Sunday, July 25, 2010

Informatics Spectrum Disorder

Tonight's post, my readers, is about the frailty of language. More specifically, I want to talk about definitions, the general problems with definitions, how challenging they are to write, and how definitions dramatically impact our daily lives.

If you're reading my blog, you might be a front-line clinician who just got your first job in informatics, and you're looking for helpful tips or advice. If you're new to the field, welcome! You may quickly notice a problem, however - Very few people understand what you do. The reason why : Most hospitals (and hospital administrators) still have a hard time understanding what exactly an informaticist does.

You might also be a healthcare administrator, trying to figure out what informatics is, because you've read about how important it is to have an informatics platform in place to make your EMR implementation run smoothly, for Meaningful Use and other reasons. You are coming here asking, "What is informatics and why do I need to hire informatics people to help with our EMR and meaningful use?".

One of the challenges both the front-line clinician doing informatics, and the healthcare administrator looking to build an informatics platform will both face is the definition of the term "Informatics" itself.

(Front-line clinical informaticist : Good luck explaining what you do to an administrator!)
(Healthcare administrator : Good luck hiring someone who "does Informatics"!)

Why is this position hard to explain, and so hard to hire? The answer lies in the definition of Informatics itself.

Let me explain.

First, there are various positions in healthcare, that all apply to the term "Informatics" loosely. Some of them include :
  1. Clinical Informaticist
  2. Nurse Informaticist
  3. Physician Informaticist
  4. Physician Champion
  5. Chief Medical Informatics Officer
  6. Chief Medical Information Officer
  7. Embedded Informaticist
  8. Bioinformaticist
... or another similar-sounding position. Curiously, each of these positions has a wide variety of job descriptions at different hospitals. The reason that these job descriptions vary so widely is because of the definition of the term "Informatics" and how it applies to "What an informaticist does".

Wikipedia tries to define health informatics, medical informatics, and health care informatics under their page titled "Health Informatics" : .

You'll notice, however, it's not much of a definition : "... is the intersection of information science, computer science, and health care. It deals with the resources, devices, and methods required to optimize the acquisition, storage, retrieval, and use of information in health and biomedicine..."

(In other words, this generally tries to define Health Informatics as the 'thing you get by crossing information science, computer science, and health care'.)

I, myself, tried to define informatics in older posts by what it is not : Information Technology. (One of the biggest mistakes you can make is confusing informatics with IT - Mixing the two will result in a budgeting problem that will prevent you from having adequate resources for your informatics platform.) But anyway, I acknowledge that my definition also lacks substance.

So why is it so hard to hire? Because of the challenge of defining Informatics, most popular job salary sites (like have virtually no data about informatics positions. And when they do, the job descriptions are often very wide, and as a result, the salary data is usually very poor. Consulting groups like Premiere are then challenged with giving labor data for positions that have very poor definition. Their reports are also limited by the poor definitions of informatics.

Q : So... I get it, Dirk - Informatics is limited by the poor definitions. So why hasn't the informatics community come up with a better definition yet?

There are a few reasons.
  1. The field of Informatics, although it's been around since the 1960s, is still really in its infancy, and is relatively new to healthcare.
  2. It's relatively new to healthcare because insurers, pay-for-performance, and Electronic Medical Records have pushed healthcare to develop a wide-spread informatics platform.
  3. There's no such thing as a perfect definition.
  4. As a result, there are a LOT of people who have the job title "Informaticist", who have very different skill sets, and perform a wide variety of different jobs, and...
  5. As a result, there are wide salary distributions when looking for "Informaticist", and...
  6. As a result, there are poor job descriptions for many "Informaticist" positions.
That's not to say that people haven't tried to develop better definitions - AMIA, HIMSS, various standards organizations and professional organizations (e.g. CMIO Magazine, Healthcare Informatics) have collected and published definitions and various articles about these issues, but the problem remains : There is still a pretty wide distribution of people who call themselves "Clinical Informaticist".

To draw an interesting linguistic parallel, Autism research suffers from a similar definition problem.

Q : Dirk - Really???!?!

As a person who thinks about the intersection of information and culture, I think I can make a convincing case for this.

In the 1980s and 1990s, researchers realized that one of the biggest problems for people with Autism was that they often didn't get the early intervention and resources they needed to help them. So the term "Autism spectrum disorder" was created, to help front-line physicians to make a diagnosis that had higher sensitivity and lower specificity. (That's generally what you want to do when you want to get resources to people that need it, earlier - Make the label easier to apply.)

The problem with such a broadening the definition of Autism, however, is that it hampers research into the causes of Autism. When the population of "People with Autism" varies so widely, from very low-functioning to very high-functioning, it becomes extremely challenging for scientists to find "statistical significance" between any risk factor and people labeled under the "Autism spectrum disorder".

One might argue, "Well we need to refine the definition of autism, then, to help the research!". The problem then becomes : If we refine the definition to something very specific, fewer kids will be diagnosed with Autism, and those families may miss out on the resources they need to help them.

In the end, you're always trying to balance sensitivity and specificity. As a result : There's no such thing as a perfect definition.

Establishing even good definitions can be very challenging. Linguists, interpreters, lawyers, and policy makers know how delicate and frail human language really is, and how important good definitions are. And while there are standards organizations that try to help us with these definitions (e.g. AMIA and HIMSS for Informatics, DSM-V for Autism), we should keep in mind that there is no such thing as a perfect definition, especially for something as complicated as Autism or Informatics.

Even the best organizations will publish definitions that all have linguistic limitations - It's up to us to see the costs, benefits, and limitations of the definitions and learn to work with them.

And just as important - I think both the Autism community, and Informatics community, would both benefit greatly from better balanced definitions. Namely :
  1. The Autism community could benefit from a definition which balances sensitivity and specificity a little more, and ...
  2. The Informatics community could benefit from work that increases the specificity of the term "Informaticist".
Yes, this is a pretty esoteric post tonight, but I'll keep working at defining both, and offer my linguistic skills to both communities as much as I can.

In the meantime, good luck explaining informatics to your boss, and good luck hiring an informaticist. :)

No comments: