NPRM 4185-P RADV Provision Data Release File Summaries

CMS’ Center for Program Integrity is releasing data supporting the Notice of Proposed Rulemaking (NPRM) Medicare and Medicaid Programs; Policy and Technical Changes to the Medicare Advantage, Medicare Prescription Drug Benefit, Program of All-Inclusive Care for the Elderly (PACE), Medicaid Fee-For-Service, and Medicaid Managed Care Programs for Years 2020 and 2021  (CMS-4185-P, 83 Fed. Reg. 54982, at 55037-55041 and 55077), published on November 1, 2018.  The NPRM proposed policies regarding the use of extrapolation in Medicare Advantage Risk Adjustment Data Validation (RADV) audits and the Fee-for-Service Adjuster (FFSA).  The comment period for the RADV provisions in the NPRM was extended until April 30, 2019, via a Federal Register notice published on December 27, 2018 (CMS-4185-N, 83 Fed. Reg. 66661).  Data supporting the NPRM 4185-P RADV provisions will be available to requesters who enter into a data use agreement with CMS. The comment period was extended a second time via a Federal Register notice published on April 30, 2019 (CMS-4185-N, 84 Fed. Reg. 18215) until August 28, 2019. Additional data files are now available to supplement the initial release. 

File Descriptions

FFS04_05_PROFILESPAYMENTS

This input file originated from a dataset that Research Triangle Institute (RTI) supplied (filename:y5r1s5f.sas7bdat).  It represents the calibration data that RTI used for the Centers for Medicare and Medicaid Services Hierarchical Condition Category (CMS-HCC) model version that CMS used to calculate 2009 Medicare Advantage (MA) payments. The CMS-HCC model was calibrated on a 5% sample of Fee-For-Service (FFS) beneficiaries. The file represents 1,441,247 beneficiaries and contains CMS-HCC model output variables and FFS expenditures for each beneficiary.  The file also indicates the profile for each observation, defined as the concatenated series of 0/1 indicators corresponding to the CMS-HCC model demographic and HCC model output variables.

FINAL_CID_HCC_DISPOSITION

This file contains medical record review findings from a Risk Adjustment Data Validation-like review that CMS undertook on a sample of 2008 medical records. The medical records were associated with a sample of Fee-For-Service (FFS) claims that CMS selected for Comprehensive Error Rate Testing (CERT) review. Based on medical record review by certified coders, the data file contains the disposition for each medical record reviewed, defined as whether the CMS-HCC associated with the FFS claim diagnosis was confirmed or discrepant. The file also flags other (additional) CMS-HCCs identified by medical record review coders.

FFS

The Fee-For-Service (FFS) data contains 10 datasets that represent the entire 5% sample of all final 2004-2005 diagnosis codes used for Medicare Advantage (MA) model calibrations through 2011. This data is not summarized and contains all final diagnosis codes regardless of whether the diagnosis code is included in the risk adjustment model. Risk adjustment eligible diagnosis codes are indicated with an ‘R’ in the RAS_DGNS_IND field. The source of this data is the National Medicare Utilization Database (NMUD). In this analysis, this data was used in tandem with the HCC data to map all diagnosis codes to HCCs and then count unique combinations of claims (CLAIM_UID field) and HCC codes.

HCC

The HCC file contains the mapping from International Classification of Disease, 9^th Revision (ICD-9) diagnosis code to Version 12 of the CMS-Hierarchical Condition Category (CMS-HCC) model. Diagnosis codes have been modified to remove decimals. HCC codes are numeric variables where the number represents the HCC code. The HCC description can be found in the attached documentation or on the CMS website. This file is used to map the Fee-For-Service (FFS) diagnosis codes in the FFS files to HCCs.

SAMPTB_Y13_ELIG1M

This file consolidates Medicare Advantage (MA) data for beneficiaries who meet eligibility criteria for Contract-Level Risk Adjustment Data Validation (RADV) audits from three sources: the adjusted Monthly Membership Report (MMR), the Model Output File (MOF), and the CMS Enrollment Database (EDB). A beneficiary is RADV eligible when they satisfy all of the following criteria:

1. Enrolled in an MA contract (H-number, E-number, or R-number) in January of the payment year based on CMS' monthly member enrollment files;
2. Continuously enrolled in the same MA contract (as identified in step (1) above) from January of the data collection year through January of the payment year;
3. Non-End Stage Renal Disease (non-ESRD) status from January of the data collection year through January of the payment year;
4. Non-hospice status from January of the data collection year through January of the payment year;
5. Enrolled in Medicare Part B coverage for all 12 months during the data collection year (i.e., defined as full risk beneficiaries for risk adjusted payment); and
6. Had at least one risk adjustment diagnosis (ICD-9-CM code) submitted during the data collection year that led to at least one CMS-Hierarchical Condition Category (HCC) assignment for the payment year. 

Each row represents one randomly-sampled individual beneficiary. Status flags from the MMR and EDB represent beneficiary changes in the data collection year (2010) and the payment year (2011). Payment fields from the MMR represent payments for January 2011. Risk adjustment fields from the MOF represent the final risk score for payment year 2011. All beneficiaries in this file had 24 continuous months of Medicare Advantage enrollment in a single MA contract.

SAMPTB_Y13_FULL1M

This file consolidates Medicare Advantage (MA) data for beneficiaries who did not meet all eligibility criteria for the Contract-Level Risk Adjustment Data Validation (RADV) audits from three sources: Adjusted Monthly Membership Report (MMR), Model Output File (MOF), and CMS Enrollment Database (EDB). A beneficiary is RADV eligible when they satisfy all of the following criteria:

1. Enrolled in an MA contract (H-number, E-number, or R-number) in January of the payment year based on CMS' monthly member enrollment files;
2. Continuously enrolled in the same MA contract (as identified in step (1) above) from January of the data collection year through January of the payment year;
3. Non-End Stage Renal Disease (non-ESRD) status from January of the data collection year through January of the payment year;
4. Non-hospice status from January of the data collection year through January of the payment year;
5. Enrolled in Medicare Part B coverage for all 12 months during the data collection year (i.e., defined as full risk beneficiaries for risk adjusted payment); and
6. Had at least one risk adjustment diagnosis (ICD-9-CM code) submitted during the data collection year that led to at least one CMS-Hierarchical Condition Category (HCC) assignment for the payment year.

Criteria 1, 3, 4, and 5 apply for the non-RADV eligible population, while criterion 2 does not.

Each row represents one randomly sampled individual beneficiary. Status flags from the MMR and EDB represent beneficiary changes in the data collection year (2010) and the payment year (2011). Payment fields from the MMR represent payments for January 2011. Risk adjustment fields from the MOF represent the final risk score for payment year 2011.

TABLE_10A_ORIG_0_ELIG1M

This file lists the risk scores based on the normalized coefficients for each ELIGIBLE MA enrollee in the SAMPTB_Y13_ELIG1M. This risk score was calculated by summing the relative factors that were calculated from the normalized coefficients. The enrollees can be linked back to SAMPTB_Y13_ELIG1M by the identifier field LNKEYNUM.

TABLE_10A_ORIG_0_FULL1M

This file lists the risk scores based on the normalized coefficients for each FULL MA enrollee in the SAMPTB_Y13_FULL1M. This risk score was calculated by summing the relative factors that were calculated from the normalized coefficients. The enrollees can be linked back to SAMPTB_Y13_FULL1M by the identifier field LNKEYNUM.

For file cost and availability, please see the LDS Worksheet (ZIP).

Please follow the instructions on the DUA - Limited Data Sets page to request the file.