Age, period of death and birth cohort effects on suicide mortality in Italy, 1955-1979.

Trends in death rates from all suicides and specific methods of suicide in Italy over the period 1955-79 were analysed on the basis of age-specific and age-standardised rates, and through a log-linear Poisson model to isolate the effects of age, birth cohort and calendar period. In both sexes, a large decrease in mortality from suicide was evident in the late 1950's and early 1960's. Thereafter, death certification rates showed fluctuating trends up to the mid 1970's, when steady increases became evident for both sexes. Nevertheless, overall age-standardized mortality rates from suicide in the late 1970's were still considerably lower than in the two previous decades (15% in males, 7% in females). The observed variations in suicide mortality, mostly in males, can be explained in terms of period of death effect and be related to changes in the Italian economic situation. This view finds further support from analysis of age-specific trends (e.g. mortality rates in the younger age groups started rising in the early 1970's, together with a rise in unemployment rates among the young). Cohort curves for males born in the current century were U-shaped as well, with marked declines for generations born between 1905 and 1930, and moderate increases for more recent cohorts. For females, the cohort curve was remarkably flat. Some of the changes in the various methods of suicide can be explained in terms of well-defined exogenous factors (e.g., the large fall in poisoning by domestic gas is obviously attributable to domestic gas detoxification).