In response to:
When Doctors Disagree from the April 25, 1985 issue
To the Editors:
I was pleased to read the thoughtful article, “When Doctors Disagree,” by John Bunker, MD [NYR, April 25]. In lucid fashion Dr. Bunker has pulled together concerns on which a number of research and program activities of the Department of Health and Human Services (HHS) have been focused.
The elegant small area data analyses conducted by Dr. Wennberg and supported in part by HHS demonstrate remarkable variations in the frequency with which various medical procedures are performed in this nation, even when such factors as patient age and income are statistically controlled. HHS is continuing to support research in this area—a request for grant applications to examine geographic variation on a national basis has just recently been published.
These variations have also played a role in a number of other HHS efforts, including study and demonstration of “second surgical opinions.” Patients advised by their doctors to undergo various non-emergency surgical procedures were provided the opportunity to secure a second opinion on the desirability of the surgery. Because physicians will have differing views regarding the “flat of the cost-effectiveness curve,” in Dr. Bunker’s language, it was anticipated that additional patient information, provided by the second opinion, would raise questions and doubts where appropriate, and reinforce patient confidence in proposed medical interventions elsewhere. For a variety of reasons, results of the HHS studies of face-to-face second opinions were mixed. Consequently, for Medicare patients we have moved to a variant: mandatory pre-admission reviews for selected procedures. These mandatory reviews, performed for Medicare enrollees by Peer Review Organizations, include all elective surgery.
In the same vein, Dr. Bunker makes a strong case that too many medical procedures and technologies lack an adequate assessment base upon which to judge their efficacy. Under such conditions, it is not surprising that individual physicians will examine the same incomplete data and come to different judgments. Over the last decade, HHS has moved to address this problem in several ways. The major biomedical research arm of the department, the National Institutes of Health, not only sponsors work in new areas of research, but devotes considerable effort to Consensus Development Conferences, designed to establish medical agreement on best practices, and clinical trials to determine medical efficacy.
In addition to NIH activities, the Food and Drug Administration must approve all new drugs and devices, with safety and efficacy the key criteria. Technology assessment is also of interest to the Health Care Financing Administration which manages the Medicare and Medicaid programs. Combined these programs make HHS the nation’s single largest medical “insurer.” As well, technology assessment is an explicitly legislated HHS responsibility performed through its National Center for Health Services Research and Health Care Technology Assessment.
The National Center is currently working with the National Academy of Science’s Institute of Medicine and other professional organizations to coordinate and improve the ongoing process of health care technology assessment. The common goal is to secure better, more reliable data, and to make those data widely available to medical practitioners. Indeed, my one quibble with Dr. Bunker’s article is the incomplete and therefore misleading statement in the last paragraph that the “National Center for Health Care Technology…was allowed to expire for lack of funding.”
The demise of NCHCT resulted from administrative and organizational determinations that movement of the technology assessment function into the National Center for Health Services Research would make it more effective. There has been continuity in performance of the function itself. Renaming NCHSR the NCHSRHCTA simply acknowledges, belatedly and with a terrible acronym, the changed locus of responsibility.
Finally, let me expand upon Dr. Bunker’s passing mention of prevention. Such behavioral or “lifestyle” modifications as eating three well-balanced meals a day, getting appropriate amounts of sleep and exercise, using seatbelts, refraining from use of tobacco, and proper use of alcohol certainly are not as flashy as artificial hearts. Taken together with modest medical interventions such as immunization against preventable diseases and early prenatal care, these private, voluntary actions of self-help and personal responsibility are far more effective in increasing longevity. And, it is such health education and prevention efforts that the Department of Health and Human Services is increasingly seeking to emphasize.
Charles D. Baker
Under Secretary of Health
and Human Services
To the Editors:
I am very much in accord with Dr. Bunker’s view…. I believe strongly that the medical system in this country is distinctly overused. We have made enormous progress in medical diagnosis and treatment in the forty plus years that I have been out of medical school (I am a contemporary of Dr. Bunker’s), but as he points out, we tend to overuse technology, and new techniques and procedures very often become widely used before they are adequately evaluated. Inevitably, some turn out not to be useful, but by then a considerable amount of money has been wasted, not to mention the fact that patients have not been benefited.
