Sunday, April 29, 2012

Big Bad Legacy EHR Products

IBM "cloud" computing - circa 1975
There is no self-respecting innovator in Health Information Technology (HIT) who has not spoken or written about the horrific state of Legacy EHR products, which are slowly but surely being deployed in more and more health care facilities as a result of Meaningful Use incentives and changing reimbursement models. A couple of months ago I saw an EMR in a small practice. They’ve been using it for 15 years and it was a DOS based system with the ubiquitous neon green text glaring on a black and sometimes blue background. Aha! That must be a Legacy EMR, and sure enough the doctor was looking to replace it with a more modern product, but which one should he get now? After all, the last thing you’d want is to have him buy yet another Legacy EMR.

According to, a Legacy system is a “computer system or application program which continues to be used because of the cost of replacing or redesigning it and often despite its poor competitiveness and compatibility with modern equivalents. The implication is that the system is large, monolithic and difficult to modify”.  Well that little DOS EMR was anything but large and monolithic, but nobody was going to invest a penny in redesigning it, and competitiveness wasn’t a term that came to mind when you looked at it, and replacing it is sure going to be an expensive proposition. The DOS EMR is definitely out then. The only question remaining is what it should be replaced with. Which EMRs in the marketplace should be avoided since they are truly Legacy EMRs sold under false premises to unsuspecting buyers? Well, it depends on who you ask.

You could separate the various EHR constituencies based on programming technology (e.g. MUMPS vs. .NET), based on promotional labels (e.g. Cloud hype vs. everything else), based on software architecture (e.g. integrated vs. modular), and a host of other technical criteria, most of which are overlapping to various degrees. A much clearer and natural separation occurs if you divide Health Information Technology (HIT) companies into two groups: those who have lots of customers and those who don’t. According to the latter, the former are all peddling Legacy systems. It seems that a veritable tsunami of innovation is building up outside the infamous walled gardens of existing, Legacy EHR vendors, threatening to bring those walls down any minute now.

As with any worthwhile technology innovators, the newcomers to the EHR marketplace have brilliant Silicon Valley pedigrees and beautiful Web 2.0 style websites, along with iPhone/iPad/Android native (i.e. client/server proprietary) apps to complement, or even supersede, the web offering. Actually using an “old fashioned” computer or laptop is starting to feel a bit Legacy in and of itself. The innovative products themselves can be divided into two categories as well: full-fledged EHRs and a variety of self-contained pieces, or modules, of what is currently considered a complete EHR. [Note: I am not including products like athenahealth here, since they are not new, do have a respectable customer base, and had no disruptive effects on the rest of the market.]

The innovative new EHRs are all Cloud-based, intuitive and easy to use, built from scratch by user-centered designers, and are offered at a fraction of current prices, or so the ads say. These are the Southwest Airlines of health care, coming in below market pricing, with bare-bones, friendly solutions for the non-customer segment and they have two insurmountable problems. First, almost none of them are actually below market price, which in the ambulatory sector stands now at about $500 per provider/per month for a fully loaded, gold-standard integrated EHR and practice management solution. This is an extremely difficult number to beat. Second, even a bare-bones solution should have all the bones. My guess is that Southwest Airlines would not exist today if their first flight service consisted of boarding passengers in Houston and then proceeding to cheerfully shove them out of the aircraft 150 miles outside of Dallas, expecting them to arrange for their own transportation into the city. And yet, this seems to be the preferred model of our innovative HIT products, and as ePocrates (a household name in health care), painfully discovered, there are no customers lining up for this type of experience no matter how innovative it is touted to be.

In the meantime the Legacy EHR market seems to be thriving, and no, the recent Allscripts misfortune (or mismanagement) is not an indication of an impending disaster any more than this year’s snowfall in Texas is a sign of global cooling. The reason for this seemingly inexplicable prosperity is threefold: a) the government is subsidizing EHR purchase b) there are no viable alternatives to existing products c) innovation is occurring within the established market leaders. Let’s look for example at one of the more popular ambulatory EHRs, which shall remain unnamed. A few short years ago, the product consisted of a basic integrated EMR/practice management system, with very few bells and whistles and lots of bugs. Today, the product comes with a solid Patient Portal with iPhone apps for patients, a full featured disease registry, an iPad version, natural language processing, disconnected mode operations, peer-to-peer communications, and of course a much improved EHR and all sorts of other features and modules. I don’t know about other folks, but somehow this does not seem like a Legacy product to me, and there are a few more just like this one. There may be Legacy products out there, but today’s top selling EHRs do not fit the description.

A very unfortunate side effect of the forced march to HIT innovation is the confusion created by the constant barrage of misleading statements from the various Southwest Airlines wannabes. I sometimes wonder if these new folks have ever seen an EHR, let alone use one or participate in building one. A quality EHR is much more than a handful of rudimentary web pages allowing patients to communicate with providers, no matter how loud and trendy the consumer movement is. A modular architecture is much more than a collection of disparate bits of software interfaced together with duct tape, no matter how standardized the duct tape is. There is a reason why the new iEHR for the VA and DoD was allocated $4 Billion for development over five to six years. There is a reason why it took Kaiser about the same amount of time and money to take Epic from its original state to the powerhouse product it is today. There is no room for Southwest Airlines type of innovation in an industry where the routes, the meals, the fuel and the seating arrangements are regulated by the Federal Government. Innovation is coming and will continue to come from within the established systems, NASA style.

So if you are still looking to replace that little DOS EMR, or an aging and no longer supported practice management system, find a good size EHR vendor, with a hefty customer base, who develops its software in-house, instead of randomly buying shiny things, and hitch your wagon to theirs. It will not be a perfect ride because there are no perfect rides, but it will get you where you need to be. There are no miracles, there will be no miracles, and every day you waste looking for one, will make it harder to catch up, because whether we like it or not, whether it is a smart thing or not, health care is moving up the IT escalator at a very brisk pace. It’s too late for partial, gradual or “lite” solutions. The time for dabbling with a little electronic prescribing and a little email, has long since passed. You’re either all in, or all out, and your patients desperately need you to be all in.


  1. Interesting post but too arcane for a layman like me. I do see the problem, however. Having followed the PC trail from our first little Apple IIc in the early eighties through half a dozen permutations, I can relate. I decided a few years ago to simply toss whatever I was using after four or five years, come up for air, do a little homework, get a new one, and settle down for another few years until that one became obsolete. (I even knew a few durable types who went all the way back to the old Radio Shack stuff.)

    Not to be flip about it, but it looks to me like a good time to purge the old files. (I know there are legal issues to be followed, but that didn't seem to get in the way of the mortgage bankers, did it?) By the time someone went through the old files, dropping those who were dead and replacing scores of replicated data with a single copy per patient, I bet the volume would diminish by half or more. (How many times does an address or phone contact really need to be replicated?

    I'm just glad it's not my can of worms. You make me appreciate medical records people more than auto mechanics or plumbers. Their job may be tedious but it's sure not optional.

    1. You're so right, John. It is simple, isn't it?
      You do a little research and buy something that will serve you until it becomes obsolete. Everything technology related will eventually become obsolete. We would still be using card-readers if we just sat there waiting for something better to come along.

  2. The WIMP Interface is seen by too many as "intuitive" and the claim is that it "reduces training costs".

    On the flip side, there isn't anything "intuitive" about them really, especially when the mismatch between actual clinical workflows and their Information System Doppelgangers is high. As for training costs: it might possibly get end-users using the system quicker and since there is an abundance of visual cues on the screen, will help occasional (a few times/week) users of the system get things done. The downside of this is that there can be much faster ways of using a computer system versus mousing around, pulling down menus, clicking multiple times and so-forth. Full-time users of computer systems don't need so much help and can be much more efficient than this. Unfortunately, systems designed around the WIMP Interface ideas are usually very poor at providing shortcuts. Systems designed around more of a command-oriented interface CAN provide context-sensitive help when needed, but usually presume the user is an expert and knows what s/he's doing and try to simply let them do it. What little you get in reduction of training costs will be overwhelmed in production.

    System Evaluation Teams should ALWAYS include top end-users and analysts who understand this problem. Remember that everyone is looking at a brand new system and they'll focus first on what it takes to get anything at all done. They may declare it "intuitive". But after about a week, the full-time users will start comparing it to the optimized systems they were "forced" to give up.

    Well-designed green-screen systems may look old-fashioned and clunky next to full color WIMP Interface Systems but they can be very very quick for techs and others to use, and looked at from this perspective are better.

    2) The much-derided "Dumb Terminal" has a lot going for it. It is much, much easier and cheaper to deploy 100 green screen terminals or even 100 Browser-Only terminals than it is to deploy 100 "thin clients" running a general purpose operating system. It is more secure out of the box as well.

    Green Screens constrain the system designers to focus on what's really needed at the point of use as well -- a green-screen system usually isn't flexible enough for work-arounds to work. You don't hear "Oooohhh! Well, if you want to do it THAT way, all you have to do is ". The system will have one or two ways to do anything, so a lot of thought goes into matching what the systems do with what the people do.

    Flexibility is not what's needed in actual use! What is needed is one way to do the right thing at the right time by the right person. So be careful what you declare "obsolete" and don't be afraid of a system because it uses old-fashioned I/O devices.

    1. To paraphrase Albert Einstein, usability is 10% design and 90% habit. And those green screens are very habit forming when it comes to heads down data entry, hence why billers love their DOS products.
      It is also very easy to "design" usable software when the software doesn't do much of anything, e.g. the one-button Google home page.
      Other than that, all those shiny "intuitive" navigation aids, invented for websites where you order pizza or select a movie, are guaranteed to get on your nerves as time goes by and you become a habituated user. Remember the annoying Microsoft clip assistant? There should be a way to turn off all the hovering and layering and huge colored buttons that permeate what is now considered usable interface design.

      Love your last paragraph, Tom.

  3. Sounds more like Thomas Edison to me, but who knows?

    Yeah "usability" is in the fingertips of the user. There was a text editor I used back in the day that assigned commands to the 10-key keypad out to the right of the QWERTY keyboard. After a few weeks with it, I didn't even know the name of the commands anymore, or even which key it was on. For example, the "5" key in the center could have two commands; one un-shifted and the other shifted. But I didn't KNOW it was the "5" key, it was just the key in the middle that did what I wanted. All thinking about the mechanics of using the editor stopped at the wrist -- sorta like a good piano player deals with a piano.

    So yes, the first few days I used it I might have been faster using a mouse and a GUI. But after that, it was MUCH faster just to use the keyboard.

    Same thing went for anyone from straight-ahead card punch operators to lab techs using their LIMS. They want something that supports and even constrains what they actually do.

    About anyone who spends full-time with any sort of information system gets good at it, so usability should be measured in terms of the typical user of the system, not the tyro ("noob" in today-speak). This seems lost on lots of system evaluators nowadays.