Skip to main content
. 2017 Jan 17;17:83. doi: 10.1186/s12889-017-4022-4

Table 4.

Associations between alcohol outlet access and depressive symptoms from linear and log-binomial regression models (N = 995)

√CESD-10 Scorea At risk of depressionb (No/Yes)
Alcohol outlet exposurec β (95% CI) ORd (95% CI)
All outlets
 Number within 0.4 km
   Zero spike 0.098 (−0.108, 0.303) 1.341 (0.851, 2.114)
   Greater than zero e −0.003 (−0.013, 0.007) 0.972 (0.933, 1.013)
  Number within 3 km −0.0001 (−0.0004, 0.0002) 0.998 (0.997, 1.000)*
Off-site outlets
 Number within 0.4 km
   Zero spike 0.165 (−0.077, 0.407) 1.20 (0.70, 2.05)
   Greater than zero 0.119 (−0.059, 0.297) 0.73 (0.47, 1.15)
  Number within 3 km −0.0006 (−0.0041, 0.0030) 0.982 (0.967, 0.997)*
On-site outlets
 Number within 0.4 km
   Zero spike −0.002 (−0.009, 0.005) 1.29 (0.87, 1.92)
   Greater than zero −0.001 (−0.174, 0.172) 0.98 (0.96, 1.01)
  Number within 3 km −0.0002 (−0.0007, 0.0003) 0.998 (0.997, 0.999)**
On-site (excl. late night) outlets
 Number within 0.4 km
   Zero spike 0.105 (−0.143, 0.353) 1.64 (0.95, 2.86)
   Greater than zero 0.003 (−0.017, 0.023) 0.99 (0.94, 1.05)
  Number within 3 km −0.0003 (−0.0008, 0.0003) 0.997 (0.995, 0.999)**
Late night on-site outletsf
 Number within 3 km
   Zero spike 0.0450 (−0.1105, 0.2005) 1.360 (1.000, 1.851)
   Greater than zero −0.0002 (−0.0009, 0.0005) 0.996 (0.994, 0.998)**

*p < 0.05, **p < 0.01, ***p < 0.001. Models adjusted for education, employment, household income and urban/rural classification

aSquare root transformation of CESD-10 to deal with skewed distribution

bCESD-10 score of 10 or more classed as ‘at risk’

cSeparate models were fitted for each of the alcohol outlet access measures (i.e., number within 0.4 km and number within 3 km)

dOdds Ratio (OR) and 95% Confidence Interval (CI)

eGreater than zero is the continuous predictor for all observations greater than zero

fFew participants had a late night on-site outlet within 0.4 km so only the 3 km distance was considered for this exposure