Mr. H says:

I’m not convinced Epic is changing strategy at all just because a couple of unnamed consultants speculated as such (Epic has always sold ambulatory-only deals), but if they are, I’d infer the opposite. Epic has not hit the predicted wall on scalability, customers keep giving its products industry-leading KLAS scores, nobody is de-installing or grumbling about value, and prospects keep signing up in droves despite high project costs.

First, I want to be clear that I have no evidence that Epic is refocusing on the ambulatory EMR market. I’m not saying that Epic has hit a wall, that its customers are not happy, or that they are grumbling about lack of value. My post was a speculative response to an HIStalk post about an alleged shift in Epic’s sales strategy. My intent was to say “If in fact there has been a shift, here are some potential reasons why a market leader would have made that decision.”

Mr. H continues:

Each time Epic sells an ambulatory-only deal, it (a) deprives a competitor of a new sale, and (b) plants a flag that has a decent percentage chance of yielding an easy inpatient sale down the road.

We’re in agreement on this point: Whether Epic considers its primary competitors to be high-end (Cerner, Allscripts, Siemens) or low-end (eCW, PracticeFusion, Amazing Charts), it makes sense to claim as many ambulatory customers as is economically feasible in the near term.

If anything, I suspect Epic is gaining confidence given the near absence of significant competition and is willing to ramp up sales, which by definition means they will be selling to smaller hospitals and practices.

The problem with this thinking is that in chasing smaller customers a vendor acknowledes saturation or diminishing returns in the upmarket segments. There is a fixed cost to landing a deal with any customer, plus the opportunity cost of the potential customers you aren’t chasing. It just doesn’t make sense to take 90% less revenue against those costs if you have better options on the table.

The company’s favorite statistics involve not the number of hospital customers it has, but rather the percentage of physicians and patients using its systems. I think they want that number to keep rising for reasons beyond financial, and any change in strategy can be attributed to unchallenged dominance rather than newfound desperation.

The interesting point here is that these percentage metrics tend to favor a vendor whose strategy is to pursue the largest customers, those who have the most patients/physicians/beds per hospital. A conscious decision to move downmarket means that each win moves these metrics less at the margin. And I don’t think I cast a downmarket move as “newfound desperation”. It could be a rational business decision based on feedback from the market under the four scenarios I originally described.

Finally, I’d caution against leaning too heavily on data from KLAS or other common sources to make inferences about current vendor strategies. The problem with most third-party sources of customer information is that they are only lagging indicators from a strategic perspective. A vendor makes a strategic decision, they operationalize the strategy, they close some deals, they do the implementation work, and 6-12 months later those new customers fill out some surveys. Cycle time: 3-5 years. These reports may contain clues about vendor strategies from 2008, but not from today.

From Barry: “Re: Epic. A market research report suggests that Epic is backing off its push for inpatient installations and going with an ambulatory-only sales approach to plant the seed for future inpatient sales.” That was reported by two consultants quoted in the report, with an additional consultant saying that Epic is getting some pushback from customers who question whether they’re getting their money’s worth.

Why might the market leader choose to pull back? A few possible reasons come to mind:

