If clinical data moves along an assembly line, it’s a lurching and circuitous one, with stops and interruptions along the way. The final product of that assembly line is almost always a SAS file. Many in the industry are wondering which stations on the assembly line are eternal, obligatory—and which may supplanted by new technology.
The people at Majaro Infosystems have a different way of looking at the situation. Why not leave the data in SAS, period? The company is a contract research organization with a heavier emphasis on technology than most. Founder and president Martin Rosenberg is a biostatistician by training, having gotten his start in the industry in the mid-1980s. In 1987, he started his company, which is based south of San Francisco. He can see Intel’s headquarters from his office and wax poetically about demigods like Jim Goodnight or Steve Jobs.
Back in the day, he reports, computers were curiosities in the industry and at FDA. He participated in the first computer-assisted new drug application (CANDA) for Syntex. Procter & Gamble was one of his early customers, using Majaro’s ClinAccess application for its Olestra submission. The latest version of the Majaro software allows customers to take advantage of built-in business intelligence functionality in SAS.
“When I got into this, no one submitted computer files to FDA,” says Rosenberg. That’s changing, especially in a drug-safety landscape that includes less trusting regulatory bodies all over the world. He adds: “The actual data we collect are becoming a more important part of what gets submitted to FDA and other agencies.” The company believes that leaving the data in SAS can speed all submissions and e-submissions in particular.
Rosenberg says the advantage of using his company’s approach is simple: speed. It’s not necessary to move the data from System A to System B and, finally, into SAS. “We cut those steps out,” he says flatly.
“When you’re done,” says Rosenberg, “and the last patient is entered and clean, you have your files that are ready to analyze. [With other systems] it sounds like it should be a simple task to move data from one system to another. For practical purposes, it is much harder than it seems. We don’t see SAS going away.”
Out of the box, ClinAccess links images of case report forms (CRFs) to the underlying trial data. That helped Majaro support one of the largest medical device trials, with 680,000 pages of case report form data for about 90,000 patients. The pages were manually scanned into the system.
Keeping the data in SAS, he says, is especially appealing to smaller companies without vast IT or data management staffs. “We’ve gone into real startups and successfully implemented the product,” says Rosenberg.
Role of CDMS
ClinAccess does something that many in the industry have been talking about —merging systems that collect and manage data, electronic data capture and clinical data management. “Clinical data management systems are not going to disappear. They will evolve,” he says. “We are still going to collect data. We are still going to clean it. You are going to need database management capability to do that. It is not just about keystroking the data.”
When it comes to using and deploying computers, Rosenberg agrees that the life sciences may not be in the same league as other industries such as financial services. But what research-driven companies do is fundamentally more complex, he says, more difficult than keeping track of whether someone bought 100 shares of stock at a certain price. “That’s how most corporations use databases,” he says. “Data that we collect never becomes obsolete. As long as that drug is on the market, you need access to that data.”
In some respects, ClinAccess has elements of an electronic data capture system. Data can be keyed directly into ClinAccess. But Rosenberg feels that despite all the attention, EDC remains in the early phases of a promising growth curve. Most trials still use paper. “We are not anti-EDC,” says Rosenberg. “I just don’t think it is a panacea.”
The current promotion of “hybrid” EDC systems is something that Rosenberg finds fascinating. Says Rosenberg: “It is interesting that the fact that these guys are promoting paper-electronic hybrids. This is a recognition of the short term reality. Short term could be years. The early predictions that EDC would replace everything that came before it haven’t happened.”
Thoughts On CDISC
Rosenberg supports automation, computers and the data standards that may come with them. But he sees the laudable CDISC standards as better suited for moving data than storing or analyzing it. “My crystal ball isn’t necessarily perfect,” he says. “But I view CDISC as positive for SAS and positive for us.”
“The industry is finally talking about standardizing things that could be standardized without restricting the ability to collect quality scientific information,” says Rosenberg. “It opens up a lot of possibilities we’re really excited about. As these standards evolve and are proven, it opens up the door to writing software to analyze those things that are common to what everyone in the industry does.”
But he does not think that the industry’s preference for SAS will wane in the coming era of standards.
“You can’t analyze your data in Oracle. You can analyze it in SAS. SAS has a Proc CDISC,” he notes. “They also have a SAS CDISC data engine. You can move from SAS to XML and from XML to SAS data. It’s bidirectional. We think we’re ideally positioned.”d9A2t49mkex