Rural/Urban and Regional Variation in the 2019 CMS Hospital Wage Index


The Medicare hospital Inpatient Prospective Payment System is designed to pay hospitals for services provided to Medicare beneficiaries based on a national standardized amount adjusted for the patient's condition and related treatment. Further, Social Security Act Section 1886(d)(3)(E) requires that the standardized amount be adjusted for differences in hospital wage levels among labor markets, which the Centers for Medicare & Medicaid Services (CMS) implemented through the wage index system.

In November 2018, the Office of Inspector General (OIG) issued a study entitled "Significant Vulnerabilities Exist in the Hospital Wage Index System for Medicare Payments." The OIG identified four significant vulnerabilities in the wage index system: (1) CMS lacks the authority to penalize hospitals that submit inaccurate or incomplete wage data and noted instances where some hospitals had submitted inaccurate wage data and inadvertently been overpaid by Medicare; (2) Medicare administrative contractor limited reviews do not always identify inaccurate wage data; (3) the rural floor decreases wage index accuracy; and (4) hold-harmless provisions in Federal law and CMS policy pertaining to geographically reclassified hospitals' wage data decrease wage index accuracy. The OIG concluded that these vulnerabilities mean that wage indexes may not always accurately reflect local labor prices and, therefore, Medicare payments to hospitals and other providers may not be appropriately adjusted to reflect local labor prices. Several U.S. legislators have also criticized hospital wage index for adding to rural disparities.

This brief describes the geographic variation of the 2019 CMS hospital wage index by rural/urban definition, census region, Frontier and Remote Area (FAR) codes and state. Results show rural hospitals' medial wage index is lower than the median urban hospital index; rural hospital wage index medians are lower than urban hospital wage index medians in every Census region – lowest in the South Census region; and lower for hospitals in more remote and less populated areas.

Rapid Response to Requests for Rural Data Analysis
Kristie Thompson, H. Ann Howard, George Pink