Novartis Seeks Hidden Cures in Machine Learning, AI
Novartis' Global Drug Development has completed the first phase of a digital transformation that has enabled it to better leverage one of its best assets -- data.
Big companies may not be as fast moving as small ones, but they often have an asset that smaller, younger companies can never have -- years' worth of historical data. But even though they have that asset, are they really able to turn it into value? Can a scientist or marketing pro in any given company query their data for insights that add to the top or bottom lines?
At pharmaceutical giants such as Novartis, the stakes are even higher. Give doctors access to the data they need and maybe, together, they can perform drug discovery faster. Maybe they can optimize drug trials to make them more efficient. Maybe, eventually, they can apply machine learning and artificial intelligence to their vast stores of data sets to surface and develop a new drug using just a computer.
That was the directive given two years ago to Novartis' Achim Plueckebaum, Global Head of Drug Development IT at the global pharmaceutical company. It is one that may be the cornerstone of future drug development at the company. It kicked off under Dr. Vasant Narasimhan, who was then the global head of drug development and chief medical officer. Narasimhan was appointed CEO of Novartis In February.
"He's a big proponent of technology and how technology can transform businesses," Plueckebaum told InformationWeek in an interview. "He also saw that our systems were absolutely not up to par. We were using the phrase: 'The technology is not as good as the people we have.'"
Plueckebaum set out to change all that in a big and dramatic way. In fact, he calls it "open heart surgery" for Novartis' Global Drug Development (GDD) IT infrastructure. His team gave themselves a challenge: rip out all the transaction systems for the GDD group and put everything together under one program within the timeframe of 18 to 24 months. Internally, the program was officially called STRIDE -- Systematic Transformation Initiative for Development Excellence.
"It was such a big project that we had to split it into multiple areas and subprograms," Plueckebaum told me. The new infrastructure included a number of new systems, including a brand-new submission and document management system, a brand new investigative portal, a brand new high-performance computing environment, and a new clinical trial management system (CTMS), among other new systems. But the centerpiece of the strategy went beyond the new transactional systems. The team connected all of these new best-of-breed systems with a data hub that enabled them to talk to each other. Why best of breed?
"We want to focus on what we do well, running trials," Plueckebaum said. "We are not a tech shop per se…We recognize these technologies are evolving fast."
Novartis deployed these systems both on-premises and in the cloud. Plueckebaum didn't have an estimate on how much data was stored, but he said it was the data behind 500 to 550 active trials at the company and was one of the largest development data sets in the industry. The company runs these trials across about 60 countries in parallel. Novartis has been collecting this data over the last 20 years. The total data comprises about 2 million patient years.
"It's a huge machine," Plueckebaum said.
Now that the IT team has completed the STRIDE transformation, Plueckebaum has identified two new programs to take Novartis' drug development IT into the future.
The first is called Nerve Live, and it's a predictive analytics platform that applies analytics and machine learning algorithms to clinical trial operations. Novartis is working in partnership with US machine learning company Quantam Black on Nerve Live.
Among the problems Nerve Live is tackling is to optimize the physical location of drug trials. Plueckebaum said that it can be a challenge to recruit the right patients at the right time to drug trials. Ideally they should be located close to your trial site. By analyzing data about the performance of sites in the past, this tool can make predictions on future site performance.
"You can optimize your trial up front, you can avoid waste for patients, and you can make the whole trial more effective," Plueckebaum said.
The second future project is Data 42 -- what Plueckebaum calls the company's "moonshot." The name is a reference to The Hitchhiker's Guide to the Galaxy's where 42 is "the answer to the ultimate question of life, the universe, and everything."
The idea is to put all of Novartis' data sets together, virtually, to enable queries on all of it.
"What if we were to combine all our data sets together, access the data, and make it specific to disease areas so that scientists can ask the questions they weren't able to ask before," said Plueckebaum. This project will give researchers access to the data and the tools to do it. It's the kind of large data set where machine learning and artificial intelligence can accellerate the efforts, too.
The ultimate goal is to one day design a drug "in silico" -- completely develop a product not in the lab but instead on the computer.
"That's the big dream we are after and that's what Data 42 is going after now," Plueckebam said.
Read these articles to learn more about how healthcare organizations are leveraging IT and data and analytics to build a better business.
Jessica Davis has spent a career covering the intersection of business and technology at titles including IDG's Infoworld, Ziff Davis Enterprise's eWeek and Channel Insider, and Penton Technology's MSPmentor. She's passionate about the practical use of business intelligence, ... View Full Bio
We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.
The Next Generation of IT SupportThe workforce is changing as businesses become global and technology erodes geographical and physical barriers.IT organizations are critical to enabling this transition and can utilize next-generation tools and strategies to provide world-class support regardless of location, platform or device