Q&A: Roger Baker on the future of VistA and VLER


WILLIAMSBURG, Va. - Public debate over the electronic road ahead for the Veterans Affairs Department is often informed by a series of high-profile information technology project failures at the department. VA Chief Information Officer Roger Baker has told Congress and other officials the VA is changing its ways for the better.

FierceGovernmentIT caught up with Baker at the Executive Leadership Conference in Williamsburg, Va. on Oct. 24; below is a transcript of the interview.

FGIT: Where do things stand right now with the VistA replacement process?

Baker: So, let's be clear, in my view, VA over the last 10 years has tried to replace VistA. I don't think that's possible. It would be like Microsoft (NASDAQ: MSFT) trying to replace Windows with not an evolutionary product, but with something brand new, but it has to come out and it has to be better the day it's introduced. That, basically, was the criteria for what VA was trying to do. That program was called HealtheVet. I have stepped VA away from HealtheVet, and what we're now looking at is how do we continue the evolution of VistA.

It is the best electronic health record system in the United States, at this point, especially if you focus on it from a patient-care standpoint. So, how do we then get back to moving the innovation forward in VistA, and that's really what the whole open source campaign is all about.

Medical records systems have moved forward a tremendous amount, in the United States, since the time that VistA was started. And the private sector is doing a lot of stuff that we need to be able to incorporate into VistA. So, our thought is that by being part of an open source community based around VistA, the VA can encourage private sector folks to either directly contribute the open source--you know, make improvements. Or integrate their products with the open source, so we can very easily buy a working product, instead of having to go down the government route.

The reason that, I believe we've got to go the open source route, is that we have two important projects to integrate private sector packages into VistA going on inside the government right now--one is for laboratory and one is for pharmacy. Both of those projects are going on five years, to integrate the private sector product into VistA because we're doing it the government way.

That is far too long. We need to be able to go out and say, I'm interested in a pharmacy package, in six months I'm going to buy one that I prefer, from all the ones integrated with the open source--let's go. And when an organization like VA says it's going to buy, that could be 200 or 300 million dollars. So, you know generating the private-sector interest in it.

I just think we're going to move VistA innovation forward much more quickly if we go the open source route.

FGIT: MUMPS has been very well received from a care perspective, and as a medical informatics tool. But it's been criticized for not being very interoperable. Would a move toward open source mean a deviation from MUMPS, so something more standard could be used by open source developers?

Baker: We have 15 million lines of MUMPS code. It's unlikely that it will, in bulk, get translated into something else. But what's likely to occur is that an open source community will take a look at that and decide that there are certain things in place they'd like to improve the functionality. So, they'll ask: 'Do I want to learn MUMPS and rewrite it in MUMPS so that I have reusability? Or do I want to just use C++, or some language that I prefer to build a new version of that module?'

I mean, any language can be interfaced with other languages; there are a variety of techniques for doing that. And the community may focus on, for the core pieces, let's create wrappers that do all the hard work for the interface and make it easy for the co-writers to interface with that model. There are a variety of approaches.

One of the interesting things I've found is, this whole argument about: To MUMPS or not to MUMPS, that I've heard from folks, is indicative of folks that don't know computer languages well. There are at least hundreds if not thousands out there and in the end, they all speak in ones and zeros.

It's a question of the mode that you want to express the logic in, from a human perspective, is what the language is all about. MUMPS has been proven especially effective, as you said, on the medical side and interestingly enough, also on the financial side and the airline side.

So, there's a lot of good legacy, well-proven pieces of software out there that are written in MUMPS. In the end, I think your question is exact and the most pertinent one, both from putting new innovation into the open source platform and I don't think MUMPS is going to be the decision piece there. The question is going to be, do they see their work contributing to either the VA or the national, electronic health records debate. That's what's really going to be a good decision piece in there.

FGIT: What does open source really bring to the table? You were talking about how it will benefit development and the private sector. It'll help vendors, definitely. How does open source benefit clinicians?

