Medicare: New Directions in Quality Assurance Proceedings (1991)

Sheldon Greenfield

My charge is to answer two questions: Is the glass half full or is it half empty? Have the enormous advances in research directed at understanding how to measure and improve quality of care given us, at this time, an adequate capability to erect quality assurance systems that make judgments, or do we know so little that we should be wary of making judgments on infirm grounds? I will address our overall progress to date and directions for the future by underscoring six issues that have been raised in Chapter 11 of the Institute of Medicine (IOM, 1990) report.

Some questions can be asked now only because of past research findings. With respect to outcomes, for example, if physicians or organizations are going to interpret outcome information, can they be sure that the outcomes are not attributable to differences in case mix? Can they be sure that outcomes are not distorted by assessing the outcome too early or too late, such that new events befall patients—events that had nothing to do with the receipt of care? Will we get into the problems that we did with the mortality data, where statistical considerations make it such that one or two cases move some doctors or organizations above or below the line? When you give information back to physicians, do they know what to do with it? These are all questions that now need to be addressed and researched.

Roberts (1991) and Luft (1991) outline questions that need to be dealt with in respect to, among other things, practice guidelines. These are downstream questions that now can be asked because the state of the art of guideline development is rather sophisticated. We now have to ask, as we are asking in a study in Boston: Will physicians use guidelines? How will they use

them? Will guidelines save money? If they save money, will they hurt patients? Asked another way: Are the guidelines effective in bringing about more good than harm? Are physicians happy with guidelines? Are patients happy under a system in which guidelines are a pan of day-to-day medical care?

With respect to continuous quality improvement, the questions include: What is it, really? What are the basic invariable elements that make the most difference? What elements are more critical in different situations? Will continuous improvement, a phrase that resounds like a heavenly trumpet blast, have any impact on compliance with process and with outcomes, which is the way we know whether we have improved?

I think that the agenda and the program set out in the IOM report are the way to approach future quality issues. I emphasize that quality assessment and assurance are not a one-shot thing. If priorities are set, future directions can flow from past research, and ways to attack the problems of quality can be handled in an orderly fashion. The IOM program is going to need imaginative and not erratic funding, not only from the government and foundations, but from private sources, a point to which I will return. These private practice organizations will gain in both the short and the long run from getting involved in quality research.

Thus, point number one is the need for research. I think there is a lot of agreement about that. Many questions are settled; most are not.

Point number two relates to the issue of basic versus applied research. A table in Chapter 11 of the IOM report makes this useful distinction. I would argue that, in the future, researchers and practitioners need to get together. All too often researchers study quality-of-care methods in a vacuum. I know this from very hard personal experience; researchers perform a study in a practice organization and then disappear. Nothing ever happens to the quality measures developed in that study. Much more often, every day in this country, thousands, millions of dollars are being spent in windowless lower floors of hospitals and other places on quality assurance activities that nobody can learn anything from. The two need to be combined in a way that I will liken to post-marketing surveillance. Drugs are never completely studied even when they are approved by the Food and Drug Administration. New side effects turn up all the time, but only when, and if, good surveillance follows the approval.

Private sources such as health maintenance organizations, independent practice associations, preferred provider organizations, and insurance companies will benefit. They can put quality programs into place with the promise that results will be somewhat tentative and inconclusive now because

the methods are basically sound but need further development. Over the next five years, as these measures are in place, better answers will be forthcoming and practitioners will be able to learn from their experience, if they work with researchers who have the time and methodological sophistication to reject poor methods and improve promising ones. We need to take the field of quality and put it where it belongs, which is in practice, and to perform research and develop methods that come from practice instead of creating them in isolation.

Issue number three, as has been raised by others, involves the relationships among structure, process, and outcome. Processes and guidelines are not useful unless they have been shown to have an impact on outcomes. Outcomes are invalid and meaningless if they are not relatable to process. We all know that being dead is a health state that is in itself valid; however, to those in the health system, it has no meaning unless it is linked to something that has to do with the prior receipt of care. Despite the fact that ten years of extraordinary research has now allowed us to determine the health status of people, that health status may have relatively little to do with what we do in everyday practice until we establish the link. Finally, with respect to structure or continuous improvement, we need to know whether outcomes affect continuous improvement, whether continuous improvement affects processes and outcomes, and so forth.

