Get access

Per capita alcohol consumption and suicide mortality in a panel of US states from 1950 to 2002


William C. Kerr PhD, Senior Scientist, Meenakshi Subbaraman MS, Research Associate, Yu Ye MA, Biostatistician. Dr William C. Kerr, Alcohol Research Group, 6475 Christie Avenue, Suite 400, Emeryville, CA 94608, USA. Tel: +1 (510) 597 3440; Fax: +1 (510) 985 6459; E-mail:


Introduction and Aims.The relationship between per capita alcohol consumption and suicide rates has been found to vary in significance and magnitude across countries. This study utilises a panel of time-series measures from the US states to estimate the effects of changes in current and lagged alcohol sales on suicide mortality risk.

Design and Methods.Generalised least squares estimation utilised 53 years of data from 48 US states or state groups to estimate relationships between total and beverage-specific alcohol consumption measures and age-standardised suicide mortality rates in first-differenced semi-logged models.

Results.An additional litre of ethanol from total alcohol sales was estimated to increase suicide rates by 2.3% in models utilising a distributed lag specification while no effect was found in models including only current alcohol consumption. A similar result is found for men, while for women both current and distributed lag measures were found to be significantly related to suicide rates with an effect of approximately 3.2% per litre from current and 5.8% per litre from the lagged measure. Beverage-specific models indicate that spirits is most closely linked with suicide risk for women while beer and wine are for men. Unemployment rates are consistently positively related to suicide rates.

Discussion and Conclusions.Results suggest that chronic effects, potentially related to alcohol abuse and dependence, are the main source of alcohol's impact on suicide rates in the USA for men and are responsible for about half of the effect for women.[Kerr WC, Subbaraman M, Ye Y. Per capita alcohol consumption and suicide mortality in a panel of US states from 1950 to 2002. Drug Alcohol Rev 2011;30:473–480]