Matias Gay
The data support a decline in the proportional share of recorded suicide deaths at ages 15-24, not a demonstrated decline in youth suicide risk. Other apparent redistributions were less certain after country-clustered compositional analysis and may reflect demographic structure, reporting coverage, or changes in other age groups.
OBJECTIVE: This study described historical changes in the proportional age distribution of recorded suicide deaths and variation by sex and broad region, without interpreting death shares as age-specific risk.
METHOD: A longitudinal ecological analysis used 9,396 country-year-sex records from 118 countries during 1950-2020. Suicide deaths were grouped as ages 0-14, 15-24, 25-44, 45-64, and 65 years and older. A grouped multinomial model analyzed the five-part composition, and 1,000 stratified country-cluster bootstrap samples preserved complete national time series for 95% confidence intervals (CIs).
RESULTS: The standardized share at ages 15-24 decreased from 16.1% in 1950 to 11.3% in 2020 (-4.8 percentage points, 95% CI [-14.4, -2.3]). Point estimates increased at ages 25-44 (+3.5 points) and 65 years and older (+1.4 points), but both CIs included no change. Compositions differed by sex and region, with substantial uncertainty where few countries reported.
CONCLUSIONS: The data support a decline in the proportional share of recorded suicide deaths at ages 15-24, not a demonstrated decline in youth suicide risk. Other apparent redistributions were less certain after country-clustered compositional analysis and may reflect demographic structure, reporting coverage, or changes in other age groups.