Information on geographic variation of utilization and quality updated

As part of its efforts to support data-driven decision making efforts, the Office of Enterprise Data and Analytics (OEDA) within CMS updated the Geographic Variation Public Use File (GV PUF) in January 2016. The GV PUF is based primarily on information from CMS’ Chronic Conditions Data Warehouse (CCW), which contains 100 percent of Medicare claims for beneficiaries who are enrolled in the fee-for-service (FFS) program as well as enrollment and eligibility data. The January 2016 update to the GV PUF includes data for 2007-2014 and incorporates minor revisions to the CMS methodology.


The GV PUF is made up of 12 separate files with data designed to enable researchers and policymakers to evaluate geographic variation in the utilization and quality of health care services for the Medicare FFS population. The individual files are available from the CMS Web site and contain data on demographic, spending, utilization, and quality indicators at the state level (including the District of Columbia, Puerto Rico, and the Virgin Islands), the hospital referral region (HRR) level, and the county level. The GV PUF was originally posted in July 2011 and updated in July 2012, January 2013, May 2013, December 2013, February 2015, and September 2015.

The 12 files

The GV PUF has two files with state and county-level data, four files with only state-level data, and six files with HRR-level data. The files are presented in “Table” and “Report” formats. The “Table” files present indicators for all states, counties, or HRRs, and can be exported from Excel to another data analysis program for additional analysis. The corresponding “Report” files allow users to compare a specific state, county, or HRR to national Medicare benchmarks. The state- and HRR-level data are presented for beneficiaries under the age of 65, beneficiaries that are 65 or older, and all beneficiaries who regardless of age. The county-level data are only available for all beneficiaries.

Revisions to CMS methodology

The January 2016 update eliminated disease prevalence measures because CMS separately publishes this information in the Chronic Conditions Public Use File. OEDA produces this information to provide researchers and policymakers a better understanding of the burden of chronic conditions among beneficiaries and the implications for our health care system.

The update also made minor changes to the suppression methodology to ensure that all products from OEDA are consistent. Information is now suppressed for areas with fewer than 11 beneficiaries. Counter or secondary suppression is applied in cases where only one geographic area is suppressed for primary reasons, e.g. one county in a state has between 1 and 11 beneficiaries. In these cases, the geographic area with the next smallest count of Medicare FFS beneficiaries is suppressed as well. Also, if one sub-group (e.g. age group) is suppressed, then the other sub-group is suppressed.