by Stephanie Overby

IT Helps Pharmaceuticals Develop New Drugs, but the Tools Can’t Do It Alone

Oct 15, 200215 mins
IT Leadership

As vice president of worldwide informatics for Pfizer Global R&D, Walter Hauck had been diligent about getting just the right people on his team. He had been charged with integrating the disjointed R&D efforts at the biggest pharmaceutical company in the world. He was supposed to create the tools and processes that would allow scientists and clinicians and marketers and lawyers to work together to introduce Pfizer’s next generation of drugs.

So Hauck hunted for people who knew both IT and science and asked them to join his 1,200-person team. He assigned them to research sites around the globe so that they could learn Pfizer’s business and guide their users through the transformation. He recruited people from the research and development side to join the informatics department.

But Hauck found out that PhDs and dual degrees weren’t enough because most of the people working for him didn’t know diddly about change management. And Hauck didn’t know enough either.

That became clear during the integration of the R&D division of Warner-Lambert, the Morris Plains, N.J.-based drug company Pfizer acquired in 2000. Hauck figured that the simplest, most logical way to handle the changes was to do them all at once. His IT team agreed. At La Jolla Labs in California, for instance, Hauck’s team introduced one new application a week for six months.

Soon, the scientists and lab workers were ready to riot.

“They were saying, We’re trying to deliver new compounds and develop new drugs, and we have to spend all our time learning these new systems,” recalls Hauck, who heard similar reactions throughout the newly merged organization. “It was a lot of change to drop on people while expecting them to continue to deliver.” After all, these new technologies were supposed to improve productivity, not impede it.

That’s when Hauck decided to educate himself and his staff about managing change. “In the end,” he says, “integration comes down to nothing more than change management.”

Divided They Fail

Throughout the $418 billion pharmaceutical industry, money-making drugs such as Eli Lilly’s Prozac, Aventis’s Allegra and Pfizer’s Zoloft are going off patent, meaning the huge profits they’ve been generating will diminish. At the same time, masses of new data are piling up from the decoding of the human genome. In order to make sense of all this information and reduce the time and money it takes to bring drugs to market, pharmaceutical companies are introducing new technologies ranging from the automated screening of potential compounds by the thousands to software that speeds the process of synthesizing and then testing the most promising chemical compounds. But exploiting the full potential of these new tools requires a fundamental shift in the way Big Pharma R&D works.

“The technology itself is tricky. But once you put the technology in place, there’s the entirely different matter of getting people comfortable with and confident in using the technology,” says Roy Dunbar, vice president of IT and CIO at Eli Lilly. “That requires getting scientists to work in different ways. And that’s a huge shift in culture.”

The challenge facing the pharmaceutical industry is fairly straightforward. For years, biologists, chemists and clinicians worked in relative isolation. Integration didn’t occur among humans, let alone on a systems level. “It’s a pretty frustrating situation for a CIO,” explains Charles-Andre Brouwers, a vice president of the Boston Consulting Group who is based in New York City. “There’s so much focus on a specific project, a specific compound, a specific target. Each person is intensely focused on his own part of the value chain.”

Meanwhile, mega-mergers have created gargantuan global drug conglomerates and even more disjointed R&D cultures, “aggravating the problem at a time when shortening research cycles has become ever more a priority,” says Michael Barrett, senior analyst at Cambridge, Mass.-based Forrester Research.

As Big Pharma makes the big move from a bench-based environment of beakers and Bunsen burners and scientists working in isolation to a collaborative environment of digitized data and automated processes, the CIO finds himself at the hot center of change. And the ability to manage that change coolly will be the critical factor in his success?or lack thereof.

“Legacy behaviors are as bad or worse than legacy systems,” says Dunbar. “Systems are dumb and can be easily replaced. People are sentimental; they have reasons why they do what they do. And they can’t be so easily replaced.”

Other industries are confronting the same need for an IT infrastructure that pulls back-end systems together with marketing and customer-facing technologies. And whether the industry makes cars or drugs, the CIO is the executive in charge of the change management that must take place before this vast infrastructure can do its job.

At Lilly, Dunbar has had to struggle to meld the warring mind-sets of IT and science. For Pfizer’s Hauck, it has meant retraining his staff. And at Aventis, it has meant that users who dare to be innovative in their use of IT and are willing to participate in pilots are not only protected politically but recognized globally.

“R&D is useless unless it can be integrated into a bigger picture that helps you convert leads to drug candidates,” says Charles Cooney, codirector of the Program on the Pharmaceutical Industry at MIT in Cambridge, Mass. “The difficulty for the CIO is getting the team to work between the various disciplines needed to integrate the information in order to create knowledge.”

