Guest Post: What did the OIG work on and what did they find?

by Shannon McCall, RHIA, CCS, CCS-P, CPC, CEMC, CRC, CCDS

First, it’s important to understand how the OIG decides on issues to focus its efforts. According to the OIG, these are some of the factors that are taken into consideration:

• Management's actions to implement OIG recommendations from previous reviews
• Mandatory requirements for OIG reviews, as set forth in laws, regulations, or other directives
• Potential for positive impact
• Requests made or concerns raised by Congress, HHS management, or the Office of Management and Budget
• Top management and performance challenges facing HHS
• Work performed by other oversight organizations

Prior to exploring the findings, please note your internal audit efforts should be tailored to your facility, and not solely focused on the OIG’s areas of interest. Foresight is much better than hindsight, so knowing which particular areas are under scrutiny can be a huge benefit! The monthly updates are published on the OIG website.

Audit findings

One audit finding that I found particularly interesting was in regard to the ICD-9-CM diagnosis code for kwashiorkor. According to the December 2017 report:

• CMS targeted the ICD-9-CM diagnosis code for kwashiorkor because in 2006-2014 Medicare paid approximately $2.5 billion on claims that included this diagnosis. Given this condition’s rarity in U.S. patients, it was an obvious anomaly that warranted a closer examination. Malnutrition is commonly assigned as an additional or secondary diagnosis code and is classified as either a CC or an MCC, depending on severity levels, therefore it has the ability to affect DRG assignment.
• 2,145 inpatient claims were retrospectively reviewed for the ICD-9-CM diagnosis code 260 (kwashiorkor). The findings identified that in all but one claim, the diagnosis was not supported by the medical record documentation. The overpayments on these claims totaled $6 million.
• However, it is important to note that this coding “error” was due to a discrepancy between the Alphabetic Index and the Tabular Index in the ICD-9-CM manual that was never corrected even though CMS was acutely aware of the issue. Had CMS diligently pursued the correction, it could have saved approximately $102 million in calendar years 2006-2014, which is a rather costly result of procrastination.

Coding guidance

Coders rely on the validity of the organization and conventions included within the ICD code set. The Alphabetic Index had multiple entries that led to the diagnosis code 260. The most likely contributor was the term “Malnutrition, (calorie) protein.” Per the Official Guidelines for Coding and Reporting, words contained in parentheses in the Alphabetic Index are called “non-essential modifiers,” meaning the absence or presence of the term does not change code assignment. Therefore, if the provider documented protein calorie malnutrition or protein malnutrition, it was assigned to 260.

Not to say that there isn’t shared responsibility, but in defense of the coders who assigned this diagnosis code in error, we must consider their consistent use of encoders which apply decision tree-based logic. The decision tree logic leads to the ultimate code assignment that is based on the ICD-9-CM (and now ICD-10-CM) conventions and organization in the Alphabetic Index and Tabular List for both diagnoses and procedures.

Some coders may not feel comfortable changing the code that the Alphabetic Index and ICD conventions are stating as the corresponding code for that particular diagnosis, regardless of title. It is of note that the Tabular List has always only included the description for code 260 as Kwashiorkor. When reviewing the final list of codes assigned, the corresponding code title would state the word Kwashiorkor, not the terms protein or protein (calorie) malnutrition as likely was in the medical record documentation. The reality is that not all diagnosis code titles in the nomenclature match verbatim what is documented in the medical record, and yet it is still considered the correct code.

CMS was made aware of the discrepancy in 2007, however did not adequately address the issue to pursue the revision of the Alphabetic Index and/or the Tabular List for accuracy. CMS is one of the four Cooperating Parties that oversee the ICD coding classification system, having the ability to influence revisions made in the system. However, it seems there was discussion and deferral that it is the responsibility of AHA’s Coding Clinic to answer coding related questions. Finally, in Coding Clinic, Third Quarter 2009 (two years later), the AHA addressed that kwashiorkor (260) should only be coded if the provider specifically documents the diagnosis. Following the release of this Coding Clinic, there was a significant drop in the use of ICD-9-CM diagnosis code 260.

Aftermath

The 25 providers that accounted for the $6 million in overpayments were expected to refund the overpayment amounts. To date, approximately $5.8 million has been recovered. Coding should always be accurate and clinically supported within the medical record documentation; however, it is unfortunate when providers are penalized for a flaw in the system beyond their control. This flaw was identified and not remedied by those who had the power and ability.

Reliance on coding source authorities like the Official Guidelines for Coding and Reporting (including conventions) and supplemental resources such as Coding Clinic are vital to the coding profession and accurate code assignment. Recently, there has been a noticeable shift in opinion that if the Alphabetic Index doesn’t lead the coder to what is felt to be the correct code, they should peruse the section to determine if they feel that there is a more suitable code.

For example, per the Alphabetic Index, the term emaciation (due to malnutrition) is assigned to ICD-10-CM code E41 (nutritional marasmus), which is an MCC condition. Please notice yet again, the words “due to malnutrition” are in parentheses so they are considered non-essential modifiers per the ICD-10-CM convention. Yet Coding Clinic, Third Quarter 2017, states: “If the provider uses the term ‘emaciated’ in the absence of malnutrition assign R64 (cachexia). Emaciated is a descriptive term to describe someone as thin due to wasting. R64 is a CC condition.” The explanation concludes by stating that a basic rule of coding is that “further research is done if the title of the code suggested by the Index does not identify the condition correctly.”

This is interesting, yet dangerous advice for the coder, as this directive certainly has the potential to open up doors of scrutiny in an audit situation. Imagine the debates that will inevitably occur with an auditor as the coder tries to explain that the nomenclature was not accurate in their opinion, so they opted to use another code. This is even more dangerous when the coder’s choice may result in a CC/MCC condition.

It would seem the most logical solution versus the advice given would be to revise the Alphabetic Index to either:

1. Not have a default code for the term Emaciation but rather note See Malnutrition, providing guidance with regard to classifying the condition as either nutritional marasmus or cachexia
2. Revise the main term to Emaciation due to malnutrition, removing the parentheses

Currently, coders have a solid understanding of the ICD conventions, and know that parentheses indicate non-essential modifiers as stated in the definitions written in part by CMS. It would seem simpler to revise the entry in the Alphabetic Index, than to expect coders to know this represents an exception. Coding is always about following rules and guidelines, but remember for every rule there is opportunity for exception.

Editor’s note: This article originally appeared in its entirety in JustCoding. McCall is the director of HIM and coding for HCPro in Middleton, Massachusetts, and oversees all of the Certified Coder Boot Camp programs. Contact her at smccall@hcpro.com. For more information, see www.hcprobootcamps.com. Opinions expressed are that of the author and do not necessarily represent HCPro, ACDIS, or any of its subsidiaries. To read about the Office of Inspector General’s (OIG) new cadence for Work Plan releases and how to use them for your CDI efforts, click here.