Comprehend wants to cure data woes for pharma companies

By Richard Nieva , written on January 8, 2013

From The News Desk

As far as big data goes, the pharmaceutical industry has some unique problems. For one, it’s a multi-billion dollar market that has information and data flowing every which way, from disparate groups – doctors, regulators, company executives. It's hard enough taking a new drug to market, but having to deal with that bulk of information can make it an even tougher slog.

Comprehend Systems, a Palo Alto-based company that peddles a service called Comprehend Clinical, thinks it can help solve the problem. The service is a data visualization and analytics tool that helps to organize the data and tailor the presentation to who’s using it, from the doctor on the field examining patients to the analyst at company headquarters scrutinizing test results.

This past summer the company turned switched its product from an onsite deployment one to a software-as-a-service offering and added new investors: Box CEO Aaron Levie, former Yahoo CTO Farzad Nazem, the Life Sciences Angel Network, Easton Capital and Zorba Leiberman, who specializes in investments for information companies. Those investors joined original investors including Y-Combinator, SV Angel, and Menlo Ventures to raise a total of $2.7 million for the company.

Comprehend boasts that its customers are top 10, Fortune 500 pharmaceutical companies as well as smaller ones, like Pittsburgh-based Knopp Biosciences. Morrison said he could not name the bigger pharma players because of non-disclosure agreements.

Rick Morrison, Comprehend’s Chief Executive and co-founder, says the biggest strain is in clinical trials – the process a drug must go through before it is deemed safe and gets approval from the Food and Drug Administration. It takes over $1.3 billion to take a new drug to market, from laboratory to finally making it available to a patient, and about 90 percent of the cost is in clinical trials, says Morrison.

Morrison also says he’s heard that the clinical trial period takes about 12 years, from laboratory to finally being approved for patients.

And that period of testing is uber complicated. The type of information and data being gleaned also changes throughout the trial period, which typically happens in four stages, from seeing if the compound is safe and deciding proper dosage, all the way to testing again once the medication is on the market and a larger set of people are using it. For example, during earlier field-testing sessions by doctors, data might be focused on individual patients, like medical history, or a list of other medications they are on. Doctors can see the data on iPads and iPhones using Comprehend. By contrast, a higher-up at the company can see results from every aspect of the trial on a dashboard.

There is other, of course, business analytics software out there that helps make data meaningful. There’s old-fashioned Microsoft SharePoint, which has an integration with Comprehend. There’s also Tableau, which has been rumored to be heading for an IPO. While Morrison says he thinks both services are visually appealing, he thinks Comprehend works better when the information is coming from a number of different places.

The cost of prolonging clinical trials is real, Morrison points out. He says a typical patent that protects a company from competitors only lasts about 17 years, and is filed fairly early on in the process. If clinical trials take around 12 years, which means the company only has a small window of exclusivity before other companies can make generic versions. “Every day regulatory approval is delayed is lost revenue,” he says.

Anyone who’s ever worked with pharmaceuticals knows the clinical trial periods really are painstakingly long. I did some work while in journalism school covering a doctor testing a drug to treat Meniere’s disease, an inner ear disorder that affects a person’s hearing and balance. That drug is still in the clinical stages. Of course, there are a slew of other factors that go into how long trials might take, but dealing with data should not be one of the things slowing it down.