Bookmark and Share PrintE-mail
  ICD10 Watch
by Carl Natale


Why SNOMED cannot replace the ICD-10-CM/PCS code sets

I'm not a fan of the American Medical Association's (AMA) quest for an ICD-10 alternative because I don't believe they're serious. But I hope they don't take Jon Handler's idea to abandon ICD-10 coding and use SNOMED-CT.

What is SNOMED-CT?

The Systematized Nomenclature of Medicine — Clinical Terms (SNOMED-CT)  was created by the College of American Pathologists (CAP) to represent medical terminology in electronic health records (EHRs) It is now owned, maintained and distributed by the International Health Terminology Standards Development Organisation (IHTSDO). Here's how they describe it:

"SNOMED CT provides the core general terminology for the electronic health record (EHR) and contains more than 311,000 active concepts with unique meanings and formal logic-based definitions organized into hierarchies. When implemented in software applications, SNOMED CT can be used to represent clinically relevant information consistently, reliably and comprehensively as an integral part of producing electronic health records."

It's incredibly granular and modeled after how clinicians communicate. Sounds great. Then why don't we use that instead of any ICD code set?

Terminology vs. classification

First of all, SNOMED-CT is a terminology that is too detailed to be used for reporting. It is designed for input. ICD is a classification designed for output. It aggregates the details being input into codes designed for reporting. So when ICD-10 is criticized for not having enough specificity, that's kind of a feature not a glitch.

Second, SNOMED-CT is designed to be managed by computer. It's not just a flat list of numbers and corresponding terms. It's a complex relationship of concepts. Check out the 96-page User Guide for an idea of how it's organized. It's a database.

By storing the clinical data in a robust set of codes, it can be mapped to ICD-9 codes and eventually ICD-10 codes.

Why not SNOMED then?

Physicians are going to have to learn how to communicate with EHRs — which will be based upon SNOMED — to comply with Meaningful Use. So the transition to SNOMED-CT already is in the works.

But all that data collected using SNOMED-CT will need to be reported in a classification system. Sure ICD-9 is such a classification system. But it doesn't have new diseases, procedures, risk factors and injuries. Many are part of the ICD-10 code set because physician groups want those statistics.

And even if physicians don't believe that, specific reporting will be required by the Affordable Care Act so new quality factors can be tracked.

Basically, ICD-10 codes aren't the problem. It's the specificity of documentation that will be required one way or another. SNOMED should make it easier to document to the required specificity. It is then up to the EHR system to convert that data to ICD information. Hopefully the physicians won't know what level of ICD is being used. They will just need to know what needs to be recorded.

It is very appropriate to talk about how physicians can use SNOMED-CT to enter data. It just doesn't go far enough.

Resources

Here are some papers and pages on how SNOMED-CT should be used:


Comments

Carl, thank you for the

Carl, thank you for the comments on my post!

I completely agree that the key challenge is more specific documentation regardless of the terminology we use to encode the diagnosis. At the same time, the issue is not just about getting more detail, it’s about getting the right detail. Using ICD-10 will force clinicians to document details of low relevance to modern needs, and lead to wasteful duplication of their documentation efforts. Even if it were an offered feature, EHRs are often unable to auto-transform ICD-10 to SNOMED-CT because of mismatches between the terminologies. Use of compositional SNOMED-CT for diagnosis coding would address those problems. While ICD-10 is a better terminology for research than ICD-9, SNOMED-CT is even better. With its multi-level hierarchy, SNOMED-CT is even better suited than ICD-10 for aggregating highly specific codes into clinically meaningful classifications.

I posted a follow-up to my post that discusses and explains these points in more detail, including mention of major institutions who support this recommendation:

http://mmodal.com/articles/icd-10-vs-snomed-ct-part-2

Thanks so much for the great discussion!
 
 

Advertisement. Closing in 15 seconds.