Evidence: AstraZeneca’s ‘big data’ partnership
Throughout the week, we are taking a look at how evidence can help us identify the health and cost impact of treatments on patients, while also highlighting areas of unmet patient needs. Our first post can be found here.
Last month, McKinsey Quarterly took a look at how companies are using “big data” – large-scale data gathering and analytics – to shape strategy. AstraZeneca’s vice president of managed markets, Mark Lelinski, explained how our real-world evidence strategy will help build customer relationships that focus on improving health and managing the total cost of care.
This commentary is adapted from an interview with Sam Marwaha, a director in McKinsey’s New York office.
By Mark Lelinski
We have always designed and manufactured our products with the mind-set of “make it effective, make it safe, and make sure it meets regulatory approval.” Historically, at the early prelaunch stage, we were not thinking about the willingness of payers to pay for it—whether that’s a patient, health plan, pharmacy benefit manager, employer, or the government. We weren’t asking, “How do customers perceive our products relative to alternatives?”
But willingness to pay has obviously become extremely important in recent years—to the extent that more and more of our customers began complementing our clinical-trials data with their own proprietary data to conduct comparative-effectiveness studies. They were asking, “In a real-world setting, product X performs at this level and costs me this much. And product Y performs at this level and costs me this much. How do they compare?”
Eventually, this practice created an imbalance in our payer conversations, as the dialogue became more transactional—more about unit cost and more about the data that our customers were bringing to the table. And from our perspective, few of the comparative studies that payers were conducting focused on health outcomes.
So we decided that we needed to get beyond our single focus on the controlled environment of the randomized clinical trial and see the business from the other side as well.
The focus, we realized, needed to be on the total cost of care. Don’t just talk about the unit cost of a drug, but learn about the total cost that it takes to manage, say, a diabetic patient—including the diagnostics, the outpatient visits, the emergency room visits.
This led to an “aha” moment: if we could combine medical-claims data with clinical data collected in an electronic-medical-record system for a defined patient population, we might actually discover ways to improve health outcomes and manage the total cost of care at the same time. And why not collaborate with customers?
Prescription drugs represent about 11 percent of total health care spending in the United States. For the other 89 percent, our interests are completely aligned. By working together, we all get access to a broader, richer data environment, and we can work together on creating state-of-the-art access tools and real-world methodologies.
So we took this idea to potential partners.
From the beginning, this was about true collaboration and strategic fit, not an “I’m gonna win more than you win” mentality. When we presented our vision to HealthCore and its parent company, WellPoint, we quickly realized that their views on all of these things were so similar to ours that everyone’s jaws kind of dropped. It was a quick connect. We announced our collaboration in February 2011.
Certainly, there was some internal resistance at first. In some cases, we were asking our people to think in dramatically different ways than they had for the bulk of their careers. This is especially true in R&D, where we’re now bringing in the voice of the payer much earlier in the development process so we can “lose the losers” quickly and not take products to market that won’t be valued by the people paying for them.
And of course we still negotiate with WellPoint on individual drugs, so the increased transparency acts as a double-edged sword: if the collaboration helps us get new evidence that supports a price point we set, that’s extremely valuable. But sometimes it goes against us too.
The key to turning around the resistance and getting to where we are today has been the senior-level involvement and support we’ve received from the start. Our leaders recognized that this approach is a long-term play: there may be quick wins and short-term gains for the company, but to really have a broad impact on the company and the industry, we have to manage the complexity and growing pains.
One example was the way we brought together top-notch biostatisticians, epidemiologists, health economists, and programmers working throughout the company and created a new group focused on real-world evidence. Without the support of engaged and interested leadership, making that happen would have been like pushing a rock uphill.
While this partnership is still in the early stages, HealthCore and AstraZeneca personnel are operationally aligned and set up, and working together very well. We have a number of joint studies under way and are in the throes of completing the first one, which will be ready for discussion with payers soon.
Still, both sides see this as the first phase of a broader, industry-wide collaboration. Eventually, we expect this will include other health insurers, pharmacy benefit managers, providers, employers, other pharmaceutical manufacturers, and even federal and state governments.
It won’t be just about pharmaceuticals but about much more: Which diagnostics make sense and which don’t? Which medical devices? What leads to errors or high readmission rates in hospital settings? What key health issues need to be addressed in a given local community? Through big data, we can learn things about health care that we could never get at before. And that’s really what we’re setting out to do.
Adapted from “Seizing the potential of ‘big data,’” October 2011, McKinsey Quarterly. McKinsey & Company. Reprinted by permission.