What if patients were easy to find? What if you knew all their concommitant meds and it didn’t really matter? What if you could generate a quick-and-dirty answer to the number of people taking medicine X and having adverse event Y?

Those were the extremely juicy questions at the heart of Hugo Stephenson’s talk at the recent Society for Clinical Data Management Leadership Forum. The meeting this March in Raleigh, North Carolina, had a number of great speakers, and Stephenson was one of them.

It’s not like anyone needed to hear it, but Stephenson started out with an exceedingly blunt summation of the industry’s current predicament. Said Stephenson: “I don’t believe pharma has ever been in such a dire—I hesitate to use ‘dire,’ it means no way out—under such high pressure as it is today.”

Our Situation

In addition to ongoing operational pressures and the flameout of genomics as something that could save the industry, Stephenson said, the pooling of data by academia is a new threat. “Professor XYZ can publish data in the New England Journal of Medicine and ruin a product,” says Stephenson. “You need to be ahead of that.”

In his former duties at a contract research organization (CRO), he’s seen sponsors wrestle with difficult access to paper-based data.

Stephenson said that when the safety or efficacy of some drugs has been challenged in academia, the reaction of some data managers may be “ ’Oh my goodness, where do I have that data? If you’re sitting there saying, ‘I don’t know where that box of data is, let alone whether that data is integral,’ you’re in a difficult situation. I remember one sponsor saying, ‘I don’t even know where the study for that study site is. Let alone coming up with a short, sharp media response.’ It leaves the chief medical officer standing up there looking a little foolish.”

Hopeless Quest

As a CRO executive in the post-marketing phase of works, Stephenson said he routinely encountered minimal investigator interest, glacial patient recruitment, and massive investigator attrition as the years dragged on. “The level of interest falls like a rock,” he said. “In my view, it does not make sense for 300 sites in the U.S. each dealing with 8 patients.”

Gently mocking much of the post-approval universe, Stephenson said there are intractable structural problems with how the industry attempts to locate both investigators and patients.

His opinion? Many investigator-initiated projects can be scientifically dubious. “I was shocked at the number of grants that were doing methodically unsound proposals.  The challenge we face in the U.S. is that the people we can trust as investigators aren’t the ones who are seeing the bulk of patients,” he notes. “And the people who are seeing the bulk of the patients are not people I would trust to be investigators.”

Internet Tools

So what’s he offering? A MySpace for patients. It’s called iGuard. Owned by Quintiles, it records the medications used by each user, and then sends such people individualized safety bulletins. iGuard also serves as a conduit for people doing epidemiological and other research projects. ClinPage wrote about the Quintiles iGuard effort here.

Stephenson repeatedly acknowledged that there is plenty of statistical bias in the approach, since tech-savvy patients are predisposed to enroll on the iGuard website. Also, the site is designed to assist chronic conditions, not acute ones. “This is not FDA-submittable data,” he says. “The system is not 21 CFR Part 11 compliant. It won’t be. Neither should it be. That is not the objective. The aim was to have a pool of patients that we could target, that was very comfortable providing data into our system.”

Impressively, in its first year, iGuard has signed up 100,000 people. Perhaps a quarter have hypertension.

That patient-centric discussion is interesting and granular about which medications work and which don’t. Our jaw dropped when Stephenson, a physician who suffers from hypertension, revealed that he has changed his own hypertension regimen based on what his fellow patients are telling him on iGuard.

Working through the constraints of HIPAA, iGuard has the ability to contact patients through their personal physicians (who can also be tied into the iGuard system) and ask them to participate in post-approval research, or even come in for laboratory tests.

i-Epidemiology

What sort of people trust their health data to iGuard? All types, it appears. “We have a lot of people in their 60s and 70s,” says Stephenson. “We are missing out on younger women. They don’t think of their contraceptive pills as being a medication.”

iGuard can easily locate hundreds of thousand of patients who meet demographic or therapeutic area criteria. The site has an easy way to get people to provide additional information about their health: it pays them to answer some surveys. “The objective is to get data back very quickly,” he says.

Plenty of Bias

Indeed, the iGuard system surveys its members rigorously, making the system hard to “game” by someone intent, perhaps, on signing up phony patients to sing the praises or critique a particular drug. Among other things, even if one mischevious teenage boy signed up as 100 different individuals, it would be difficult for such a person to affect the results of an iGuard survey that only touched a fraction of everyone in the iGuard database.

Stephenson says there are instances in which the real-world experiences of patients in the iGuard database are in harmony with the scientific literature, and instances in which the iGuard data and medical journals are discrepant. Muscle pain and statins, for example, seem to occur in perhaps 10 percent of patients in the iGuard universe. That’s more than what the medical journals suggest.

Getting It Done

The system was built to preserve patient privacy, but not to withstand an FDA audit. But the sheer speed of making iGuard, Stephenson says, was enjoyable in the context of a career and an industry that does not always move technology initiatives forward efficiently. Says Stephenson: “We build something in six months. We don’t use Oracle. We use Linux. We use MySQL, which is free. We use a development team that looks more at home in a biker bar than a corporate headquarters.” He clearly had fun just making the system live.

When he shows iGuard to people in the industry, Stephenson reports, there is a split. “Marketing, in general, has hated it,” he says. “Clinical, in general, has loved it. That’s a challenge for us.”

One of the most surprising aspects of iGuard, for this correspondent, is the sheer number of people who apparently trust the system&mash;and are energetically participating in modest epidemiological research.

It will be interesting to see if other organizations can duplicate some of what iGuard has done, perhaps confining it to particular demographic populations (teenagers? Asian-Americans?) or therapeutic areas (diabetes?).