• They may have hit a capacity constraint. The critical path to revenue in this business runs straight through a long implementation process. Even if you win 100% of the deals, you have to find consultants to staff the implementation teams, and these are in short supply. Thank HITECH for moving the constraint from the health system capital purchase committee to the headhunter staffing the implementation team.
• They may have crossed the midpoint of their target market segment. The Bass diffusion model generally predicts that peak unit sales occur when 50% of the target market has adopted a product. From that point on unit sales decline year over year, or the total number of customers increases at a decreasing rate. If this is the case, it would make sense for Epic to ease up on the throttle a bit and focus their efforts on other segments.
• They may be choosing price over quantity. Epic built its reputation in the inpatient world by being very careful about choosing its early customers and maintaining an air of exclusivity. An exclusive, customer-intimacy-driven business model does well in the left end of the demand curve where the customer’s willingness to pay is sufficient to fund the attention the vendor is willing to give. But a sufficient degree of success saturates the high end of the market and forces the vendor to look lower in the demand curve for new business. Epic may be taking a look at the margins of that business, as well as other brand-related factors, and simply saying “No thanks”.
• They may be recognizing the disruptive threat posed by low-end competitors in the ambulatory market. Epic is winning because of their inpatient/ambulatory integration. They started in the low end of the market, the ambulatory side, and grew the capabilities of their offering to the point where they could begin to compete in the inpatient world, and then they began having great success displacing the incumbents. Established inpatient vendors have struggled to cram their solutions down into the lower-end ambulatory market. This dynamic typifies what Clay Christensen has termed “low-end disruption”, that low-end players are more likely to evolve and displace high-end players than vice versa. If Epic is smart, they’ll recognize that a free or very-low-cost ambulatory EMR like eClinicalWorks or PracticeFusion could evolve and do to Epic what Epic did to Cerner, Eclipsys, and McKesson.

Time will tell whether Epic really refocuses its efforts on the ambulatory market. Even if it does, chances are that in order to fend off a disruptive threat by one of today’s low-end players, Epic will need to tweak its business model to enable SaaS delivery of its products at a very low cost with little to no configuration required.

I’ve had twenty or so conversations with former HSG colleagues over the past week. Some of these folks were affected by the layoff, some were not. Most of the salespeople seem to have been unaffected. Between last week’s layoff and the attrition leading up to it, about half of the marketing and product management team is gone. I’ve heard reports that most of the physicians were affected and that up to half of the engineering team was let go. By my rough calculation, that totals 25-35% of the overall headcount at HSG. 

It’s hard to describe how it feels to hear this news. These are all extremely smart, passionate, hard-working people. Microsoft simply doesn’t hire people who don’t clear that bar. My message to each of those affected has been something along the lines of: This turn of events is not a consequence of your abilities or your work. Now is a good time to be looking for a job in this industry. Contact me if you need help networking.

On the topic of NewCo, it will be interesting to see how those invited to participate undertake their decision process over the next few weeks. I wonder how the merger of GE and HSG cultures will happen, or whether it will happen at all. HSG grew rapidly by acquisition and a pervasive culture never really had a chance to coalesce. The Azyxxi people were still Azyxxi people, and the same was true for the GCS and Sentillion people as well as the team that built HealthVault. Perhaps the emergence of a strong strategy to tie these assets together was a prerequisite for a strong culture to evolve as the teams came together to achieve a common goal.

An interesting litmus test: Will there be equity participation for the rank-and-file NewCo employees? Microsoft traditionally prides itself on having all employees participate in stock ownership, GE typically reserves equity grants for executives. Perhaps the answer to this question will foretell which way the rest of the NewCo culture will lean.

In the discussion that followed this announcement, several themes have emerged:

• This change is part of a broader initiative.
• ICD-10 and Meaningful Use precipitated a decision to rationalize the portfolio.
• The migration from Horizon’s Oracle platform to Paragon’s Microsoft platform will reduce customers’ TCO.

All of these reasons for the change make sense at first glance. But here’s what’s not being said:

• Epic is mopping the floor with these guys.
• Epic is winning on inpatient/ambulatory integration, not clinical/financial integration.
• Epic doesn’t seem to be bothered by its TCO problem.

My conjecture is that the MPT executive team is making a tactical retreat from the bloodbath it’s been enduring in Epic’s target market and is attempting to rally around Paragon, which is thriving in the sub-300-bed segment. These customers appear to be a bit more price sensitive, a bit less interested in following the herd, and a bit less dismissive of loose inpatient/ambulatory integration. If true, this approach by McKesson seems like it would have a higher likelihood of success than would standing toe-to-toe and trading punches with the champ.

A final data point that may add weight to this argument: At last week’s ASHP show in New Orleans, McKesson commanded a gargantuan booth on the main aisle in the exhibit hall. Immediately under the company name and logo on the front of the booth was a mock-up of a hospital outpatient pharmacy. The prominent signage advertising the solution? “COMMUNITY HOSPITALS”

Humedica’s offering has a value proposition is similar to that of analytics offerings from Microsoft and dbMotion: Extract data from transactional systems and make it “liquid”, apply some semantics to it, and allow users to navigate it and turn it into insight.

