ICD-10 requirements went into effect on October 1, 2015. ICD-9 codes will no longer be accepted for Medicare reimbursement and impacts ALL HIPAA compliant entities.
The Centers for Medicare and Medicaid Services (CMS) announced that for 12 months, Medicare will not deny physician claims based solely on the specificity of the ICD-10 code. This gives the healthcare community time to further prepare for the level of specificity required, but physicians are still required to provide a valid ICD-10 code.
At the point of care, providers are primarily concerned with telling the patient story – not which ICD-10 code to assign. Traditionally, this requirement has been passed downstream to coding staff who then make their best efforts to interpret the provider’s narrative after the fact.
Too often, however, these narratives don’t capture the amount of detail needed for coding staff to properly complete the ICD-10 coding exercise. For example, is the patient presenting with this diagnosis for the first time or is it a chronic condition? Which side of the body did the injury occur? These levels of specificity are known as temporality and laterality respectfully and in 2016 must also include extensions; initial encounter, subsequent encounter and sequela. If the provider doesn’t capture these details at the point of care, coding staff must track them down later or risk denial of reimbursement.
The more time elapses from point of care for ICD-10 coding, the greater the error potential for lost or inaccurate submissions. This translates into both lost revenue AND lost opportunity for capturing important data elements that can be used in population health management (e.g. diabetes and CHF patient populations)
The Solution – SayIt Code Fast
Today’s health data is estimated to be comprised of 70-80% narrative and unstructured text. This is how providers think (and talk). Without a high quality narrative as a foundation, it’s difficult if not impossible to create the necessary level of detail for ICD-10. Speech recognition solutions such as SayIt from nVoq help create these narratives in real time, at the point of care, accurately and efficiently.
How can providers bridge the gap between their traditional workflow of telling the patient story in their own words, vs. the level of granular detail required for ICD-10?
Natural Language Processing (NLP) can help code narrative text and unstructured data. The nVoq Code Fast solution utilizes SyTrue, Inc.’s (California, USA) NLP Smart Data Platform to accurately identify and create structured data for ICD-10, taking into account contextual awareness and negative findings, and automatically applies the appropriate ICD-10 code – in real time at the point of care.
How does Code Fast work?
Providers dictate their notes as usual—in a narrative fashion. In a matter of seconds, Code Fast extracts key data field concepts and returns the appropriate ICD-10 code(s). This allows providers to meet the October 1, 2015 requirement of applying the correct ICD-10 top level code(s). Even better – Code Fast goes one step further by providing a sub-menu to select more specific codes, allowing the provider to choose the appropriate detail diagnosis or problem associated with that encounter. Coded data is then returned to the system of record, ready for posting, and available for data mining and analysis.
With SayIt Code Fast, providers can bridge the gap between telling the patient story and compliance with both current and future ICD-10 requirements. Further, they can do this in real time, at the point of care – saving time, improving accuracy of documentation, and maximizing reimbursement potential.