The changing relationship between age and suicide rates: cohort effect, period effect or both?

Abstract

Massive changes in suicide rates over time have been recognized in the United States. An attempt has been made to describe these changes with age-period-cohort analyses. A variety of approaches has led us to conclude that suicide rates of non-white males, white and non-white females can be described adequately without a cohort effect. Recent suicide trends lead to the conclusion that a model based on a rising rate in more recently born white male cohorts coupled with an independent age effect could be rejected. If a cohort effect is postulated for more recent birth cohorts, it would require that the cohort suicide rate is decreasing with each successive birth cohort. Models based on high suicide rates in recent cohorts and additive age effects are probably misleading for future predictions. An association was noted between recent changes in the teenage and young adult suicide rates and rates of depression. Both may be the product of similar social influences.