One-on-One with Lutheran Medical Center CIO Steve Art

Part I

Art says VA’s open-source VistA software the answer to a cash-strapped industry’s prayers

Lutheran Medical Center (LMC) is part of Brooklyn, N.Y.-based Lutheran HealthCare, an organization which also includes the Lutheran Family Health Centers network and Lutheran Augustana Center for Extended Care & Rehabilitation. LMC is a 476-bed Level I Trauma Center providing ambulatory surgery, cardiac care, neuroscience services, obstetrics/gynecology, spinal surgery, perinatology (maternal/fetal medicine) and oncology (cancer diagnosis and treatment). Recently, HCI Editor-in-Chief Anthony Guerra had a chance to talk with CIO Steve Art about his facility’s plans to implement the VA’s VistA open-source software supported by Medsphere.

AG:Photo of Steve Art,  CIO of Lutheran Healthcare Tell me how your open-source journey started.

SA: I’ve been here eight years now and when I first got here, we did a long-range information systems plan. When I write those things, I try to go out five years. The guy I replaced, his title was CIO, but he had never been CIO before; he had been an implementer, and the guy before him had actually been a salesman and became a CIO. So they had not done a lot of the infrastructure work that needed to get done in any hospital. So my first couple of years were spent building an infrastructure, and then laying on top of the infrastructure the different systems we needed.

I’ll give you an example. The first system we did here was a payroll system. We put in a payroll system in three months because the company that we had as our payroll company was sunsetting its product and I walked into that. We really did a lot of fundamentals, building-block kind of systems. We had a McKesson STAR implementation that was half done and the operation guys were going crazy with it, and we had to finish it. There was a lot of that kind of catch-up work. The plan had in it that we were going to get all the way up to CPOE with the electronic medical record, barcode med administration, all the stuff that everyone is talking about now – that was in the plan. We built a pyramid – laid the different layers onto it, and when we got to top, we were going to have the electronic medical record.

AG: And that was eight years ago?

SA: We laid that out eight years ago. We knew that’s where we had to go. I told the boss when they hired me how much I thought this is going to cost because she was relatively new and she wanted to know. I said there is so much to do here that it’s going to cost you a lot of money, and you're going to either spend the money all at one time right now or you can spend the money over a number of years, and we’ll get to the same point either way, but it’s going to cost that much money.

I don’t remember the number of years right now, but that’s immaterial. It’s just going to cost whatever that number was. If it’s going to cost $50 million, it’s going to cost $50 million over four years or cost $50 million over one year, it depends on how fast you want to get there. So the sooner we can get the money in, the sooner we can do some of the things that we had laid out and prioritized – the IT steering committee would prioritize what had to get done first and what was most important. As the years went by, prioritizes changed somewhat, but our goal was always to get to this point.

The reason we needed to get physician order entry was for patient safety. That was the reason. We got to the point where we had most all of the other stuff on the pyramid done – we had an OR system, the lab system was in place, the pharmacy system was in place, all the ancillary departments, we got a PACS in place – we did all that preliminary work to get up to this point.

AG: Was this mostly a best-of-breed environment at this point?

SA: We did mostly best of breed, although we were a McKesson shop. McKesson had an opportunity to bid on everything that we did. They, for some reason, never could beat the price of the best-of-breed guys. Now, we didn’t do best of breed in terms of the most expensive vendor. We did best of breed in terms of the one who had a good product that we could buy and we could afford.

We ended up with AMICAS PACS. Of all the PACS vendors out there, AMICAS at that point was not the most expensive PACS and it was a relatively new player in the market. We bought them and they were giving us some great deals. They beat out the company that McKesson ultimately bought. It ended up being the two of them and AMICAS beat their price. We weren’t best of breed in the sense that we were buying the most expensive systems. I don’t want to mislead you. We were doing best of breed in terms in the sense that we did not drink the McKesson Kool-Aid.

AG: So would you say that you picked products from different vendors based on your needs and then you decided that you would deal with interfaces and integration work afterwards?

