Table 2

Univariate and multivariate results for negative binomial regression modelling of neighbourhood pedestrian injury risk by census block group (n=653)

VariableUnadjustedAdjusted*
RR95% CIP valuesRR95% CIP values
Alcohol outlet count (min: 0, max: 32)1.2111.157 to 1.273<0.0011.1421.099 to 1.192<0.001
Population density (per square mile in 1000 residents) (min: 0, max: 95.16)0.9850.975 to 0.9950.01220.9790.970 to 0.989<0.001
Percent of all lots that are vacant (%) (min: 0%, max: 49.53%)1.0201.010 to 1.030<0.0011.0161.007 to 1.0260.003
Median household income (in $1000s) (min: 0, max: 224.43)0.9920.988 to 0.996<0.0010.9910.988 to 0.995<0.001
Mean daily traffic volume (in 1000 vehicles) (spatial lag of traffic) (min: 0.07, max: 33.34)1.1001.065 to 1.138<0.0011.0761.059 to 1.126<0.001
Walk Score (range: 0–100)1.0281.022 to 1.035<0.0011.0181.011 to 1.025<0.001
Physical disorder (range: 0–12)1.0450.995 to 1.0960.144
Social activity (range: 0–6)0.9850.814 to 1.1920.893
Roadway infrastructure (range: 0–8)0.8900.524 to 1.5130.713
Intersection infrastructure (range: 0–21)1.1741.122 to 1.230<0.001
Distance from downtown (miles) (min: 0, max: 7.5)0.7920.749 to 0.837<0.001
  • Final model fit statistics: AIC=1898; Residual Moran’s I=0.0225, P=0.137.

  • *Adjusted for other covariates in the column.

  • AIC, Akaike’s information criterion.