BACKGROUND The clinical assessment and management of the risk of violence and suicide by people with mental illness may have to focus on environmental as well as individual factors.
AIMS To investigate possible associations between violence, homicide and suicide rates, population density and indices of deprivation, with particular reference to inner-city boroughs.
METHOD Coroners' statistics in London for homicide and suicide were obtained, with police-reported homicide and violence rates as a validity check. Correlations were made between these data and population density, the Jarman under-privileged areas score, and the Mental Illness Needs Index for each of the 32 London boroughs.
RESULTS Homicide rates had a 14.3-fold range, suicide a 4.4-fold range and interpersonal violence a 6.6-fold range. The variables under study were strongly correlated with each other. Rates were highest in boroughs with high population density and deprivation scores. The associations persisted when covarying for deprivation, age structure or population density.
CONCLUSIONS Because violence, homicide and suicide are so closely correlated, they are likely to be valid indices of the differences between adjacent boroughs; this has implications for the delivery of preventive and mental health services and for clinical management of risk.