SA: Correct. We have our own interface group here, so we write our own interfaces and I prefer not to develop software, just as a principle, but we do build our own interfaces. So doing interfaces was not horrible for us and that was the route we took. But best of breed implies you're buying the most expensive product or the top one of the market. It wasn’t a Cadillac. I was buying Chevy, but I was making it work.

AG: I understand what you’re saying.

SA: After that, the emergency room was a big system we had to put in, and the PACS had been a big system. Once we put those two in, we were now at the point where we were ready to start looking for something more. Because basically, the ED system had in it all the components of the electronic medical record. We were ordering online. We were seeing results online. We were documenting online. According to our vendor, I think we were the 44th hospital in the country to come up with their electronic medical record in ED.

AG: Who are you using in the ED?

SA: The company is called ECDS out of Illinois, and the name of the product is EMpoweR. That product was really working in the ED.

I realized, at some point in time, that as we started talking about electronic medical records, we were all talking about different subjects. You talk to the COO, CEO, CMO, CIO, anybody—we all had a different view of what it is and most people think it is what it was in the last place they worked. So somehow, we had to come together as a group. What I did is I brought in three or four vendors to actually give us a demonstration. They understood that we were just kicking the tires, but everybody at that point was willing to show it to us. We brought in a couple of vendors to show us what they had for what they called an electronic medical record or CPOE. It was really eye opening because we finally got a common language to talk about this stuff. Everybody came to understand that no two systems looked exactly the same, but they all do pretty much the same thing.

AG: You said you essentially found yourself having an EMR for the ED. Do you think that’s a situation that happens at hospitals where all of a sudden they have these little pocket EMRs in different units? Do you still need an enterprise core EMR when you have pocket EMRs?

SA: That’s actually a good point. Let me take it back a minute to the beginning of the conversation. When I got here, this place was full of what I call fiefdoms. Radiology ran their system. Laboratory ran their system. Pharmacy ran their system. Because of my predecessors, and I blame them frankly for this, they let each individual department have their own say about the systems and, frankly, the vendors were still trying to sell to the individual departments because the vendors have a much easier time selling to the radiologist than they do selling to the IT guys.

So early on I had to break down some of those fiefdoms and break down the walls so they started sharing data. The pharmacy system, for instance, didn’t have an ADT feed. So they were entering new patients by hand into the pharmacy system every day. That was stupid. Just plain stupid.

So we built them an ADT interface. Over time, I got their trust and I was able to break down a lot of those walls. But you’re correct, the systems grew up in different places because they had different issues or different requirements at the time that were unique to their department, but we’ve tried to incorporate them into the whole system.

Now, in the case of the emergency department, our emergency department was growing really rapidly. In the paper model, we couldn’t handle the volume we were getting, so we had to do something. That’s where we chose to spend some of our money. So there, even though long range, we were going to do an EMR, it was important in the short range to have a system for the ED. Now, in the end, I’m going to end up converting that to become the real EMR that’s going to be used in the ED also. So I’m going to replace that system ultimately when I get my EMR up. I will replace it.

Part II

Art talks about managing user expectations when replacing a niche system with core clinicals

AG: And that would be a core EMR that also is useful in the ED?

SA: Correct. It doesn’t make sense to have multiple systems. So although we needed ED system for a number of years—and I don’t remember what year I bought that thing—when we bring up the new EMR, it will replace the old departmental EMR.

AG: Is it inevitable that, because you’re going with a wider system, you will lose a bell or whistle or two from that particular ED system? Will the ED folks balk at that loss of functionality?

SA: Absolutely you are right, that was a concern of mine. I told the ED early on that I would not make them move unless I can give them all the functionality that they had today. And that’s what we did. We started out with some kind of analysis system side by side to see what the gaps were, and we realized that the new system actually was going to give them more than the old system. It had much more functionality than the old system did, but there was one piece that was missing, and it was a tracking board.

