The effect of neonatal selection on mortality in the following months

We study mortality selection in the first 90 days of life. Using micro-data for a region with internal strong mortality differences before mortality transition, we try to understand if children who survived a strong risk of death during the first three months of life (q0-2) were “selected”, i.e. they had higher probability of surviving during the following 33 months (q3-35). The neonatal selection process played a non negligible role in determining the post-neonatal mortality level in Veneto during 1816-35. For the cohorts where neonatal mortality selection was severe (q0-2>40%), the hazard ratio of surviving during the following 33 months was 20/30% lower than for the cohort where neonatal mortality selection was relatively small (q0-2<20%). This result suggests a sort of homeostatic mechanism, as the mortality differences in q0-35 are lower than the mortality differences in q0-2 and q3-35.