Today’s news about Merrill Lynch and Lehman Brothers is interesting. Trillions of dollars in mortgage securities that seemed safe a year ago are radioactive. Wall Street thought the risk could be papered over with accounting gimmicks and legal jabbering. That worked for a while.

But the tricky thing about risk is that it has a certain scent—an odor. Once investors catch a strong whiff of it, they’re gone. Good luck finding a scented candle to mask that.

John Loucks is worried about a similar shift that could affect the life sciences. As the chief customer officer at Relsys, an Irvine, California drug safety software firm we wrote about a few weeks ago, Loucks thinks that over the long term, there could be new dynamics.

“There will be a day in the future where life sciences companies will need to set aside what has been known as warranty reserves,” Loucks says. “You can also call this risk reserves. Financial people (in the sponsor community) are going to want to know, what is their exposure?”

New Bucket

Biotech and drug companies that can calculate the risk in a product or portfolio, he believes, may need to set less money aside. The tools and processes to quantify such risks, it would seem, are things that Relsys has been trying to imagine for two years. To that end, it has been preparing to handle a type of data new to the 21 CFR Part 11-regulated world: information locked inside hospital, physician and insurance firm computers.

“We see this as a huge opportunity,” says Loucks. “Consumers are desperately driving to know as much as possible about risks. Life sciences firms are craving the data from the health care side, and vice versa.”

If you’re hoping you can do this at home with a cute little Excel spreadsheet, think again. “We’re talking about orders of magnitude of more data,” says Loucks. “That is something people need to take a strong understanding about. You’ll have more data to work with. The complexity of the analysis will be impacted.”d9A2t49mkex

Data Delay

As many readers know, the industry’s reliance on clinical trial data is a blessing and a curse. It’s a good thing because such data tend to be clean, i.e. scrupulously collected. But there’s seldom enough trial data, in hand at the right moment, to support policy decisions in an era of 22-minute news cycles and academic critics who (unlike the industry) happily communicate with the public and politicians.

Once upon a time, a new trial could be ordered up to address significant safety questions. That worked in 1988, and it will probably be necessary in 2018. But in 2008, delays in answering drug safety questions lead to public impatience and regulatory mistrust. The fact that spontaneous adverse events are so erratically reported only adds to the problem and confounds the analysis of the signal, because media attention itself can generate additional reports.

As best we can tell, Loucks doesn’t have a Relsys product or service to announce with health care data. But it’s clear the company is thinking about combining data from electronic health record (EHR) systems and clinical trial systems. He would like those connections to be straightforward. “The linkage today between EHR and safety systems is manual,” he says. “If anyone is doing it today, they are doing it very inefficiently.”

Converging Standards

In some cases, he says, sophisticated companies may base their “statistical analysis” on quickly eyeballing the corresponding cells in two separate spreadsheets. Even rudimentary drug safety risk assessments may require programmers to work for months or, yes, years. Loucks’ ideal time line is shorter: “You would take something that normally takes two months to do and streamline that to something that takes half a day.”

“What other data could contribute in the process to identification of potential problems, beyond what is being done today?” Loucks asks. “The people in the industry are very much in desperation to understand any potential risk to the drugs they are developing and marketing.”

Loucks acknowledges that there are issues of incompatible systems in the universe of industry-sponsored research and the physician/hospital landscape. “There are a variety of standards that are not in alignment between the two industries,” he concedes. But he’s optimistic, largely because of the integration project that is under way between the Clinical Data Interchange Standards Consortium (CDISC) and HL7. That project began in 2004.

Adoption Curve

Indeed, Loucks predicts that the adoption of EHR systems will occur more quickly than the gradual uptake of electronic data capture (EDC) systems. “EHRs will be adopted much more rapidly,” he says. And the usage of EHR systems, in turn, will allow the harmonization of CDISC/HL7 standards to bear fruit, aiding firms that seek to tap into new data sources like insurance claims or EHRs.

We hope he’s right. But the U.S. is hardly a big and tall version of the UK or Scandinavia, where nationwide databases have been pioneered. The American medical data landscape is extremely fragmented. Despite grand promises being made for EHR systems, most U.S. doctors want federal handouts to purchase them. With EDC, it was clear who had to cut the check and why.

Loucks insists the EHR technology is farther along, and will be easier to connect to pharma’s systems, than a journalist may suspect. In a trade show demonstration project with Allscripts and Pfizer, he says, Relsys helped capture adverse events inside an EHR system, and zapped them into its industrial-strength safety system.

Wider Net

Once the data are in the system, of course, the work has only begun. “You need the analytics, but you also need a disciplined process behind that and a dedicated workflow,” says Loucks. That’s something Relsys can help develop. 

Some of the analytical algorithms may be from third parties, whether academic or commercial. “Many of these analytical techniques that are used are very sophisticated. People need a Ph.D. to really understand it,” he says.

But once the data are in the system, there may be scientific insights of a new caliber than could have been harvested in the past. Says Loucks: “The beauty is linking not just standard adverse events but other data sources and other parts of the population you may not have captured” in an old-fashioned, randomized clinical trial.