Key points are not available for this paper at this time.
Individual level risk factors for violence have been widely studied, but little is known about country-level determinants, particularly in low and middle-income countries. We hypothesized that income inequality, through its detrimental effects on social cohesion, would be related to an increase in violence worldwide, and in low and middle-income countries in particular. We examined country-level associations of violence with socio-economic and health-related factors, using crime statistics from the United Nations Office on Drugs and Crime, and indicators from the Human Development Report published by the United Nations Development Programme. Using regression models, we measured relationships between country-level factors (age, education, measures of income, health expenditure, and alcohol consumption) and four violent outcomes (including measures of violence-related mortality and morbidity) in up to 169 countries. We stratified our analyses comparing high with low and middle-income countries, and analysed longitudinal data on homicide and income inequality in high-income countries. In low and middle-income countries, income inequality was related to homicide, robbery, and self-reported assault (all p's < 0.05). In high-income countries, urbanicity was significantly associated with official assault (p = 0.002, β = 0.716) and robbery (p = 0.011, β = 0.587) rates; income inequality was related to homicide (p = 0.006, β = 0.670) and self-reported assault (p = 0.020, β = 0.563), and longitudinally with homicide (p = 0.021). Worldwide, alcohol consumption was associated with self-reported assault rates (p < 0.001, β = 0.369) suggesting public policy interventions reducing alcohol consumption may contribute to reducing violence rates. Our main finding was that income inequality was related to violence in low and middle-income countries. Public health should advocate for global action to moderate income inequality to reduce the global health burden of violence.
Wolf et al. (Sat,) studied this question.