Baker: Three pieces of it. The easy one is much broader availability of commercial products for use on top of the VistA platform.

Goal number two is the way VistA was developed was with the clinicians in the loop. One of the things that has slowed down innovation is that about 10 years ago, VA centralized development of the VistA applications in a software development group.

It kind of pushed the clinicians out of the future of VistA. This will allow them to get back into it and contribute their own work into the open source for consideration by the governance board.

The third piece and the part, if you will, is in the nirvana of how successful this could be is if clinicians across the country start using VistA as a way of discussing what's the right process, what's the best evidence-based process for us to follow is treatment of some disease, or some human issue, then that will have achieved the ultimate.

If you looked at VistA the way IT people look at software development processes, VistA could be a CMMI Level 5 type process. It really is a learning system, and we're constantly improving its processes by feedback from the folks who are managing those processes.

If we can do that not just at VA, but let's say across government or nationally, even if hospitals aren't using VistA, if we end us saying 'Okay, we've agreed on process X and exemplified in the VistA open source,' and the private sector incorporate that into their product, that's a great solution. That means we're getting optimal care studies and analyzed across the country, argued by clinicians across the country, based on evidence across the country and figuring out what the  right approach to care is. I think that's when we might make a real leap forward in the result of patient care.

FGIT: Speaking of VistA as an evolution, what are the inherently good aspects of VistA you want to be sure to maintain in this next iteration?

Baker: There's one fundamental goodness in there, and that is that VistA was designed by the clinicians, for the clinicians. It is an example of how doctors want to work with an electronic medical record.

As far as I'm concerned, probably everything else is up for grabs--you know, as alternatives, making changes. But, if we fail to incorporate that, if we turn this into an IT system built by IT people, then it will absolutely out in the general marketplace. If it retains its basic goodness of it's first a medical system and it's second an IT system, then we will have succeeded in where we need to go with it.

FGIT: There's been a lot of talk about AHLTA in DoD, and that overhaul--is that something you're taking notes from?

Baker: We're working very closely with the DoD on how do we have a joint way forward for the modernization of the two electronic health record systems. The missions are somewhat different. VA does not have a requirement on the warfighter front. DoD doesn't have a requirement on the elder-care front.

But, by and large, 80 to 90 percent of what the systems do is exactly the same thing. There's a real argument for us to be working together, because it'll provide better care through the lifetime of the service member and veteran--and because it just makes financial, logical sense. But in the end, we both have to meet our missions.

And so, while we want to have as much commonality as possible, I'm not having a presumptions about how that's going to end up. We're just going to keep working together.

FGIT: Switching gears away from VistA, I'm curious what's going on with personal health records and measured use of that. Is it just on Blue Button downloads that you're basing that on?

Baker: The VA's version of personal health record is the MyHealtheVet website. Over a million veterans register on that. They can see a lot of their medical record on that. We're just continuing to add new things they can see of their medical record directly out of the VistA system. It's also a personal health record. They can upload lots of information on their own, so that it's available both to them and their doctor.

What the Blue Button does is take everything that's on the MyHealtheVet website and turn it into an ACSII file that allows you to download it. The best part of that is the number of private sector organizations that are not saying 'We're going to take that download, and turn it into interesting things that either the patient or the doctor can do with that information.' It's a fundamental example of what can happen when you free the information.

A patient downloading an ACSII text file of their information doesn't sound really interesting. But when another application developer can pick that up and do interesting things with it, that can turn into something amazing. From our standpoint, we're not going to develop those apps, that's not our specialty. But allowing that download should turn into a very powerful application.

FGIT: Where does Virtual Lifetime Electronic Record fit into this?

Baker: The way to look at it is that VLER is the umbrella for all of this. The president's vision and the two secretaries vision is that from the moment a service member raises his or her hand to swear in, to the moment when they're honored in National Cemetery, there is one record of their information. At the point of service, when a service provider has a service to provide them, they're able to access all information that could be pertinent to that.

