What's missing here? How about a mention of the electronic medical record that the Office of the National Coordinator for Health Information Technology (ONCHIT) has decreed as a ten-year national imperative?
ProSanos Talks Safety Tools, Predicts New Controversy
Forget, for a moment, the media hyperbole and legal stakes surrounding drug safety. The related scientific paradox is even more discouraging. Profoundly so.
On the one hand, virtually every seasoned participant in or observer of clinical trials understands that such research will never predict many serious adverse events. These only become apparent once a drug is approved by the FDA and sold in every Walgreens in the country.
On the other hand, sound approaches to analyze a drug's safety profile after its commercial launch remain clinically and mathematically controversial – techniques of art as much as science. When the FDA proposed research to fix this, Congress emitted a bored sigh.
So pharma is stuck with immaculate data from clinical trials – results surprisingly free of noise and confounding factors like other medications. And yet after a decade or more, when a drug emerges from clinical trials and reaches national distribution, the data become filthy, ambiguous, and unsuitable for the same precise analysis that persuaded the FDA to approve the stuff in the first place.
Is there a way forward?
Jonathan Morris thinks so. He's the president, chairman, and CEO of ProSanos, a 30-employee firm out of La Jolla, Calif. He spoke at the Bio-IT World Expo this year, and we chatted with him recently. ProSanos specializes in the analysis of drug safety data. It is helping pharma and biotech companies listen for subtle signals in the data available after a drug has reached the market.
The services, data sources, and software that ProSanos provides are varied. The company may clean your data, design your case report forms, or blend data from your trials and the real world. The math and the IT underlying the company's judgments are cutting-edge, but the company doesn't get mired in arcane theoretical or statistical issues. In some cases, Morris says, ProSanos may have an answer to a drug-safety question in 48 hours.
Morris predicts another Vioxx is on the horizon. "Absolutely," he says. "It's coming." He notes cardiologist Eric Topol of the Cleveland Clinic was not the only person to detect the Vioxx safety signal in 2001, years before the lawyers began circling. "We saw that," Morris says.
His chief concern now? The industry is simply not as sophisticated as it should be in studying the links between drugs and adverse events that arise after a drug has been approved. "We have to be able to understand the associations, the relationships," he says, between drugs and problems allegedly caused by drugs.
... He's blunt about the predictive limits of many clinical trials. "Today, in 2005, the best-designed prospective Phase III studies will not pick up all of the potential safety problems that will occur when the product is being used," says Morris. "Once a product is launched, you have to be able to follow it over time. That's where the unexpected things begin to occur."
Of all the Phase IV studies that sponsors are obliged to conduct, Morris notes, just 27 percent are actually completed. "Many companies do not fulfill their obligations," he notes. "If only a quarter of all the committed studies are done, it's hard to say anyone is holding up the gauntlet for good behavior."
New thinking and new tools are needed. "Conventional statistics and conventional approaches – the way you approach Phase III data – cannot be applied to the association data," Morris argues. "How do you deal with the association between the drug and the event? That we don't have a handle on."
Some customers, he says, want to go deeper. "There is a need, once you integrate the trial data you have, to go to other collections and see if the things you're seeing are there as well." That could mean data from a pharmacy benefit manager. It could mean health insurance claim data, or even reports from emergency rooms.
There are no paradoxes. Only false assumptions and stubborn resistance to new thinking.
Indeed, new thinking and tools are needed. The current thinking in the pharmaceutical industry seems stuck in the "if you don't want to find a fever [and harm sales], don't take a temperature" psychology. Pharma industry progress in understanding, encouraging and exploiting the EMR as a source of adverse drug effects data seems to be close to zero.