At Eli Lilly The Sound of Two Hands Clapping

When Dunbar took over as CIO at Eli Lilly in 1999, the Indianapolis-based drugmaker was in the doldrums. Lilly’s leading moneymaker at the time, Prozac, was set to go off patent in 2001, bringing a major erosion in revenue right around the corner. The $12.5 billion company had been increasing its investment in R&D (up to $2.2 billion that year), but the money wasn’t boosting productivity.

Dunbar, who had previously worked in marketing and product development at the 126-year-old company, knew that simply pumping more money into R&D would not fill Lilly’s pipeline. For more than a century, Lilly’s divisions had operated virtually autonomously. But now Lilly wanted to connect the dots of the R&D work being done at dozens of the company’s sites by 8,000 employees throughout the world.

“The goal is to have information flowing from tool to tool so you have a kind of wall-to-wall carpeting of technologies along the value chain,” says Dunbar. “That will give us a huge amount of productivity.”

During the past three years, Dunbar and his team have introduced a host of new technologies. They include a searchable database of molecular information that allows researchers to locate all relevant experiments in a specific area, a bioinformatics application called Game (Gene Anatomy Made Easy) that enables researchers to perform genomic searches in minutes, and a tool from Internet Database Solutions called ActivityBase that cuts the time it takes to analyze compounds from five hours to 30 minutes. Dunbar has made those and other applications available through the $4 million Elvis (Eli Lilly Virtual Information System) portal launched last year. Dunbar’s team is also working to integrate the data in the disparate applications.

The master plan is to extend the integration to all parts of the business. “We have a new area called pharmacogenomics that looks at how my DNA may result in me responding differently to a drug than another person,” Dunbar explains. “In the future, our marketing people are going to be interested in understanding what subsets of the population are going to respond best to a particular product so they can market it directly to them.”

Dunbar says the ultimate success of those tools will depend on what he calls the two-handed clap?the collaboration between IT and scientific business partners.

“It’s one thing if you have a fantastic IT organization but you have a constituency uninterested in using IT,” he says. “It’s another thing if the users are demanding real IT solutions and the IT department is not up to the task. What we have is the coming together of a capable IT team and a scientific team that sees IT as a partner. One without the other would be as futile as a one-handed clap.”

That kind of collaboration wasn’t happening when Dunbar joined the IT team three years ago.

“Traditional IT thinking is based on the engineering mind-set: You describe a process, break it down, build one solution and use it thousands of times,” Dunbar explains. “The scientific mind-set is more experimental: You discover a problem, ask a question, conduct an experiment, and once you get an answer, you move on to a new question.”

In order to work on successful IT solutions, the two groups had to adjust to each other’s native thought processes. “The IT group needed to start thinking about building a small set of systems that allow for some experimentation,” says Dunbar. “The scientific folks needed to get a sense of a more structured approach to doing their work that could be modeled in IT.”

To facilitate such a meeting of mind-sets, Dunbar doubled his department to 2,700 (600 of them devoted to R&D), and he has increasingly hired people with dual computer science and biology or chemistry degrees. Such interdisciplinary fluency enables Dunbar’s staff to talk to scientists about the value of trying out new systems.

At Lilly, “scientists and technologists literally sit together in a room and engage in conversations about the future,” says Dunbar. They first convene to discuss resources and prioritize a portfolio of projects each quarter. Then a small group of scientists chosen to champion a project acts as liaisons between the IT project leader and team of programmers and the scientists. They discuss requirements, and the programmers develop a prototype. Then the real talking starts, as the two sides figure out what has to happen for the new technology to work. Does an application need to change? Does a business process need to be tweaked?

“It’s never just a question of putting the technology in place,” says Dunbar.

The discussions seem to be working. Scientists are actually asking to be trained on new systems, like ActivityBase, now being rolled out to thousands of researchers on three continents. “They’re hearing about tenfold cycle time reductions on getting information, and the demand for training and process change is increasing. It’s a good sign,” Dunbar says.

No more than two weeks go by without a meeting between IT and R&D. “Even as the projects move into production, it’s senior- level scientists who are championing the new tools and evangelizing about their benefits,” Dunbar explains. “From the vice president level down, they’re leading the change.”

At Pfizer Teach Your Techies Well

It’s hard to imagine that Pfizer?the number-one U.S. drugmaker with seven blockbuster drugs currently seeing annual sales of more than $1 billion each and 130 new drug candidates in the pipeline?would be under pressure to streamline development. But it is.

“We have to deliver the next wave of compounds?15 in five years?and those will come only through integration,” insists Hauck. Money certainly is not the issue. Pfizer Global’s R&D budget in 2002 was $4.7 billion (the highest in the pharmaceutical industry and the third highest, it says, of any company in any industry worldwide). “We need to double our R&D productivity, and that’s not going to happen by making researchers work 80 hours a week.”

