Background Age- andgender-specific suicide rates in England andWaleshave

Declaration of interest None. There have been marked changes in ageand gender-specific suicide rates in England and Wales over the past 50 years (Charlton et al, 1992; McClure, 2000). A feature of the recent trends has been the rise in young male suicides. An important public health concern is that birth cohorts could carry with them their increased predisposition to suicide as they age. Such cohort effects could arise because of the exposure of particular generations to factors during their development or early adult life that have a longterm impact on their risk of suicide throughout life. Sixteen years ago an analysis of suicide trends in Britain up to 1980 found no evidence for birth cohort effects (Murphy et al, 1986). This finding contrasted with similar analyses that identified such effects in both Canada and the USA (Murphy & Wetzel, 1980; Solomon & Hellon, 1980). The years covered in the earlier analysis of data for England and Wales included only the first 10 years (1970–1980) of the period over which young male suicide rates increased most rapidly. Subsequent analyses of suicides up to 1985 (Surtees & Duffy, 1989) and 1990 (Charlton et al, 1992) found some evidence of birth cohort effects in younger males and females. In males, each successive 10-year birth cohort, born in the post-war years, experienced higher suicide rates at all ages; in contrast, each successive cohort of females appeared to experience lower suicide rates. Here we update and refine the earlier assessments of cohort effects in England and Wales to investigate evidence for cohort effects on suicide between 1950 and 1999. As suicide rates in the post-war years have been influenced by changes in the lethality of commonly used methods of suicide and changes in the coding of suicides (period effects: Kreitman, 1976; Charlton et al, 1992; Gunnell et al, 2000), we assessed the effects of both these factors on apparent trends. METHOD