Could the pharmaceutical industry manage clinical data more intelligently, using software to go toe to toe with Steve Nissen of the Cleveland Clinic? Would such tools ensure more consistent communications with regulatory bodies? Properly executed, could better software even deflect a meta-analysis cooked up in academia?

The answer from SAS is yes. At this year’s annual meeting of the Drug Information Association (DIA), SAS is unveiling a new pharmacovigilance product, SAS for Patient Safety. Here’s a press release.

The idea is to have an engine for standardizing the curation of drug safety signal detection throughout the life cycle of a drug—before and after approval by regulatory agencies.

Informational Preparedness

The product has been designed to anticipate higher scrutiny of drug safety and forthcoming changes in how the FDA reviews and processes drug safety data. “It’s a top of mind issue,” says David Handelsman, principal product manager for clinical R&D at SAS. “Safety, safety, safety. It seems like not a month goes by without a previously approved drug that has some new concern. It’s much better to find it yourself and develop a strategy for dealing with it than read it on CNN.com. The pharmaceutical companies want a quick and effective way to respond. Days mean a lot.”

Regulators also have rising expectations for the industry that existing drug safety software may not easily address. As an example, Handelsman points to this good review practices guidance document at the FDA website. The idea is to have more consistency within FDA across agency divisions and therapeutic areas.

“This is how FDA will review decisions internally,” says Handelsman. (There’s more background from the FDA here.) “The government is taking a much more active and aggressive position around safety. They want to get to the point where the cardio-renal unit shouldn’t do things differently than the respiratory unit. That’s not helping anyone. There is a lot of attention at the federal level in terms of safety.”

’Holistic View’

Sponsors of clinical trials, Handelsman believes, are already searching for additional tools to have more enterprise-wide responses to drug safety analyses under both ordinary circumstances and during crises that seem to erupt with growing frequency.“It’s about getting a complete view of the process, pre-approval and post-approval,” he says. “SAS has new packaged capability that is designed to address the new guidances that provide a sophisticated approach to pharmacovigilance and signal detection. They provide a more holistic view.”

In short, changes at FDA are going to trickle down to industry. “There has not been good consistency in terms of how they review new drug applications,” he says of the industry.  “The tools they have are dated. They have been doing the same thing for a number of years. They need a way to approach the new guidance.”

Data Exploration

As with other drug-safety applications, SAS for Patient Safety allows the data to be presented in a heat map view that shows the relative magnitude or intensity of a variable being measured. “We are bringing together a lot of different data sources,” says Handelsman, who says other safety analytics software cannot embrace the same multiplicity of data sources. “We have the means of taking the data from whatever source and mapping it to the model of what we’re trying to do.”

In a demonstration of the software, much of the data in the application is clickable—a particular cluster or list of adverse events leads to a particular subpopulation, which leads to a particular organ system or even an individual patient’s electronic case report form book. “This type of visual interface lets you dig around and determine which subpopulations are more intriguing from a safety perspective.”

Software for drug safety, SAS believes, should be able to incorporate new statistical algorithms and external data sources as they become proven in the scientific literature or industry practice. And such systems should not necessarily honor the quirks of Mary in Regulatory who refuses to use sans serif fonts, or who puts all of her safety tables inside a large gray box for reasons no one can quite remember.

“You don’t need the Pfizer-specific analytical summary or the Glaxo specific-analysis,” Handelsman says. “Standard data leads to standard reports. That has not been something the industry has been close to. We’re looking at this as a comprehensive safety analysis system. There is more value in being able to roll this into a single workflow, a single business process.”

Madness, Formatted

Precious time and resources, he says, have been wasted on cosmetic and superficial formatting of reports. If we’re submitting this in Europe, is this box always blue? Handelsman says smart companies will worry more about the content of the report and less about whether the left margin of the page is 1 1/4 inches or 1 1/2 inches wide. “The content was 95 percent the same,” says Handelsman.  “We wasted so much effort making it ‘just so.’ The days of ‘just so’ are fading away. The time is now to support tools that support that standardization.”

He’s quick to say he’s not just referring to a clinical data warehouse in which everyone makes an electronic contribution. “They can’t do anything,” he says. “Stuff goes in [to the warehouse] and never comes out.”

Broad View

Instead, he notes, the new SAS safety product can analyze both a company’s internal data and data in the literature, even data from competitors. “Let’s look at the clinical trial data and start applying signal detection earlier in the process,” he says. “The signals will be harder to find, but they can be found.”

The industry wants to be able to respond both to sharp, sudden critiques from academia, and more substantive questions from regulators; the software is designed to anticipate both types of needs.

Advocating Standards

Once a particular way of slicing and dicing the data has been prepared (a subgroup of men over 60 taking Advil, say), it can be saved and applied to other compounds.

Is there a catch? The software will need some structured data—data that may need to be translated into industry-standard formats. There is a growing arsenal of CDISC data utilities tools to do that, both from SAS and other firms. But part of the chore is designing studies and data management systems to be CDISC-aware from the start. In many organizations, of course, those data conversions will require extensive discussions that triangulate the views of regulatory, biostatistical and data management colleagues.

Handelsman says: “We would use our data integration technologies and provide the means of taking whatever you had, and smoothing it out and making it work with our CDISC-friendly reporting tool. We can’t dictate what your standards are. But we would strongly encourage you.”

The goal is to automate and make drug safety analysis more consistent across medical disciplines and geographies before the FDA makes similar changes. “There’s nothing wrong with what companies are doing today,” says Handelsman. “It’s just much more manual than it should be.”