What you don’t need to know about ICD-10

Last blog I talked about why the number of ICD-10 codes is not a big deal. This blog describes two other aspects of ICD-10 you don’t need to spend time or money to learn in detail.

Under normal circumstances I could have covered this information in a few short bullet points. These details are about as relevant to your practice as learning the Library of Congress book-cataloguing scheme when what you actually want to know is the title of a book. But opponents of ICD-10 have made a big deal of them, so I am going to spend some time here demonstrating why they are not such a big deal.

First, you and your coding staff do not need to be trained in the ICD-10 Procedure Coding System (ICD-10-PCS). ICD-10 procedure codes are the hospital inpatient coding department’s responsibility, because it is the only place they are required to be used. The hospital coding staff will be trained to use the operative report just as they do now with ICD-9: Take the operative report provided and assign the correct PCS code or codes based on the classification system’s definitions and guidelines. If they have questions, they will query you just as they do now.

Next, ICD-10 opponents have made a big fuss about the fact that codes begin with a letter and are up to seven characters long. The next few paragraphs are about how ICD-10-CM is structured, and why that should give it a more useful shelf life as a disease classification system. As a direct descendent of the World Health Organization’s International Classification of Disease, Tenth Revision, ICD-10-CM (the version customized for the U.S.) divides the universe of disease and health-related conditions into 21 very broad chapters that correspond either to a body system like the digestive system, a type of disease like infectious disease, or a supplemental piece of information like family history of disease.

All ICD-10 diagnosis codes begin with a letter simply because this gives the system a “larger frame” for sensible expansion of the classification within the chapter. It’s like a big set of dresser drawers that has enough drawers to begin with, instead of having to use the same drawer for socks and underwear.  The 21 chapters are spread across 26 letters instead of 10 digits.  For the most part, a body system or type of disease is allotted one letter each. (Which is, incidentally, much easier to remember than some range between 001 and 999. I never knew the associated ranges for any of the chapters in ICD-9, but without any conscious effort I know these things in ICD-10: Respiratory system codes begin with J, obstetrics codes begin with the letter O, injury and poisoning codes use S and T, infectious disease codes use A and B, and so on.)

A complete ICD-10 diagnosis code can be any length from three to seven characters long. Much has been made of the fact that the additional detail in ICD-10 is required to be used, as if with ICD-9 you can leave off digits of a code if you want to. This has never been the case for ICD-9 and it is not true for ICD-10. The HIPAA definition of a valid code includes all of the detail contained in the classification for a particular category, sub-category or sub-sub-category.

Let’s look at examples from both ends of the code detail spectrum:  Parkinson’s disease and pressure ulcers.

Parkinson’s disease is coded to G20. Parkinson’s disease has its own category, and that is the complete code. No further axis of classification has been added, so the complete code is three characters long.

Pressure ulcer codes are six characters long. There are three additional axes of classification for pressure ulcers: body site, laterality and stage.  (The detail in each additional axis of classification is tacked on successively to the basic disease category, so the whole thing maintains its integrity as a statistical tool.) The pressure ulcer category is L89; pressure ulcer of heel is L89.6; pressure ulcer of right heel is L89.61; and pressure ulcer of right heel, stage 2 is L89.612. Statisticians and analysts can comb their databases for all pressure ulcers by using category L89, but the code on an individual patient record must contain all six characters.

This is not to say that you or any coder who works for you needs to memorize codes, axes of classification and the like. Just know that as with ICD-9, code lengths differ in different areas of the classification, the length of the code depends on how much detail has been added to a particular disease category, and if the level of detail is there, it is in your interest to document for that level of detail. In the case of pressure ulcer codes, I would wager that you already document to that level of detail. If you do not document the stage by number, you describe the stage clinically, which a coder can correlate with the correct stage number.

Rhonda Butler is a senior clinical research analyst with 3M Health Information Systems. She is responsible for the development and maintenance of the ICD-10 Procedure Coding System since 2003 under contract to CMS, and for the development and maintenance of the ICD-10 General Equivalence Mappings (GEMs) and Reimbursement Mappings under contract to CMS and the CDC.  She leads the 3M test project to convert the MS-DRGs to ICD-10 for CMS, and is on the team to convert 3M APR-DRGs to ICD-10. Rhonda also writes for the 3M Health Information Systems blog.