Changes in infant mortality and related rates by health service area: 1969-73 to 1974-77.

The uses of infant mortality data for health planning at the local level were discussed in Statisticril Note for Health Planners No. 2.1 Its associated data supplement provided infant mortality rates by health se~ice area for 1969-73 and 1974-75. These data have been updated through 1977 for the present Note, which provide: additional perspectives on the geographic ~tifi~ri characte=tig infant mortality in the United States over this time period. Examples are presented to tilustrate the strategies healti plan-'m ners cm employ to identify geographic areas witi extreme rates and assess changes in these rates over time. The usefulness of infant mortality data for small areas is limited by the stability of the rates. As an estimate of the underlying rate, the observed inf~t mortality rate is always subject to some degree of chance variation, depending on the number of births and deaths under consideration.: Aggregating data for several years, as done in this Note, augments stability but tends to diminish the timeliness of the data. \ A recent Note (No. 12) explored the potential for using data based entirely on the birth distribution (i.e., maternal characteristics such as low education and late prenatal care) to evaluate infant mortality. For health planning at the local level, this approach generally employs estimates that are more stable than direct measures of mortality. Natality indicators are also useful for assessing the matemal-tisk dimension of infant mortality. However, analysis must eventually encompass measures of outcome—for instance, mortality in addition to variables that are associated with outcome. It should be noted that mortality is not the only useful measure of pregnancy outcome. Certain characteristics of the infant can be considered as outcome variables. Low birth weight and gestational age are both highly correlated with infant mortality and morbidity. Any analysis of infant health should consider the role of these factors on pregnancy outcome. A forthcoming Note covering data on birth weight will further expand the analytic potential of the data presented here. Thus the data and methodology presented in this Note can be viewed as a component of a broader plan for evaluating infant mortality with small-area data. The Rand algorithm on infant mortality provides details of a step-by-step approach that carries analysis from problem recognition through intervention and evaluation. Three measures of infant mortality are shown on the attached computer printout: the infmt mortality rate, the neonatal mortality rate, and the postneonatal mortali~ rate. All …