Page images

STATEMENT OF AVEDIS DONABEDIAN, M.D. Dr. DONABEDIAN. I am Avedis Donabedian. I teach medical care organization at the University of Michigan at Ann Arbor.

The contribution I would like to make to the deliberations of the subcommittee has to do with the quality of medical care. This is an important subject because we are all committed to providing the best possible care. Moreover, quantity and quality are intertwined, and both are related to cost.

It is obvious that we cannot have quality unless, as a first step, care is made available to those who need it, in the quantities that are required to meet their needs. Therefore, by simply making care more equitably available, we improve the quality of the system. But this may not be enough. People should receive the right kind of care, and neither too little nor too much. The wrong kind of care can be either ineffective, or harmful, or both. Too much care is always wasteful, and frequently harmful as well. This means that as we make care more available to people, they are exposed not only to its benefits but also to its dangers. If we want the most in benefits with the least in dangers, we have to pay attention to quality as a distinct issue.

The relationship between quality and cost is also complicated. More and better care is, obviously, costlier. But there are two kinds of savings. First, we reduce direct costs by cutting out inappropriate and unnecessary care. Second, we have the benefits of better health. Unfortunately, the benefits of better health are, in a sense, temporary. This is because people live longer with an existing disease, or develop new diseases that we cannot cure, but which require large amounts of

As I see them, the policy implications of the complex relationship between quality and costs can be stated in general terms as follows:

1. As a nation, the average level of quality that we can buy for a given amount of money can vary greatly. There are ways of getting more for our money. However, in my opinion, we don't know enough to specify what is the best way.

2. Still further improvements in quality will almost certainly mean more expenditures for care.

3. Beyond a certain point, further improvements in quality will yield smaller and smaller additional benefits. Unfortunately, the costbenefit relationship is so poorly understood that we have no way of saying at which point additional improvements in quality are simply not worth the additional expense.

So far, we have spoken about quality as if we know what it meant and were agreed upon the meaning; and to a certain degree, this is so. But there are also differences of emphasis. Physicians and other health professionals tend to stress the full application of the technical procedures of medical science. Given the opportunity, their tendency is to spare no expense and no effort on behalf of each patient as long as some benefit can be expected from the care given.

As a result, the standards of care formulated by professionals tend to envisage more care, more complex care, more costly care. Consumers are poor judges of the technical quality of care. They accept the health professional's formulation of that, but they add a special emphasis of their own, which is the emphasis on the interpersonal relationship.


Consumers look for personal interest and commitment, mutual respect, information about their condition and a share in the decisions that affect their lives.

It follows that the quality of care that we seek as policymakers should include both of these viewpoints: The professionals and the consumers'. Moreover, we should not hesitate to challenge the professional to show us to what extent his standards are based on a demonstrated relationship between the technical procedures he wants done and the kinds of outcomes that consumers find worthwhile. If we do so, we can expect wholehearted support from enlightened professional opinion.

The public and the health professions have always shared a concern for the quality of care. Accordingly, our society has developed many mechanisms to safeguard quality. These include standards of medical education and training; licensure and specialty certification of health professionals; licensure and accreditation of hospitals; controls by hospitals on who may join their staffs and under what conditions ; and, more recently, mechanisms for monitoring the performance of professional personnel, especially in hospitals and other organized health care settings.

We can be justly proud of what we have done and what we have achieved in safeguarding quality. However, we also need to ask whether we need to do more and, if so, what? Both of these questions are not easy to answer.

One way of deciding whether we need to do more is to compare our national record of mortality and longevity with industrial nations that are similar to our own. It is true that we do not do well in such comparisons, which should be a matter of deep concern to us. However, it is also true that we do not know how much of our relatively poor showing is due to deficiencies in medical care, and how much due to other social factors. Similar differences by social class are to be seen within the United States itself; and, again, it is difficult to say how much of these differences in mortality and disability is due to differences in medical care. We believe that medical care plays a role, but we do not know how large it is.

We have no national surveys of the extent to which care conforms to professional standards. However, we have the reuslts of many local studies spanning several decades. From these we can conclude that whenever the quality of care has been examined, a significant proportion of cases have fallen short of reasonable professional standards. This does not mean that health professionals are particularly prone to error due to ignorance or carelessness; only that their work has perhaps been more closely scrutinized, and that there is less tolerance for error in these professions than others. And this is as it should be.

In order to remedy the situation, we need to understand what factors encourage or discourage good performance. As one reviews a succession of studies two factors stand out: One is training and specialization, and the second is the influence of an organized setting such as a hospital or group practice. It appears that general practitioners who have had more training after completing medical school perform better than those who have had less training; and that physicians who have had enough training to qualify as specialists do better than those who have not, provided the specialists confine their practice to their specialty


The type of hospital is also important. Hospitals that are affiliated with medical schools tend to stand out as distinctly superior in quality. There are additional, but smaller, distinctions which favor hospitals that are larger, that have some training function and that are nonprofit.

The characteristics of the hospital and the characteristics of the physician reinforce each other for good and for bad. The best care is provided by highly qualified physicians in well organized hospitals, and the worst care is provided by poorly qualified physicians in poorly organized hospitals.

Our knowledge of the effects of group practice is not as extensive, and the findings not as clear-cut. My interpretation of the evidence is that a well organized group practice has the capacity to improve the quality of care. However, just as all hospitals are not similar, not all group practices are alike in having a good influence on the quality of

An additional important feature of the setting in which physicians work is the fee-for-service method of payment. The fee-for-service incentive encourages the provision of service, which is good. However, fee-for-service, especially when coupled with health insurance, also encourages the performance of unnecessary, costly, and sometimes harmful procedures. Thus, we have a mixture of good and bad effects.

Before we go on to draw conclusions as to public policy, we should emphasize some limitations in the findings we have described. First, our measures of quality are still quite rough. Second, very few of these measures pay attention to the attributes of the interpersonal relationship. Third, all the findings refer to averages; and there are always exceptions to the rule. In fact, there are so many exceptions that the differences in quality that remain unexplained are probably larger than the differences that one can attribute to known factors.

All these findings lead to the commonsense conclusion that the best levels of care are obtained when we have highly qualified physicians working in the best settings.

But what makes the setting, whether a hospital or a group practice, good? We do not know for certain. It appears that the teaching and training function is important, perhaps because it provides a constant intellectual challenge and opportunities for continuing education. Also important are selectivity in recruiting staff, matching their practice to their areas of competence, continuing to obtain information about the quality of their work, rewarding good performance and taking steps to correct poor performance.

Does this mean that, ideally, every physician must be a highly trained specialist and every hospital a university-affiliated teaching and research center? If people are to be free to see any physician they wish, and every physician is to be free to undertake whatever he believes hé is qualified to do, the answer would tend to be, “Yes."

In other words, the goal would be to make every physician and every hospital competent to deal with a very high proportion of cases that present themselves for care. Obviously, this is very costly.

An alternative, and less costly solution, is to recognize that some medical care tasks can be performed just as well, if not better, by less trained people and in less highly developed institutions. This means that professionals and hospitals are organized in such a way

that each medical care problem is treated by the person or persons with the appropriate qualifications in the appropriate institution. But to accomplish this will require a move even further away from a free medical care market. Because there is a great deal of resistance to such a move, we are now searching for compromise solutions.

The PSRO's can be regarded as one such compromise measure. The PSRO scheme takes the medical care marketplace as it exists, but imposes upon it a formal system of monitoring. As a further concession, it delegates the implementation of the monitoring system to the physicians themselves, with some degree of external control.

In my opinion, monitoring, both internal and external, is a necessary part of any medical care organization, because it is the only means of obtaining continuing systematic information about how the medical care organization works and the degree to which it attains its own objectives. However, I believe that major reliance on a monitoring mechanism to police the marketplace is only a temporary phase, and that fundamental solutions will involve reorganizing the medical care system itself.

