Separation of Data from Apps: the First Step to Modular EHR

When your personal computer crashes, what do you worry about most: the data that you have saved on the disk, the time it will take to reinstall and reconfigure your apps, or the cost of repairs?

I worry about the data. The data, what I have done with it, am doing with it now, and might do with it in the future is why I use the computer in the first place. If I was a gamer, I might have a different view, but I'm not. The value is in the data. The cost of obtaining that value is primarily the hours of time I have devoted to its creation and maintenance. In my opinion the value proposition for a personal computer, or a multi-billion dollar EHR, is in the data and its use now, and in the future. Without the data the apps are useless, and that even goes for games that can be suspended in mid-play and resumed later.

I've written in the past about how Microsoft, by periodically changing the internal format of Office files, has denied me the use of documents that I created in the past and would very much like to use today. As for the old files, my current MS Office apps are worthless - money down the drain. Were I to continue using Office this "document stranding" would likely occur again and again. The value is in the data, not in the apps. I no longer use MS Office. I only use apps that work with standard, non-proprietary formats that I can use with any of a wide variety of apps.

The concepts of separating apps from data and of choosing data formats that are durable (can last for decades and still be accessed and used) are behind the sudden popularity of cloud services like iCloud, Dropbox, and SugarSync. People no longer want to be tied to a specific machine and a specific proprietary app. They want to be able to open their photos in their favorite image editor, use the images in documents and slide presentations, email them etc. using any computer they choose.

This way of using personal computers is an example of a modular approach. Various data sources - some private and some public - reside in various locations coupled with an assortment of application modules, which access one or more of your data sources. The phenomenal success of the iPad is due in part to the availability of 700,000 application modules and cloud services.

Anyone who uses a cloud service or a tablet has working knowledge, whether explicit or implicit, of what a modular system is. Some of these people (perhaps you) then go to work and participate in buying and using EHRs that violate practically every principle of modularity.

You probably don't have a modular EHR. You can't buy them today although you could simulate one. To create a modular EHR as a product offering, a developer would have had to break the 50-year-old mold from which EHRs are cast and start from scratch by focusing on the medical information itself and formulating a clear idea of who needs it, what they intend to do with it now, what ways it might be reused in the future, and the time span that future implies.

Modular refers to a system composed of separate modules, each with clearly defined functional goals and distinct boundaries separating their concerns from those of other modules. Most organizations wish for a single, integrated, do-everything-for-everyone system. There is a price to be paid (pun intended). Each year the EHR sprouts additional integrated "features" in response to "user demand." Since the modules have internal connections, the complexity of the overall system increases geometrically until, at some point, it becomes unmanageable.

This is an argument for starting small, adding functions gradually and separating functions that do not share concerns. This requires that data is treated as your sharable resource, not the property of a particular vendor's product. For a technical but cogent explanation of this principle, slog through Rich Hickey's video discussing functional databases. You will also get a glimpse of just how far behind health IT is when compared with the outside world.