Epic is famous for its intense focus on interoperability across its own systems coupled with its conservatism regarding interoperability with other EMRs. In 2012, KLAS said Epic has the “deepest data sharing of all the vendors” across its own practice and hospital EMRs (see this example in which Cleveland Clinic and neighboring system MetroHealth — both on Epic — have put interoperability in place). But when it comes to non-Epic systems, customers must work through defined “exits” to the Epic system (“we don’t let anyone write on top of our platform” per Epic’s CEO) and interoperability – where it has been put in place – seems to be the result of specific customer request (and presumably specific customer expense)– rather than built into systems. (David Shaywitz and I have commented on the potential system consequences of Epic’s strategy and success here).
In this context, the recent eClinicalWorks-Epic announcement looks very innovative. The partners promise real time data transfer between systems configurable in the software rather than custom-engineered for a particular customer (such as the Greenway/Lancaster deal). Of course, it is hard to know what reality will emerge after the press release. Let’s assume, however, the deal will yield genuine interoperability (e.g. sharing cross-organizational referral management – something Epic has explicitly said occurs only if both organizations have Epic). Under what conditions would this signal a fundamental change in strategy or just a change in tactics?
Why a walled garden?
What makes walled gardens attractive strategically (quite distinct from the potential clinical impact – see interview with Epic CEO here on safety and IP concerns)?
For Epic: Suppose we have a delivery system with multiple legacy EMRs across sites and the flagship hospital considering an Epic implementation. The less interoperable Epic is, the more likely the delivery system will opt for total replacement across all sites; the more easily interoperable it is, the more likely the flagship might buy Epic for itself and allow legacy sites to stick with their current systems. Thus: less cross-system interoperability = more sales.
For Epic customers: CIOs will tell you “easy interoperability” is an oxymoron. EMR vendors plugging into each other create potential faults and complexity which no one vendor owns. Also, the more easily patient records are shared outside the organization, the easier patients and plan sponsors can cherry-pick services. As Bruce Friedman and David Shaywitz have noted, less interoperability = less business risk and preserved patient “share of wallet”. (And if the vendor can take the “heat” for the walls, all the better).
This implies three key assumptions for the walled garden approach to be attractive are:
- You are the EMR market leader among systems/hospitals (something not going away for Epic anytime soon)
- Interfaces with other EMRs are expensive, unreliable and of questionable strategic value for the flagship
- Targeted system/hospital CEOs are able to persuade the entire system (including ambulatory) to adopt their preferred EMR.
Given this, how can we interpret the eClinicalWorks-Epic move?
Scenario 1: Tweaking the value proposition to move down-market
Having largely exhausted the large hospital segment, Epic has been
moving down-market. By some counts, there
are about 400-500 hospitals operating on legacy systems or have made no EMRcommitment and are therefore sales targets.
This tail of the market is likely smaller hospitals (e.g. <300 beds),
not part of an integrated system and dependent on an array of independent and
fragmented practices for their referrals.
The ability for the target CIO therefore to influence collaborating clinicians
will be very limited (breaking assumption #3 above), sharply reducing the value
of reduced interoperability. In fact,
interoperability between the hospital system and the referring providers could
be a strategic advantage for the hospital.
To meet the needs of these customers, therefore, Epic needs to
find ways to enhance its cross-system interoperability (particularly with
referral management into the hospital system) without placing a large
incremental cost burden on its customers.
eClinicalWorks is the logical place to start the process. Per some studies, eClinicalWorks has the
strongest share among ambulatory practices, and is highly regarded among physicians (specialist and primary) and
is rated particularly strongly for smaller practices given the price point. It also is a purely ambulatory solution and
has stayed neutral toward the Commonwell Alliance which means it does not threaten Epic on the inpatient side.
Under this scenario, Epic made this deal to tweak its value
proposition for the smaller, less integrated portion of the market. It may cut a few additional deals with
ambulatory-focused EMRs or it may stop with eClinicalWorks given that this is a
game of 80/20 (remember the CIO anxious about too many interfaces) for a
targeted set of customers. For the
broader Epic installed base, however, walled gardens will remain in place.
Scenario 2: Shift in strategy to
address competitive pressure
An alternative view is that Epic sees the writing on the wall
regarding cross- EMR interfaces. Potential
signposts are:
- Launch
of Commonwell (the interoperability
alliance among Epic competitors Cerner, Allscripts, McKesson and others
announced at HIMSS this year) - Improvements
in the ability of vendors offering private health information exchanges services
to lower the costs and increase the reliability of exchanging data across EMRs. - Narrowing
of the sales advantage of Epic vs. the more “interoperably-minded” Cerner (from
the extraordinary 5-1 lead in sales in 2010 to simply impressive 2-1 lead in sales in 2012. - Other
business drivers making interoperability more important for core customers
(e.g., expanding penumbras of clinical and business collaboration between core
Epic clients and non-Epic sites; mobile patient engagement tools which pull
from multiple EHRs such as eClinicalWorks Healow)
Under this scenario, the second assumption for walled gardens –
the expensive and limited strategic value of interoperability – is under
attack. Epic’s move with eClinincalWorks
would represent, therefore, the beginning of a longer term effort to a more open
strategy through many more deals to build interfaces into the base software.
Scenario 3: The walls are fine, the
garden just needs to get bigger
Perhaps the deal represents no real change in strategic
direction. All the assumptions behind
walled gardens still apply. However, under
this scenario, Epic has taken close stock of its clients and found this:
- Many systems
are affiliating with more and smaller practices - These practices do not have Epic
- The practices cannot afford the Epic ambulatory price point and the system is unwilling to subsidize
Perhaps these have eClinicalWorks already in
place, or, eClinicalWorks has a tolerable price point for an upgrade. Either way, tighter, easier integration with
Epic offers an attractive (and perhaps interim) hybrid model for the flagship
customer and a lot of comfort for the CIOs.
In this case, I would expect longer term to see Epic tightening its grip
on eClinicalWorks as it becomes an more critical part of serving these hybrid systems, perhaps acquiring it altogether (ambulatory practices are
perhaps still fragmented enough for such an acquisition to pass regulators
especially given Commonwell). The
garden, then, is expanded, but remains walled.
I don’t know which strategic scenario Epic thinks it is in as the
eClinicalWorks deal makes sense as a starting point for either frame. We will know more as we start to see what
other deals Epic strikes and – especially – if the players on the other side of
these deals have EMR footprints with hospitals vs. just ambulatory. Interestingly, Epic may not know itself which
scenario it is in either. The best way
to deal with strategic uncertainty is to buy options and the deal with
eClinicalWorks is just that: more time to evaluate whether its strategy is
right as it is, needs a tweak or a real change of heart.