Managing Reference Data in the Era of Analytics and Big Data

RHIA,Global Healthcare Industry Ambassador, IBM Information Management

Managing reference data in a healthcare system or a payer organization is not easy. While trying to wrap my head around this vexing requirement, I pondered how many coding systems I’ve learned in my health information management career. 

Despite still being 39 (I celebrate the anniversary of my 39th birthday every year), I’ve learned three ICD coding systems and must now address ICD 10. This doesn’t even count all the specialty terminology systems that exist for radiology, lab, pharmacy and the like.  And if you really want to appreciate the challenge of all this reference data, add the code tables that must be managed across systems to represent demographics, financial and regulatory data sets. 

My head spins when I think of all the spreadsheets and processes commonly used to map all this data.  Here’s an illustration of demographic and financial data.

The “aging” and mapping of ICD, LOINC, RxNorm and other clinical codes/terminologies are much more complex than the basic demographic above. Standards organizations and regulatory bodies may govern when updates are issued (perhaps even annually) and the dates they must be applied. Financial or other penalties may be imposed if these codes are not accurately maintained. 

Let’s not forget the patient perspective. The stakes may be higher for this use of reference data, since insights regarding individual and population health will be derived from the analytics using the clinical reference data. 

While instinct might be to just add another spreadsheet to manage the latest requirement, or change 12 spreadsheets housing the data being updated, this is not the prudent approach, given a few key healthcare pressures.  

Control costs.  Pressure to control costs has never been higher. Yet, hundreds (if not thousands) of hours are spent each year updating spreadsheets of reference data. Maintaining today’s cumbersome approach can annually cost tens or hundreds of thousands of dollars. Efficiently managing reference data can save money and allow an IT organization to acutely predict reference data management costs.   

Use precious IT resources wisely.  The IT resources that may manage your spreadsheets or do mapping within systems may be the resources you desperately need for installing a new electronic medical record (EMR), upgrading your existing EMR to meet Meaningful Use requirements, or preparing data for important analytic initiatives. These resources are simply too precious to be using them on mapping reference data in spreadsheets. A far better approach would be to master your reference data and make changes once as terminology or requirements change, and propagate this to the consuming systems. 

Become data-driven healthcare organization.  Several forces drive the push to become a data-driven healthcare organization , including Meaningful Use, ACO establishment or analytics. Organizations must blend historic and new data to generate insights about delivering better, more cost-effective care. Advancing this strategic need creates even more demand for efficient, accurate reference data management. Woe to the healthcare organization that begins using data to advance a strategic imperative and then discovers that months of planning are squandered because the reference data is not accurately mapped as terminologies, nomenclatures or financial data have changed. 

I may relinquish celebrating the anniversary of my 39th birthday after I learn ICD 10. However, I know that I’ll have more terminologies to master, and the demands for accurate, efficient reference data management will continue to increase. So, perhaps I’ll just stick with 39!