Cover Image


View/Hide Left Panel

and outcomes of care and to support a systematic effort to develop clinical practice guidelines and standards of care.


Let us examine how we might address this recommendation through improved decisionmaking. We can start with what we might refer to as first-order research: the need for better knowledge of the relationships between treatments and outcomes of care. Some generic issues warrant attention. For example, better markers of outcomes of care are needed. Fortunately, most patients survive their treatments, and the majority do not have complications. The current markers that we have of morbidity and mortality are not sufficiently sensitive to give us the full spectrum of information needed to judge outcomes. The work by Greenfield and Ware1 on measures of quality of life and functional status is a good example of current research that is helping to address this problem, and the IOM report indicates that more work of this kind is needed. Such markers immediately suggest opportunities to generate clinical data of a longitudinal nature, that is, observations over time and across settings. Disease-specific issues also arise. Given more than one approach to the management of a specific clinical problem, what are the outcomes associated with each approach, those that are good and those not so good? We will refer, in a moment, to the work that people such as Jack Wennberg2 have done that points to the lack of knowledge of effectiveness as an explanation for the tremendous variation in patterns of medical practice across settings throughout the country.

How do we achieve such knowledge? We cannot expect to conduct classical randomized clinical trials (RCTs) to address very many of the outstanding questions; RCTs are too expensive and take too long. The report points to another alternative, and that is to take advantage of what David Eddy refers to as experiments of nature. The opportunity exists to look at the outcomes of care rendered to a large population of patients with a given condition in such a way that the outcomes of various treatment approaches can be compared among subgroups after adjustments for important variables.


Editors' Note: The reference is to a long-running project, the "Medical Outcomes Study," begun at the RAND Corporation and now located at the New England Medical Center. The principal investigators include Sheldon Greenfield and John E. Ware, Jr.


Editors' Note: The reference is to the body of research in geographic variations in use of health care services and outcome and effectiveness research pioneered by John Wennberg of Dartmouth Medical College. See also the papers in this monograph by Mulley (1991) and Wennberg (1991).

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement