This article is more than 5 years old. Information may no longer be current.
New GAO report claims self-referral increases costs: Is it right?
On Sept. 28, 2012, the Government Accountability Office issued a report entitled “Higher Use of Advanced Imaging Services by Providers Who Self-Refer Costing Medicare Millions.”
The Government Accountability Office (GAO) tried to compare the number of scans ordered by physicians who were ordering imaging supplied by the same entity that employs them with the number of scans ordered by physicians who sent patients to scanning equipment at outside entities. The report also attempts to focus particularly on physicians who acquired imaging equipment and began “self-referring” (the report refers to them as “switchers”). The GAO asserts that when physician “switched” from non-self-referral to self-referral, the number of scans they ordered increased 67%, while the number of scans ordered by those who “self-referred” the entire time decreased 3.4% and the scans ordered by non-self-referrers decreased 6.8%. The report asserts that Medicare spent $109 million on scans that would not have been ordered if the physicians did not self-refer.
It is certainly possible that physicians who self-refer order more unnecessary scans than those who do not. But this report doesn't prove it. It doesn’t even come close.
Examine the scans ordered
To determine whether physicians are unnecessarily ordering scans, you need to examine the scans ordered. A good study would take scans ordered by physicians who own imaging and physicians who don’t, and then have outside experts review the situations on a blind basis to consider whether the patient’s symptoms supported the need for the scan. The outside reviewer could possibly review the scan to see whether the suspected problem is less likely to occur in patients who are “self-referred.” That would be a meaningful study, but it isn’t what the GAO did. Instead, they looked at the number of scans ordered in aggregate by physicians when the scan was provided by the entity the physician is employed by to the aggregate number to the number of scans ordered by physicians when another entity provides the scan.
To see the fallacy of what the GAO has done, it is useful to consider an analogy. If you wanted to determine whether people were embezzling, could you determine it solely by looking at their wealth? Of course not. People who embezzle might have higher wealth than those who don’t, but the presence of higher wealth doesn’t suggest that the wealth was improperly obtained. Wealth correlates with embezzlement, but that doesn’t suggest wealth is caused by embezzlement.
One would expect a significant correlation between physicians who order many scans and have ownership of scanning equipment. A physician who orders more scans is far more likely to seek to have the equipment available. In fact, it would be bizarre if physicians who rarely ordered imaging purchased equipment, while those ordering frequently did not.
GAO’s methodology
If I understand the GAO’s methodology, it made absolutely no attempt to examine the physician’s compensation methodology. Physicians who received a fixed salary regardless of the number of scans performed were considered “self-referrers” in exactly the same way that a physician who received every dollar obtained from imaging. For the reasons outlined above, I think any study that fails to examine the medical necessity of the scans is fatally flawed. But if you are going to review the impact of “self-referral,” then it seems that, at a minimum, you need to examine whether there is, in fact, an economic benefit to the physician.
The bottom line is that we have yet another “study” that will cause people to claim that we should prevent imaging in the clinic. And once again, the study doesn’t actually examine medical necessity. The influence of self-referrals is worth studying. But before someone suggests making every patient leave their physician to go elsewhere for imaging, it would be wise to have a decent study suggesting that the benefit outweighs the cost.
Reference:
www.gao.gov/products/GAO-12-966