Association between gross domestic product throughout the life course and old-age mortality across birth cohorts: parallel analyses of seven European countries, 1950-1999.

Mortality levels of national populations have often been studied in relation to levels of gross domestic product (GDP) at time of death. Following the life course perspective, we assessed whether old-age mortality levels for subsequent cohorts are differentially associated with GDP levels prevailing at different ages of the cohorts. We used all-cause and cause-specific mortality data by sex, age at death (65-99), year at death (1950-1999), and year of birth (1865-1924) for Denmark, England and Wales, Finland, France, the Netherlands, Norway, and Sweden. Trends in national GDP per capita between 1865 and 1999 were reconstructed from historical national accounts data. Through Poisson regression analyses, we determined for each country both univariate and multivariate associations across five-year birth cohorts between mortality and GDP levels prevailing at time of death, and at earlier ages of the cohorts (i.e. 0-5, 6-19, 20-49, and 50-64). For the subsequent cohorts, levels of GDP at time of death were strongly inversely associated with all-cause mortality, especially among women, and among men in England and Wales, Finland, and France. In most countries, stronger associations were observed with GDP levels prevailing at earlier ages of the cohorts. After control for GDP at time of death, these associations remained. An independent association of GDP at earlier ages of the cohort was also observed for cause-specific mortality. The associations were negative for ischaemic heart diseases, cerebrovascular diseases, and stomach cancer. They were positive for prostate cancer, breast cancer, COPD (women), and lung cancer (women). GDP prevailing at ages 20-49 (men) and ages 50-64 (women) had the largest associations with old-age mortality. These findings suggest an independent, mostly negative effect of GDP prevailing at earlier ages of subsequent cohorts on old-age mortality. Socio-economic circumstances during adulthood and middle age seem more important in determining old-age mortality trends than those during infancy or childhood.