There is a board that when you go in the ED and an outside attendant comes in, he can find out the bay or room the patient who he sent is in. So if you want to know that your patient is in 812, you go to the board, you find your name, you find the patient’s name, and you go to 812 and find your patient. So that component was not part of the system I bought. What we’re doing is we’re developing with my open-source vendor, the tracking board here, and then I will give it to them, they will QC it, and incorporate it into the product for all of their clients. That’s the cool part about open source; I have the software and I can develop it.

AG: So it sounds like CIOs have to be sensitive about taking away functionality from departments when they replace niche systems with larger core roll-outs?

SA: Yes and the only caveat to that was all EMR systems require you to use their pharmacy, because pharmacy is the one piece that is a whole closed loop. You have to make sure that you’ve got the right drug to the right patient at the right time and that all the orders that flow into the pharmacy flow out correctly to the patients. I believe all of the vendors require you to use their pharmacy.

So this whole decision turned on the fact that I had to get the pharmacy to agree that the new system was no worse than their current system, or I couldn’t have done it at all.

AG: And were they on McKesson? The pharmacy?

SA: No, they were on Mediware. They’re still on Mediware. They are on the Mediware Pharmacy and they were the only ones I actually had to change because I could have figured out interfaces for all the others. The other ones were manageable. Pharmacy was absolutely critical to the process.

AG: So you’re saying they’re not on Medsphere pharmacy at this point?

SA: No, I’m not on Medsphere yet. I’m going to it. I’m going live. We don’t have the final (re)lease of it yet.

AG: Okay, but you will be going to Medsphere pharmacy?

SA: We will be going to Medsphere pharmacy, yes.

AG: Because you’re saying you can’t go to a new system and not take the pharmacy?

SA: Yes. In our case, our labs guys objected, they didn’t like the lab system that Medsphere had. So we’re writing interfaces for them so they can stay on the Sunquest pharmacy product. The old Sunquest pharmacy is what we had here, and they're staying on that lab system. We built an interface between the lab system that’s in Medsphere and the lab system I have here, so people will order labs in Medsphere and labs and orders will go over to the lab system and the results will come back.

But pharmacy I couldn’t do that with. Pharmacy, we had to use the Medsphere Pharmacy. Like I say, I think that’s universally true of all EMRs, you have to use their pharmacy.

AG: Do you think it’s a losing argument for CIO to say to a department, ‘I’m switching you off your current system to another system, it’s better for the organization, but you’re going to lose some functionality?’ Is that too hard to sell to make?

SA: I always say it’s their decision. I always try to lead every department to the correct decision. I think I know what the correct decision is in terms of how these things all work together, but I try to give them the information that they need to make the decision, because if you tell them what you want them to do, most people won’t do it. You have to let them decide it. I try to give them the information they need and that’s what the process is, to bring these systems in, to let the departments make that decision. If it’s not a good decision for the department, then it’s not a good decision for anybody.

AG: That’s a good point.

SA: It’s really true. And so I can't dictate that. I don’t feel I have that much pull. I just can’t dictate that. Being in the business as many years as I’ve been in the business, I do know the hot buttons and I know how to sell this stuff to them and I know how to get the right guy in front of them and frankly, if it was rational, I wouldn’t have even brought somebody to them. They have to be rational.

AG: Let’s go back, because I took you off track a little bit. We’re talking about the fact that you had the ED system in place, and you were looking for a core system. Now, take me back to that point of looking for that core system.

SA: We had all these other systems in place. We’re ready to go. We then started getting educated about what the systems were and what they did so we could all start talking about it in the same terms. I’m going to take you back a second.

Some people thought that an EMR system was when you scan a paper record and index it into an optical system, where you can call up the old record because you received the paper and print it like a fax jukebox or something. There were people in the room that actually thought that’s what the EMR was. So to get them off that and say, ‘No, it actually has data and there’s a database.’ There was a big gap there in their understanding. So that took a bit of doing.

