Regulation or Autonomy in Transplantation: A Debate

Transplant surgeons are among the most innovative of physicians, and they have to be: placing an organ from one person into another is risky, even under the best of conditions, and the shortage of organs has meant that surgery is often performed with less-than-ideal organs. Although that willingness to take risks has led to life-saving operations for many patients, not every outcome has been ideal, and not every transplant program matches the best centers in the level of patient care it provides.

Given the tension between the benefits of innovation and the costs of poor outcomes, what is the proper role of regulation? That was the topic of debate between Thomas Hamilton, director of survey and certification for the U.S. Center for Medicare and Medicaid Services (CMS), and Dorry Segev, MD, PhD, associate professor of surgery at Johns Hopkins University, speaking at the American Transplant Congress held in Boston in June.

“When there is a high level of complexity, and a high degree of trust is required, these are environments that make regulation useful. That is true of the banking industry, and it is also true of organ transplantation,” Hamilton said.

Segev countered, “If the system for identifying consequences is not good, but the consequences are severe, then risk-averse behavior will undoubtedly ensue.”

The CMS has regulated transplant centers since 2007, stepping in after many years of a hands-off policy because of “a number of headline articles” involving either questionable ethical practices or poor outcomes, Hamilton said. “The problem is there are always a certain number of outliers. That’s what the regulations are designed to address. And this is not just in the public interest. CMS is the primary purchaser of organ transplantation services in the world,” and the regulations are the means for insisting on a certain basic level of outcome to get the most value for public resources.

“There are bad regulations and there are good regulations. At CMS, we are talented in both directions,” Hamilton said. “Smart regulations tend to be outcome focused, rather than dictating every single step, and they recognize unique circumstances, and they promote self-governance and learning.” Those, he said, are the kind of regulations CMS uses to improve programs. A center can be “flagged” for outcomes below the expected range, after consideration of the many factors involved, including patient and organ characteristics. But flagging doesn’t close a center. “We provide the time for programs to improve.” A “mitigating factors” provision “allows a program to come forward, and demonstrate how it has turned the ship around,” Hamilton noted.

Some centers, he said, “are unacquainted with their own data,” and once they understand the data better, they improve. “Ninety percent of programs [which have been flagged] have improved,” some dramatically so. “Ten percent, we invited to voluntarily withdraw.”

The result, he said, is that “for every organ type, since 2007, we’ve seen a continued overall increase in survival. “So despite acceptance of riskier organs and recipients,” which has been a national trend, “the survival is going up. That is absolutely tremendous.”

Making the case for autonomy, Segev stressed a distinction that is often lost when outcomes between centers are compared. “A transplant candidate is someone who does better with a transplant than without one, not someone who does better than someone else,” he said. “The question is, will this patient benefit from this procedure?”

Pressure to improve, he said, can and does come from multiple other sources besides the CMS, including one’s self, one’s fellow workers, and one’s colleagues in the field. “This field has been doing this for 50-some years, without a tremendous amount of regulatory pressure, and outcomes have become better every single year.” Segev said. And the system of regulation, he said, may not be very effective at identifying which centers are doing well and which are not. There are certainly centers that perform below par and are flagged. Even in these cases, Segev said, the consequence could be that fewer patients benefit from transplants. Remember the definition of a transplant candidate, he said: someone who does better with an organ than without one. Studies have shown that the effect of flagging is to reduce transplant volumes at the institution, reducing the number of patients who benefit.

More troubling is that the model used to flag institutions is imperfect and will inevitably lead to flagging centers that are performing at the norm. This, he said, is a statistical artefact of the model. What happens when an artefact makes it seem as if there is a problem when there isn’t? “What happens when we try to improve outcomes that don’t need improving? The answer is we become more restrictive, because we don’t know what else to do.”

To explore the risks of this scenario, Segev, who has a master’s degree in biostatistics, set up a simulation model. His model included the same number of transplant centers as there are in the United States, with the same size distribution, and gave every center the nationally expected outcomes. “Everybody did OK in this simulation,” he said. “Then we ran the [Program Specific Report] methodology to see who gets flagged.”

“When everybody is doing OK, when nobody is working outside of the window we want them to work in, there is an 11 percent chance that a center will be flagged by the methodology at least once,” with higher-volume centers far more likely to be flagged than low-volume centers, even if they are performing exactly the same.

In another simulation, he programmed some centers to have higher risk, with enough poor outcomes to trigger flagging, but he found they were not being consistently flagged. “When there are centers with worse outcomes, there is a16 percent sensitivity for identifying those centers.” And if a center was flagged, there was less than a one-in-four chance that it was in fact one of the poorly performing ones.

Finally, he said, the appropriate level of risk may not be captured, leading the center to appear to have worse outcomes than it actually does, given the risks. “Our ability to predict outcomes in transplantation, based on the data we’ve collected, is just not very good at all.” but barely better than a coin toss, he said.

August 2012 (Vol. 4, Number 8)