The mortality of very low birth weight infants: the benefit and relative impact of changes in population and therapeutic variables.

Objective: Very low birth weight (VLBW, <= 1500 g) infants' mortality rates have decreased markedly. We aimed to quantify the relative contribution of changes in the distribution of population characteristics and changes in specific mortality rates on the decline in mortality rates of VLBW infants. Study design: A population-based observational study of the Israel national VLBW infant database. The study population comprised singleton VLBW infants of 24-32 weeks' gestation born during the epochs 1995-2000 (n = 3728) and 2006-2010 (n = 3246). The Kitagawa methodology was applied to determine the contribution of changes in demographic and perinatal characteristics and changes in specific mortality rates on the decline in mortality between the periods. Results: During the study epochs, VLBW infant mortality rates decreased from 19.7 to 13.8%. Of the 5.9% decrease in mortality, 60.6% was attributed to the decrease in specific mortality rates and 39.4% to changes in the proportions of population characteristics and therapies, predominantly early initiation of prenatal care (8.1%), antenatal steroids (25.1%), and cesarean delivery (8.1%). For most of the demographic and perinatal categories considered the relative contribution of changes in their proportions was <3%, whereas >97% could be attributed to changes in the specific mortality rates for these characteristics. Conclusions: The decrease in preterm VLBW infant mortality was attributable predominantly to changes in variable specific mortality rates whereas changes in the proportions of demographic, perinatal risk factors, and therapies had a limited impact on VLBW infant mortality. Future assessment of determinants of VLBW infant mortality data should be dissected by discriminatory models.