Wednesday, August 15, 2012

BigData needs Outcomes

I've had some interesting reading assigned to me by Twitter and G+ over the past couple of days.  The first of these was this artcile by John Glaser: From the Transaction-Based EHR to the Intelligence-Based EHR.  In that artcile, Dr. Glaser talks about how the next generation Health IT solutions need to focus less on what was done (the transactions), and more on the processes.

Next up is this article from O'Reilly's Radar: Solving the Wanamaker problem for health care.  In this article, Tim O'Reilly and crew talk about being able to use data to discover the most effect treatments.

There's a major challenge with these ideas, and it means a fundamental shift in how physicians practice medicine.  I've looked at the structures used for thousands of different types of clinical documents (really, and not just some of the standard ones).  Not once do I recall seeing a section for the critical piece of data that will enable the capabilities discussed by these thought leaders.  To change healthcare, and really make use of big data, we need a place for it.

So, where do outcomes go on the patient's chart?  


Think about it.  If you don't record success or failure, how will you ever know what works?


1 comment:

  1. Isn't success and failure recorded in the temporal relationships in a patient's EHR? If the patient presents with high blood pressure and is given an antihypertensive, then that patient comes in two weeks later, reporting that they are taking that medication while presenting with lower BP, isn't that an outcome? Now one might question whether there is a clear cause-and-effect between the antihypertensive treatment and the lower BP, and maybe more information about the patient's original health characteristics and their new status, can provide some insight, especially given other patients with similar pre-treatment and post-treatment status. So I'm not getting what other information is needed to assess success and failure.

    ReplyDelete