In recent weeks, the Centers for Medicare & Medicaid Services (CMS) has released the Calendar Year (CY) 2024 Advance Notice for the Medicare Advantage (MA) and Part D Prescription Drug Programs, as well as the Medicare Advantage Risk Adjustment Data Validation (RADV) final rule. Both documents contained significant changes that CMS expects will result in $11 billion in savings to the Medicare Trust Fund in 2024 as well as a 3.15% decrease in risk scores.

Key Takeaways: What Do These Changes Mean for…

• Payers? Changes to the RADV audit—particularly the removal of the Fee-For-Service (FFS) adjuster—mean that payer organizations will need to do everything in their power to validate codes in provider claims. This means chart audits and claims data validation.
• Patients? Many of the changes proposed or implemented by CMS are designed to encourage accuracy and specificity in patient diagnosis. To achieve that specificity, providers will need to focus on patient engagement. This should result in more comprehensive visits with providers and overall better outcomes for patients.
• Clinicians? With these proposed changes, HCCs would be more specific and tightly defined. This will help providers more accurately depict the health status of the patient and their cost of care. In addition, documenting clinical severity will be increasingly important.
• Medical coders? Coders will need to become fluent in the proposed new HCC coding structure. Specificity and documentation of clinical severity will be more imperative than ever before.
• Risk adjustment analysts? From an analytics perspective, the proposed updates to the Part C HCC model will have a significant impact on suspecting and the associated logic.

Changes to RADV

CMS uses RADV audits to validate the payments made to MA organizations for the member diagnoses submitted. In the RADV final rule, released on Jan 30, a few days before the Advance Notice, CMS confirmed its intent to extrapolate the results of these audits for the audited plan’s entire population.

Moving forward, CMS has also decided to eliminate the FFS adjuster that has been used to offset the error rate in the data. With this change, organizations will need to provide more documentation to CMS.

Understand the underlying principle of CMS’ decisions. CMS is driving to ensure all diagnoses are accurately documented.

Make sure you are doing your part to ensure accuracy. If you are a healthcare provider, take the time to understand the nuances of coding (more on that later) and code correctly. If you are part of a payer organization, consider implementing a proactive quality review as a best practice.

Finally, the MA RADV final rule also included the announcement that CMS will begin extrapolating data with the RADV audits for the 2018 payment year. Originally, CMS indicated that it intended to extrapolate data going all the way back to PY2011.

Changes to the HCC Model

With the Calendar Year 2024 Advance Notice, the big news concerned the proposed new Part C Hierarchical Condition Category (HCC) Model. The HCC Model was last updated two years ago, using FFS claims from PY2014 and PY2015. The revised model will use diagnosis codes from PY2018 and expenditures from PY2019. In addition, the denominator year used to determine risk adjustment factors has been updated to 2020 and the new denominator is $10,402.34.

CMS worked with a panel of outside clinicians on a clinical revision that was used to rebuild the condition categories. The idea here is to reflect diagnosis coding under the ICD-10-CM diagnosis classification system and ensure that diagnosis codes map to condition categories with similar clinical characteristics and cost. For example, major depression now has its own category instead of being lumped in with bipolar depression.

Also, discretionary diagnostic categories are being excluded from payment models. Diagnoses that are particularly subject to discretionary coding variation or inappropriate coding, as well as codes that are not clinically or empirically credible as cost predictors should not increase cost predictions. By excluding these diagnoses, the model will be less sensitive to coding variation and coding proliferation. This should help the HCCs to serve as a more accurate predictor of cost.

Similarly, CMS removed several condition categories that do not accurately predict projected cost. Eliminated categories include protein-calorie malnutrition (47), angina pectoris (230), and atherosclerosis of arteries of the extremities, with intermittent claudication (265).

The new model also constrained some HCCs so that they carry the same weight in the risk score. All diabetes HCCs (36, 37, and 38) and congestive heart failure HCCs (224, 225, and 226) have been constrained. The implications here could be significant. For example, diabetes with complications had a risk adjustment factor (RAF) of 3.07 in the old model, but will have a RAF of 1.66 in the new model.

Despite these removals and consolidations, the number of HCC categories actually increased from 86 to 115. The increase is a result of some new HCC codes being added and existing HCCs being split. While the number of HCC categories shot up, the number of risk-adjustable codes decreased from 9,797 codes to 7,770 codes.

While all of this shuffling of codes may seem confusing, keep in mind that the new HCC Model was designed to create HCC codes that directly align with the diagnostic classification system that is currently being used in the industry. This is the first time CMS has used the ICD-10 classification system to create HCC categories. Over the long term, this shift should create consistency, enable interoperability, and allow for cross-comparisons across quality and value-based care programs.

Conclusions

The overarching themes for all of CMS’ recent changes seem to be specificity and documentation. Providers, payers, and outside vendors are all being asked to pinpoint diagnoses as accurately as possible, substantiate those diagnoses, and submit correct codes to CMS. Recognize your role in this process and seek out the education and tools you will need to supply CMS with complete, accurate data.