Another significant factor tending to increase medical costs is the threat of malpractice. Dr. Bunker alludes to this, and I think it needs emphasis. Patients obviously should be protected when they have been victimized by incompetent or ignorant physicians, but it is painfully clear that the situation is out of hand. Many able physicians do more tests and more procedures than might otherwise be the case because they are practicing “defensive medicine.”
Increases in costs without improvement in quality also result from the inequity in the payment system. One gets much more recompense for doing a procedure, often an unnecessary one, than one would get were he to sit down and spend some time talking with a patient. Given that so many patients do not have significant disease as a basis for their complaints, more talk and less procedures would be cost effective.
As Bunker points out we can expect more and more new technology and more and more new laboratory tests. Some of these may, if properly used, help to maintain quality at a lower cost or at least be cost effective. One cannot gainsay the fact that certain procedures, as for example, correction of a congenital heart defect, may be expensive but in terms of the patient’s subsequent economic productivity and freedom from illness may be highly desirable. On the other hand, instituting on a broad scale the treatment of irreversible disease, as for example, end stage renal disease by dialysis, without realistic consideration of the ultimate cost, does not make sense, at least as a policy. In my view, the current interest in artificial hearts is another example of a bad trend. Lacking a fail-safe power source that can be implanted, the quality of life for a patient with an artificial heart, even if he avoids infection or other complications (and it is doubtful that he will for very long), can hardly be considered desirable. Given the costs involved, it is difficult to justify the expenditure when further attention to simple, preventive measures such as more widespread immunization, would do a great deal more for the populace as a whole.
Robert J. Glaser, MD
Lucille P. Markey Charitable Trust
To the Editors:
Dr. John Bunker is surely correct in nothing that “overtreatment,” deriving at least in part from inadequate knowledge on the efficacy of many medical procedures, is a serious problem that demands serious attention by the federal government. Doubtless it is, as he says, responsible for much of the variation in patterns of treatment among regions in the US noted by, e.g., Dr. John E. Wennberg—and thus, inevitably, for a good deal of waste.
I submit, however, that lack of knowledge cannot explain the overall increase in health costs that has been going on for at least thirty years, and especially during the past twenty. Since 1965, these costs have outstripped the rate of inflation by over 50 percent, rising from about 7 percent of our GNP in 1965 to 10.7 percent last year.
Like anything else, overtreatment can be properly understood only within its social setting. As I have pointed out elsewhere,1 the economic structure of the American health “industry” actively encourages overtreatment—and waste.
One reason is that the industry was, until very recently, almost totally immune to the influence of supply and demand—because “demand” (the amount of health services delivered) was largely determined by the suppliers (physicians and hospitals). Doctors, not patients, make most of the decisions on who shall be admitted to hospital, what is done to them there, and how long they will stay.
An even more important—I am tempted to say pernicious—influence has been exerted by the system of payment, not by the patient but by “third parties”: Blue Cross–Blue Shield, private insurance companies, and the federal and state governments that finance Medicare and Medicaid. Such payments cover over half our doctors’ bills, and more than 90 percent of our hospital costs.
For complex historical reasons, the third parties were (and often still are) unconcerned with costs: they paid the bill, whatever it was, on the assumption that the services it covered were (1) necessary and (2) fairly priced. Thus neither doctors nor hospitals had any incentive to scrutinize the actual value of what they were doing, and plenty of incentive not to: the more they did, the more they got.
I am not suggesting that most doctors or hospital administrators are venal—merely that they are human. As the distinguished surgeon George Crile, Jr., put it some years back, the best interest of the patient is often “in sharp conflict with the financial interest of the surgeon,” and so, mutatis mutandis, with other practitioners.
