Abstract
OBJECTIVES: We use new data and robust methods to estimate mortality levels and trends for US states and Washington DC and quantify cohort life expectancy for birth cohorts between 1941 and 2000. DESIGN: Observational retrospective cohort study. DATA: Mortality data and population estimates for cohorts born between 1941 and 2000 at the state, regional and national levels were obtained from the US Mortality Database. MAIN OUTCOME MEASURES: We estimated cohort life expectancy at birth by sex and geographical level. We used the coefficient of variation to assess levels and trends in convergence in cohort life expectancy by geographical level. RESULTS: In contrast to recent findings, estimates suggest that all states and regions experienced cohort life expectancy gains between the birth cohorts of 1941 and 2000 for both sexes, typically more than 5 years; the range included gains of 7-8 years in Oklahoma and Arkansas to over 14 years in Arizona, New Mexico and South Carolina (men). States sharply converged in the 1940s due to the South making substantial gains in under-five mortality, and since the 1960s there has been overall stability in geographical variation that is a combination of continued convergence by the South and modest divergence in the Midwest. CONCLUSIONS: Cohort life expectancy estimates across US states over the birth cohorts born between 1941 and 2000 have shown important gains in all states and regions. The lack of progress estimated in other work is not replicated; for example, recent work estimated that between 1950 and 2000, at least 15 states' female populations experienced cohort life expectancy losses; however, the use of better data and robust methods show no losses by any state. Our results also point to early life public health and medical interventions as key explanations for regional convergence in cohort life expectancy around 1950, however, the results suggest only modest additional convergence since then.