Info & Opinion
February 19, 2019
With news about big data, Aetna, Covance, GNS, Scrip, Quintiles, PPD, Icon, BioClinica, Merge, Medidata and GSK.
With news about FDA, CTMS, PMG Research, Inclinix, EMA, Hemofarm, Parexel and the Korea Drug Development Fund
David Underwood of Quanticate says some firms are giving short shrift to the basics of clinical trials
With news about Roche, Quintiles, Allscripts, Janssen, SGS, Oracle, TriReme, OpenClinica and FDA
What are appropriate turnaround times as images pass through the different stages of a clinical trial? Is the pharmaceutical industry applying outmoded thinking?
Here's the context for such questions. A few weeks back, we ran an article on benchmarks in imaging. In brief, Perceptive Informatics, a division of Parexel, had adopted some standards in consultation with a nonprofit organization called the Metrics Champions Consortium. It all seemed noncontroversial.
Then Jon DeVries, general manager of the eclinical business at Merge, got in touch. He competes with Perceptive and suggested, in a gentle and polite manner, that the ClinPage editorial pencil may not have been at its sharpest when that article was written. Specifically, DeVries said, the clinical trial world can ask more of imaging core labs in 2010. His view? Digital systems to move and manage images offer such big savings in time and money that the MCC benchmarks may already be obsolete.
Our previous article had included a chart of a few metrics. Some call for no more than 20 or 30 days to elapse when acquiring images or notifying research partners. "Perceptive is measuring those in days," says DeVries. "A contract research organization (CRO) or core lab can use our platform to take [those metrics] from days to minutes. All of those things get done automatically."
As many readers will recall, the Milwaukee, Wisconsin based Merge acquired the software of etrials. Apart from that transaction, Merge had built up its own body of intellectual property around data integration and imaging in the health care industry. Merge's imaging technology is used in 1,500 hospitals and 2,200 imaging centers.
Setting The Bar Too Low?
Within an hour of their creation, images in Merge's system for clinical trials are available to the next person who needs to see them—whether that person is working at a CRO, a site, or a sponsor.
DeVries says that any viable company in the broader teleradiology landscape must be able to turn around images in less than 24 hours. One top firm, Nighthawk, has a 20-minute turn around time. We don't know the comparable numbers for other leaders in the imaging sector of clinical trials, companies such as Bioclinica, Synarc and Radpharm.
But here are a few questions all sponsors can ask of imaging core labs and technology specialists:
1. Are all of your clinical trial images viewable on the internet?
FedExing a physical CD with research images will always involve a certain amount of delay in re-entering metadata about that CD in one or two additional systems. Perhaps two thirds or three quarters of modern images in clinical trials, DeVries estimates, are still transported using overnight couriers. But Merge and other clinical trial providers are working with AG Mednet to transport images electronically—an art unto itself. (Here's a previous ClinPage article on AG Mednet.)
Additional time may be lost if it is necessary to de-identify patient data from the images or transmit the massive pictures. And of course there are significant costs to using FedEx, UPS or other couriers to move packages all over the globe.
Finally, DeVries notes that medical devices firms are especially interested in allowing their research and development experts to look at how a particular stent or artificial joint is fitting inside an individual patient. All allied participants can have easy online access to all images.
2. Do the physicians working on the images automatically know what they have to do?
Some radiologists working for industry wait a day or a week for enough images to accumulate. Some person may need to manually compile a list of what they need to do. That eats up time.
The Merge system, according to DeVries, automatically drops imaging exams into Merge's web-based queue so that each radiologist knows what needs to be done. That lets a radiologist handle 4 or 5 images at the end of a busy day and not wait for more cases to arrive. "As soon as the exam comes into our system, it automatically comes onto that radiologist's work list," says DeVries, "instantaneously."
3. Does your imaging system scale? Can it allow radiologists to have the same productivity in trial-related practice as they enjoy in the rest of their work?
When its activity in the larger health care arena is considered, Merge's system handles approximately 20 billion images annually as part of some 90-100 million patient exams. "Some of the core labs are managing a couple hundred thousand images a year," notes DeVries.
For a radiologist who needs a full hour to fully complete a read for industry, the clinical trial workloads and fees seem unappealing. But if the right technology allows a radiologist to read just as efficiently for industry as in community practice, one barrier to working on clinical trials could fall.
4. Does your system reduce queries from the start?
By forcing some fields of data to be entered into case report forms initially, DeVries says, Merge has been able to reduce the number of imaging-related queries. As one example, DeVries sites the system's ability to automatically convert linear radiological measurements to the RECIST criteria in oncology. Fewer errors at the start, he says, mean fewer to resolve.
5. What is the structure of your imaging database?
The company's proprietary workflow and database structure, Merge says, enable it to build and launch imaging trials especially quickly. "We've cut the startup time for these trials by a factor of 3, if not a factor of 6," he says.
One aspect to the company's database helps reduce the general migraine potential for data managers working on oncology trials. The company's system can easily record the radiologist comments on specific lesions from one time point to the next, DeVries says.
6. In an end to end fashion, can users rely on one 21 CFR Part 11 compliant system to do all their imaging-related work in the trial?
Some imaging core labs and CROs require their radiologists to use multiple databases and systems. If the radiologist is working on multiple projects, this can lead to a certain level of chaos and confusion as the physician bounces between different systems, passwords and user interfaces.
The Merge system handles all of the following: image import; de-identification of patient data; image submission; short- and long-term storage; viewers and image analysis tools; quality control, review and adjudication; outcomes analysis and reporting; and workflow customization for individual studies.
7. Does your core lab autopopulate the case report form with the measurements from the radiologist?
"Many core labs have their own process, separate from the EDC process," says DeVries. "They have to manually double-data enter it or send it to the CRO. There is kind of a two-path process." That creates uncertainty about where the image is along the clinical trial workflow. Thus is born one set of inefficiencies for sponsors, and another for radiologists.
On the latter score, DeVries was kind enough to forward an email from an anonymous, frustrated radiologist. The physician was describing the Rube-Goldberg-worthy route by which images were logged and reviewed in his trial: "...the vendor in turn sends me an alert when 3-D reconstruction is complete, I then must alert a group of radiologists who had to be trained on the vendor's imaging system, and in turn on our vendor database so that they could enter the data onto a CRF, then they have to take a screen shot of the measurements to provide back to the sponsor, very tedious and time consuming .... If there are endoleaks or migrations, I have to then get the imaging to my CEC [clinical events committee] reviewers. This has entailed having the vendor send me CDs in triplicate, because the imaging data is so large, only one scan will fit per CD, so if there are multiple imaging modalities (KUB, CT and Angiograms) I could have 5 or 6 CDs per subject ..."
Is it any wonder some radiologists don't want to work with the pharmaceutical industry?
Except, according to DeVries, that could be changing. He cites the recent selection of his technology by Intrinsic Imaging as a sign that some U.S. radiology groups sense new opportunities in industry research—if the hassles of managing the work flows and IT challenges can be loaded on to someone else's shoulders.
"We're going to provide a better platform for radiologists so they will participate in more trials," he vows. Some core labs may focus on their clinical insights, and leave the IT to companies like Merge. Says DeVries: "The core lab model worked very well when it was created 15-20 years ago and film was mailed around. The world has moved on."
Indeed, he adds, the core lab itself may not have a placid future, especially if radiologists would prefer to work with sponsors and CROs directly. Says DeVries: "Our platform is allowing the radiology groups to leapfrog the core labs and go directly to the sponsor and CRO."