Good news about ICD10.
ICD-10: New Codes, New Opportunities
The adoption of ICD-10 is a gigantic leap into the unknown for medical coders and healthcare organizations on either side of the payment process. The comfortable, time-tested relationship with ICD-9 codes will be shelved for the new, vastly unfamiliar codes that comprise ICD-10.
Not surprisingly, the mere thought of
diagnosis codes jumping from the thousands to more than 100,000 in October 2011 is already causing administrators and coders alike to reach for the aspirin. Keeping up with the codes will be a tremendous task for coders and healthcare organizations that have spent more than 30 years using the ICD-9 system.
But the conversion to ICD-10 – directed by the Department of Health and Human Services' Centers for Medicare and Medicaid Services – is about more than replacing one set of codes with another. It represents an unprecedented opportunity to improve the quality of information, which will – over the long run – benefit every constituent of our healthcare system.
ICD-10 gives the healthcare industry the chance to improve the way it handles documentation and coding operations through technological capabilities and human expertise and, at the same time, to elevate the power of biosurveillance and pharmacovigilance.
Considering, however, that even small changes to the current ICD-9 system can net 80 percent to 90 percent denial rates initially, the overwhelming scope of the disruption stemming from ICD-10 cannot be understated.
Given the diversity in size, specialty, and payer mix within the healthcare industry, the complications of the change are nearly impossible to properly measure. But so, too, are the new codes' possibilities in terms of technology adoption, research, and discovery.
A Computer-Assisted Lifeline
The enormous increase in the number of codes available will likely cause many healthcare organizations to seek technological solutions to help deal with code selection. While the use of computer-assisted coding (CAC) is not new, the adoption of ICD-10 may cause adoption rates to increase considerably.
In fact, a recent survey of medical practices shows that two-thirds expect to purchase software to manage the change. CAC will offer organizations a lifeline during a chaotic time – automated code selection from a sea of unknown new codes. At its core, sophisticated CAC helps organizations harness incredible amounts of information and streamline the coding process.
Through natural language processing (NLP)-driven software, CAC applications “read” physicians' dictated reports and, to varying degrees, extract clinical language to assign appropriate and accurate codes for patient encounters.
The technology already reduces the demands on human coders by removing their involvement in simple and repetitive coding tasks, letting the technology tackle those tasks instead. This frees human coders to focus on the more complicated and challenging aspects of coding.
CAC will do the same thing – and more – with ICD-10. Rather than leaving human coders to dig through ICD-10 coding manuals, the technology will mine medical reports, assign relevant codes, and present them to human coders to review and approve. Through this process, coders will be able to learn along the way and expand their familiarity with ICD-10.
This alone will be a tremendous benefit for organizations that fear the effects that cost, training, and time investments in ICD-10 will have on their staff. Healthcare organizations should heed warnings of a Band-Aid® approach with CAC, as all systems are not created equally.
Term-matching or pure rules-only systems employ an “if, then” approach: If such software detects the phrase “diabetes,” it will code for diabetes, but it may not catch contexts that could change the diagnosis (such as “negative for diabetes”), nor information elsewhere in the record that supports a more refined diagnosis (i.e., reference to malnutrition).
More advanced systems use NLP to evaluate the entirety of a medical report, scanning for all possible diagnosis information. This is especially relevant for ICD-10, where the volume of codes will increase to about 155,000 from ICD-9's 17,000, in order to tackle diagnosis coding at a more granular level.
Organizations should also be aware that some systems are configured to send reports directly to billing after simply matching codes to a list that a claim scrubber would approve, without more sophisticated validation of the codes given the evidence in the dictation.
This lack of proper review introduces significant compliance risks, especially in light of an entirely new code set – trying to predict what a scrubber would approve with ICD-10 is an unknown gamble.
CAC applications should verify accuracy based on national coding guidelines, coding results from a myriad of organizations, input from certified coders, and statistical analysis, in addition to advanced coding and statistical technologies.
The technology should also be able to offer continual reporting and aggregate analysis of the changes that review coders make to its output, to highlight possible areas for technology refinement and coder education – a vital element as human coders begin to learn the new ICD-10 code sets.
Operational and Medical Insight
Beyond the opportunity for broader adoption of coding technology, the additional promise of ICD-10 in the United States is coming into focus. The current ICD-9 system was developed more than 30 years ago and simply cannot support today's healthcare, much less that of the future.
The codes are already stretched to cover current diagnoses, because today's practice of medicine has grown to involve concepts unthinkable in the 1970s. Medicine, technology, and diagnoses have advanced incredibly in the last three decades, while coding remains mired in outdated systems, and essential healthcare information too often remains locked away in unanalyzed clinical dictations.
With ICD-10's new organization of codes, medical coding can catch up to today's healthcare because ICD-10's granularity offers the industry a real chance to revolutionize the way it gathers and processes information. Driven by vast amounts of data, rules, and analysis, ICD-10 dives deeper into the building blocks of diagnosis than ever before.
At the same time, NLP technology excels at extracting such structure from unrestricted medical language. By pairing deeper standards of description with NLP technology's ability to automatically map language to structured information, healthcare organizations can discover patterns, identify outliers, and create flexible and powerful new windows into their operations.
NLP's capacity to map from language to structure has already proven its value in other domains. Automatic information extraction serves as the foundation for surveillance, interactive search, and knowledge discovery applications in business intelligence, homeland security, and biomedical research. What healthcare could do with the same kinds of insight is boundless.
ICD-10 gives the data to support – among other things – biosurveillance, pay-for-performance initiatives, safety improvements, quality measurements, and more accurate reimbursement rates. In short, the new code set offers a better way to organize data, and ultimately provides higher-quality information to gauge the safety, efficiency, and quality of care.
By embracing ICD-10, the American healthcare industry will adopt a richer, more systematic, and more extensible approach to clinical documentation. Doing so will reveal unprecedented opportunities to link documentation and diagnoses to new knowledge and deeper medical understanding.
— Andy Kapit, MBA, is CEO of CodeRyte Inc. (
www.coderyte.com), based in Bethesda, Md. Questions and comments can be directed to
editorial@rt-image.com.
