How Physician Report Cards Can Enhance Covert Rationing

Posted on July 23, 2007
Filed Under General Rationing Issues |

The state of California recently published a report showing risk-adjusted mortality rates for coronary artery bypass surgery, tabulated according to specific hospitals and specific surgeons. For instance, Dr. Charles Hoopes, a prominent and highly regarded cardiac surgeon, and director of the heart and lung transplantation program at the University of California, San Francisco Medical Center, received a “worse-than-average” rating. (The full report itself can be found here.) Reporters from the San Franciso Chronicle note that Dr. Hoopes’ statistically “worse-than-average” surgical mortality can be specifically attributed to his operating on two very high risk patients who subsequently died. If he had declined to try to help these patients, he would not have been rated worse than average. One wonders what Dr. Hoopes will do the next time a high-risk patient comes to him for a shot at long-term survival.

A study published in 2005 in the Journal of the American College of Cardiology reported that doctors practicing in states that require the public reporting of outcomes data are already holding back potentially life-saving medical care from patients who are at high risk. The authors compared the use of stents for coronary artery disease in New York, where reporting occurs, to Michigan, where reporting does not occur. They found that significantly more high risk patients in Michigan received stents than in New York. The mortality rate of stent patients was likewise higher in Michigan - until the data was recalculated to account for the underlying risk of the patients. Once this risk-adjustment was made, the outcomes were equivalent. So, the actual performance of physicians in these two states is equivalent - but doctors in New York are apparently optimizing their outcomes (and buffing up their reports) by avoiding the highest-risk patients.

Public reporting of outcomes data has several positive attributes. Knowing that the public is watching has caused many institutions to employ new and intensive quality control measures. Some doctors and surgeons who probably should have chosen another career have subsequently had to choose new careers. And patients have a right to know this data, so they can make more informed decisions.

But how do you suppose colleagues of Dr. Hoopes are feeling about now? Here is a prominent and respected surgeon whose colleagues (one suspects) believe him to be the victim of his choosing to do challenging surgeries that others would walk away from. They see his name on a public “worse-than-average” list, and prominently mentioned in newspaper articles (and on blogs); his reputation, his professional standing, and possibly his career in jeopardy.

An object lesson in spades.

The kind of high-risk cardiac patient we’re talking about has a high risk of mortality with or without the procedure - but it’s marginally better with the procedure. Importantly, if the doctor declines to do the procedure - declines to at least try for an improved outcome - the patient’s subsequent death is not publicly attributed to him/her.

If there’s a choice between pulling out the stops to try helping a patient in real trouble or safeguarding their careers - well, they’ve got to protect their careers.

So let’s imagine how the payers feel about this result. The patients who are reasonably likely to die anyway are now kindly doing so without the added expense of an invasive cardiac procedure. The payers get to champion transparency and the public’s right to know, to advertise their aggressive quality-improvement measures - and they save a lot of money besides.

The trick is for the payers - like the state of California, for instance - to maintain control of the “risk-adjustment” methodology that is used in these report cards. In theory and if applied with great care, such risk adjustments are supposed to prevent what seems to have happened to Dr. Hoopes. If not applied rigorously, then not so much. But the payer still gets to say,”Sure he treats high-risk patients, but our statistical measures take that into account. Worse-than-average is worse-than-average.”

Doctors who don’t trust the payers to do accurate risk-adjustment, in an era when payers are desperate to find new ways to covertly ration, will do their own “risk-adjusting.” The doctors in New York have demonstrated how that works.

The bottom line: if you are in a normal risk category, you may be marginally better off in a state providing such physician report cards. Some below-average cardiologists and cardiac surgeons in those states are now either treating warts or the illiterate. But if you are high-risk, consider moving to Michigan.

And once again we have demonstrated the Fourth Corollary of the Grand Unification Theory of Healthcare - Covert rationing corrupts everything it touches. Even physician report cards.

Comments

2 Responses to “How Physician Report Cards Can Enhance Covert Rationing”

  1. Erdos56 on July 24th, 2007 11:40 am

    About a month back, I was in a heated online debate contending that there were real differences in the risk profiles for auto insurers versus health insurers. It was getting deeply technical and I searched about and found your GUTH.

    I was drawn in and spend all day reading. Spectacular piece of work and I want you on my cardiac team should I ever need one!

    On this post: VA now has a health informatics system in place that provides complete history at a glance *and* is also used to manage VA physicians at the level of individual patient contact times. Already there are resources on the web devoted to helping physicians maximize their throughput/points. The urge to game the system (even mildly) is not overridden by ethics or responsibility.

    Sp how can risk really be assessed impartially without independent review boards (expensive and time consuming for a limited resource: physicians) or some other mechanism?

  2. DrRich on July 30th, 2007 9:57 am

    Statistics for “risk adjustment,” aimed at really trying to rigorously assess quality of care, could probably be achieved without requiring independent review boards to assess each case. Today, the data used risk assessment usually comes from Medicare diagnostic codes (which is an extraordinarily imprecise source of data from the get go).

    So, for instance, two patients might have a code for “congestive heart failure” (CHF) and the patient’s risk is adjusted accordingly. But one of these patients had a single episode of shortness of breath two months ago and now feels normal. But the other has had 5 hospitalizations in the past 6 months despite maximal medical management, and even when “stabilized” is greatly incapacitated. Both of these patients would get the same statistical “risk adjustment,” but which one would you be more likely to offer angioplasty to if you were a cardiologist?

    It is possible, by taking into account objective measures (such as number and length of recent hospitalizations, various blood lab results, and several others), to assign a relative risk to each of these very different patients. Risk adjustments done in such a way would be fair, would give a much more accurate notion of a physician’s quality, and would limit what we’re seeing now - doctors avoiding the riskier patients who might benefit from therapy.

    However, by doing accurate risk adjustments we would be limiting the ability of the centralized authorities to use this technique for covert healthcare rationing. So I don’t see it happening.

    - DrRich

Leave a Reply