“Visionary organizations already know the train has left the station,” Yele Aluko, MD, says.
Healthcare executives may have been slower than most to harness the power of predictive analytics, but they have a good strategy for expediting their entrance onto the big data playing field: look to the role-model industries who’ve done it right.
Jean Drouin, MD, CEO of Clarify Health, recommends taking a page from the book of industries, such as online retail, financial services, and shipping, that have adopted predictive analytics early and are now reaping the benefits. He consistently looks to FedEx.
“In the background, they have very sophisticated workflow optimization software to route each truck precisely every day, based on what’s packed inside it,” he said. “I began with the business plan that we were going to bring exactly that kind of capability to, for example, the way a hip replacement journey works today—or a coronary bypass or asthma, if you will.”
On the academic side of the equation is Marco D. Huesch, MBBS,PhD, vice chair of radiology research and an associate professor of radiology at Penn State University's Milton S. Hershey Medical Center.
“If Santander or Ally bank have a chatbot integrated with core banking to tell me everything I need to know and guide my consumer decision making, we need to get with the program and do that, too,” he said.
The goal of predictive analytics in healthcare is to learn how to lower costs, increase efficiency, and produce the best outcomes. Achieving those goals means studying the delivery of care throughout its entire journey. Having a plan to make it happen is crucial for any forward-thinking health system.
The reporting stage of the Medicare Access and CHIP Reauthorization Act (MACRA) took effect this year, and the incentives and penalties within its Merit-Based Incentive Payment System and Advanced Alternative Payment Models will begin to be doled out in 2019. The more control providers have over cost and outcomes, the safer they are from MACRA’s consequences.
Drouin, a doctor and former longtime consultant with McKinsey, saw the stars aligning a few years ago. In classic start-up fashion, he left his job at the firm, packed a truck full with his belongings, and drove cross-country to Silicon Valley. He said he wouldn’t have done it if he didn’t believe that the transition from fee-for-service to value-based care was really happening. Still, cultural resistance is built into the conversation his firm must have when approaching providers.
“When you get down to a more micro level and you try to work with hospitals to become more efficient and provide more satisfying journeys for patients, you very quickly realize that the tools of the trade today are often still Excel and Post-it Notes,” he said. “We didn’t create the payment model, and we understand that you would’ve liked to remain in the old fee-for-service system. But now that you’re under these new sets of rules, would you like information about what you need to do to win or to beat the new system?”
A recent Ernst & Young survey highlights the gap in commitment to value-based models. Their 2017 Health Advisory Survey of 700 medical executives found that 67% of the smallest health systems (revenues of $100 million to $499 million) and 61% of the next group ($500 million to $999 million) had no value-based reimbursement initiatives planned as they entered 2017. By contrast, 92% of the largest health systems (revenues of $5 billion or more) had such initiatives for 2017. Only 26% of respondents placed analytic insights and new technologies among their top 3 priorities for the year.
“Visionary organizations already know the train has left the station,” said Yele Aluko, MD, MBA, an author of the report, in an interview.
The decision to invest in advanced analytics seems like an easy one to make, but Huesch said that isn’t always so. There are several complicating factors: the very reason that healthcare is behind in using predictive analytics strategies is largely cultural.
Fee-for-service has been a lucrative model in the American healthcare system, and staff at every level of the system are trained to operate in this entrenched environment. Plus, according to Huesch, plenty of hospital systems operate in markets where there are “cozy duopolies or oligarchies” of regional payers and providers, where little urgency is felt to make any fundamental cultural changes. Despite the fervor around MACRA, he said he thinks the stakes might not be high enough given CMS’ recent rollback of bundled payment requirements.
“I kind of think of it as a perfect storm: lack of history, lack of overall competences in the industry, lack of people with the skills, lack of means to purchase the skills, and just a very aggressive, repetitive sales pitch from a lot of the vendors who make a cogent case,” Huesch said.
It may not just be the culture of American medicine that shapes the discourse, but the larger culture of the United States that dictates the culture within its hospitals. Huesch mentioned the example of Ontario, Canada, consolidating all cardiac surgeries to a single medical center, sacrificing convenience for thousands to raise the quality of care by shuttering subpar facilities. He said he sees the abdication of convenience in the name of better overall outcomes as a nonstarter in the United States.
Besides marking a tremendous cultural shift, hiring and training analysts can be expensive. “I think it’s important to be fair to some hospitals where it may legitimately be a bridge too far,” he said. “You might need $750,000 a year to support 1 data scientist and a few analysts. It is a lumpy expenditure, and I don’t want to knock the people that are balking at making that.”
Drouin believes the healthcare industry is still in the early-adopter phase. “The adoption of both value-based payment models and the technology and the analytics we’ve been talking about to effectively operate under those models [is] slower than, obviously, some of us on this side of the table might have wished for.”
Although he believes some third-party analytics firms are reliable, Huesch sees the future of analytics adoption in medicine occurring in-house. In a commentary that he co-authored with colleague Timothy J. Mosher, MD, for the New England Journal of Medicine’s Catalyst publication, the 2 reference a common example of analytics success. “When Amazon and Jeff Bezos anecdotally claim never to throw data away, the unstated corollary is that they actually use their data in-house to drive actionable insights in their field.”
When asked about looking outside for analytics work, Huesch said there’s a tradeoff. “What you get with external vendors is some reliability, some baseline sense of trust because they have done it before...but even the best vendors can fail,” he said.
For academic medical centers, on the other hand, Huesch sees it as near necessity that they do their own analytics in-house to foster their culture of innovation. “You become a learning system. You set up a flywheel in which parts of the institution that were originally flying blind now fly with some data and then appreciate it and want more and generate an increased desire to consume and generate more data. You have a very positive, virtuous circle going on in terms of your learning,” he said, mentioning the success of Johns Hopkins University as a prime example.
Drouin also thinks momentum will be key to the spread of data analytics. “If you get 30% or 40% of your book of business being on these models, that creates a tipping point where we’ll rapidly accelerate to more like 70% or 80%,” he said. “If you had to ask me when that tipping point is going to occur, it’s probably more in the 5-to-10-year time frame.”
If experts are looking externally for inspiration for their analytics programs, it’s also where they’re finding many of their analysts. Huesch’s department is staffed by people who all came up in other industries, and that is not uncommon. He also believes it’s possible to create a quick reversal in narrative, but only through centralized approaches, such as subsidized healthcare analytics training programs or additional pressure on the industry.
“I would say that a sober, feasible approach would be the same way we’ve spent upwards of 2-dozen billion dollars supporting [electronic medical record (EMR)] implementation across the board, right? If we were willing to spend just a little bit more to put political pressure on Epic and Cerner, etc, to make analytics more of a low-cost add-in to the EMR stuff they already have, as opposed to a high-cost add-on, I think that would be a very efficient scale approach to doing this,” he said. “Call it ‘Part X’ Medicare for Analytics.”
Transitions may be difficult, but hospitals have several places to look for inspiration in their approaches to new data uses. Although there is no debate about the potential and necessity for analytics, there remains plenty to be had about what they mean, who should provide them, what they should cost, and what can be expected of them.