By Mark D. Uehling
November 15, 2003 | It's an open secret. In the 1990s, wags say, Merck sank $100 million into developing software to collect and store clinical trial data. On the back end, the project’s data management system is used to this day. But Merck’s homegrown, pre-Internet electronic data capture (EDC) software to gather clinical data went nowhere. Now, Merck tells Bio-IT World that it is trying to put that fiasco behind it and hammer out which 21st-century technologies it will use to gather clinical data.
Rivals such as Lilly, Bayer, and Novartis have broad, aggressive, and effective EDC strategies, even if their solutions' cost and performance will never be subjected to third-party analysis. Strangely, however, Merck's hesitation about EDC could prove astute. Its strategy holds the promise of leapfrogging beyond EDC, anticipating a broader trend in physician offices. The catch: Merck's approach depends on a hard-to-synthesize injection of clarification from the FDA.
For Briggs W. Morrison, Merck's executive director of worldwide clinical data management operations, the crux of the matter is not, purely speaking, IT. Everything comes down to not getting in the way of doctors who can recruit patients. "Data management is way, way down our list of critical business problems," Morrison says. "We think of data management as more like the electric company -- it's something you have got to do. But it is not a critical business problem. The strategy is to partner with [principal investigators] and say, 'What works for you?'"
Letting Doctors Choose
To accommodate physicians, Merck will offer a menu with three dishes. The first offering: paper. "There are people in the world who still don't know what a computer is, even among American physicians, even among American study coordinators," Morrison says. "We don't want to put a computer in their offices and spend hours trying to get them to be computer literate."
The second dish on the menu: EDC. Merck is now evaluating vendors, Morrison says, providing no specifics. The winner might not be chosen until October 2004. Eventually, all clinical sites doing research for Merck will be obliged to use that solution. "We're trying to find a vendor that has a simple, scalable, intuitive tool," Morrison says.
His tea leaves are difficult to read. But they would seem to hint that Merck is less interested in larger, more established companies (such as Phase Forward and Oracle Clinical) and more interested in newer, nimbler, Web-based upstarts (such as Medidata and LifeTree). Another tiny newcomer, DataLabs, which will be handling Phase IV trials for Merck, would also appear to meet Merck's criteria. DataLabs says only a few vendors are in the running.
What if a particular investigator requested an EDC product not blessed by Merck? Would he or she be allowed to use it? No, Morrison says. Merck hopes to select an EDC platform that, by sheer ease of use, presents no learning curve for physicians or the clinical staff who do most of the work. "This [software] should be so simple that if you used those other ones, you really shouldn't have any problem with this one," Morrison says.
But Morrison is most excited by the third dish on Merck's technological menu. The company hopes to siphon trial data from electronic medical record (EMR) systems, now blossoming in the most technologically savvy hospitals and physician offices. Simplicity is part of the appeal. Why force doctors to learn a new system for clinical trials when they are already grappling with other software on a daily basis -- and that software already sends its data into Oracle?
Feed the Database with EMR?
Says Morrison of the EMR systems: "The structure of that database doesn't look a whole lot different than the Clinical Data Interchange Standards Consortium (CDISC) models. Demographics are demographics. Medicines are medicines. Doses are doses. Can't we just take it out of this computer system and move it into our database for analysis?" In fact, he already has. In a large Scandinavian trial completed a few years back, Morrison says, doctors working for Merck used an EMR system to gather clinical data about an unidentified cardiovascular drug. The FDA reviewed those data; the drug is on the market now.
The catch is that the Scandinavian sites did not e-mail or electronically transmit their data to the idyllic rural campus that is Merck's U.S. headquarters. The participating Nordic doctors printed out their EMR files, and (after they were transcribed onto worksheets) Merck entered the worksheet data into an Oracle database. Why were such contortions necessary? Because the Scandinavian clinical sites were using an EMR system that was not in compliance with 21 CFR Part 11.
Despite Merck's heft (annual sales: $52 billion), Morrison says EMR vendors feel no need to offer adherence to 21 CFR Part 11. Just 1 percent of all physician offices are engaged in clinical research. A pragmatist, Morrison does not believe he or the pharmaceutical industry can persuade EMR vendors to change.
What would be more helpful, he says, is for the Department of Health and Human Services (HHS), or its most prominent fiefdom, the FDA, to clarify how strictly 21 CFR Part 11 will be enforced at clinical sites. "If I go to Dr. Smith, and he uses EMR, and his EMR is not CFR Part 11 compliant, if you strictly interpret the FDA rules, I cannot use Dr. Smith as an investigator," Morrison says. Would other clinical trials using EMR data be as warmly received at the FDA as Merck's Scandinavian trial? "It is an open question," Morrison says.
Ever scientific, Merck is now surveying its clinical sites to research how many use EMRs. "What would simplify everybody's life is some guidance from the agency," Morrison says. "In the case of an EMR that is a source document, are they acceptable data or not? They had better not say no, because they'd shut down a lot of sites."
Such a ruling, Morrison predicts, could affect every company doing clinical trials. Patients would be even more difficult to recruit. And one stated goal of HHS -- to speed the adoption of EMRs -- would be undermined. Morrison wryly notes that HHS has pushed the Snowmed medical terminology at the same time that the FDA has endorsed another vocabulary, MeDRA. Comparable bureaucratic squabbles about IT systems for trial research and physician offices, he suggests, could impede U.S. drug research.
Beyond that, Merck's key insight seems to be that EMRs are always going to be a much stronger wind blowing across the medical landscape than EDC. Thus, it could be prudent for the company to invest in EMR adoption, and piggyback its trial technology on top of that trend. For other companies in the clinical trials tools market, Merck's plan raises questions about how quickly to absorb EMR data -- or to bet that the FDA will be unwilling to budge in the enforcement of 21 CFR Part 11 at clinical sites.