The CMS HCC V28 Coding Mistakes That Cost Health Plans Millions (And How to Fix Them)

CMS HCC V28 has been fully implemented for over a year now. Most health plans think they’ve adapted. They’ve trained their coders. They’ve updated their systems. They’re submitting codes under the new model.

But internal audits and early RADV results are revealing systematic V28 coding errors that are costing millions in lost revenue or creating significant audit risk. Here are the mistakes that keep showing up.

Mistake #1: Still Coding Diabetes Without Complications

Under V24, diabetes without complications mapped to HCC 19. Under V28, there is no HCC for uncomplicated diabetes. You need to document and code diabetes with chronic complications to get any HCC value.

Sounds simple. But I’m still seeing health plans code diabetes (E11.9) without any complication codes and wonder why their risk scores dropped. That code is worthless under V28.

You need diabetes with chronic kidney disease (E11.22), diabetes with neurological complications (E11.4x), diabetes with circulatory complications (E11.5x), or diabetes with other complications. Just having diabetes isn’t enough anymore.

The coding teams that adapted fastest created systematic checks: any time a coder assigns a diabetes code, the system flags it if there’s no complication code. The coder either needs to find documentation supporting a complication or accept that this diabetes diagnosis won’t contribute to risk adjustment.

Mistake #2: Missing the Pressure Ulcer Stage Requirements

V28 completely changed how pressure ulcers are valued. Stage 3 and stage 4 pressure ulcers now map to much higher HCCs than they did under V24. But you need to code the stage correctly.

I’m seeing organizations code “pressure ulcer” without specifying the stage. That maps to an unstageable pressure ulcer, which has minimal HCC value. If the documentation clearly states “stage 3 pressure ulcer” but the coder only codes L89.XXX without the stage digit, you’ve lost significant revenue.

The documentation requirements got stricter too. Under V24, coders were more relaxed about pressure ulcer staging based on descriptions. Under V28, you need explicit stage documentation. “Deep tissue injury” or “wound to the bone” isn’t the same as “stage 4 pressure ulcer” for coding purposes, even if they’re clinically equivalent.

Mistake #3: Overlooking the Protein-Calorie Malnutrition Split

V28 split protein-calorie malnutrition into severe (E43, E44.0) and other (E44.1, E45, E46). Severe malnutrition has significantly higher HCC value.

Most providers document “malnutrition” without specifying severity. Under V24, that was fine. Under V28, you’re leaving money on the table if the patient actually has severe malnutrition but it’s not documented specifically.

This requires provider education, not just coder training. Coders can only code what’s documented. If the provider writes “patient is malnourished” without specifying severe versus moderate, the coder can’t assume severity even if the clinical picture suggests it.

The health plans succeeding with this have implemented systematic provider queries when malnutrition is mentioned. “Your documentation indicates malnutrition. Based on the patient’s BMI, albumin levels, and clinical presentation, does this meet criteria for severe protein-calorie malnutrition?”

Mistake #4: Not Capturing Chronic Kidney Disease Stage Properly

V28 changed the HCC values for different CKD stages. The financial spread between CKD stage 3 and CKD stage 4 is now larger. Accurate staging matters more.

I’m seeing two problems. First, coders are coding “CKD” without a stage when the documentation includes a GFR value that would clearly indicate the stage. If the note says “GFR 25,” that’s stage 4 CKD even if the provider didn’t write “stage 4” explicitly.

Second, coders are using outdated GFR values to determine staging. CKD staging should be based on current GFR, not a value from two years ago. If the most recent GFR is 48 but the coder is using an old GFR of 28 to code stage 4, that’s not defensible.

Clear coding guidelines help: use the most recent GFR value from within the past 90 days to determine CKD stage. If no recent GFR is available and the provider documented a specific stage, code that stage. If no recent GFR and no provider-documented stage, query the provider.

Mistake #5: Continuing to Code Morbid Obesity Alone

Under V24, morbid obesity (E66.01) had HCC value. Under V28, it doesn’t map to any HCC by itself. You need morbid obesity due to excess calories (E66.01) combined with either BMI documentation or obesity-related conditions to get HCC value indirectly through other pathways.

Health plans are still coding morbid obesity as a standalone diagnosis and expecting HCC credit. It doesn’t work that way anymore.

Mistake #6: Missing the Vascular Disease Documentation Requirements

V28 tightened documentation requirements for certain vascular conditions. Peripheral vascular disease needs to be documented with more specificity. Just “PVD” isn’t enough in many cases.

Atherosclerosis of the arteries needs to specify which arteries (extremities, native arteries, bypass grafts) and whether there are complications like claudication, rest pain, or ulceration. The code specificity required under V28 is higher than V24.

Coders who don’t query for this specificity are either overcoding (assuming complications that aren’t documented) or undercoding (missing HCC value because they can’t code specifically enough).

What Actually Fixes These Problems

Fixing CMS HCC V28 coding errors requires three things working together.

First, coder training needs to be specific to the changes that matter financially. Don’t just explain that V28 is different from V24. Show coders the exact coding scenarios where V28 requires different approaches: diabetes, pressure ulcers, malnutrition, CKD staging, vascular diseases.

Second, systematic quality checks need to catch these errors before submission. Build automated flags for diabetes without complications, pressure ulcers without stages, malnutrition without severity specifications, CKD without staging.

Third, provider education needs to address the documentation gaps that V28 exposed. Providers need to understand that “diabetes” isn’t specific enough anymore. “Malnutrition” needs severity. “Pressure ulcer” needs staging.

Most health plans did surface-level V28 training and thought they were done. The organizations that are actually succeeding with V28 treated it as a multi-year operational change, not a one-time training event. They’re still finding and fixing V28 coding errors 18 months into full implementation.

If your risk scores dropped more than expected when V28 implemented, or if your internal audits are finding systematic coding issues, you probably have one or more of these problems. Fix them now before they show up in RADV audits.