Now that half of the U.S. Congress consists of experts on drug safety, we thought it might be a good time to have a chat with Chan Russell. He’s the founder of Phase Forward’s safety division, Lincoln Technologies. Russell sold the company in August 2005 but stayed aboard as president of the unit.
You can read Russell’s full bio for yourself. But the gist of it is simple. Over the years, he and his team at Lincoln have been involved in more drug safety tools taken up by the FDA and industry than anyone else we can think of. Russell’s early years in the industry included stints at BBN and Belmont Research. Besides the current crop of Lincoln applications, Russell had a hand in the original development of Clintrial and Clintrace.
At the moment, Russell says, many of Phase Forward’s safety customers and electronic data capture (EDC) customers are considering the company’s safety products which, of course, can be integrated with other EDC solutions. But it sounds to us like that synergistic sales cycle for the two Phase Forward product lines may not be a rapid one.
With a few decades of perspective on the industry, Russell says the market for drug safety software has shifted in just the last five years. “The market has evolved,” Russell says. “It was a missionary sell. Companies took the point of view that the less they knew, the better. That was not an official reaction. That was the informal reaction.”
Earlier resistance was based on the anticipation of additional work generated by safety systems, or potential exposure to punishment from regulators thanks to something turned up by such a system. “There were concerns about wild goose chases,” he says.
Sea Change
Now, Russell says, more safety customers genuinely want to detect the signals as early as they can. He credits flagship Lincoln customer Glaxo SmithKline for helping to show the industry what the Lincoln tools can do. “GSK has been quite public in working with us,” says Russell. “They’re real leaders in genuinely wanting to find safety problems early.”
Is the current political fixation on drug safety a fleeting one? Russell concedes the news media may exaggerate the significance of isolated clusters of patients. But he doesn’t think interest in the topic will fade. Says Russell: “Drug safety is here to stay. It’s much more significant in drug development than it used to be. Safety is a real business problem. It’s worth some investment, some thinking and some process change to work more effectively with it.”
With the hypothetical value of drug safety software no longer debated, Russell says, “people have gotten beyond the ‘is it a good things to do stage,’ and into ‘how do you really roll it out into your pharmacovigilance group?’ That was another place where GSK took the lead.”
Hearing The Signal
Almost two years ago, as some readers will recall, GSK and Lincoln won an award for an Online Signal Management (OSM) system.
More recently, Lincoln updated one of its crown jewel applications, WebVDME. It’s a tool for visual data mining and quantitative analysis and prediction of drug safety signals. The software can be turned loose on proprietary data or public databases. It is used within the FDA, the company says.
The new version of WebVDME is 6.0. Russell is especially keen to note that the software now facilitates two ways to slice and dice drug safety data more finely. He cites two examples.
WebVDME
One is that this release allows the use of “custom terms.” In plainer language, that means that a large group of patients taking both a generic and a brand name version of a drug (or the same drug by two routes of administration) could be assessed separately. Customers, Russell says, “want to get as much focus as they can.” To quote the company’s news release on what the functionality entails: the custom term “can be created by collapsing terms, combining terms or applying complex logic to analyze the safety ramifications of specific prescribing behavior.”
Another technique: logistic regressions. Many ClinPage readers do logistic regressions on the weekend, just like crossword puzzles. But for those who don’t, it’s a Bayesian statistical technique that helps tease apart two related but distinct signals. Says Russell: “It does much better in unscrambling cases of polytherapy. If your drug is prescribed with another drug, but not always, and you want to see the effects of your drug and not the drug it is prescribed with, logistic regression is a good technique.”
So how are customers using the Lincoln software in general? Russell says he isn’t aware of any instances where something connected with newspaper headlines was debunked or confirmed with WebVDME or any other Lincoln application.
New Directions
But there does appear to be growing ability to tune out drug safety noise, and hear a signal, in much the same way that a faint radio station can be heard with a more powerful antennae. “There have been situations where a concern has gotten less acute or more acute,” Russell says. “There are clear cases where people have changed what they were focusing on.”
The company’s relationship with the FDA is ongoing, he says. It’s clearly an outgrowth of trust and more than one person at Lincoln interacting with more than one person at the agency. Russell can’t say much about what’s happening. “Some of the more creative people [at FDA] can do a huge amount with our tools. We have some activities with the FDA that are new capabilities,” he allows. “We have not figured out whether they are extensions of products or a new thing.”
The company is also exploring the analysis of medical claims and records data, Russell says. Such records “are of reasonable quality. Not of the quality of a clinical trial, but decent. Clinical trial data is astonishingly clean compared to health care data. But you can get a fair amount out of dirty data if you have a lot of it and the biases are not systematic.”
d9A2t49mkex

