The IT Side of the ICD-10 Implementation Equation

November 11th, 2013 / By Jason Mark

Last month I attended the AHIMA Convention in Atlanta, and everywhere I looked it was all about ICD-10. With less than a year left to implement, there will be an increasingly frenzied push to make the Oct. 1, 2014 date. I hope organizations won’t be short-sighted when planning for the IT side of the implementation equation. All too often, solutions are put into place with the goal of meeting an immediate need without a lot of thought about long-term implications. Next year the immediate need will be success at processing claims coded under ICD-10. I worry that within many organizations once that is done the “box will be checked” so to speak and the project will be seen as a success. Obviously in the short-term this part has to go well, but what comes next?

ICD-10 will dramatically expand the universe of coded data from healthcare. I understand that there is debate in the industry about the value of, say, ICD-10 coded data vs. SNOMED coded data, but regardless of where one stands on that issue, the fact of the matter is there WILL be more data, period. In today’s world of Big Data and analytics, the more relevant data that is used in algorithms and models that analyze and predict future events, the better those analyses will be. With the increased specificity required by ICD-10, groups of patients may be identified with higher degrees of similarity. Similarly, cost data can be analyzed with more direct correlation to particular conditions or combinations of conditions. A wave of transformations for the healthcare system will likely result from analytics built on top of these ICD-10 datasets.

However, in order for any of that to happen, the data has to be available in a way that makes such analysis possible. IT teams should be approaching ICD-10 with the realization that once the immediate need to process claims is met, one of the next steps will be to enable analysis of this new data. A little extra planning to that end today will simplify things in the future. For any IT professional working on ICD-10 implementation, I think it’s important to anticipate how to best make that data accessible for future analytical uses. Data sets that need to be available will likely be larger and for a period of time there will likely be a need to maintain data coded in two schemes (ICD-9 & ICD-10) for comparison to historical reporting. It will also take some time before enough ICD-10 data has been collected to be useful. Given the higher specificity of codes, it could well turn out that a longer time period of gathering data is needed. For some purposes, such as identifying a sample of patients with a very narrow coded condition for instance, it may take significantly longer to have enough cases to prove data is useful or statistically valid.

In technical roles such as software development and IT, we can often see ways to improve our architecture and systems. Rarely do we have a mandate such as ICD-10 that gives us the opportunity to really stop and make the changes we need to. Too often there are time restrictions or resource constraints. While these constraints certainly do exist around ICD-10 implementation, we have an opportunity to better position ourselves to support future needs. Smart organizations will see this and take advantage of the opportunity to position themselves for life “after” the ICD-10 implementation and not simply “check the box” on Oct. 1, 2014.

Jason Mark is Director, Emerging Business Technology with 3M Health Information Systems.