As for hospitals, more than twenty years ago Dr. Milton Roemer of UCLA framed the medical version of Parkinson’s law: “Hospital admissions multiply to fill the beds available”—which in fact has long meant substantially more beds than are needed. For example, Brian Hendricks, of the New York State Office of Health Systems Management, recently estimated that one quarter of the hospital beds in New York City are superfluous.
Since hospitals, like hotels, lose money on empty rooms, they have a powerful reason to fill them—if necessary, by admitting patients who do not really need hospitalization, or keeping them longer than is medically necessary. And any number of surveys have shown that this is precisely what most of them do—sometimes must do, on pain of insolvency.
The amount of overtreatment and waste generated by the prevailing third-party system is suggested by the financial performance of a very different system: the so-called HMOs such as Kaiser-Permanente and Group Health Cooperative of Puget Sound. For HMOs, the incentives run the other way: since they are paid a flat annual fee per patient, they have every reason to deliver only such services as are actually needed. And in fact their patients undergo 50 to 60 percent less short-term hospitalization than other Americans.
In Dr. Bunker’s terms, physicians in such an institutional setting are evidently much less prone to overtreat—not because they know more than other doctors, but because they have a constant incentive to ask themselves, “Is this operation (or test, or extra day in the hospital) really necessary?”
I have had occasion to observe overtreatment personally, as a patient. Some years ago, I spent a week at a large New York City teaching hospital, with a pulmonary embolism—a rather common and easily diagnosed condition. My laboratory bill at the end of that time was $1,500—equivalent, I suppose, to nearly $2,500 today. Since I am fairly sophisticated medically (I make a good part of my living writing for doctors), I took the trouble, with the aid of some physicians of my acquaintance, to check out the bill point by point.
My conclusion was that between one-half and one-third of the tests it covered were unnecessary: that is, they did not provide information really needed for diagnosis or treatment. (A definitive diagnosis was in fact made—not by tests but by radiology—the second day after my admission.)
My personal experience also suggests that third-party payments are not unconnected with the steady rise in physicians’ fees (in constant dollars) over the past twenty years. About five years ago, I consulted an eminent New York specialist who, in two visits to his office, gave me less than ten minutes of his time. The bill was $125. When I suggested that this was a little steep, he exclaimed, in honest bewilderment, “Aren’t you insured? If you aren’t, we can discuss it.” Of course I was insured, so I—or rather the insurance company—paid.
In a setting where the bill will usually be paid regardless—and at little or no additional cost to the patient—most busy doctors (and most doctors are busy) will naturally think first of whether a given procedure will do the patient harm rather than of whether it will or won’t do him (or her) any good. As someone once remarked, man’s social existence determines his consciousness, and doctors are men—and women—like the rest of us.
Certainly we need more information on the efficacy of medical procedures. But let us also devote some thought to what sort of institutional incentives will encourage doctors to make full use of that information as it becomes available.
New York City
To the Editors:
John P. Bunker, in his “When Doctors Disagree” falls into the common error that decreasing smoking would reduce total medical cost. As a rule of thumb, smoking a great deal tends to develop diseases such as lung cancer, for which there is substantially no treatment, and which cause premature death. If the hundred thousand people who may die of lung cancer in some period had not contracted lung cancer, they would have lived on and probably contracted a number of diseases of old age, the treatment of which is massively more expensive than lung cancer.
Smoking, of course, causes other diseases, some of which are, in fact, expensive to cure. But the net effect is a reduction in total medical expenses. Needless to say, this is not an argument for encouraging people to smoke: Pol Pot sharply reduced the cost of medical care in Cambodia. But people arguing against smoking should not use false arguments. There are, after all, enough correct ones.
George Mason University
John P Bunker replies:
I welcome the foregoing several endorsements of the views expressed in “When Doctors Disagree.” To summarize those views briefly, I have argued that widespread medical and surgical overtreatment in this country is the result of professional optimism coupled with inadequate evidence of the effectiveness of many, perhaps most, medical technologies. I welcome the additional views expressed concerning other contributing factors. I agree with Mr. Claiborne that inadequate knowledge is not the sole cause of overtreatment. It is, of course, widely recognized that there is substantially less short-term hospitalization under prepayment or capitation systems of reimbursement than when physicians are paid a fee for service. Elective surgery for subscribers of prepayment schemes, such as Kaiser-Permanente and other HMOs, is also less than for patients who reimburse their surgeons by fee for service, but the difference is much smaller, and I believe that there is considerable overtreatment even with prepayment. That the method of payment is not the only culprit, however, and perhaps not even the principal one, is strongly suggested by the very high rates of elective surgery reported for physicians and their spouses when they are patients, as discussed in my article.
