they have to monitor the safety of patients and make sure investigators are following good clinical practice. But the difference is the time on-site isn't spent in query resolution and cleaning data, because that's happening before they go to the site through the data managers.

TIEDE: Yes, the time on-site is much more productive and focused on the relationship between the monitor and the investigator—time spent data cleaning has decreased significantly. It's all accessible online, so they can see what the status is, where the issues are, if people are getting behind. To the extent one builds as much front-end loading as you can, what we've seen is up to a ten-fold reduction in queries. That's where the greater efficiency comes in. They want to be able to look at all the trial and patient information on a dashboard, which to me is great. I love that people are beginning to think that way.

BROWN STAFFORD: We're changing the way we train our monitors. We ran our first trial in 2000 and now we've just put into production our 200th EDC trial. We have an enormous amount of experience, but through that phase for clinical monitors, EDC was still an initiative. They came into Quintiles and wanted to receive the general CRA training within Quintiles, which really had no EDC component. It was only after you were allocated to an EDC study that you were trained. So what we did was put an e-CRA certification process together. This has changed recently because it's not an initiative anymore. It's mainstream, so we have to certify CRAs to be e-ready and not wait until they're assigned to a new study.

There's the axiom, "If it ain't broke, don't fix it." I think some of our readers, from pharma execs to investigators, need to understand why role changes matter. How do you convince them that EDC is the next wave of the future when the paper process has worked for them in the past?

TIEDE: Sometimes you have to put [in] lots of effort to convince the skeptics early on, but once they see the benefit, they become the strongest advocates. That's what helped Centocor move from doing everything in paper to essentially putting everything out with EDC in three years.

Steve Olsen

Building success and being able to relate it as a story to the new teams helps get others on board. People get concerned when out of the 40-odd trials we've done, we've only made nine submissions. I don't want to do this on a submission study because there's too much at stake, so I build a story on the fact that the nine submissions we have made have all been successful. Early on, you really have to be out there advocating and really trying to convince the people we have our ducks in order, and that we're going to follow through on our commitments and make it a success.

OLSEN: Last December, we had a global team over with our clinical operations group for the first time in several years. We had to work hand-in-glove with them to get it going, but once the project manager stood up in a meeting of 600 and said, "I just entered information for first patient/first visits and I already have data in my database," it was fabulous. That's the reward that you get. At a meeting where a peer said it was worth the pain? You can't buy that publicity.

In the end, we wound up with an internal government structure that we've been able to tap into. It's not just the EDC team that needs this. The entire organization needs it. All of a sudden, the light went on that indicated we were serious about this—it's now baked into the process and allowed us to move to the next level. We've progressed 25 to 75 percent this year, and teams are coming to us because they want to be part of this. Last year it was pulling teeth just to get the early adopters on board.

It seems that what you're trying to do is reinvent a lot of existing pieces by leveraging the technology and bringing the knowledge work to the forefront. In the long run, what parts do you think will benefit your company?

VAILLANT: One of the things I experienced is that implementing EDC forces the organization to look at protocol development very carefully and the actual data you're collecting. In the paper world, it's okay to collect everything "just in case." And we hear that all the time: "Well, we want the data just in case." In EDC, it's a little more difficult to make that just-in-case scenario work. You need to be very, very careful about what data you collect. You need to be very discreet.

TIEDE: One thing we did post-implementation that really helped was to create an e-clerical steering committee. We had people from the therapeutic area, data management, and trial management. This was kind of a standing group that served as a clearinghouse for issues, and also has been used to promote further development.

CHIN: When it really comes down to building a clinical trial, it's quality by design. Now people are being forced to think up front, design up front, invest up front. So you're front-end loading the entire process. We have to hit this timeline, and push a lot of activity not only in the trial design, but also in training our staff and supplying the sites.

TIEDE: I've actually challenged my group to say, "As for deploying EDC, if we have a standard on-the-line data architecture, based on what the investigator does on the screen, does it always have to be the same? Or can we meet the customer needs by customizing the screen for them so that they can get at the information in a way that's friendly and useful for them?" In the future, we have the potential to make this investigator-friendly. It's going to be a challenge because it puts a burden on our shoulders, but at the end of the day, it can well increase the acceptance and the overall use in the investigator community.

The phases of clnical trials are blurring. For example, certainly at FDA, Phase II is sort of splitting into IIa and IIb. What is it going to take to get us to the point where you can use EDC to have visibility across the entire development process?

