A Health Affairs study posted and discussed on the Health Affairs Blog yesterday concludes that — based on one insurer’s data for 2004 California — self-referral, though banned, is "surprisingly common."
Researchers traced referral patterns for MRI, CT and PET and found a significant volume of referrals that fit within Stark II exceptions. The author more than hints at a skepticism about the bona fides of the arrangements. The article touches on the financial arrangements as well, but I’d like to focus here on the structural arrangements.
Two figures summarizing the data collected from the article are reproduced below. (Click to enlarge.)
In the first figure, the proportion of MRI "self-referrals" among all MRI referrals by non-radiologist physicians was pegged as about 33% — with lower figures for CT (22%) and PET (17%).
The second figure shows that the majority of "self-referrals" for MRI and CT involve equipment not owned by the referring physicians which is, instead, part of a shared equipment arrangement (with equipment not located in the referring physician’s office).
This second point is held up as a "gotcha" because, while earlier on in the Stark II regulatory development process the in-office ancillary services exception was developed in order to accomodate physicians’ desire to supervise more directly the diagnostic tests being ordered for their patients, the exception also includes certain shared facilities.
The conclusion the author draws from this data is that physicians have used the Stark II exceptions to inappropriately self-refer, thus increasing utilization and supplementing their incomes.
In my humble opinion the data doesn’t quite support such an indictment. The studies that supported the initial passage of the Stark law comapred referral patterns of physicians in a position to self-refer with referral patterns of other physicians, and noted some significant differences. The data presented show that physicians and their advisors are familiar with the regulations and have structured business arrangements based on the regulatory landscape.