Violence interruption programmes are a common type of community violence intervention (CVI), but there is no empirical evidence on programme impacts on youth violence specifically. This study evaluates the effectiveness of Safe Streets, a violence interruption CVI programme in Baltimore, Maryland, in reducing homicides and non-fatal shootings (NFS) among youth ages 15–24.
The synthetic control method was used to assess site level and programme level average treatment effects for 11 Safe Streets sites between 2007 and 2023. Models compared smoothed monthly rates of youth homicides and NFS in treated areas with their synthetic controls. Models were estimated using multiple inference approaches, treatment times and control pool eligibility definitions.
Safe Streets overall was associated with a 42% reduction in youth homicides and a 21% reduction in youth NFS, though neither estimate was statistically significant. Uncensored site-specific models estimate large reductions in youth homicides in five sites and youth NFS in seven sites and large increases in youth homicides in two sites and youth NFS in two sites. No site-specific estimates were statistically significant using primary inference methods.
The rarity of outcomes and sensitivity to modelling decisions warrants cautious interpretation of findings. Differences in programme effectiveness across sites suggest opportunities to strengthen youth engagement in some sites. Changing youth violence dynamics underscore a need to augment and adapt CVI models to match these cultural shifts. Continued work to integrate Safe Streets with other violence prevention approaches, support implementation fidelity and invest in structural change may improve programme outcomes further.