Once we got there and they understood what it could do, then we had to go through that fear stuff of, ‘Is it going to practice medicine? Systems don’t practice medicine; people do.’ We had to get rid of that stuff too. When those things were done, we then went out looking with an RFI and said to a number of vendors, ‘Please bid on this.’ What we’re looking for first was just CPOE. We wanted somebody who had a big system, but we wanted to only buy the CPOE part of it because we thought that was our biggest need, and we didn’t know that we had enough money to do it. So we got quotes back from the vendors and it just knocked our socks off.

AG: And these would be the usual suspects?

SA: The usual suspects—Epic, Eclipsys, McKesson, QuadraMed, MEDITECH—all the regular suspects. We sent this thing out to 10 vendors or something and they were all happy to bid, of course, but the numbers were outrageous. There was absolutely no way we were going to be able to afford it. We came to (the) conclusion that either the systems were going to (get) cheaper because more people were going to buy them and it was going to be easier, or we were going to have to do something else. That was our decision. And we pretty much were walking away from the whole idea of buying anything soon.

And then out of the blue we got a rate appeal from some charges that we put in years ago, and we got some money. We got a pot of money. When we got that pot of money, then the boss said to me, ‘Start looking again and see what you can find.’ Now, in the meantime, and I don’t remember what this year was – I think it was end of ’05/beginning of ’06, I read an article somewhere that under the Freedom of Information Act, the Veterans Administration software was available at no charge.

Part III

Art says he doesn’t need CCHIT telling him what systems he should buy!

AG: And that knocked your socks off?

SA: That knocked my socks off. I and two other guys started downloading the Veterans software because there was no charge, and there’s a Web site where anybody can download it. I started downloading, but the software is so big that you would download for two hours, and the connection would break and you had to start over again. I can't tell you how big these files were – they're just huge, huge files. So I was doing it at home and I was doing it on weekends and there are three of us – literally around the clock. I would have three downloads going on all at the same time but we never could do it. It was an absolute nightmare.

Then I read that there was a place you can send a check, and I believe the check was $47.16 to get the software on disks. I sent my personal check because – and this is a good one – I couldn’t get the hospital to write me a check to the government for that much money. It was just too much paperwork here to get that done. I wrote my own personal check, which I got reimbursed for, sent it off to an office of the government (which I can give you the address if you need it), and they sent me the six or eight CDs, DVDs, whatever they were that had all the software on it. I said, ‘Oh my God, I've got it all.’

My head tech guy and I sat down and we loaded the software up and, sure enough, we got it working. It was perfect. It was gorgeous. It had everything in it. Think about it – the VA spent, over the course of 16 years or something between $6 and $8 billion building the system for all of the VAs. That was the basis for the Freedom of Information release, taxpayer money had paid for it, so therefore, we should get it for free, and that’s what a guy won on (in court).

This thing had everything in it because they had a lot of money to spend and they spent it wisely. They actually built and bought a good system. And (the) way it was built was, in the early days, every VA had to have the basic product, but anybody in any VA could develop a new product to tack on and have it used only locally, or they could take that add‑on product and send it to some central place where they would then send it out to everybody and make it part of the main product. It was almost like an open-source sharing arrangement in VA. That’s how they built the product. And when people come up with a new template or a new whatever, they would send it in, get it part of the product and the product grew and grew and, of course, from then had vendors working on it at the same time. So this thing was gorgeous. I can’t say anything bad about it. It is just a gorgeous product, and it looks as good as, or better than, any product we saw on the marketplace and it’s free.

The problem was we didn’t know how to use it. How do you get data into it, how do you get patients into it, how do you hook it up to your ADT system, how do you hook it to your billing system, how do you do this, how do you do that. It created more questions than we had answers for. So we said to ourselves, ‘Who is going to help us put this in?’ One option was to find an ex-VA guy who knows the whole system, but there aren’t any because the VA guys knew the piece that they worked on and they knew how to use it, but not how to install it. That’s because it got installed over so many years, it wasn’t a single group that installed it. It came in pieces over time and you just added to it.

