Why Absolute Value-Based Purchasing is better than Relative Value-Based Purchasing

The Center for Medicare & Medicaid Services (CMS) is gearing up to mandate some form of Relative Value-Based Purchasing (rVBP, see my previous post). rVBP, as currently proposed, states that the bottom one-fourth of hospitals will have their CMS risk-pool dollars entirely withheld. The withhold will likely be implemented yearly in an attempt to average out seasonal fluctuations, increase total sample size for the various metrics used to determine where hospitals will fall on the curve as well as to minimize the effects of the 30-90 day lag time in the reporting of many of these metrics. The magnitude of this withhold is significant and will likely push many already financially strapped hospitals into insolvency.

The problem with rVBP implemented in this way is that it imposes these financial penalties to the bottom one-fourth of hospitals every year. Remember, every year there will be a new bottom one-fourth. In a simple model where each hospital in the bottom one-fourth is forced into insolvency within one year of falling into the bottom one-fourth, it would only take about 20 years to go from the roughly 5000 community hospitals in the U.S. as of 2004 to less than 20 community hospitals in the U.S. What is even more frightening is that within 3 years more than half of all community hospitals in the U.S. would insolvent.

This is admittedly a simplistic and severe model. But even if we assumed the actual rate of rVBP caused insolvency was only 10% per year, half of all U.S. community hospitals would be insolvent within 6 years; and at 5% per year, half of all U.S. community hospitals would be insolvent within 14 years.

I think that it is clear that rVBP is a disaster.

Even if rVBP doesn’t make sense, VBP does. Hospitals should be pressured and rewarded to provide the highest value care for the most reasonable price. So what is the solution? Absolute Value-Based Purchasing (aVBP).

In aVBP, hospitals are not graded on a curve (where someone gets 100 and someone gets 0, regardless of actual performance), but are graded relative to an absolute score. Let’s take the example of a 100 question test where the highest scorer has 99 correct and the lowest scorer has 94 correct. In rVBP, the highest scorer would be graded as 100, A+, and the lowest scorer would be graded as 0, a most miserable F, despite having gotten 94% of the questions correct. In aVBP, both the highest and the lowest scorers would be given a solid A.

Obvious weaknesses with aVBP are how do you set a passing score and how do you set the sliding scale of rewards and punishments. I will spare you any detailed discussion of the statistical methods and organizational processes which can address these issues, the important thing to note is that these issues do not invalidate the concept of aVBP, they do not necessarily lead to a collapse of the U.S. hospital infrastructure as rVBP does. There are myriad ways to implement aVBP and whether one is any better than another is not as important as the fact that the system can be continuously tweaked and refined to reflect changing standards of care, macro- and micro-economic realities and public health infrastructure concerns.

Disclaimer: The opinions expressed herein are my own personal opinions and do not represent my employer's view in any way.