Infant mortality equivalence values assigned to specific morbidity category/age group combinations do not differentiate with respect to gender, race, ethnic origin, socioeconomic class, place of residence, occupation, or lifestyle. Often, however, diseases occur more frequently in some parts of the population than in others. The size of these high-risk groups may play a significant role in determination of the national disease burden. For example, if disease A attacks 10 percent of the white population of the United States and disease B attacks 60 percent of the black population, and if the sequelae of the two diseases are equivalent, then on a national basis disease A will produce a greater disease burden than disease B because whites are so much more numerous. Examples of diseases with higher attack rates or more serious sequelae for particular groups are presented in Table 8.1.

One method for going beyond a single, national burden-of-illness comparison is to construct individual disease burden profiles for specific groups, for example:

• American Indians

• women

• persons whose income levels fall below the poverty line

• children in their first year of life

• hospital patients

• homosexual men.

These multiple burden-of-illness profiles might not lead directly to new public policy recommendations, but they would serve as a reminder that the national profile is an aggregate that obscures differential effects in definable population groups. The multiple profiles also could be used to deal openly with issues such as the extent to which one’s behavior places one in a vulnerable group (e.g., drug abusers), and whether or not society wishes to devote special attention to meeting the needs of the more vulnerable groups, however those groups are defined.

The adoption of special campaigns for delivering vaccine(s) to high-risk groups might be one solution to the problem of differential risks; another method to meet the needs of these groups might involve the portfolio approach, described below.

In establishing consistent methods for setting priorities for vaccine research and development support, some consideration should be given to ranking within a group (or portfolio) of candidate vaccines. It may not be desirable or appropriate to have a single priority list based on final numerical (expected benefits) scores for all possible vaccines that need developmental support. This concept is reflected in the committee’s charge to rank vaccines separately for use in the United States and for use in developing countries. Perhaps grouping of