So we looked around and found a couple of companies that were in this space that would help implement it and eventually Medsphere (Carlsbad, Calif.) came through to me as the one that was going to help, and that was history. We struck a deal with them that they would make the changes I needed to the system and, of course, some of the changes were fundamental. For example, the VA uses social security numbers to track patients. We, in the commercial space, use medical record numbers and account numbers. What we call a nursing system, the VA calls a ward, so some of that terminology had to change. And then I needed a bunch of interfaces to work with the systems I was still keeping. My dictation system and my PACS system and my lab system and then those things had to get done and integrated with the system.

So they signed up for all that stuff. We came up with a price and they’ve been helping me implement and literally building all that stuff for me now. They’re also loading some of the back data that we need to go in, and we’re getting closer and closer to it. We’ve built all of our templates and all of our screens and we’re pretty much ready to go now as soon as I get all of those other interfaces they’ve done. At that point, we’ll do our testing, and then training and go live.

AG: What’s your timeline coming up?

SA: We’re expecting them to give me the last of the pieces they are doing beginning in June. And if they make that date – which I’m hopeful they will, we are looking at going live toward the end of the year. We have to do a lot of testing of stuff first, but I’m assuming that testing goes okay and training goes okay.

AG: Okay, so we all know we’re in the ‘meaningful-use’ environment, waiting for the clarification.

SA: Yes we are.

AG: I know CCHIT has made some overtures regarding the certification of open-source software, but what would happen if they are deemed the one and only source of certification and open source doesn’t qualify?

SA: You’ve made a correct description of the situation. I hope CCHIT does not become the body that does the certification. Because heretofore, their certification has been around doctor’s office EMRs/small practice EMRs. And they were formed because it was thought that doctors are too busy to figure out what they should have in the EMRs, so somebody would just say, ‘Okay, this is a good EMR and this is a bad EMR.’ It was built as a service to the small practitioner – one-man, two-man, or three-man shops. It has taken on a different life. It then became a political process where the big vendors were paying CCHIT lots of money to further certifications each year and because it’s an annual certification, CCHIT is making money off this stuff, so they’re happy to be doing it every year, and the criteria has been more and more restrictive to exclude vendors from the market as opposed to making it more inclusive.

I’ll give you an example. At the HIMSS conference, a lady from WorldVistA got up – I don’t remember her name now – and said to them, ‘Why do you have a requirement that is so restrictive that you require passwords to be in both upper and lower case?’ And I said to myself, ‘Oh my God,’ because in the VA software, every password has to have a special character, a number, and a letter, at least one of each. So for passwords, a lot of times, people have numbers and letters which may be inconvenient, but the numbers and letters and a special character is totally over the top, but even more secure. So by CCHIT’s criteria, a password scheme that OpenVista has, which is more rigid, more than what they’re asking for, would not pass because there’s not an upper and lower case.

To my view, that is restricting this vendor, not helping me – not helping the consumer. And, from my point of view, when you talk about a system that’s going to be sold to a hospital, it’s different than talking to a physician running his own practice. I understand when you’re in doctors’ offices, somebody needs to worry they have security and all the other issues that CCHIT is concerned about. But when you sell to me, I’m a big enough consumer that I know what I need. I know how to evaluate a system. I know whether or not the security is valid or not valid. I mean, having three different kinds of characters is valid enough for me on two levels – username and password – and that’s fine for me, because you have to get on my network to start with, you have to log into the network and then log on to the application. So I've got multiple levels of authentication here.

Part IV

Art says being part of an open-source community can be a beautiful thing

AG: So you don’t need a chaperon leaning over your shoulder to help you select a system.

SA: I don’t need them to tell me how my system did or what should be in the system. They don’t even need to tell me about security; I wouldn’t buy a system that didn’t have security. I’m not a fool. That would not escape me. So their requirements have become restrictive, so that the big guys who pay them lots of money have a bigger say in getting their system certified, and they’re excluding the little guys from playing.