One of the things we recognized very early on--we started with the philosophy of what the president didn't say is as important as what the president did say. It's not for some veterans, it's not for only active duty, it's not for only DoD and VA, it includes private sector.

When we look at that virtual lifetime electronic record, the word "lifetime" takes on a very strong meaning. You can look at all of these things we're doing--working with the DoD on electronic health records, bringing information over and strengthening commonality, getting rid of the chasm between you leave the DoD and when you come into the VA and making that seamless--all of those things are components of the entire VLER strategy, including component of the Nationwide Health Information Network, to reach out to the private sector and bring that information back in to be part of the veterans record, from our standpoint.

We look at VLER as a guiding philosophy, an architecture, and in the end, the ultimate goal is that lifetime electronic record. There are a lot of components that go into feeding that.

FGIT: Besides data sharing, is there anything further VA is doing to make VLER a national program, as far as reaching out to physicians' offices, small hospitals and clinics, things like that?

Baker: The Direct Connect piece is really oriented around--if you think of the NHIN, it is really oriented around large scale entity to large scale entity and an exchange of records like we're doing with Kaiser Permanent and Health South and various other large entities. But there are tens of thousands of small practitioners out there that we've got to bring into this. That's what the Direct Connect piece is--it's not oriented around the heavyweight implementation of the NHIN. It's oriented around information of patient X, and I know they're seen at the VA, and here's that information. And then we can incorporate that into that patient's health record, and vice versa--getting authorization from the patient to deliver that information back to that health provider, potentially by fax or by some secure method that we trust for protecting the information.

The important thing is that the provision of healthcare is really a community thing. We're part of that community. For many veterans, we're not the sole provider, and we've got to work with their health care provider, both to get information back and to provide information, to make sure they get the maximum quality healthcare.

FGIT: If VLER is the all-encompassing standard of where VA wants to be, where do things stand right now with all these pieces that are under development?

BAKER: I'm amazed at the progress so far. We're in pilot on the NHIN. By 2012, we'll have it roll out across the country. Every VA hospital will be connected to the NHIN by 2012.

We've done things like move to a common individual identifier between DoD and VA and implement that throughout our systems.

Once we took the view of 'We're going to do this' and the president and the secretaries are no longer going to accept the 'It's too hard,' a lot of things just started falling over at that point.

The progress on VLER has been very substantial over the last year. The e-Benefits portal has been stood up, where every service member leaving the DoD has a log into the e-Benefits portal, and when they leave the DoD, they retain that login. They don't have to change their ID and password to come in and look at their VA benefits and their legacy DoD benefits. They can come into that same portal--it's a big piece in that seamless transition.

If you add all those things up, we see a tremendous amount of progress for the first year on what has to be a very, very long term project. When you talk about the breadth of something like the lifetime electronic record, it's huge.

I think we started out down the right path: Let's remember that the 'V' stands for 'virtual.'  We're not going to try and replicate all this data all over the place, we're going to know where the sources and uses of the data are and be able to map those together and make them available to the service provider, wherever they are. That's the philosophy for every system going forward. The system will eventually get all the way to that, but in the beginning, you start with the small steps and you move to the completeness a few years down the road.

FGIT: Anything you'd like to add?

Baker: We are making an incredible amount of progress on the benefits side, too. The Veterans Benefits Management System, which will turn the whole compensation and pension process into an electronic system is moving forward well. We're managing it under, if you will, the VA way: Short milestones, managed tightly to the dates. It has the promise of being a key part of meeting one of the secretary's and president's main goals and that is breaking the back of the benefits backlog that veterans have had to put up with for so long.

I feel very, very good about that one and it's just fun to see the progress.

Related Articles:
Guest Commentary: Tom Munnecke on VistA lessons learned
VA improperly shared veteran health information with DoD 
VA looks to IT improve its reputation, from the inside out 
VA CTO: Disability claims process will be paperless within days
4,000 Social Security numbers potentially exposed in VA mismailing
Audio: VA CIO Roger Baker's September IT report
Baker: No increase in VA IT spend in fiscal 2012