The title for this column is a partial quote from the first of a series of articles in the December 2010 issue of Health Affairs. The full title is, “The Self-Referral Boom and the Ongoing Search for Effective Policies to Contain It,” and the article is written by Bruce Hillman and Jeff Goldsmith. This first article provides a good summary of the issues of self-referral in diagnostic imaging. It is a fascinating story and worth reading because the underlying dynamics are evident in a much broader scope of medicine than diagnostic imaging. The studies and authors identified below are all from this issue of Health Affairs.
Most of you are familiar with this common situation. A physician practice with no training in diagnostic imaging, usually a medical or surgical group, purchases imaging equipment for their office. In the past, this was limited to ultrasound and plain x-ray films for examining chests or bones. The advantage of this arrangement was convenience for the patient who could get the examination during a visit to the doctor's office and not need to return or go elsewhere for the film. An advantage for the physician was the efficiency in his/her time in wrapping up an informed medical approach. A second advantage was the extra income to the practice for the imaging.
From the start, there were concerns about conflicts of interest—i.e., that doctors might order more imaging procedures, some unnecessary or excessive, when they had a financial stake in them. But the plain films and ultrasound were not a huge source of income and the rationale seemed to make sense.
That scene changed dramatically when CT, MRI, and even PET scans entered the picture. Manufacturers aggressively pushed for this equipment to be in doctors' offices and structured deals that dealt with the substantial capital needed to purchase, install, and maintain this sophisticated equipment.
So the arrangement became financially feasible and a lucrative source of income. As the cash on the table skyrocketed, the risk of overusing the tests also increased.
Studies demonstrated unequivocally that increases of as much as 25% or more in usage occurred when compared with a matched patient population in the same practices before the new equipment was installed. The charges for CT, MRI, and PET scans were an order of magnitude greater than for the older equipment, as were the profits.
The study by Laurence Baker showed that the number of MRI procedures after a first visit to orthopedists increased 38% after the latter acquired the equipment. Not only did spending increase for MRIs, but for other aspects of care as well.
Is This Really a Problem?
Policy makers and Medicare have tried to stem this tide and have had some success, but new financial approaches to skirt the Stark laws often popped up in response. Is this really a problem? Shouldn't the convenience and efficiency of this arrangement call for flexibility?
Laying out the findings of studies of this issue over the past several years may help us decide. The results and references for these findings can all be found in the Health Affairs articles and their bibliographies.
Convenience and efficiency are cited as a major rationale for this arrangement. However, an analysis of Medicare data by Sunshine and Bhargavan in the Health Affairs series showed that, in practices with their own equipment, same-day imaging was obtained for 74% of plain films but only 15% of more advanced procedures such as CT and MRI.
The authors say they believe that even this number may be an overestimate because other data have shown that many of these procedures may actually have taken place at another location under what CMS calls “abusive leasing,” by which the patient's doctor leases time on equipment located elsewhere and bills for the service. So the convenience and efficiency argument is, at a minimum, brought into question.
Does the self-referred imaging result in shortening the duration of illness and lowering costs, as claimed by some proponents? The study by Hughes et al found that self-referral was associated with substantially higher costs per episode of care. There also was no decrease in the duration of illness with one exception—when doctors self-referred patients for plain x-ray imaging.
2 Other Relevant Points
The conclusions from this and other supporting data are: excepting plain x-rays, self-referral by physicians to imaging facilities in which they have a financial interest results in substantially more images and higher cost, little or no advantage in convenience and efficiency, and no shortening of the duration of illness. But there are two other relevant issues to consider.
First is the ethical issue of a conflict of interest. Hillman and Goldsmith state that, “Physicians' conflicts of interest strike at the heart of the integrity of the physician-patient relationship.” Attempted remedies for this problem include the posting of signs in a doctor's office stating that he/she has an ownership stake in the services he refers the patient to. But that is like closing the barn door too late; what is the patient to do, walk out?
Not likely, because the patient has already taken a long step toward trusting the physician and will be reluctant to consider modifying that trust over an issue that does not seem to affect them directly.
In my opinion, it is the physician who is fully responsible for avoiding conflicts of interest whether the patient knows about it or not. Ethical norms are of a higher order than legal norms and require one to avoid putting oneself in a position where conflicts of interest are likely to occur.
However, I condemn the practice of self-referral with caveats: there are circumstances when it may be tolerated— for example, when the lack of referral services in a community leaves no alternative. There is one other possible exception that I struggle with intellectually and remain ambivalent. It has to do with medical oncologists.
The Medicare Modernization Act of 2006 led to the severe reduction in income of medical oncologists from the purchase, resale, and administration of chemotherapy to their patients. It was widely believed that the drug reimbursement system was being abused. At one point it was reported that 65% of the income of medical oncologists was coming from the resale of chemotherapy.
But the reduction in income forced some medical oncologists out of business and others to join larger groups or become employed by hospitals because they could not support a modern practice with income from examining and managing the patients.
The other effect was that some practices, usually large ones with access to capital, sought other sources of income. This included purchasing diagnostic imaging or radiation oncology equipment and hiring the appropriate specialists into the practice or contracting with the specialists for services. Imaging and radiation oncology are among the more lucrative medical services and helped practices sustain the income they felt they deserved.
The Fuzzy Line
There certainly are opportunities for conflict of interest abuses in these arrangements, which I have no problem in condemning. But there is a Catch 22 that makes me ambivalent: the current US reimbursement system is awful and unjust. It encourages the use (and overuse) of technical procedures and fails to pay enough for non-technical services so physicians can sustain a practice. So the CMS and other insurers spend enormous energy on policing self-referral abuses while throwing gasoline on the fire by their counter-productive reimbursement system.
There is a fuzzy line between abusive conflicts of interest and potential conflicts that may be risked in order to keep a practice in operation at the highest possible level. It is relatively easy to identify the egregious offenders, as have some of the papers in Health Affairs. But I am not smart enough to know how and when to distinguish among those on the fuzzy line.