Comparison of rates of suicide methods used by the elderly in England and Wales.

the following United Nations websites: household size – www.unstats.un.org/unsd/demographic/ sconcerns/housing/housing2.htm (median year for these data from different countries was 1991); density per room – www.unstats.un.org/unsd/ demographic/products/socind/housing.htm;and general population density – www.unstats.un.org/ unsd/demographic/products/dyb/dyb2006/Table03. pdf. Spearman’s correlation coefficient (ρ) was used to examine the relationship between elderly suicide rates and the household size, the general population density and the density of the number of people per room. There were significant negative correlations between the average number of people living in a household and suicide rates in males aged 65– 74 years (ρ=−0.39, P= 0.006), males aged 75+ years (ρ=−0.46, P= 0.001), females aged 65– 74 years (ρ=−0.53, P ;< 0.00001) and females aged 75+ years (ρ=−0.51, P< 0.00001) for 47 countries. There were no significant correlations between elderly suicide rates and general population density (data for 78 countries) and the density of people per household room (data for 36 countries). The negative correlation between elderly suicide rates and household size confirms the findings of a recent study (Shah, 2009), which used a different data source for household size. However, the negative findings do not allow further clarification of the relative contribution of emotional and geographical proximity explanations to elderly suicide rates. The negative findings may be the result of the well-described methdological issues in ecological studies of elderly suicides (Shah, 2009). The emotional and geographical proximity explanations are both likely to contribute simultaneously to elderly suicides rather than be “either/or” explanations. Conflict of interest