Medicare adjusts payments to Medicare Advantage (MA) organizations for cost variations in providing healthcare to beneficiaries based on various risk factors, including health status. Risk adjustment aims to properly reimburse plans for the risk of the beneficiaries enrolled, including those with chronic conditions. The Centers for Medicare and Medicaid Services (CMS) uses the diagnosis codes submitted by MA plans for their enrollees’ medical conditions to determine enrollee risk scores
Risk-adjustment data validation (RADV) occurs after the final risk adjustment data submission deadline for the MA contract payment year. RADV safeguards risk-adjusted payment integrity and accuracy. CMS reviews medical records from the hospital (inpatient and outpatient) and physician providers to validate enrollee CMS-HCCs that were assigned based on risk adjustment diagnoses submitted by MA organizations for payment. As a basic risk-adjustment rule, all risk-adjustment diagnosis codes submitted by MA organizations must be supported by medical record documentation.
RADV Objectives
RADV has several critical objectives, including:
- Verify enrollee CMS-HCCs used for payment
- Identify risk adjustment discrepancies
- Calculate enrollee-level payment error
- Estimate national and contract-level payment errors
- Implement contract-level payment adjustments
How is CMS RADV performed?
CMS determines which plans will be audited and which patients are included in the audit. The health plans must obtain all relevant medical records for the patients selected. An initial review of these medical records is performed to determine which contain documentation required to substantiate the HCCs that have been captured in the CMS system for these patients.
After the initial review, additional analysis is performed to determine the medical records that should be submitted to CMS to support the HCC values sampled. When CMS completes its secondary review and analysis, an error rate is determined, which is extrapolated to the premiums for the entire patient population for that health plan.
Extrapolation means that if an error is found during a RADV audit on a particular HCC, not only are the overpayments recouped on that member, but payment is recouped on all plan members who are in that HCC category.
Stages of CMS RADV
Per the CMS RADV protocols, there are six stages in the RADV process:
- Stage 1: Sampling and medical record request
- Stage 2: Medical record review (MRR)
- Stage 3: MRR findings and contract-level payment adjustments
- Stage 4: Documentation dispute
- Stage 5: Post documentation dispute payment adjustments
- Stage 6: Appeals
In CMS RADV, CMS selects a random stratified sample of patients to audit. The sample is:
- 1/3 of patients with a high RAF
- 1/3 of patients with a medium RAF
- 1/3 of patients with a low RAF
Only Part C HCCs are audited in RADV. The health plans must submit up to five of their best records from an acceptable treating provider, demonstrating those diagnoses as current in the year being audited that support the HCC values that were paid.
Types of CMS RADV
CMS conducts two types of RADV audits
- Annual national-level audits to estimate the national MA improper payment rate
- Contract-level RADV audits to identify and recover improper payments from MA organizations
CMS uses two main RADV types—Random and Targeted RADV
Random CMS RADV uses a selection process in which an MA plan is randomly selected for an audit
Targeted CMS RADV is applied to MA plans that have raised concerns, such as a significant increase in risk scores
What are the documentation requirements for CMS data validation?
With CMS proposing to expand the scope and scale of RADV audits to reduce payment errors, providers must ensure their documentation complies with the necessary standards. Adhering to the following documentation guidelines will promote compliance with CMS data validation requirements:
- The medical record documentation should be legible.
- The record should be from the correct calendar year for the payment year being audited.
- CMS accepts only medical record documentation from a face-to-face encounter.
- The date of service should be present for the face-to-face visit.
- The patient’s name and date should appear on every medical record page.
- The physician’s signature and credentials should be included in each patient encounter, and the responsible provider should authenticate the electronic signature.
- Correct ICD-10-CM codes support medical record documentation.
- Accurate coding of all documented conditions that coexist with chronic conditions at the visit and require or impact patient treatment. Use historical codes if the historical condition affects current care or influences treatment.
- State specificity of condition – specify if the condition is chronic, major, recurrent, and type, e.g., chronic renal insufficiency, major depression, and chronic hepatitis.
How is HHS-RADV performed?
To ensure the integrity of the risk adjustment program, CMS, on behalf of HHS, performs risk-adjustment data validation, also known as HHS-RADV, to validate the accuracy of data submitted by issuers for risk-adjustment transfer calculations. HHS-RADV aims to promote confidence and stability in the budget-neutral transfer methodology used by the HHS-operated risk adjustment program by ensuring the integrity and quality of data provided by issuers.
Steps of HHS-RADV
Per the CMS RADV protocols, there are six steps to the RADV process:
- CMS creates a sample of a health plan’s enrollee records for audit.
- The health plan must select an initial validation auditor to audit demographic and enrollment data, prescription drug categories (RXCs), and health status data submitted on the health plan’s External Data Gathering Environment (EDGE) server for the selected sample enrollees.
- A second validation auditor performs a quality assurance audit on a subsample of the initial validation auditor’s data to verify the accuracy of the findings.
- CMS performs error estimations and calculates the health plan’s risk score error rates using the failure rate of each HCC.
- CMS administers the second validation audit findings attestation and discrepancy reporting process, the error rate attestation and discrepancy reporting process, and an administrative appeals process.
- Final results are used to adjust the risk adjustment risk scores and transfers.
HHS applies a sampling methodology to choose a statistically valid sample of enrollees based on each issuer’s enrollee-level risk score distribution. HHS designed the sampling methodology to ensure that the sample covers critical subpopulations of enrollees for each issuer by dividing each issuer’s population into 10 strata, representing different age and risk score bands, and sampling from each stratum.
Difference between CMS and HHS RADV
| CMS RADV | HHS RADV |
|---|---|
| An audit is performed two to three years after payment | An audit is performed six months after year-end |
| Choose health plans for an audit by random sampling or targeting efforts | Is an annual requirement for all health plans |
| Typically involves approximately 30 health plans | Involves all participating health plans |
| Uses a stratified sample of three strata | Uses a stratified sample of ten strata |
| Allows any face-to-face encounter for audit support | Allows only those DOS that were submitted on the EDGE server with an exception to allow encounters that would normally be accepted on the EDGE server |
| Allows for up to five best records to support an HCC | Allows for as many submissions as the health plan would like to submit, provided they were all processed on the EDGE server |
| Only Part C HCCs are audited | HHS HCCs and no HCC patients are audited |
Author