The big and little argument isn't exactly accurate; it’s the haves and have-nots—but some vendors are trying to exclude other vendors. Every hospital knows how to buy a system. We don’t need a CCHIT on the inpatient side.

AG: The last time I looked (as of June 1) there was only one 2008 certified inpatient system—Epic.

SA: Yes. The other thing is that, when I went looking for this in ’05 or ’06, there was no vendor that had both inpatient and outpatient. That’s in 2006. There are people who say they have. Epic started in one side of the business and Eclipsys started on the other side of the business, but neither of them had both, although they both say they do. So VistA was probably the only system that had both inpatient and outpatient back then in ’06. So, today, we still have outpatient and inpatient. The other guys are catching up, but in terms of certification, let the buyer beware – let me pick my system.

AG: You make some good points there. Let me play devil’s advocate and ask you about the limitations that people say VistA has. The first thing would be that it was designed initially or mainly for clinical use and therefore it does not do billing, especially to deal with a multiple payer environment.

SA: That is correct, it does not do billing. It is not a billing system. I didn’t buy them to be my billing system. I’m using the McKesson STAR for billing and McKesson STAR for registration. I send a registration to VistA, VistA knows where the patient is, we treat the patient. It sends charges from VistA to the billing system and the billing system bills. I have not changed my billing process because, frankly, I don’t want to lose a day of billing. That may be a downside for a doctor’s office that needs to do billing as part of their medical record, but I don’t see it that way.

AG: I haven't heard anyone suggesting that this is a good solution for a practice.

SA: Right, exactly. So from my point of view, anybody who is going to buy this thing, they're buying it because they want a clinical record, and they want to have a record that they always have and they never lose.

AG: Second thing would be that going with this type of solution requires a sophisticated, mature, robust IT shop within the hospital because there’s quite a bit of work that must be done after the download.

SA: I’m pretty robust. I consider my shop okay. We’re not silly, but we needed more than to be robust. We needed someone who had product knowledge and understood how to put the thing together. So I had to go out and hire Medsphere to do this. I could not do it myself, and I don’t think anybody is making the argument you should download it yourself and put it together yourself. The only people I think that can possibly put it together themselves is somebody who actually worked in the VA and is a programmer.

AG: Let’s stipulate that you would need an organization like a Medsphere, for example, but in addition to that, would you still need a decent-sized IT shop within the hospital?

SA: No. Medsphere will do everything for you. You need somebody who knows how to build templates, which is a tool like when we’re processing a document on building forms or something, you need somebody who knows how to build the forms so that somebody can add eye color to a form that wasn’t on it before, and that kind of stuff. That way, you don’t have to go back to the vendor after every one of those little things you want to do.

If you have a development shop, your development shop has a source code so you’re not locked to any vendor. If you learn it well enough, you can get away from your vendor, but I don’t expect that’s ever going to happen for me frankly, and I’ve got a pretty decent programming group. I think that almost all of that work is done by the vendor. You still have the same requirements. You still have to do your backup every day and all the normal systems things you do for any other system. But it’s no different than owning a system from any other vendor, I don’t think, except that you have the ability that you can add to it and change it because you have the source code.

AG: And if you do make some changes, do you have to send those changes back in some way to Medsphere or is that only if you want those changes to be part of the core open-source software?

SA: Right. You don’t have to send them back, but if you do send them back, they’ll make it part of the product. The upside to that is, if it’s anything big, they will QC it for you to make sure it works. So they put their QC department on it. Frankly, it saves me time and energy to let them QC my work instead of having my guys QC my work, and I know that we didn’t make some boneheaded mistake with it.

AG: What about the idea that it doesn’t contain some of the more department-specific solutions that the VA would have little need for, such as obstetrics and maybe pediatrics, and things like that?

SA: We have pediatrics and obstetrics here, and we are building the templates we need to support those practices. In fact, they have large psych departments in VAs and we re-tooled the psych part of it to match what our psych guys wanted. So all of the data collection and templates in the system are in the control of the user anyhow. We literally took our existing paper forms and did—that’s how we started the project—did a side-by-side analysis of what was on the form now and how we needed to change and put it in the system. For instance, every form has your name, account number, medical record number, date of birth, all that kind of stuff—we don’t have to ask all those questions on forms in the computer because the computer already knows who you are. So we had to literally pull stuff off our forms.

The other thing we did is we standardize the pain scale. It’s on 42 different forms and we made sure that the pain scale module in OpenVista looks exactly the same no matter which form it’s called up on. The forms became much cleaner and looked more consistent by putting them into the system. I don’t need a vendor for that, and I also don’t need the vendor to understand the application. They don’t have to understand peds or OB in order for me to build forms for it.

I think that it’s accurate that it doesn’t come with it natively, but as we get to be a bigger community, we will share (and we already are sharing) our templates between hospitals. We all share and then we give them to Medsphere and they post all of this stuff on their online site. I just got an e-mail last night from them. They said, ‘Do I know the hospital that’s near me in Long Island? They keep going to the Web site and pulling down the forms that we’ve all posted, and they want to know if they would be a candidate to have a salesman call on them.’

I said no, that hospital happened to be owned by North Shore Long Island Jewish and they bought Eclipsys. I know that. I don’t know how much they spent – $600 million or something for Eclipsys for all their hospitals. But that hospital found the Medsphere site, and they’re downloading the forms that we’re contributing to the open-source community. Which is fine, I have no objection to that; I just think it’s neat that OpenVistA figured out how to share stuff, and all the proprietary vendors won’t do that with their clients. If you want to buy forms from them, they’ll sell it to you.

Part V

Art says that with just a few more satisfied customers, open source will change the HIT landscape!

AG: I’ve heard a few commonly cited reasons for why open source hasn’t taken off. First off, I’ve heard that CIOs too closely aligned, or dependent, on the traditional vendors—they’ve either worked for them, or may want to work for them someday and so they want to stay in bed with the usual suspects. Or because it’s easier to have a huge cadre of people to be able to call, you're not going to get fired for hiring Epic or Eclipsys.

SA: Yeah, the IBM syndrome is true. That is really true, and I think it’s wrong. I tell the story a little differently. I have McKesson here as my vendor, and every time I would need something done by McKesson that I know was just a little, little thing, they charge, no matter how big or little – they charge me to give me (a) quote. They charge me $174 an hour just to give me a quote. Whatever it takes to them to build a quote, it costs me money. Then, if I decide to do the work, then they charge me additionally for that.

I recently put in a work order to make a change to an existing interface. I wanted to add one piece of data to the working interface. They came back to me and the programmer said it was going to be a zero charge, as it was just adding one data element to an existing interface. McKesson sent me a bill for $261 for them to tell me that there was no charge.

This is on Nov. 4, 2008, the bill came out and the due date was Dec. 5, 2008. So I said, ‘Why do I have to pay $261 for somebody to tell me that this isn’t going to cost me anything?’ And that’s the relationship that we have with our vendors. Yeah, their argument is that we have to keep it all closed so that we keep control over everything, and I appreciate that they have to keep some control over it. But every time I asked for any little thing, it costs me money. The thing I like best about open source and—anybody who has got any brains is going to figure this out over time, and I think that these guys eventually are going to change the industry—is that when I want a little change done like this, adding an element to an existing interface, my programmer can go ahead and make that change for me, and I don’t have to pay anybody, I don’t have talk to anybody, I don’t have to think about it. And that alone is worth a fortune because we make these changes all the time.

AG: So why do you think it hasn’t taken off?

SA: Because the world doesn’t believe there are enough of us that have done it yet successfully for them to see that it’s safe. I think that’s the issue. Medsphere now is in a bunch of places; they’ve got 200 Indian Health Service sites and about 11 hospitals now. Most of those places are pretty small. There’s one big guy in Midland, Texas, and me in Brooklyn, and I think when we get up and there are a couple more up, I think what everybody is going to see is that this is really doable and it really works and there is no reason to be fearful of it. The guys right now are fearful because there aren’t enough of us that have done it.

AG: Once people get over the fear, isn’t it still more work for the CIO to go this route?

SA: No. The money, the affordability, is the driving thing in open source. Remember, the software is free here. There aren’t a lot of hospitals that are making enough money that can spend $10 million for a system. How many players want to spend the money on that stuff? The stimulus is making it easier because you get back some of it, depending on who you are, but I think the big guys are going to stay with those big systems because of the IBM issue. But I think the majority of people who think will buy open source. You haven't seen it yet, but this system is as good or better than any other system out there. I've seen them all.

AG: So companies like Medsphere—and I don’t want to just trump up Medsphere—but companies that, would you say they take the risk out of this for the CIO because you do have that support?

SA: I've said this already, it’s the same relationship I had with my vendor as I had with all other vendors. They are doing all the other stuff that normally a vendor would do and I’m doing the stuff that I can do, and if I want to take on more, I can. The relationship is I can do more if I want to, I don’t have to. And for some of their small clients, the client does nothing, literally does nothing. I’m doing more because I’m able to do more and I want to do more because I want to see this thing go. That’s why I’m building the tracking board for ED and that kind of stuff. But there is no requirement that I do that. I don't see this is as being any more difficult than putting in McKesson or Eclipsys or Epic or anybody else.This is no different.

AG: Would you call this decision a no‑brainer?

SA: This is a no‑brainer. I was at the Marcus Evans event a couple of weeks ago, and I made my presentation on EMRs and put up the slide and said, ‘This thing cost me $37.16,’ and you could see everyone’s mouth open. I’m talking to CFOs. Their mouths were hanging open. Someone asked, ‘Why does my CIO want to buy Epic or Eclipsys or the other big guys if they are free to take this chance that you’re taking about?’

I said, ‘I think that is exactly it, because it still is a chance in their mind,’ but the fact is, if they got into it and learned it, they would see what I saw. I mean, I see it.

One of the really interesting things I saw at that meeting in San Antonio is that we asked the CFOs how many of them know how much stimulus money they’re in line for and they didn’t, and that just knocked me over. If you haven’t started thinking about an EMR system yet – I’m in the middle of implementing it, and I’ll get it up before 2011 – but if you haven’t bought one yet, you are going to be hard‑pressed to have it up by 2011. These things don’t go in 15 minutes. Imagine – you're potentially replacing lab, rad and pharmacy, which are all nine month installs to start with. You're going to put in CPOE and EMR. I mean, this is not something that can be shoved in.

AG: Especially if you’re talking about training and workflow change…

Absolutely. You have to have doctors putting in orders too. We never finished that conversation, but I think what’s going to be measured is how many are using the system and how many orders are going through it. If that’s what meaningful use becomes and these guys haven't started thinking about it, they’re in trouble. They also need to know about the money.

My place, the Greater New York Hospital Association, came out with a guide for all the New York hospitals, and we’re in line for something like $13 million over four or five years. So we know what's at stake for us—that money will pay for this thing over and over again. This system will be paid for with the stimulus money, but only because I started when I started and because I’m going to be live by 2011.

AG: And it’s not going to pay for it if you spend $50 or $100 million …

SA: Exactly. Most places won’t be able to afford the $50 million to start with, and they’re not going to be reimbursed for it anyway. Medsphere will charge you a quarterly fee, and the fee is based on how big you (are), the number of beds you have . . . And then they show you when the stimulus money kicks in and you get your money back, and the return for most hospitals is less than four years. That means that the stimulus money for most places, will be enough to reimburse for the system completely. And they’ll pay it out slowly to the vendor.

When you see those graphs on your hospital and you say, ‘God, I’m going to get $6 million from the government, and it’s only going to cost me $3 million to put this thing in,’ you say to yourself, ‘What am I doing, why am I looking at $50 million?’ That is just eye opening to the CFO.