Birth outcome measures and infant mortality at different time periods

Background: Recent studies have proposed alternatives to the traditional approach of using low birth weight as the sole means of assessing infant mortality risk and other outcomes. Nevertheless, there hasn't yet been an integrated analysis of these newer approaches.

Method: This study reviews thirteen alternative strategies applicable to vital statistics and survey data, including various combinations of birth weight, gestational age, fetal growth rate, and Apgar scores. Using the NCHS linked birth/infant death file for 2001, the study then constructs multivariate logit models, accounts for covariates including parental and child characteristics, and assesses the associations between each of the 13 key measures and three outcomes (early neonatal, late neonatal, and postneonatal mortality).

Results: All evaluated birth outcome measures are strong predictors, but Apgar scores are the strongest among all models in all time periods, independent of birth weight and gestational age. Apgar scores' predictive power is stronger for Mexicans, whites, and girls than for blacks and boys. All birth outcome measures remain significantly associated with mortality, but their predictive power reduces drastically over time, particularly for Apgar scores.

Discussion and conclusions: These results show the importance of risk factors and birth outcome measures varies by race/ethnicity, gender, and time. They also suggest a rule of thumb for predicting infant mortality odds: when Apgar score is available, it should always be included along with certain classifications combining birth weight and gestational age.

Policy implication: This study suggests that tailored prevention and education efforts could reduce mortality, especially during the postneonatal period.