OLSEN: I spent a lot of time in our Phase I team. I actually was leading the charge to bring EDC into that realm. At BMS, we're using the same EDC tools across Phase I to Phase IV. It was the data visualization that was the biggest challenge. People were used to looking at the data in a paper format, so when we came in with EDC, they didn't know what to do with it. Once we got into it and brought the data in near real time and were able to block a database at the speed of light, they finally said, "Wow, this really works."

Still, although we're progressing, some people are still skeptical because they still haven't seen what the results might be at the end to become believers. But there are so many experiments, you can show the types of powers that EDC can do. That's one of the hats I've been trying to wear in that space.

EDC is really good at stopping errors from entering the system. So why are there still delays in getting results?

OLSEN: What we have found, in some cases, it's another dataset that is now on a critical path. It's not the eCRF dataset. You've been able to lock that, but you're still waiting for some loaded data that's coming from an analytical laboratory. You've exposed another weakness in that activity that went on to get to data log.

CHIN: It still comes back to integrating all the relevant data, you know, at the end. It's not just the patient data.

BROWN STAFFORD: Well, because we're waiting to integrate the data at the end, that's when we're doing some of the cleanup. That's what takes time, because we locked an eCRF database in nine days after last patient/last visits. And then we had to sit around a few days, because we ended up having to clean up some other database and then put them together and do a merge. You find a few errors here or there, things that didn't match. So even though you had the "wow" out of the EDC, you didn't have the overall impact you wanted because of the other things we found.

Clearly there are still many challenges as you continue to implement EDC in your companies. There's a challenge for you as leaders to help your team make a collaborative effort toward your end goal, and there's a challenge to tap into that enthusiasm. Can you share some lessons with other companies to help them move these initiatives forward? That's so necessary in encouraging change.

VAILLANT: If history is any predictor of the future, three to five years from now we'll be very happy if we're successful in our role for EDC. Looking idealistically and more optimistically in the future, the different integration of data sources is key. Moving forward with that integration through data warehousing and some other mechanism is going to be important, and that will be the next step forward.

CHIN: This is a business imperative to the pharmaceutical industry. We're under tremendous cost pressure, litigation, and safety, so it's not a matter of if we do it. It is a question of when we do it and how we do it. The time is here and it will take time, but people have to be able to get through this steep transition. There will be other issues to face, but it will be a little more clear sailing by then if we can enhance the user experience.

One thing to look at is the four pillars: people, process, technology, and support and services. Guess what? You can't work at them in series. You have to work with them concurrently. You have to balance all of these pieces and that's the challenge for an organization if you work on aggressive timelines.

BROWN STAFFORD: We need to see that EDC is not a data management position. It is mainstream, and it is enterprise wide. I still don't feel that, across the industry, there's complete buy-in, but it does impact the protocol design and clinical monitoring. I think it is also about clinical informatics in the end, and it is about patient data not just within a study, but a cross study. I don't know if it will end up being pharma companies doing that or not, but it will likely be a partnership between sponsors and CROs.

CHIN: In the next three to five years, I see us trying to truly manage the continuum from the time we do pre-clinical throughout the life cycle. The question will be: How do you leverage large databases, epidemiology, and genomics? We're all having issues in increasing patient recruitment, but with EDC, we should be able to tap into these large databases we've created. We have to be able to look at our safety and integrate it further upstream and wonder how to change things.

TIEDE: One of the things I observed is what I think the deployment of EDC is: It opened people's minds to looking at alternative ways of doing things. We've had some success and we started with this one area, and it's worked, but now, where else can we go? What else can we do? I think it's going to be related-type things branching out from there. That's the direction innovative work is going to take. The informatics and the large data warehouses are going to be really important.

We're in the process of conducting a trial in asthma, and we wanted to collect daily spirometry and diaries of patients, so we found a device that's a combination spirometry and diary. That's a whole new way of doing things now. I think there's so much going on, we're probably not even aware of. Now that we've deployed EDC, it says we can do other things since we've been able to demonstrate the success.

To me, EDC is not a panacea. It is not going to solve everything. Early on people felt we were going to save millions and millions of dollars by cutting out so much time. EDC is a cost-effective tool, but it is not going to cut months out of our development time. Recruitment will still be an issue. But there are still huge significant values that EDC brings in terms of the way we work, the efficiencies that we can gain, accessibility, and so on. We also have to keep expectations realistic in terms of what this tool can and will do for us, and what it can't do for us.

TYSON: This really represents a very sort of interesting inflection point for clinical organizations because it's this whole idea that the way we used to do things is mutable, it's changeable. There is new technology and we need to start thinking about what we do, not in terms of our jobs, but in terms of what the ultimate end goal is to get this data. This legacy being left behind by EDC is this concept and this experience set of all these terrific people understanding the common embrace of technology and using it to gain benefits. It's next generational thinking.