Shouldn’t you be able to quickly tell when healthcare codes don’t match up? That’s the question Andrew Boyd, an assistant professor of biomedical and health information sciences at the University of Illinois at Chicago, asked when he found out about the U.S. healthcare system’s upcoming transition to a new coding system.
On October 1, 2015, the $3 trillion U.S. healthcare system will begin transitioning from the ICD-9 to the ICD-10 coding system. While ICD-9 provided about 14,000 codes for doctors and hospitals to choose from when billing insurance companies for various patient treatments, ICD-10 includes more than 80,000 codes. Further complicating matters, the code definitions are not apples-to-apples. For example, when using the ICD-9 coding system, a patient can be described as having either a “depressive episode” or “clinical depression.” In the new coding system, both types of depression are classified as “depressive disorder.”
These changes mean big challenges ahead for IT administrators. Come October 1, when they run reports on things like the best treatment for a specific condition or to evaluate the safety of a particular hospital, they’ll have to compare two sets of data with different collections of codes and definitions.
Boyd quickly realized what was at risk with this coding transition. He was concerned that the switch would invite costly errors in coding and patient information, as well as inaccurate reports or reports that were easily skewed. For example, a hospital looking to paint a misleading picture when it comes to safety could potentially leave out important sets of data from previous years, just because the new coding system doesn’t have similar definitions.
Boyd knew there had to be a solution to this dilemma; a way to simplify the transition to the ICD-10 coding system. So, he set out to create an algorithm to identify when definitions are convoluted and when certain reports are near impossible.
What Boyd and a team of researchers came up with is a free web portal tool and translation tables designed to provide IT administrators guidance on ambiguous and complex translations, and to reveal where analyses may be challenging or impossible. The tool identifies bi-directional conversions and their level of complexity, which includes identity, class-to-subclass, subclass-to-class, convoluted, and no mapping.
“Every single healthcare encounter is assigned one of these codes,” Boyd says. “Anyone who wants to understand what’s going on runs reports. Even small practices have incentive payments based on these codes.”
Boyd adds that, “The coding system is used to make many decisions in healthcare, from specific conditions that are affected seasonally to identifying problem areas in hospitals, such as when more nurses are needed or where supplies are falling short. There was too much at risk to not address the differences in these two coding systems.”
With Boyd’s algorithm, an IT administrator can quickly determine problem areas, predict potential errors and improve productivity when running reports. If an IT administrator is looking to run a list of 10 codes related to depression, for example, all they have to do is plug data into Boyd’s algorithm and they can quickly find out which code definitions are convoluted. The IT administrator can then sit down with a domain expert or a professional encoder to see if the report would still be possible, or if they need to decide to map to other codes.
The concept behind Boyd’s algorithm is applicable to many different scenarios. For example, it can also apply to data collected from dating websites, when looking at male and female groups and nonreciprocal relationships between the two groups.
Boyd is hopeful that his algorithm will help solve a complex problem. “In 3 or 4 years, billions of dollars in Medicare payments will be tied to these new definitions,” he says. “We hope this tool will help IT administrators determine which areas within the two sets of data are worth investing the time or money in to hopefully find a solution.”
Boyd adds, “In other cases, it might be okay to go with, say, a 10 percent error when running certain reports, especially if it’s intended to be used internally. It helps knowing that’s the case, so practices can take that into consideration.”
For more information on the future of healthcare IT go to: http://healthinformatics.uic.edu/
Renee Morad is a freelance writer and editor based in New Jersey. Her work has appeared in The New York Times, Discovery News, Business Insider, Ozy.com, NPR, MainStreet.com, and other outlets. If you have a story you would like profiled, contact her at [email protected]
The IT Innovators series of articles is underwritten by Microsoft, and is editorially independent.