A few weeks ago, the Office of Medicare Hearings and Appeals (OMHA) announced what it described as a way to clear up the current federal backlog of appeals and move the process forward more efficiently.
Of particular interest is the Statistical Sampling Pilot Program. Also of some interest is that the Centers for Medicare & Medicaid Services (CMS) has been conducting statistical audits for the past umpteen (I love that number) years now, and statistical sampling and extrapolation have become a way of life for many healthcare providers.
The basic idea is this: If you have more than 250 claims subject to an appeal, you can be the lucky contestant who gets to play the CMS statistical sampling game. If you meet all of the criteria and choose to participate, OMHA will assign an independent statistical expert to assist the Administrative Law Judge (ALJ) in “carrying out the statistical sampling in accordance with Medicare guidance.” Sounds very much like the 100 or so audits I have worked on over the last couple of years. The auditor uses its statistical expert to pull a statistically valid random sample of claims (or beneficiaries or lines or whatever) and then extrapolates the results to the universe of claims from which the sample was drawn.
This sounds great, except that, in those cases in which an appeal has been filed, 80 percent of the overpayment findings were reversed in favor of the provider. In my experience, the majority of the extrapolations also ultimately have been vacated because the “statistical expert” utilized by the auditor failed to actually produce a statistically valid random sample.
See, the problem is that Chapter 8 of the Program Integrity Manual (PIM), while touching on the concepts of statistical sampling and extrapolation, allows for a huge amount of latitude when it comes to the actual process of conducting the analysis. Recently, I received a letter from a Medicare Administrative Contractor (MAC) discussing my objections to an extrapolation that was imposed on one of my clients. It was a vitriolic and personal attack that basically said that the auditor could do whatever it wanted, even if it wasn’t the best statistical practice, as long as it abided by Chapter 8 of the PIM. That’s crazy! If there are no consequences for using poor or erroneous techniques, then what is the sense of appealing on the basis of the extrapolation?
I point out that, based on the methodology, the sample was not drawn from a universe with independent units. I was told that this was my opinion, and that this issue wasn’t covered in the PIM. My opinion? I don’t think so. The facts speak for themselves. There were other issues, also, for which the “statistical expert” reviewing the case repeated that my objections were based on my own opinions (apparently the auditor had opinions, too). And the deference was given to the auditor. In the end, the letter stated that the auditor had done “everything” right. No one ever does everything right, and it was very clear to me that the expert reviewing the expert either had some skin in the game or was biased in some other way – not because they didn’t agree with me, but because of the manner in which the objections were handled.
So now, OMHA, another branch of CMS, is saying that we can streamline the ALJ process by allowing one of their experts to draw a statistical sample from what already may have been touted as a statistical sample and use the results to extrapolate back to the universe. And even if extrapolation was not originally involved, it will be, on this smaller scale. Now, I have an opinion and it is that this is going to add complexity to the equation, not simplify it. I hope I am wrong, but based on my experience in challenging these statistical audits on a daily basis, I have little to no confidence that CMS, OMHA, or any other entity assigned to this process is going to do any better a job that what is done now. And as a caveat, I am not attacking the quality, honesty, or integrity of the CMS statisticians. I have come up against many who were exemplary in their work and their work product. They, like everyone else in this arena, are often overwhelmed and simply don’t have the necessary time to validate their samples for representativeness (or opine outside of the PIM guidelines). And the rules even allow for the provider to bring in its own statistical expert – and here we are, exactly back to where we started.
This is like treating the symptoms of a disease rather than treating the underlying disease itself. ALJs are overwhelmed with appeals because the appeals have been so darn successful for the providers. And the only reason I can think of regarding why they are so successful, being as a disinterested arbitrator is making the decision, is that the audit findings are poorly constructed, with the majority based in error. Hey, this isn’t me talking; it’s the numbers. Somewhere in the neighborhood of 80 percent of appeals are overturned in favor of the provider, and that is just way too high a figure to be a statistical anomaly or a coincidence. So either there is a grand conspiracy between the judges and CMS, or the model is just so broken that it needs to be reengineered – not fixed with Band-Aids.
What we need to do is to diagnose the problem and find a cure. We need to treat the underlying cause of the two-year delay, not keep trying to fix the symptoms. I just don’t get how sending in the fox to guard the henhouse is going to do anyone any good.
But hey, I’m open-minded and I hope that I am wrong.
About the Author
Frank Cohen is the director of analytics and business intelligence for DoctorsManagement. He is a healthcare consultant who specializes in data mining, applied statistics, practice analytics, decision support, and process improvement. Mr. Cohen is also a member of the National Society of Certified Healthcare Business Consultants (NSCHBC.ORG).
Contact the Author
Comment on this article