In the interim, the PSRO's pose two kinds of dangers: One is that they will not be implemented or that they will be implemented halfheartedly; the other is that they will be implemented using the wrong kinds of standards, so that their effect will be to encourage a great outpouring of routinized medical care procedures with proportionately little benefit to health. Possibly both eventualities will occurthat the PSRO's will be implemented half-heartedly using the wrong standards and methods, so that they will produce neither much harm nor much good, except that they will become a large and expensive, bureaucratic fixture.

If I had to make a bet, this is the course of events that I would predict. But, to be fair, one should recognize that the PSRO's also have much potential for good. To achieve this potential, the PSRO's should be regarded as a great social experiment to achieve quality assurance in a pluralistic and still largely private medical care system. As such they should attract the best thinking in medicine and medical care organization; they should be implemented in stages, with careful study of their effects; they should encourage variability in methods rather than uniformity; and they should reduce the emphasis on policing and punishment, and increase the emphasis on self-examination, learning and appropriate reform in the medical care system.

In closing, let me return to the question implicit in my opening remarks: To what extent should all this be the concern of the subcommittee, and why? To begin with, a change in the method of financing the purchase of medical care will almost certainly bring about changes in what care consumers will seek and health professionals provide. As we said, this will bring about changes in the quality of care and these, in turn, will have an impact on cost. What I am saying is that, at a minimum, the subcommittee should assume responsibility for the consequences of its own recommendations. More broadly, the design of a national health program presents many opportunities to strengthen those tendencies in the medical care system that serve the public interest, and to weaken those that are harmful to it. I have no doubt that the subcommittee, in everything it recommends, will remain watchful over the quality of medical care, and, in so doing, will earn our respect and gratitude.


Mr. ROSTENKOWSKI. Thank you, Professor. We will afford the panelists an opportunity for interchange. If any member of the panel would like to question another member of the panel about any statement that was made, we so encourage that.

Mr. FELDSTEIN. I noticed as I listened to the others, one theme that kept coming up over and over again and it comes back to what I said earlier.

Mr. Cohen said there is no limit to the amount that individuals will want or use of health care or rather the only limit is the limit imposed by supply. A similar statement was made by others. Dr. Donabedian said doctors are all committed to providing the highest quality of care.

While I was quite interested in the schedule and scenario that Wilbur Cohen presented to you, nowhere did I get a sense of how the fundamental question of the level, the style, the quality of care would be determined. I want to emphasize the nature of that problem.

If you consider the kind of care that you would expect to get if you became ill, seriously ill, the kind of hospital services that would be available to you as a Congressman, is that the kind of care that ought to be extended to the entire population ? Is that the level that would be appropriate under national health insurance?

If that cost $50 billion would that be the best use of $50 billion of additional tax revenue?

Think about the kind of care that might be available 10 years from now-better than today. We could speed it up. Would that be a good use of resources ?

I am afraid all of the discussion and all of the bills presented do not squarely face the problem of how the quality of care would be determined. The term "prospective reimbursement" came up a number of times. I think a look at the Canadian experience causes one to be skeptical about how effective that is in controling costs. Even if one could, and that is ultimately a political consideration, at what level would you control them? Prospective reimbursement might give you a lever but it would not yield the appropriate level or a mechanism for deciding what the appropriate level is.

I think whether costs continue in the future to go up at the astronomical rates we have observed, 600 percent in two decades will depend upon whether you find an alternative financing mechanism which makes the cost of hospital care a reflection of the actual preferences of patients and their physicians. What we would really like to be aiming at is a financing and insurance mechanism with deductibles and co-insurance that reflects the tradeoff between the desire to protect ourselves against unexpected bills on the one hand and the desire on the other hand not to have a distorted care package.

For the vast majority the opportunity to buy insurance through employee groups affords a way of making that tradeoff, of deciding how much additional care, how much additional protection one wants against the extra costs. Unfortunately, as I indicated in my statement, taxes distort that choice and distort it very substantially—a subsidy of some $4 billion.

I think what you want to be aiming at in national health insurance is either a restructuring of the private insurance system so it corsponds more to what is actually desired or the provision of national health insurance with the deductibles and coinsurance that corresponds to what individuals would want if they were buying it directly, allow

« PreviousContinue »