In an era when Apple gives away its Mavericks OS X operating system for free, when completely free open-source operating systems like Linux and BSD are available, when even Microsoft is considering giving away its Windows operating system for free, one has to ask the question, why is medical software, in particular EHR (Electronic Health Records) systems, so expensive? The software industry appears to be headed towards a business model of low-cost or free software, with profits generated in other ways. Linux distributions (“distros”) sell the Linux operating system at no or minimal cost, but generate profits by providing support services. Apple gives away its software but profits from hardware sales. Google gives away its software in exchange for advertising revenues. Of course not all software is free. Oracle still charges an arm and a leg for its database systems, but viable open-source alternatives such as Postgre, MySql exist. Many companies have moved to these lower-cost or free alternatives due to this cost. But even the cost of enterprise software like Oracle pales in comparison to the cost of EHR systems.
Case in point, Epic System’s EPIC EHR. The costs of this system can only be described as astronomical. Duke paid $700 million for their system, while UCSF managed a bargain-basement price of $120 million (see this Forbes article). And what do you get for this price? As a former user, I am not in the minority in my opinion that EPIC sucks. The problem is, the other EHR systems suck worse. EPIC is replete with useless mouse clicking to satisfy bureaucratic ideas of what doctors need to document. It is the epitome of software written with a total disregard for the end-user. It is difficult to go into more specifics. Epic Systems is touchy about their software. I know from personal experience that EPIC employs full-time staff whose sole responsibility is to scan the Internet for EPIC screenshots and have them removed. Isn’t it remarkable to spend $700 million and not even get a screenshot to look at?
Medicine is big business, so it attracts the big business types. I remember a simpler time. Back in the 1980s, prior to the advent of electrophysiology recording software (like the GE Cardiolab system), I wrote some software (in Turbo Pascal — oh the nostalgia!) that used a graphics tablet to enter measurements such as A1A2 vs A2H2 intervals into an Apple II or IBM PC computer. The software constructed tables of intervals and graphs, and could calculate such now-a-days neglected functions as the anterograde functional refractory period of an accessory pathway. Today we don’t measure this value, we destroy it using radiofrequency ablation, but in those days of yore when we were trying to figure out why our patients placed on class I antiarrhythmic drugs for PVCs were dying suddenly, figuring out the effects of these drugs on the electrophysiology of the heart was important. The software I wrote helped ease the pain of doing these repetitive measurements for the poor EP fellows who were wearing out their calipers otherwise, and they at least were thankful for this early example of medical software.
The American College of Cardiology or the American Heart Association (I don’t remember now which) had an abstract session back then devoted to medical computer software. I submitted an abstract for the program and it was accepted. I went to meeting where the ACC or AHA (darn which was it?) provided the hardware and I got to demonstrate my program to the participants milling through the exhibit area. It was similar to showing a poster, which, if you have not done it, is one of the most fun things you can do at a medical meeting. People who come by are nice, interested in your work, and ask good questions. This is diametrically opposed to the nervous tension of presenting an abstract in front of a large group. So presenting my program at the meeting was a blast. People came by, both health care workers and from industry, and had a lot of nice things to say. Of course a few years later the program was obsolete due to the launch of computerized EP systems as mentioned above. But I learned a lot from the experience, not the least of which was how to code properly linked lists.
Flash forward to modern times, about 2 years ago. The Heart Rhythm Society (HRS) at their upcoming meeting planned to showcase medical software relating to electrophysiology. I had written the mobile app EP Mobile for Android, and somehow HRS had gotten wind of this and invited me to show the app at their meeting. Distant happy memories of my earlier experience with software and medical meetings came back to me, and so I eagerly replied positively to the HRS email. Then in a follow-up email came the fine print. HRS wanted me to pay $3000 to show my app.
A quick email exchange followed. EP Mobile for Android was a free app. EP Studios, my little company I set up to try to protect my personal assets from liability, though not technically a non-profit company, was in a very real sense more non-profit than any so-called non-profit company, and I could not afford to contribute $3000 to the Heart Rhythm Society for the honor of displaying my software. I explained this to the contact at HRS, but he was not able to wrap his head around the concept of free medical software, even in the form of an mobile phone app, so the deal fell through.
Winston Churchill purportedly said: “You can always count on Americans to do the right thing – after they’ve tried everything else.” The use of monolithic, proprietary medical software has been foisted on us by our government, at the behest of the multi-billion dollar medical software industry. Everyone is happy about this except the doctors and their patients. Is it too late to fix this? Probably. However, maybe there is still a way out of the current disastrous situation. There is little doubt that, if done properly, codifying medical data is good thing. The problem is the current interfaces are terrible. There are too few vendors and they don’t care about user feedback. I would go back to basics. Medical data is after all just data. There needs to be a standard data format for medical data. I know that HL7 exists, but it seems to be pretty messy, and only last year opened itself up as a free standard (it required a license fee before). If all the energy devoted to making mutually incompatible EHR software systems was instead directed towards defining a good, open, free medical data standard, to be used by all computerized medical hardware, imaging, and so forth, I believe we could break down the medical software oligopoly. Domain specific languages and APIs (application programming interfaces) could be developed and with these building blocks anyone could create medical software. It wouldn’t have to free, but it would have to be affordable to sell, unlike the present situation, so costs would go down. User interface quality should improve. And Judy Faulkner, billionaire CEO of Epic Systems, might finally have some competition.