What appears to be different, from my outsiders perspective:

• The solution appears to be delivered via a SaaS model that requires no hardware or software purchases by the customer.
• The business model appears to be free to providers, with the revenue coming from the provision of de-identified data to researchers (pharma, etc).

While these differentiators may appear to be straightforward, they will have an important effect on the reach of Humedica’s business.

The value proposition for noodling on one’s data has near-universal appeal, but there are two classes of buyers: Those who “get it” and are willing to pay handsomely for tools that allow them to supercharge their existing efforts at process improvement, disease management, research, etc.; and those who don’t “get it” but just think it’s cool to be able to do these type of things. The latter outnumber the former by about 10:1 in the provider space, meaning that the majority of the provider market for this type of offering is extremely price sensitive.

Humedica has the potential to navigate around this problem by making analytics effectively free to the provider in both time and money. And if they can build all the internal plumbing for each provider on the dime of research sponsors in pharma etc., they should be able to turn around and monetize that asset by delivering additional use cases at extremely low cost and charging handsome fees to providers and payers.

I’ve ranted about business models before, it’s nice to see some smart people who seem to have done the right thing by delivering an innovative product and service with a commensurately innovative business model.

Klein quotes a book called “Prevention vs. Treatment” edited by Haley Faust and Paul Menzel, specifically a chapter written by Louise Russell who in turn “draws heavily on research” done by Joshua Cohen, who along with three co-authors appears to have rolled up evidence from “hundreds of studies over the past four decades”.

(How many times can one person’s original work be repackaged by others? Good grief!)

The centerpiece of Klein’s post seems to be the following chart, originating from Cohen et al:

The chart specifically states that it shows a distribution of cost-effectiveness ratios of various prevention and treatment strategies for different disease states, presumably “hundreds” of them. It absolutely does not show pairwise comparisons between specific treatment and prevention strategies as substitute goods within the context of particular disease states. Perhaps one of the editor-authors misunderstood the rather obvious symmetry between the prevention and treatment distributions to imply that prevention and treatment are roughly cost-equivalent. Oops.

Full disclosure: I have not read Cohen’s original work, nor any of the derivative works mentioned above. But from what I have inferred about the study design (a meta-analysis of published QALY data from hundreds of different prevention and treatment studies over four decades), it’s simply impossible to use such a study as the rational basis of an argument about the relative merits of treatment vs. prevention in general.

The debate over healthcare policy is complex enough without having to sift through heavily-repackaged summaries of summaries of other people’s research. Didn’t we learn enough about the perils of repackaging from subprime CDOs? My gut tells me this one doesn’t deserve a AAA rating.

At the end of the day — I want to assure you that Microsoft — and I — remain 100% committed to the HealthVault business. I believe that together we’re creating a health system for the 21^st century, and I’m proud to be sharing that journey with you. Our new venture with GE will simply add to our ability to deliver.

It’s no surprise that Sean is in the lifeboat. Microsoft generally doesn’t send DEs packing.

What I find surprising is the assertion that HealthVault is a business. Technology? Yes. Product? Okay. Ecosystem? Check.

Businesses generally have a business model, one that generates revenue, and hopefully profit. Microsoft has tried a number of different business models with HealthVault:

• An ad-funded model. Couldn’t attract enough eyeballs to make this work. (Update and correction: Thanks to Sean for reminding me in his comment below that this model was for MedStory, a separate health search business that was folded into Bing after HSG originally acquired it. HealthVault does not and has never attempted to monetize its users via advertising.)
• Selling the HealthVault platform to telecom companies to develop healthcare apps for their own subscribers. Didn’t scale beyond a couple of customers in Canada and Europe.
• Selling HealthVault connectivity to Amalga customers through HealthVault Community Connect. The economics of Amalga’s long install process killed this one.

None of these attempts at a business model produced much revenue or any profit.

The problem is that Microsoft is not very good at business model innovation. The company generally tries to follow a sustaining innovation approach of extending existing business models with incremental features and new product releases. When it builds new businesses, it is most successful when it can copy the business model of a competitor. Neither of these options apply to HealthVault. The product has struggled largely because the company was unable to craft a unique business model tailored to its value proposition, preferring instead to follow an existing model from either inside or outside the company.

The most ironic thing about Sean’s note to HealthVault partners is his invocation of Clay Christensen’s work. If Microsoft understood anything at all about disruptive innovation, it would understand that great technology in a legacy business model is sustaining, not disruptive innovation.

I’m not surprised by the deal, it has been in the works since before I left Microsoft. The two companies are very well connected: Ballmer and Immelt are old friends, they have existing JVs (e.g. MSNBC), and many of Microsoft’s HSG people are GE Healthcare alumni.

The deal makes sense because:

• Both Microsoft and GE have struggled to come up with innovative product strategies that scale beyond early adopter customers.
• Neither company’s culture is conducive to innovation in health IT. GE is an expert at getting into any business where it can dominate on a global scale, and dominating health IT on a global scale is harder than winning a land war in Asia. Microsoft is an expert at being a fast follower in businesses that leverage its dominance in the desktop OS market, but it chose to try to create a new product category rather than follow a health IT incumbent. Both lack the organizational structure and the right processes, people, and incentives to take a risk on something new and innovative and really disrupt the industry.
• Disruptive innovation tends to emerge at the low end of the market, and giants like GE and Microsoft are not well-suited to start new businesses serving only a few customers that they have never heard of and generally don’t care about.
• These two companies burned well over a billion dollars between them over the past 5 years trying to solve very similar problems in the same market. If you are Ballmer or Immelt, teaming up with your trusted friend to cut your collective burn rate in half seems like a no-brainer.

The deal concerns me because:

• As one of my mentors at Microsoft once said: “Two limping men don’t make a running one.”
• Microsoft had assembled quite a health IT brain trust over the past five years, and I hear rumors that this spin-out was accompanied by a significant layoff. I’m sure the same is true for GE. Still trying to get in touch with a few insiders to learn more.
• HealthVault is not included in the JV. Will it serve as anything more than a lifeboat for key people who Microsoft wants to retain?

My hope is that this new entity will be able to transcend the barriers to innovation that its people faced inside of Microsoft and GE. A smaller, stand-alone company may be a better fit for the risk tolerance required to disrupt the market. But the new management team will have to work to counter the brain drain that has preceded this announcement, that has accompanied it, and that may yet happen as the lifeboats fill up.

I left Microsoft a few months ago and went to work for CareFusion. My new work environment has provided perspective on some new topics that are worthy of treatment in this forum. Stay tuned!

The term moral hazard is generally used within the insurance industry to describe the tendency of the insured to increase their risk tolerance once their risks are hedged by insurance. For example, a consumer with a homeowner’s insurance policy might decide against returning home to double-check that she had turned off the stove or locked the door before she left the house. Or a consumer with a comprehensive auto insurance policy might not worry so much about parking his car in a bad neighborhood at night. In general, reducing people’s exposure to risk seems to increase their tolerance for risk, and the corresponding increase in risky behavior is what constitutes the moral hazard.

The concept extends to other domains. Professor Kent Smetters of The Wharton School taught that US Department of Transportation rules requiring seat belts and airbags in new vehicles have led to increasingly-reckless driving as drivers have felt more secure. And while driver and passenger fatalities have diminished due to the direct effects of these safety devices in the years since these regulations took effect, fatalities of motorcyclists, bicyclists, and pedestrians have increased proportionately with drivers’ reckless behavior during the same time period. Smetters joked that the most effective automobile safety device may not be a seat belt or an airbag, but rather a large dagger mounted on the steering wheel and aimed directly at the driver’s heart, as it would give the driver and any pedestrian roughly equal odds of dying in a minor collision!

Recent events in hospitals illustrate how moral hazard applies to healthcare technology: