Halamka Speaks: athenahealth & the Future of AMCs as Tech Innovators

It’s always interesting to talk with John Halamka, and last week–after athenahealth bought the IP but apparently not the actual code of the Beth Israel Deaconess Medical Center (BIDMC) web-based EHR he’s been shepherding for the past 18 years–I got him on the record for a few minutes. We started on the new deal but given that had already been covered pretty well elsewhere we didn’t really stay there. More fun that way–Matthew Holt

Matthew Holt: The guys across town (Partners) ripped out all the stuff they’ve been building and integrating for the last 30 years and they decided to pay Judy Faulkner over a billion dollars. And you took all the stuff that you’ve been building for the past 15 to 20 years and sold it to Jonathan Bush for money.  Does that make you a better businessman than they are?

 (Update Note 2/11/15: While I’ve heard from public & private sources that the cost of the Partners project will be between $700m and $1.4 billion, Carl Dvorak at Epic asked me to point out less than 10% of the cost goes to Epic for their fees/license. The rest I assume is external and internal salaries for implementation costs, and of course it’s possible that many of those costs would exist even if Partners kept its previous IT systems).

John Halamka:  Well, that is hard to say, but I can tell you that smart people in Boston created all these very early systems back in the 1980s. On one hand, the John Glaser group created a client server front end. I joined Beth Israel Deaconess in 1996 and we created an entirely web-based front end. We have common roots but a different path.

It wasn’t so much that I did this because of a business deal. As I wrote in my blog, there is no benefit to me or to my staff. There are no royalty streams or anything like that.  But sure, Beth Israel Deaconess receives a cash payment from Athena. But important to me is that the idea of a cloud-hosted service which is what we’ve been running at Beth Israel Deaconess since the late ’90s hopefully will now spread to more organizations across the country. And what better honor for a Harvard faculty member than to see the work of the team go to more people across the country?

MH: There’s been a lot of debate about the concept of developing for the new world of healthcare using client server technology that has been changed to “sort of” fit the integrated delivery systems over the last 10 years, primarily by Epic but also Cerner and others. In particular how open those systems are and how able they are to migrate to new technology. You’ve obviously seen both sides, you’re obviously been building a different version than that.  And a lot of this is obviously about plugging in other tools, other technologies to do things that were never really envisaged back in 1998. You’ve come down pretty strongly on the web-based side of this, but what’s your sense for how likely it is that what has happened over the last five or ten years in most other systems including the one across the street we just mentioned is going to change to something more that looks more like what you had at Beth Israel Deaconess?

John Halamka: My belief, and I think it’s probably one that you and many of your Health 2.0 folks share, is that we should have an app store for health–That we should have those functions that are the best based on value inserted into our workflows. And if that’s two or three different vendors, that’s fine.

I understand that in the status quo the reason that a monolithic vendor makes sense is that if you want to really beautifully document an encounter within a single healthcare system, it probably works pretty well. But the future belongs to continuous wellness, not episodic sickness.So having an app on your iPhone for the patient and having multiple apps for data visualizations and care management and unique dashboards of event-driven medicine, that’s the sort of the future. So yes, I hear that based on this FHIR work we’re doing through the Argonaut initiative that you’ll see APIs coming out of MEDITECH and Cerner and Epic and McKesson and others, and it might foster an ecosystem. But what you sort of hope comes out of this Athena work is that that ecosystem of web-based services is more rapidly accelerated.

MH: Well, that’s the other half of the question, right? So you talked about FHIR and Argonaut and these other initiatives.  Well, many people might be cynical about the fact that it’s another one of these things–we’ve heard a lot about interoperability before and we’ve heard a lot about meaningful use and getting stuff out of the patient record with Blue Button and Direct.  It doesn’t seem to have happened yet.  Is your sense that the way this is going to be sped up is with this new arrangement — but and I saw Jonathan Bush quoted — it’s going to be five or six years before this thing becomes available for the commercial market.  Won’t those other guys have caught up by then?

John Halamka: It could be.  Yeah, that’s what I put in my blog, let the best man win. But if you look at what issue I’m facing today, Beth Israel Deaconess like many organizations is merging and acquiring other entities.  Sometimes it’s a hospital, sometimes it’s a primary care practice.  As a CIO I’m told, “Oh, by the way, in two weeks we’re acquiring X, Y, and Z.  You’ll have all the IT systems ready, won’t you?”

And of course, we’re not looking at 18-month lead times. We’re looking at turning functionality on and off in 18 days. So you ask this sort of question. If the business driver, the ability to handle a merger and acquisition in a subscription kind of fashion, can the existent vendors really thrive in such an environment? You hope that as new products enter the marketplace and organizations are shown that IT doesn’t have to be this capital-intensive lengthy time delay but can be a service that you add almost analogous to the way you had an app to your phone, that it has commercial demand and different approach.

And sure, Carl Dvorak, who is a lovely man, will probably say they are developing a data center in Verona, Wisconsin and they will provide web-hosted Epic in the next couple of years. Could be.  And so, as I said in my blog, as we have a license back to our own technology for 20 years, we will continue to innovate and build. But we will also look to migrate to vendor products when they are good enough. And if it’s MEDITECH or Epic or Cerner or Athenahealth, we’ll just watch who the agile service-oriented web mobile is and go with them.

MH: You put out in 1999 one of the very first PHR tethered web-based access virtual medical records for the patients? And we’re now in 2015, I’m still giving talks about how impossible it is to get my kid’s data out of the various practices and systems that he has gone around in his short life. And you hear much, much worse stories from people who have really serious illnesses. Where do you think we are?

Oh, and just to give you some commentary.  I was railing and ranting about this at a meeting the Robert Wood Johnson Foundation had out here and Roni Zeiger said, “That’s great, Matthew, you go and you march to Washington about this lack of a patient access to data. You’ll be one of five people.”  Where do you think we are in the process of getting easy patient access to most of that data?  And to tie back to what you just said, is this type of activity that you’re doing with athena going to help?

John Halamka: Let me give you a couple of answers to that. As you know, my wife was diagnosed with Stage 3A breast cancer in December of 2011. And she managed 35 rounds of chemotherapy, radiation, and surgery because she had access to 100% of her records.  Because as you’ve seen from OpenNotes and such things we’ve done, the patient and the doctor actually see the exact the same record.  There is no idea of an EHR to PHR.  It’s a shared medical record. She told me she could not have made it through her breast cancer treatment without that seamless access to 100% of the data. Her care preferences and care plans were jointly developed and mutually visible. So I think you’re seeing huge patient demand for this sort of thing. And there are some enablers, meaningful use Stage 2.  Although as you know, our eligible professionals haven’t done that so well, but it does have that requirement for that sort of shared medical record.  And I really like the idea of Apple Health Kit which is a middleware application to take data from all of the devices in your home and share them as you wish with apps that might include a care manager or your doctor.

So you know, there will be in another interview, probably, we can do in a few weeks.  Beth Israel Deaconess has finished an app that will pilot shortly. It will sit on your iOS device and allow you or enable you to say, “Yeah, share my scale. Share my blood pressure. Share my glucometer with the Beth Israel Deaconess doctors.” It’s all coming, and the patients are driving it. So that person who told you five people, are you kidding me?  Have you had a sick person in your household recently?

MH: Just to defend Roni a little bit, as you know, he has these scars with the Google Health experience and there were many reasons for that failure. And secondly, there hasn’t been overwhelming consumer demand–apart from where people can get it easily and understand it–to break down the extra barriers that get put in your way.  And if you are ePatient Dave or in your wife’s situation using a system that’s using OpenNotes, that’s a bit different from, frankly, the experience that 98% of Americans go through.  Most of them, I think, kind of give up rather than fight for it and we’ve got to make it much easier.  I’m on your side, I think it’s much more important than some people think it is, anyway.

John Halamka: Roni is a great friend  And you know why Google Health had its struggles. There were two reasons.  They were ahead of their time because you hadn’t yet had accountable care organizations in wellness, care management and all the rest that patients cared about. But second, I hear — I mean Roni would tell you better — that Larry and Sergey didn’t really like the idea of being a covered entity or at least auditable so that you would have to disclose how Google uses its data. And that’s the reasons they really killed off their PHR.

MH:Yeah, I heard the latter as well. And also, I think frankly, the other thing was it was one of the first things Larry got to do when he was made CEO. And when you’re a new CEO you have to do something decisive.  I actually think it might be as simple as that, but maybe I’m being too simplistic.

But I kind of agree with Roni that that hasn’t been the kind of level of consumer outrage.  And frankly, personally, I only found out when Jacob Reider told me at our Health 2.0 meeting that “view, download and transmit” was actually “view, download OR transmit” for the providers so long as you can view. So there was no real reason why UCSF or Sutter in my case actually had to put the Blue Button up there on my chart so long as I could see whatever limited amount of stuff they were letting me view, which makes me feel that we have a long way to go.

John Halamka: It’s funny. As you know, because I get to be that federal advisory committee guy making recommendations, anytime we put our “or” in regulations, it actually becomes an “and” for certification.  That means Sutter must actually have the transmit capability, they just chose not to implement it.

MH: And it gets even more fun, right?  I met Paul Tang on a rental car shuttle.  So I was whining about my experience at Cal Pacific, and he said — well, he’s also part of Sutter, although a rather special part of Sutter at Palo Alto Medical Foundation.  He said, “We’ve done Blue Button.”  I said, “That’s great you have 85% consumer usage or whatever in Palo Alto?”  But my kid goes to a clinic that’s in the poor part of San Francisco and consumer access isn’t  1%, in fact I think I’m the one person who actually has got proxy access to their Epic record and there was no way they’re putting Blue Button up there.

John Halamka: But to finish answering your question, and I don’t endorse any vendor, I’ve got to be very careful of that. But the fact that athena has its More Disruption Please program and open APIs means there actually should be vendor third-party apps that allow you to seamlessly download your record and visualize your data and that should be the norm.

MH:  Let’s hope we get there shortly. Last question about the athenahealth folks and where they’re heading. Now they’ve got something that works that is very, very big, medical center.  They just bought the RazorInsights guys, sort of like the Clayton Christensen disruption model of someone selling a just good enough model to smaller hospitals. Do you think for them, their panoply of inpatient and outpatient services, is this enough that they can build on or do you think there’s a hole in the market that they need to have a MEDITECH-like system for the middle market somewhere as well?  What’s your guess about that?

John Halamka: It’s really interesting to watch the next 18 months as Athena builds products that initially will alpha test by layering into our MEDITECH systems that exist in our community hospitals. Our alpha tests will not displace MEDITECH, it will add to it. And so, there’s a huge amount that Athena could do as you know. I mean whether it’s in patient engagement, care coordination, care management, analytics, they could simply layer into and on top of existent incumbent systems. And you could imagine that over time, it’s a bit like the Epic strategy.  They will continue to develop modules and capability. And at some future time, they might be able to displace those systems.

Athena would better answer this question than I, but I would imagine in the short term, you’re building up functionality to target the critical access hospitals and the smaller organizations that either have no IT or not good enough IT. And then you move to the gap-filling, and then eventually, a self-sustaining complete application. And that might take five years. But the idea that it’s web, that it’s mobile and that it’s available in a multi-tenancy model to everyone with the same version at the same time with a focus on APIs and interoperability certainly seems like a good direction to me.

MH: Two last questions, mostly about you. Do you think that the role of medical centers, which as we started this conversation with, has been in part to build a lot of information technology which in some ways would be like a large corporation building its own IT and not using technology vendors, do you think that that role will continue or do you think that your move now and the move of Partners and Mayo go to an outside vendor is a signal of that that’s going to stop? And then coda to that is what are you going to be doing next?

John Halamka: Sure.  I have told my staff that I believe in five years it won’t be the IT department, it would be the cloud services department. That is our job will be to empower our users with services.  And we might provide some of them, but probably, we’ll procure them. That is sort of the nature. Does anyone really want to run an email server anymore?  Not really, right? I mean you should be able to as long as the quality is high and the cost is competitive, to find a service provider that will give your users document sharing, email, social media that’s secure, and EHR. I mean that’s really where we should be heading.

And so, my role, whether it’s standards committee role, the Argonaut role, the HIE role, or the CIO role, is how do I empower that transformation that healthcare IT across this country recognizes that it’s an app store.  And whether it’s standards that needs be filled, philosophy and policies that needs to be changed or an alpha site that needs prove it’s possible, that’s my next big thing.

4 replies »