Purpose � To determine whether neighborhood disorder changes an individual's risk of being injured with a firearm.

Methods � A one-year, population-based, case-control study set in Philadelphia.� Information from residents who were shot (the cases) was collected from a dozen local, state, and national data partners.� Based on this information, an age, race, and sex-matched resident was randomly selected (the controls) and asked their location at the time of each shooting.� Neighborhoods were defined as 1,801 block groups and disorder was separated into both social and physical dimensions.� Odds and 95% confidence intervals were calculated with conditional logistic regression.

Results � A total of 472 cases and 485 controls were used for analysis.� Shooting cases were located in neighborhoods with significantly higher levels of both physical (Map 1) and social disorder (Map 2) (p<0.001) and the risk of being shot increased significantly and monotonically with neighborhood disorder (p<0.001).� The most socially disordered neighborhoods were 6.96 [2.49, 19.46] times more likely and the most physically disordered neighborhoods were 3.97 [1.40, 11.26] times more likely to experience an interpersonal shooting.� The most socially disordered neighborhoods were 1.15 [0.21, 6.19] times more likely and the most physically disordered neighborhoods were 1.08 [0.17, 5.52] times more likely to experience a self-directed shooting.

Conclusions � Social disorder was more strongly associated with increased interpersonal shooting risk than physical disorder.� Similar associations were not found for self-directed shootings.� Efforts to modify neighborhoods with identifiable characteristics of social and then physical disorder will go far to prevent interpersonal shootings.