developing countries and the assumption that extrapolation to the entire developing world was reasonably valid.

Certain general procedures and assumptions were adopted to promote consistency in the derivation of estimates:

• Cases were included under chronic categories (D through G) only if the condition would persist for the remainder of the individual’s life; convalescence or protracted initial illness (even possibly leading to death) from which the individual eventually might recover was not considered chronic disability.

• To simplify implementation of the scheme, acute episodes of illness usually were assigned entirely to the morbidity category representing the most severe signs and symptoms present, although the episode might include periods of recovery at less severe levels.

• Category C was interpreted as morbidity for which hospitalization was desirable, even if probably not accessible.

• For diseases in which the pathogen produces a broad spectrum of illness severity rather than reasonably discrete conditions, estimates of the portions falling into different morbidity categories were obtained from individuals familiar with the disease’s clinical symptoms and epidemiology. In some cases, estimates were made from the most recent epidemiologic surveys of the disease; in other cases previously reported incidence rates were applied to 1984 population figures (Population Reference Bureau, Inc., 1984).

• It was judged that trends in the patterns of diseases under consideration (see Appendixes D-1 through D-19) were generally not of sufficient magnitude to obscure differences among diseases, that is, that the relative impact of diseases when vaccines were likely to become available would be similar to that in 1984. This is amenable to verification. The effect of trends in population numbers and disease incidence on future vaccine benefits is discussed in Chapter 7. The impact of certain diarrheal diseases (especially mortality) probably will be decreased by the increased use of oral rehydration therapy. Two scenarios are therefore included in the calculations of disease burdens and vaccine benefits for these pathogens; they are described in Appendix C.

• For epidemic diseases, the approximate average annual incidence was calculated using epidemic incidence and the average length of the inter-epidemic period.

Limitations on the accuracy of estimates included in Appendixes D-1 through D-19 need to be recognized. The extent to which the estimates represent true disease patterns varies among diseases for the following reasons.

• The quality and availability of data on specific diseases vary.

• The types of data from which estimates were made vary. (For some diseases, infection rates for certain populations could be coupled