Posts Tagged ‘martin fowler’

Contradictory Observations and Electronic Medical Records

Tuesday, March 3rd, 2009

Martin Fowler has an interesting discussion in his ContradictoryObservations post.  This little slice of medically related software design insight is particularly relevant because it highlights (at least for me) the complexity of the use of electronic medical records and their interoperability.

In a broader sense I suppose it also shows some of the underlying difficulties that face the Obama administration's new EMR adoption push.  But I'm not going there.

The concepts of observationsrejection, and evidence are good, but they're just the tip of the iceberg:

rejected and evidence

Even after you've modeled the data interactions, how do you effectively communicate these concepts to the user?  Or to another EMR that doesn't know about your model or how it's used?

Martin's view is that:

Most of the time, of course, we don't use complicated schemes like this. We mostly program in a world that we assume is consistent.

Unfortunately, many of the issues facing electronic medical records do require complex solutions. And even when the world is consistent, how you implement a solution may be (actually, will probably be) very different than how I implement it.  Either way, interoperability will always be a challenge.

We're going to need lots of good software design tools to solve these problems.