Bringing a new drug to market takes an average of 15 years and $1 billion dollars, according to trade group PhRMA. Factoring in the cost of failure—just one in 10,000 potential medicines makes it to approved patient use—other industry analysts estimate a single drug can cost $4 billion or more. Yet despite the $125 billion spent annually by public companies on pharmaceutical research and development, much of the industry’s business development processes remain rudimentary.
Dr. Brigham Hyde, cofounder of Relay Technology Management, wants to change that. Working in partnership with Attivio, Relay has developed a software-as-a-service business intelligence tool to help drug companies validate their R&D opportunities more effectively and benchmark their product pipelines against the competition. The system is built on Attivio’s unified information access platform, which brings together structured and unstructured data.
Say, for example, someone in business development at Pfizer or Merck decides she wants to acquire an Alzheimer’s treatment in development. She looks up relevant assets using one system. Then she might dig up clinical data using another tool. And she relies on Google for information on business deals by drug companies and investors in the space. “Manual data curation is a huge part of people’s jobs still,” says Hyde.
Relay’s developers write a single SQL query to search both structured databases and unstructured information like press releases or SEC filings. Relay’s BDLive! tool then delivers the information to users in a searchable way
“You can think of it as taking every piece of information about an asset and laying it out flat on a table,” says Hyde. “We can ask questions from anywhere on that table.” Hyde estimates it reduces business development research time by 80 percent. The data-driven tool is hitting the drug industry at an important time, says Dr. Kenneth Kaitin, director of the Tufts Center for the Study of Drug Development and a member of Relay’s advisory board (advisers are paid in company stock).
“With a large number of drugs that have already lost or will lose patent protection within the next few years, and a dearth of new drugs in company pipelines, the research-based pharmaceutical industry is increasingly looking to strategic partnerships with academic research centers and other organizations to replenish portfolios and spread development risk,” Kaitin says. “Making sounds decisions about which external drug candidates represent potential commercial value and should be licensed in is critical.”
Relay also uses both TIBCO Spotfire and Tableau Software to create custom data visualizations and dashboards for its customers—25 so far, including three of the five Big Pharma companies. “Spotfire and Tableau have very specific differences in performance and capabilities related both to their ability to connect to different types of data and their format of data structures,” says Hyde. “We use Spotfire for some things and Tableau for others.” Hyde’s analysts can produce dashboards—a disease-specific snapshot or visualization of investment or regulatory trends—on the fly and incorporate them into users’ BDLive! log-ins. “It is this blend of custom and out-of-the-box technology that enables us to focus on the customer.”
Partnering with Attivio was critical for Relay. “As a young company, we didn’t want to build and own a search and relevancy model,” says Hyde. “We wanted to spend our time on what we do best.” That, Hyde says, is Relay’s ontological system and natural language processing, which takes into account scientific naming issues, disease hierarchies, and how health conditions relate to each other. “It’s our secret sauce,” Hyde says.
Every drug has a name but also various synonyms. Ditto with diseases, genes, and proteins. Diseases can have subset classifications. “Cancer’s not just cancer. There are subvarieties of it,” Hyde says. “We know what a user means when they type ‘chronic lymphocytic leukemia’. We build that logic into the front end to give the users what they actually want.”
Tools from industry stalwarts like Elsevier and Thomson Reuters incorporate some unstructured data the old-fashioned way, says Hyde: “They have teams of people in India reading documents and typing the information into spreadsheets or databases.” The drug makers themselves have failed to develop systems to integrate varied sources of business intelligence. “I argue that they should develop their own systems, but their efforts in this area generally fall down,” Hyde says. “They have their engineers build a system, but they don’t understand the context and it fails to serve the users.”
Since launching BDLive! in May, Relay has moved into higher-end analytics, performing statistical analysis on data simultaneously in SQL and R to uncover trends and patterns. One result is The Relay Relative Value Index (RVI), essentially a stock price for pharmaceutical assets. The RVI puts a dollar value on treatments in development from 500 academic researchers along with phase II and III clinical trial candidates from 2,500 biotechnology companies based on factors including strength of intellectual property, market potential, and scientific trends. “I’m careful when I talk to clients about it,” says Hyde. “They’re not going to say, ‘Relay says this is number one. Let’s go do the deal.’ But they can use it as a proxy and a way to prioritize.”
Stephanie Overby is a Boston-based freelance writer. Follow her on Twitter: @stephanieoverby.