Will Silicon Valley lead health care's next revolution -- or miss it?
(Editor's Note: the following commentary was co-authored with Tory Wolff, a founding partner of Recon Strategy, a healthcare strategy consulting firm in Boston.)
Silicon Valley entrepreneurs and investors have never quite been able to figure out health, and they know it.
For years, the clever technology fixes dreamed up by engineers have largely failed to take hold, their well-conceived rationality no match for the complexity of medical care, the persistence of clinical habit, and the counter-intuitive impact of existing incentives. Many of the Valley's most audacious VCs have become leery of the space, electing instead to pursue innovation elsewhere.
The new battlefield is on the technologists' home turf: information systems for electronic medical records (EMRs). Will this time be different? Are technology entrepreneurs finally ready to disrupt medicine?
Here's the context (please see our last commentary, available here, for more details). Most of the nation's largest and most prestigious medical centers seem headed towards a relatively closed health information system, driven by a single dominant private company, Wisconsin-based Epic, which excels at the near-flawless, customized installation of their client-server platform in big hospitals.
While Epic is meticulously working its way through the largest hospitals, the long tail of stand-alone ambulatory practices operate largely on a jumbled mess of EMRs, using many emerging vendors (such as AthenaHealth and PracticeFusion) with a multi-tenant model, similar to salesforce.com.
Since medical care as a whole is consolidating, the basic question is whether emerging EMR vendors will gain enough traction and offer enough capability to enable stand-alone practices to remain independent. Or will platform fragmentation put unaffiliated practices at such a competitive disadvantage that they'll be even more motivated to join up with larger hospital systems (the most important of which will rely upon Epic)?
What makes Epic particularly interesting is that its success seems to fly in the face of how so many of us -- Silicon Valley technologists in particular - have come to view innovation; it also contrasts with the much-celebrated, widely accepted strategy of open innovation.
"No matter who you are, most of the smartest people work for someone else," Bill Joy's law goes, and so much of the current Silicon Valley innovation ecosystem relies upon the ability to leverage the insight and wisdom of others. Let good ideas bubble up, find ways to capture creativity from everyone.
From the perspective of most technologists, Epic epitomizes the exact opposite of how a health information system should work (which is also why many of these same technologists feel it's bound to fail). Ideally, according to the technological experts, there should be a common, robust and open set of standards governing healthcare information, easing its accessibility. Companies would then compete for the most effective way to exploit the information, almost certainly via a multi-tenant platform.
According to the smartest people in the Valley, Epic shouldn't be winning. But it is. How is this possible, and what does it mean?
Epic's success suggests that it has locked onto something that its key clients - academic medical centers and large health systems -- need most right now. This burning need, it turns out, isn't the capacity for bubbled-up innovation. What they need is the quick and flawless imposition of structure - pushed down from above, and proprietary if necessary.
The tertiary hospital is a vast enterprise with incredibly complex array of care delivery, with a wide web of participating - and very vocal, idiosyncratic - stakeholders involved. It faces long-term pressures - e.g. care shifting from the hospital to ambulatory settings - and shorter-term pressures with more uncertain endpoints - e.g. reimbursement changes, regulatory requirements. They need someone to step in and define the information sharing processes for the system, providing a reliable way to capture, transport, receive, and use information, as a path towards measuring and incrementally improving the quality and efficiency of care.
It turns out Epic has been paying attention -- a lot of it -- to successful technology enterprises, and appreciates better than anyone else the most important lessons of two of the nation's most successful entrepreneurs, Amazon's Jeff Bezos to Zappos' Tony Hsieh: know your customers.
"Epic is successful for many reasons, but possibly chief among them is their CEO's laser-like focus on the customer. Judy Faulkner knows what large health systems need, and she gives it to them. Often, she knows what they need before they do. Further, Epic figured out long before its competitors that the vendor usually knows best -- not the customer. While this sounds paternalistic (and probably is), the truth is it works."
Epic's credibility -- built up from installations across many of the biggest brands in hospital care - allows it to say what can and can't be done. Even places like Partners in Boston - as unique a medical system as any, and a long-time investor in home-grown solutions - has recently announced that it, too, will move to Epic.
Competitors earnestly contend that Epic's approach locks clients in to a platform that is neither agile nor open, hence it lacks adaptability. Epic is viewed by many in the health information space as a canonical example of an established player who will be disrupted by nimble innovators -- and there are no shortage of eager contenders who lay claim this mantle.
Is the story so simple? How much of it is wishful thinking (or sour grapes) on the part of Epic's competitors?
The great paradox here is that by imposing a specific and relatively strict structure on a large health system, Epic provides an "industrialization" function that enables not only process efficiency but also the comprehensive capture of clinical decisions and the assessment of quality and costs, which can then be iteratively improved; the availability of a robust clinical data set can also support the research mission of an academic medical center - or at least, can help those investigators who are able to obtain access to the data. This can accelerate progress in several key areas:
Evidence-Based Medicine : The ability to collect and easily share extensive medical information from the large patient populations served by major medical centers means that hospitals may be able to develop a robust dataset upon which to guide evidence-based decisions - they can start to segment and track both patients and physicians, and develop, and iteratively improve, best practices
Medical Research : Access to a coherent, longitudinal medical record will enable inquisitive physicians to generate hypothesis which can be explored in silico, providing an important opportunity to accelerate knowledge development and advance clinical science. These observations can also form the basis for both basic molecular research and rigorous prospective clinical studies. (See here for recent discussion of phenotype as medicine's next frontier.)
Importantly, these are two types of innovation AMCs are relatively good at today. Well structured systems like Epic will make them much better.