Abstract
Mortality data were collected for eight of the nine study countries in the BIOMED‐II concerted action “Alcohol Consumption and Alcohol Problems among Women in European Countries”;: Finland, France, Germany, Italy, The Netherlands, the United Kingdom (Scotland), Sweden, and Switzerland. An analysis of mortality that is limited to causes of death that are fully attributable to alcohol (DAA) results in a large underestimation of alcohol‐related mortality (ARM) rates in all study countries and especially among females. These estimates can be improved by including an analysis of selected causes of death that are indirectly attributable to alcohol. This produces geographic and time‐trend variability of ARM rates that are more congruent with alcohol drinking levels and trends. From this study, it is evident that when German data are analyzed separately for the former German Democratic Republic (GDR) and the former Federal Republic of Germany (FRG), two distinct ARM patterns result. The authors underline methodologic limits of this study and recommend procedures for a more reliable calculation of European ARM estimates.
Notes
U.O. Epidemiologia, Azienda Sanitaria di Firenze, Firenze, Italy.
Institute for Medical Informatics, Biostatistics & Epidemiology, Free University of Berlin, Berlin, Germany.
To whom correspondence and review proofs should be addressd at UO Epidemiologia, Azienda Sanitaria di Firenze, Villa Margherita, Vle Michelangelo 41, 50125 Firenze, Italy. Ph: ++39 055 6577300; Fax: + +39 055 6577533; e‐mail: [email protected].
With the collaboration of Salme Ahlstrom, Allaman Allamani, Marie Choquet, Gerhard Gmel, Béatrice Janin Jacquat, Ronald Knibbe, Ludek Kubicka, Thérèse Lecomte, Patrick Miller, Moira Plant, and Fredrik Spak.