Background. Geographic variations in the incidence of deliberate self-harm (DSH) and suicide have been shown to be associated with area-based measures of socio-economic deprivation and social fragmentation. Previous studies have been subject to methodological limitations. None has investigated whether ecological associations are reflected in characteristics of individuals involved in suicidal behaviour.

Methods. DSH patients presenting to a general hospital between 1985 and 1995 and suicides (including open verdicts) from the same catchment area were studied. Mean annual rates of DSH and suicide by gender were calculated for electoral wards. The wards were amalgamated into 20 groups according to their ranking for socio-economic deprivation (Townsend) and social fragmentation scores. Associations of these variables with DSH and suicide rates were investigated. Characteristics of DSH patients living in ward groups with the highest and lowest socio-economic deprivation and social fragmentation scores were compared.

Results. Socio-economic deprivation was associated with DSH rates among males (r = 0·89) and females (r = 0·87). After controlling for social fragmentation the associations remained relatively strong, particularly in young males. Associations with social fragmentation in both genders (males, r = 0·83; females, r = 0·86) were attenuated after controlling for socio-economic deprivation. For suicide, the only significant association was with socio-economic deprivation in males (r = 0·79), but this was attenuated after controlling for social fragmentation. The characteristics of individual DSH patients reflected those of the areas where they lived.

Conclusions. Reducing socio-economic deprivation and its associated problems may be an important strategy in the prevention of suicidal behaviour, especially in young men.