the potential impact of the candidate diseases on health care systems in developing countries.

The possibility that a disease might be eradicated has not been incorporated into the calculation of potential benefits, but it could be considered in the ultimate selection of priorities for vaccine development. Early investment in the attack on a theoretically eradicable disease might speed its ultimate elimination and hasten future savings. However, because the current prospects for global eradication of the candidate diseases are remote, this issue probably should not be a major component in the selection of priorities in the near future.

A further consideration in establishing vaccine development priorities is the extent to which particular disease consequences exacerbate (or are exacerbated by) other illness.

The interaction between diarrheal disease and measles is probably the most significant example (Feachem and Koblinsky, 1983). Other interactions are suspected, for example, between viral respiratory infections and bacterial pneumonia, and between various diseases and malaria. In addition, all diarrheas pose the threat of nutritional debilitation, with exacerbation of other illnesses as a consequence.

Because the committee judged that knowledge of these phenomena would not permit reliable quantification of their consequences, no adjustments were made of disease estimates. The phenomena should, however, be recognized and the strength of the evidence considered in the ultimate selection of priorities.

Various legitimate perspectives exist of the relative undesirability of death and morbidity in different age groups. The method proposed for calculating expected benefits from various vaccines allows decision makers to observe the effects on the final rankings of adopting various perspectives. The determination of which weights to use is inherently a political or public policy decision, not a scientific one.

Scientific evidence can be valuable, however, in assessing certain value judgments. For example, epidemiologic and demographic studies suggest that mortality reduction may be a prerequisite for fertility reduction in developing countries. If a long-term view is taken, this finding contradicts the suggestion that programs to reduce infant and child mortality should receive low priority because they contribute to undesirable population growth and increase pressure on scarce resources (although some growth may occur in the delay between mortality and