Aaron Berdofe, MHI – Health Informatics

Last night I attended a Minnesota HIMSS event reviewing the Quality Measures for Meaningful Use Stage 2.  Much of the content matter was review for those of us who are well entrenched in executing Meaningful Use requirements, but it was a good overview of some of the more technical aspects of how the quality measures are being structured.  One question from the audience though turned my thoughts to how Meaningful Use is playing an integral role in the slow march of the much hyped and oft-overused term of Big Data in healthcare delivery.  A physician innocently asked, “Are we beginning to see some results or ideas for improvement surrounding all of these quality measures we are submitting?”

The answer of course is no, not yet,  but that is by design of the whole Meaningful Use project.  Stage 1 is focused on getting everyone to capture data in the same format.  Stage 2 is focused on the transmission of that data in a standardized format and it isn’t until Stage 3 that there is supposed to be an analytical component.  However, that got me thinking about how this whole Big Data thing is going to play out in the area of healthcare delivery.

You might remember previously I have written a warning on how vague analyses of big data sets can lead to some incorrect conclusions.  In other words, we as humans really, really want to see a pattern so sometimes, we just pretend one is there.  But what about the patterns that do exist and how can we harness this knowledge to improve healthcare delivery?  How will the upcoming onslaught of vast amounts of data effect the industry?

Recently, Felix Salmon over at Wired wrote a wonderful article entitled Why Quants Don’t Know Everything.  In it he describes how this quantification movement has played out in various industries thus far and has identified four stages in which this progresses.  I’ll let you link over to the article for a richer explanation, but in short, here are the stages:

1. Pre-Disruption: No standards exist and there are no big data sets to be mined.  Success in the industry is achieved by those who do so in isolation based on individual principles or theories.
2. Disruption: Standardized data is collected and individual consumers are targeted based on analysis of the big data sets. Metrics are prevalent.  Seemingly prophetic conclusions are able to the be reached by pushing the data into robust mathematical models.
3. Overshoot:  Now that metrics are in place, enterprising individuals find ways to game the system to achieve high marks within these metrics.  This is known as Campbell’s Law.
4. Synthesis:  In order to stave of the gamers of the system, some metrics are painted with a bit of old fashioned common sense.  In other words, we realize that the data we collect may not tell the whole story so we apply a bit of human intuition to enhance the insight.

An industry like healthcare delivery is definitively in the Pre-Disruption stage.  We are just beginning to standardize and collect data on the quality of care alongside general vital measurements that will be curated, sifted, and sent up the chain for analysis.  The disruption will come when someone can harness this data to modify or even subvert the current healthcare delivery model and hit the metrics better than anyone else.  Inevitably, there will be overshoot when we realize that by simply meeting the metrics, we aren’t necessarily providing better healthcare, because good health requires some immeasurable components like listening compassionately and physician insight.  The question is will this third stage be reached without a major negative impact?

Finance is arguably in this third stage of overshoot right now and moving into the fourth stage of synthesis as is evidenced by the “Common sense” laws that are now being attempted to be put in place.  However, we had to enter into a period of major economic decline that had a largely negative impact on the financial health of the population.  How does healthcare delivery avoid the potential negative impacts of creating a system that can be gamed? This question will no doubt be in the back of my mind over the coming years as the data sets become bigger and they begin to be mined by enterprising individuals.

Honestly, I look forward to the upcoming disruption that will occur since it is desperately needed.  I’m sure the growing bio-surveillance movement will mesh nicely with this push and it will be exciting to see what is developed.  However, my hope is that we can proceed into this new data awareness with the notion that sometimes good healthcare delivery requires something that is not quantifiable or measurable: the human touch.