HIT Think

Why providers still must take unstructured data into account

We have arrived at the digital transformation era in healthcare. We have full-throttle EMR use in multiple settings, database sizes that rival our on-premise storage assumptions, we are checking our boxes to meet meaningful use metrics and electronic workflows that decimate the paper-pushing of the past.

Yet through this transformation, discrete data abounds. Clinicians now trade the completion of paper forms and disjointed information-sharing for the clean “click-on-the-drop-down” form of creating documentation.

EMR purists will tell you:

  • “The only way to survive in a value-based reimbursement environment is to leverage analytics and BPM tools in more aggressive ways.”
  • “The best way to predict successful outcomes is to use evidence based-practices.”
  • “The only way to manage population health is through the analysis and aggregation of discrete data.”
  • “Evidence-based analysis needs to occur more often at the point of care, tailored to patient-specific variables.”

I agree with these positions and have witnessed the transformation of documentation practices first-hand on the frontlines of the HIM profession. However, we need to see where we stand technologically and to challenge a few of these paradigms.

The burden of data entry on providers of all types is real. Notice my reference in the first paragraph to the “clean” form of documenting—this does not always equate to the “quick” form of documenting. In a recent study featured in Annals of Internal Medicine, it was estimated that for every hour spent with the patient, providers spend two hours documenting patient care. Even Vindell Washington, our new National Coordinator for HIT, contends that we are still in evolution when it comes to the usability of electronic records.


Has the pendulum swung too far? While it’s true we need discrete data, it is also true we are sitting on a wealth of advancement in natural language processing technology along with very mature ways to manage and orchestrate the integration of unstructured healthcare data.

Admittedly, the value is in the eye of the beholder. Ask any physician, “what is most important and when?” and you will most likely hear, “access to the right information at the right time in the patient care process.”

Patient-centered care is all about benefiting from this access. Aside from the interoperability aspect, the ability to get to information you need from your systems for direct patient care trumps the broader strategy of compiling mountains of discrete data.

In light of this new reality, we must begin to rely on technologies that help us garner value from unstructured data, combine it with discrete data and formulate analytics sources from the resulting unified body of data.

I am reminded of a popular quote that says, “Just because you CAN automate something doesn’t always mean that you SHOULD.” We have the solutions needed to move away from paper. We also have the technology to deconstruct unstructured data. These must be woven into our next phase of clinical documentation evolution.

For reprint and licensing requests for this article, click here.