Differences that Impact MS-DRGs: Esophageal Hemorrhage Doesn’t Just Happen

In my first blog for 2013, I laid out what I plan to write about for the next little while, and why:

  • What—information that gets lost in translation from ICD-9 to ICD-10 and impacts MS-DRGs
  • Why—because words matter, because the connection between codes and patients needs to be kept clearly In mind, because uncritically analyzing ICD-10 “data” that is the product of ICD-9 translation as if it were real and authoritative risks missing important clues when a prediction of impact has strayed from reality. Because, because…

For this blog I am going to start with a straightforward example of a codeable distinction made in ICD-9 and discontinued in ICD-10—something you can “say” in ICD-9 that you can’t say in the same form in ICD-10. This difference can cause predicted ICD-10 MS-DRG reimbursement to shift—down. This particular shift is ranked in the top ten for impact (overall national impact across all DRGs is predicted at pennies per hundred dollars), though I chose this example because it’s a good example and looked up the ranking later.

You can “say” esophageal hemorrhage using a single, unique diagnosis code in ICD-9, 530.82 Esophageal hemorrhage. When coded as the principal diagnosis, 530.82 is assigned to MS-DRG group 368-370 Major esophageal disorders. You cannot say “esophageal hemorrhage” using a single, unique code in ICD-10. Instead, esophageal hemorrhage is included in the “garbage code” K22.8 Other specified diseases of esophagus, along with a bunch of other conditions (e.g., cyst of esophagus, esophagocutaneous fistula, leukoplakia of esophagus ). I say “garbage” because as coded data, it is noise rather than real information. The variety of things that could be wrong with patients who have this code on their record is too diverse. All the code tells you is that a patient doesn’t have a diagnosis captured by a more specific code.

Code K22.8 as the principal diagnosis is assigned to ICD-10 MS-DRG group 391-392 Esophagitis, gastroenteritis and miscellaneous digestive disorders, lower in the DRG hierarchy and therefore paying less than the ICD-9 coded version of the same record. Using translated records from two large data sets, on a national average, there are 394 shifts per million predicted for the Medicare population and 149 shifts per million predicted for the non-Medicare population. How much money this predicted shift means for your organization depends on both your case mix index and the frequency of this type of coded record in your own data.

Why is K22.8 assigned to a lower-paying DRG in ICD-10? Because there is a similar garbage code in ICD-9—530.89 Other specified disorders of esophagus—and it is assigned to the same lower-paying group in the ICD-9 MS-DRGs. Since the stated goal of the initial test version of ICD-10 DRGs is to replicate the ICD-9 DRGs insofar as possible, the two matching garbage codes must be assigned to the same DRG group, even though the ICD-10 garbage code encompasses a slightly larger set of patient records that includes people in the hospital for treatment of esophageal hemorrhage.

Lost in translation, then, is the fact that patients with esophageal hemorrhage from an unspecified cause were put in the same statistical pigeonhole—and payment systems designed around the distinctions made in ICD-9 naturally took account of it. Is this loss of information significant? That is for others to judge. But I can imagine why a code for esophageal hemorrhage all by itself was taken out of ICD-10, and why it wouldn’t matter in the long haul—“long haul” meaning the ability to determine fair and appropriate reimbursement in payment systems designed for ICD-10 specificity, or the ability to effectively track treatment and outcomes. After all, esophageal hemorrhage doesn’t just happen. It is caused by some other underlying disease; most often it is caused by bleeding from esophageal  varicose veins. Esophageal varicose veins don’t just happen either. They are most often caused by liver disease. So it can be argued that a unique code for unspecified esophageal hemorrhage is more like a second-order garbage code than real information.

Of course, this whole exercise in predicting impact is about nothing if not the short haul. In the short haul, everything that changes projected reimbursement matters because it tends to freak people out. So let’s get back to our ICD-10 DRG shift. Now that you understand where it is coming from and why, you can explain it. It is probably someone’s job (maybe yours) to decide the importance of this shift to the organization and what if anything can be done about it. Is this is an unavoidable shift, or is it more a matter of using the ICD-10 code set more effectively with complete documentation and correct coding? In this case I think it’s the latter.

For example, if the underlying cause of the esophageal hemorrhage is varicose veins in the esophagus, two ICD-10 codes unambiguously describe this condition (I85.01 Esophageal varices with bleeding, or I85.11 Secondary esophageal varices with bleeding.) Either code as the principal diagnosis is assigned to the ICD-10 MS-DRG group 368-370, same as ICD-9. If the cause of the bleeding is an ulcer in the esophagus, the correct code is K22.11 Ulcer of esophagus with bleeding and the MS-DRG group is 380-382 Complicated peptic ulcer (the same code exists in ICD-9). So the deal here is, the underlying cause of esophageal bleeding needs to be fully documented and correctly coded for ICD-10 specificity, and MS-DRG assignment will take care of itself.

Rhonda Butler is a Senior Clinical Research Analyst with 3M Health Information Systems.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s