Like Dunbar, Hauck has found that forging ties between IT and the New York City-based company’s 20 research sites is critical to the integration effort’s success. His 1,200 full-time IT employees and 800 contractors are organized into development teams working elbow to elbow with 12,000 scientists trying to discover the next Viagra or Zoloft. “It’s important to me to have people at each site, affiliated with single users and spending years to learn their business processes,” says Hauck.

“I’m a software professional by training, but I need to know the business cold,” he adds. Since he became vice president of worldwide informatics in 1995, he has been working to move R&D applications to the Web so that everyone at the company can access them.

But two years ago, the ante was upped with the acquisition of Warner-Lambert. The merger made Pfizer the fifth largest company in the world. It also presented a huge integration challenge.

“In the end,” Hauck recalls, “we changed pretty much everything at once. And the sheer volume of change was a lot to digest.”

In the process, Hauck found that his employees would “trip over simple things,” like forgetting to coordinate training so that users wouldn’t have to sit through the same presentation several times. Or they wouldn’t wait to see how well a new application was being adopted before introducing the next one.

“What I’ve decided to try to do,” he says, “is to stop making it an experiential thing in which we have to relearn the same lessons, and instead begin to codify change management practices and teach them to the staff.”

To that end, Hauck created a standard change management strategy and checklist for every informatics project. Did the project leaders assess business readiness for change? Did they adequately resource the first pilot? Did they check for feedback?

“What’s most important,” says Hauck, “is to teach my staff what it takes to make change a reality.”

At Aventis Encourage Innovation

At Strasbourg, France-based Aventis Pharmaceuticals, the name of its R&D operation?Drug Innovation and Approval (DI&A)?certainly suggests an environment that embraces change. But as Peter Loupos, vice president of DI&A information solutions, discovered, introducing new collaborative systems to the division’s 5,000 scientists and support staff in the United States, Germany, France and Japan?and actually getting them to use the systems?required more than just a cool name.

As was the case at every other drug company, Aventis (formed by the merger of Rhone-Poulenc Rorer and Hoechst Marion Roussel in 1999) researchers focused on their own tiny piece of the puzzle. “Chemists just worked with chemists,” says Loupos. “Biologists just worked with biologists.”

But today, with the goal of introducing two to three new drugs a year, DI&A is devoting itself to chemical biology, an approach that, as its name suggests, involves bringing together once separate disciplines.

For Loupos and his team of 250 IT staffers, that has meant creating Web-ready databases, visualization tools that combine information on chemical structure and biology, portal technology to enable researchers to access common data, and collaboration software so that far-flung individuals can work as a team. This integrated IT platform allows an expert on ion channels in Bridgewater, N.J., to contribute to a Parkinson’s disease project in Paris.

But getting that ion channel expert to explore this new way of working can take some coaxing.

“You need to signal that risk-takers will be rewarded rather than penalized,” says Forrester’s Barrett. “If the leader of a team volunteers to use a new technology, he should be protected from the downside that always attends innovation. A researcher trying to get a certain therapy to market in four years isn’t going to want to risk a six-month delay. The remedy for that is political support.” What that means is that executives and department heads must let it be known that project delays due to technology innovation will not reflect poorly on participants.

The other part of Loupos’s effort involves publicizing success. For example, last year Loupos was conducting a pilot of a Web-based knowledge management tool called KnowledgeMail. At the same time, a scientist researching degenerative thrombotic and joint diseases in Frankfurt, Germany, was beginning a project to isolate and culture large white blood cells that ingest and destroy microbes and other foreign substances. Using KnowledgeMail, he discovered two scientists in Bridgewater who could help him out. One provided culturing protocols, the other offered information on magnetic cell sorting. The result: four weeks shaved off his project.

Not only did his success generate positive word of mouth outside the pilot group, it gave Loupos something he could take to Chief Scientific Officer Frank Douglas, who approved a full-scale rollout of the system. “We’re using technology to push a real cultural shift here,” says Loupos. “We’re getting to the point where people get excited when they discover something new using IS and they want to share it with their colleagues.”

Loupos admits that the attitude change at Aventis hasn’t occurred overnight. “Obviously, we had our early adopters and others that have seen what’s happening and eventually came on board,” he says. “And there are some others who may never jump on the wagon.

“I think it’s more important to spend time with the innovators,” he adds, “which is where you’re going to get the biggest benefit.”

CIOs will need to get the word out about what the innovators are working on, Barrett says. “One of the real problems is that genuine innovation with a given group isn’t communicated through the rest of the organization,” he explains. “The CIO has a key role as facilitator. He or she stands at a crossroads of information. He can showcase successful IT applications companywide.

“There are very few genuine innovators, but there are massive numbers of fast followers,” Barrett adds.

And that’s the way change happens. Leaders experience success, and everyone else wants to share in it.