Objective: To identify the potential impact that different definitions of live births and practice patterns have on infant mortality rates in England and Wales, France, Japan, and the United States.
Methods: United States data were obtained from the 1986 linked national birth-infant death cohort, and those for the other countries came from either published sources or directly from the Ministries of Health.
Results: In 1986 in the United States, infants weighing less than 1 kg accounted for 36% of deaths (32% white and 46% black); 32% resulted from fatal congenital anomalies. These rates were much higher in both categories than in England and Wales in 1990 (24 and 22%, respectively), France in 1990 (15 and 25%, respectively), and Japan in 1991 (9% for infants weighing less than 1 kg, percentage of fatal congenital anomalies unknown). These cases are more likely to be excluded from infant mortality statistics in their countries than in the United States.
Conclusions: In 1990, the United States infant mortality rate was 9.2 per 1000 live births, ranking the United States 19th internationally. However, infant mortality provides a poor comparative measure of reproductive outcome because there are enormous regional and international differences in clinical practices and in the way live births are classified. Future international and state comparisons of reproductive health should standardize the definition of a live birth and fatal congenital anomaly, and use weight-specific fetal-infant mortality ratios and perinatal statistics.