AMA's Opposition to Bill on SGR Fix and ICD-10 Delay Is a Winning Ploy
It is time for the U.S. to begin implementing health IT smartly

Ask Not What ICD-10 Can Do For Healthcare, Ask What Healthcare Can Do With SNOMED and ICD-11

ICD-10 is so “last century”.    The United States did not adopt ICD-10 twenty years ago when the standard was first developed.    The current version of ICD-10 that the United States is designated to adopt is based primarily on the international version of ICD-10 that the World Health Organization (WHO) published in 1990.    The international version was drafted by committees that began their work over thirty years ago in 1982 (see 2nd Edition of ICD-10 by WHO).    In other words, our version of ICD-10 is based on work done before use of the rich information space called the Internet became common and before the human genome was mapped.

ICD-11 is “this century”.    According to an article in Healthcare Financing News, Christopher Chute who is one of the leading informatics experts and a Chairman of an ICD-11 Revision Steering Group at the World Health Organization stated:

“ICD-11 will be significantly more sophisticated, both from a computer science perspective and from a medical content and description perspective…. Each rubric in ICD-11 will have a fairly rich information space and metadata around it. It will have an English language definition, it will have logical linkages with attributes to SNOMED, it will have applicable genomic information and underpinnings linked to HUGO, human genome standard representations. ICD-10, as a point of contrast, provides a title, a string, a number, inclusion terms and an index. No definitions. No linkages because it was created before the Internet, let alone the semantic web. No rich information space.”

ICD-x codes are used by non-clinicians for important administrative and financial purposes.    SNOMED-CT, on the other hand, is what physicians will actually use to communicate information about patients in their electronic health records (EHRs).    In fact, physicians must use SNOMED vocabulary in their EHRs, not ICD-x codes, for their problem lists in order to achieve Stage 2 Meaningful Use for incentive payments and to avoid Medicare penalties in the future.    Unlike ICD-10, ICD-11 is based on SNOMED.  And SNOMED includes over 311,000 concepts with unique meanings, making it more granular than ICD-10 or ICD-11.  

One way to think about the relationship is that SNOMED is the input and ICD-x is the output.  SNOMED is used by clinicians to input clinical information into the EHR at a high level of detail.  ICD-10 and ICD-11 aggregate that data into less detailed classifications that are more useful for output purposes such as quality reporting.    They really cannot replace each other.   But we could and should require EHRs to map in the background the SNOMED codes used by physicians into the ICD-x codes used by others.    No need to engage physicians in ICD-x debates or to learn new vocabularies each time WHO does their thing with the U.S. traditionally following way behind.

So what the HIT are we thinking?    Do we really believe that healthcare quality will be significantly improved based on ICD-10 that was developed out of work done over 30 years ago before the Internet was commonly used and before human genome coding was completed?    Or do we believe that we need to adopt ICD-11 for output purposes and to use SNOMED–CT in EHRs for input purposes in order to move the quality needle in the right direction?

I for one believe that we need to get to ICD-11 as soon as possible.     And I believe we should cut the umbilical cord to ICD-10 right now because:

  1. There is currently no information showing that a conversion to ICD-10 is required before ICD-11.
  2. It is intuitively obvious that the costs of going to ICD-11 directly from ICD-9 would be less than incurring the remaining costs of implementing ICD-10 in 2015 (or later) and then implementing ICD-11 sometime thereafter.    And that includes the sunken ICD-10 costs.    If you believe that this is an outrageous assumption, then prove it to be untrue.    Show the comparative costs of both pathways.    But don’t just comment or blog that it’s ridiculous without providing some kind of evidence.     Sometimes it’s wisest to go with intuition.
  3. The ICD-10 implementation has been so painful that it is unlikely the industry will have the stomach to move on to ICD-11 within a decade.    This will result in an excessively long delay to ICD-11 and an excessive period of time using a classification system from the previous century.
  4. There is consensus among leading informatics experts that ICD-11 is superior to ICD-10

 

Matt Murray, MD

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Agree with sound argument. ICD-11 and the ICD-11CM version for the US whether it is available makes sense to adopt rather than ICD-10. According to WHO, ICD-11 will not be available until 2017 (they have enjoyed their own series of delays) and we could, and must use SNOMED-CT for meaningful use anyway. Question: how does one convert SNOMED-CT to ICD-10? The NLM crosswalk is available (last update July 2013) to map diagnosis across nomenclatures, but what about laterality and "initial" vs "subsequent" encounter? There is no component of the SNOMED-CT code that speaks to encounters.

Those are good examples of how SNOMED alone would not be able to support the detail that is currently expected to be required for billing in the future. From the physician perspective I believe we should use SNOMED in our EHRs and that we will learn to enter somewhere in our documentation enough information (laterality et al) for the coders to be able to enter the right ICD-x codes. Pick your version of ICD-x, and I will still enter the same documentation notes and SNOMED codes. And an opinion for another post is that the amount of detail that is going to be expected for billing is actually more important to address than what coding set is going to be used for billing.

The comments to this entry are closed.