If overtreatment is caused by many factors, there is not apt to be a single solution, and I enthusiastically endorse changing the current pattern of reimbursement. Physicians should be reimbursed for their “cognitive services”—that is for their advice and their decisions—as well as, and on par with, reimbursement for procedures, as Dr. Glaser apparently favors. Prepayment for services will certainly also lessen the incentives to overtreat, and, indeed, many physicians warn that prepayment will result in undertreatment. Given the unlikelihood that the exact optimal level of services can ever be determined with precision, I would argue strongly that the possible error of undertreatment is to be preferred to the current reality of overtreatment, an argument that I develop in detail in my article.
Whether reimbursement is provided through prepayment or by fee-for-service, we will need a much-improved effort to evaluate the effectiveness of medical and surgical procedures. I was startled, therefore, to note Undersecretary Baker’s “quibble” with my reporting of the demise of the National Center for Health Care Technology. It is true, as Mr. Baker states, that the administration has maintained “continuity…of the function” of technology assessment—but just barely. The 1984 budget of the National Center for Health Service Research, before addition of responsibility for technology assessment, was $18.9 million. Congress has authorized a 1985 budget of $21 million for the Center expanded to include health-care technology assessment, but the President has instead recommended a reduction of over $2 million to $16.7 million, and with no increase for fiscal year 1986. Some relief is available by way of the so-called set-aside of I percent of all Public Health Service funds to be used for evaluation and to be distributed among the several Public Health agencies. Although this might have resulted in $5 million of new funds for technology assessment—i.e., evaluation—only $1.49 million has actually been made available, leaving the Center with half a million fewer dollars for the current year. In reality, therefore, the National Center for Health Services Research and Health Care Technology Assessment is expected to assume its new responsibilities with a considerably reduced budget. The substantial national commitment necessary for an effective program of health care technology assessment that I, and others, have called for, unfortunately remains an unrealized goal.
Professor Tullock raises an interesting and important point, albeit one somewhat peripheral to the main argument of my article. He points out that a decrease in smoking would increase, not decrease, “total” medical costs. Whatever difference he has with me is more apparent than real. The yearly cost of medical care of nonsmokers is certainly less than that of smokers, and a decrease in smoking could be expected, therefore, to lessen annual per capita medical expenditures. Total medical costs—that is, medical costs over a lifetime—are, on average, apparently greater for nonsmokers, as claimed by Professor Tullock, simply because nonsmokers live several additional years and in so doing incur several additional years of medical expenses.2 These additional medical expenses are small, however, compared with the large societal costs of providing Social Security benefits for the many additional years of postretirement life for the average nonsmoker; indeed, it was for this reason that the British government chose, in 1971, not to accept a recommendation by the Royal College of Physicians to mount a national antismoking campaign.3 The British government has more recently reversed its position and presumably for the “correct” reasons that the professor, quite properly, urges we rely on when making policy concerning smoking.
“The Great Health Care Rip-off,” Saturday Review (January 7, 1978); “Why We Can’t Afford National Health Insurance,” Saturday Review (May 13, 1978). ↩
Robert E. Leu and Thomas Schaub, “Does Smoking Increase Medical Care Expenditure?” Soc. Sci. Med., vol. 17 (1983), pp. 1907–1914. ↩
Howard Leichter, “Public Policy and the British Experience,” The Hastings Center Report (October 1981), pp. 32–39. See also A.B. Atkinson and J.L. Townsend, “Economic Aspects of Reduced Smoking,” The Lancet, vol. 2 (1977), pp. 492–495. ↩