Social and health care determinants of area variations in perinatal mortality

Abstract

Perinatal mortality rate (PNMR) variations between the English health areas can be resolved into two components, namely: (a) a variation in expectations (EPNMR) based upon differences in birthweight distributions, and (b) a variation in the ratio between observed and expected (i.e. weight-standardized) numbers of deaths and stillbirths (SPNMR). Data for the 90 English areas, for the years 1974–76, were related to a series of social indicators obtained from census and other sources, and a set of statistics describing the facilities, staffing and throughputs of perinatal and other health care services. Birthweight standardization accounted for more than half of the initial PNMR variance. The social descriptors accounted for a large part of the variation of the weight-standardized expectations (EPNMR). The social descriptors also accounted for a large fraction of the mortality variance which remained following standardization by birthweight (i.e. of SPNMR). Variations in the health-care descriptors showed little evidence of an effect upon either component, hi particular, the effects of the social variables upon SPNMR do not operate to any discernible extent through associated under-investment in health care services. The implications for policy formulation, and for further research, are discussed.