A hospital in New York City faces higher labor costs than a hospital in Billings, Montana. To take into account these cost differences, Medicare adjusts hospital payments to reflect these cost differences using a hospital wage index. As currently constructed, however, many hospitals petition to be included in labor markets where they would receive a more generous wage index value.
A number of reforms to the CMS hospital wage index have been proposed. In a recent Acumen report to which I contributed, we evaluate whether some reforms proposed by MedPAC would improve the wage index’s accuracy. Below is an excerpt from the executive summary. The full report is available here.
“The Medicare statute requires that per-discharge payments to hospitals in the inpatient prospective payment system (IPPS) reflect geographic differences in the cost of labor. As a result, Medicare’s IPPS payments are adjusted by a hospital wage index that seeks to reflect the average price of labor facing each hospital. To construct the index, Medicare clusters hospitals into metropolitan statistical areas (MSAs) and residual areas (“balance-of-state” or “rest of state”). These geographical areas approximate hospital labor markets, and average wages are calculated for each using wage data from an annual survey of IPPS hospitals’ labor costs. However, accurately representing a hospital labor market is not a simple task, and inaccurately specifying a hospital labor market can create two problems.
The first problem occurs when hospitals in the same MSA (or county) but located a significant distance from each other receive the same wage index value, even though they face different labor costs. In this situation, the hospital that faces higher labor costs is at a disadvantage. The second problem, called a “cliff” or “boundary” problem, occurs when neighboring hospitals face the same labor prices but receive significantly different wage index values because they are located in different MSAs. For example, these hospitals could be on opposite sides of the same street yet in different MSAs. In this case, the hospital with the lower wage index value is at a disadvantage relative to its neighboring hospital.
These two situations lead to a problem of incentives: in each case the disadvantaged hospital will have an incentive to seek reclassification or an exception that increases the hospital’s wage index value. These reclassifications and exceptions aim to compensate for the wage index inaccuracies resulting from using MSAs and balance-of-state areas as representative hospital labor markets. Although reclassification and exceptions may in many cases improve the match between the wage index value and the prevailing the existing patchwork of reclassifications and adjustments on the wage index has created a very complicated and convoluted system. In the Tax Relief and Health Care Act of 2006 (TRHCA), Congress required the Medicare Payment Advisory Commission (MedPAC) to develop recommendations for revising the wage index and required the Secretary of Health and Human Services to respond to these recommendations.
In June 2007, MedPAC recommended repeal of the existing wage index statute, including the elimination of reclassification and exceptions, and proposed an alternative index.1 MedPAC’s proposed hospital compensation index changed both the data used to construct the hospital wage index and the method of its construction. In an earlier report, “Revision of Medicare Wage Index: Final Report, Part I” (April 2009), Acumen evaluated the Bureau of Labor Statistics (BLS) occupational wage survey data proposed by MedPAC. In this report, “Revision of Medicare Wage Index: Final Report, Part II”, we analyze MedPAC’s proposed method of improving upon the definition of the wage areas used in the current Medicare wage index. This method first averages or “blends” MSA and county-level wages and then implements a “smoothing” step which eliminates large differences in index values among neighboring hospitals. As proposed by MedPAC, smoothing would limit differences in wage index values between adjacent counties to no more than 10%.
Since it is possible to separately analyze MedPAC’s method of defining wage areas and the wage data, Acumen applied the blending and smoothing methodology to both a wage index that uses current Medicare wage data and another that uses BLS wage data. We isolated the effects of the blending and smoothing method on the two sets of underlying wage data. This approach gives a detailed assessment of the advantages and disadvantages of MedPAC’s blending and smoothing method.”