DECISIONS about the allocation of resources in health care are being made throughout the world and promise to become more obvious and difficult as budgets become tighter. The issue is how allocation decisions are to be made, and what kinds of information and techniques are to be marshalled to guide decision makers. Michael Drummond’s paper introduces the issue of using quality of life indicators beyond clinical encounters to formulate policy, especially public policy. His paper presents an insightful overview of the growing use of economic evaluations for this purpose. He defines the problem of resource allocation; describes the components, processes, and uses of economic evaluations; and raises key issues about incorporating quality of life assessments into these evaluations. My comments center on the different frameworks of economic analysis and the use of results. They are intended primarily to clarify and less frequently to take issue with aspects of Drummond’s exposition. The role of quality of life indicators in economic evaluations can be clarified by examining the continuum of analytical frameworks that can be used. Consider the question of whether to use extracorporeal shock-wave lithotripsy or open surgery for certain stones in the upper urinary tract. Possible health effects to consider include mortality; days of illness, discomfort, and distress; and long-term consequences, such as damage and eventual loss of the kidney. If the probable health effects of the alternative medical interventions are the same, the problem becomes one of cost minimization, that is, managing the treatment of urinary stones with the least costly use of resources. But if, as is more likely, probable health effects differ among the alternatives and health effects go beyond mortality, one needs an index to combine health effects, so alternatives can be compared. This task calls on the expertise of biomedical researchers, clinicians, health services researchers, and policy analysts. And, as Margolese pointed out, health effects in additional to mortality are of greater importance with chronic disease than with acute infectious disease. Cost-benejit analysis addresses the problem of combining different health effects by using monetary units to express all effects. As Drummond explains, reduced morbidity can be valued by the extra productivity gained from increased working time or by the patients’ willingness to pay to avoid the discomfort associated with a medical condition. Despite the initial use of cost-benefit analysis in the health field during the 1950s and 196Os, there has been a decided trend away from using the technique and toward using cost-effectiveness analysis [ 11, probably because of the distaste for attaching a dollar value to health and life. Cost -e@tiveness analysis examines differences in monetary cost and health effects from using one medical intervention instead of another [2]. The cost-effectiveness ratio expresses the net cost to be expended to achieve a net change in health effects. Health effects may be measured in years of life, in cases of disease, or by a health index, such as quality-adjusted life years. Such an index adjusts changes in years of