Violence reduction, revisiting a public health approach

Abstract

Objectives

Propose and test an alternative method for identifying population targets for public health model derived violence reduction programmes.

Study design

Quantitative and qualitative focused on neighbourhoods in the 75th percentile of violence or higher, using Lower Super Output Areas (LSOAs) as the data collection/integration focus.

Methods

Cluster analysis to group similar LSOAs together. Significant factors in the regression analysis plus violence rates were entered into a k-means cluster analysis creating five groupings and a short list to include in the quantitative and qualitative arms.

Results

Local Government Area (LGA) or a city-wide perspective, masks locations with high violence rates and misses potential solutions. Crime and violence are more prevalent in areas with high deprivation, poor design of housing and space associated with isolation and fear. Less considered violence vectors were identified.

Conclusions

Deprivation was the strongest predictor of violence, but there are highly deprived areas that do not have high levels of violence and conversely high-violence neighbourhoods in non-deprived areas. A granular and dynamic understanding of these patterns should form the basis of future investment and intervention efforts.