In the IOM End-Stage Renal Disease Study, we are trying to put together some quality indicators. I will give you one illustration of how process and outcome need to be united. It might be asserted, for example, that with new drugs and transfusions, patients undergoing dialysis should have a hematocrit of 30 as a proximate outcome. They might also feel better and function better. If those outcomes are not achieved, the process needs to be examined to see whether the four or five other competing causes of anemia are present, whether the patient refused treatment, or whether care was suboptimal. Process also needs to be examined because putting someone on erythropoietin can cause depletion of iron stores. That consequence of care would not show up in an ordinary outcome assessment, because it is relatively rare.

What I have just described is a set of potential indicators of process and outcome that can be validated only by putting them in place and seeing how they work. There is no other way; it is not magic. We cannot, as the definition of quality implies, go beyond the state of our professional knowledge.

Point number four is generalizability: what works in one place does not necessarily work in other places. It is not that the principles are not valid; it

is just that the tailoring of methods from one situation to another will in itself make some methods invalid, and then they need to be retested. This is not a trivial matter. Taking methods off the shelf and applying them requires more than just a simple application. It has to do with understanding both the methods and the specifics of the context in which they are applied. That needs to be done for poor people's care; it needs to be done in outpatient settings and inpatient settings, long-term care, and so forth. In my view, for example, outcomes are going to be a better measure of quality of care in the office practice of medicine than process. That assessment must be tested to determine which is better in which circumstances.

Points number five and six are ones that, as my four-year-old Nintendo-playing son says, take us into some new worlds: World V and World VI. World V for me is the new world of the patient that has been amply and, in a visionary way, alluded to in the IOM report. In a recent book chapter, two of my colleagues discuss patients as (1) the raters of their health (that is, satisfaction), (2) the reporters of health (health status), and (3) the participants in their health care (preferences) (Kaplan and Ware, 1989). All three of these roles, and others that may arise, must be fleshed out, quantitatively and understandably. It is not just a matter of doing what the patient wants. None of us ever gets exactly what we want at the time we want it, and patients are not going to get that either. Negotiation or compromise first demands understanding the patient's role and how to measure and agree as to when the optimal role of the patient is fulfilled.

Finally, World VI is the doctor's world. There is a great deal of work to do here. Many programs including, I would argue, measures of cost- effectiveness and cost-containing initiatives have foundered because physicians have not been brought into the equation. That has been amply emphasized in the IOM report.

I think we need to test all kinds of incentives and to refashion the training of physicians so that they incorporate quality assessment into everyday practice, much as they have integrated billing into their lives. Many of you remember the days when office practices were not as high powered in terms of their automated billing systems. I think that we have to be imaginative about ways to get physicians involved in the quality assurance effort, without making them sullen, difficult, and unhappy.

These are just six issues that I chose as an overview to this discussion of research and capacity building. I would like to see the kind of program proposed in the IOM report take hold. I know it is somewhat vague, but to

me it has, in fact, quite a lot of teeth in it compared to other proposals that we have seen over the years. I would like to see this program implemented over an extended period. I think it will not only serve to integrate a lot of the disparate and fragmented players in the health care field, but also result in the improvement of quality of care not only in the Medicare population, but for all of us.

Institute of Medicine. Medicare: A Strategy for Quality Assurance. Lohr, K.N., ed. Washington, D.C.: National Academy Press, 1990. (See especially Volume I, Chapter 11.)

Kaplan, S.H. and Ware, J.E., Jr. The Patient's Role in Health Care and Quality Assessment. Pp. 25-68 in Providing Quality Care: The Challenge to Physicians. Goldfield, N. and Nash, D.B., editors. Philadelphia, Pa.: American College of Physicians, 1989.

Luft, H.S. Research and Capacity Building: Issues Raised by the Institute of Medicine Report. Pp. 130-139 in Medicare: New Directions in Quality Assurance. Donaldson, M.S., Harris-Wehling, J., and Lohr, K.N., eds. Washington, D.C.: National Academy Press, 1991.

Roberts, J.S. Translating the Institute of Medicine Report Strategy Beyond Medicare. Pp. 116-126 in Medicare: New Directions in Quality Assurance. Donaldson, M.S., Harris-Wehling, J., and Lohr, K.N., eds. Washington, D.C.: National Academy Press, 1991.