Skip to main content
. 2016 Feb 22;26(2):141–148. doi: 10.1136/tobaccocontrol-2015-052724

Table 2.

Results of the segmented regression analysis for current smoking presenting the trend in smoking previous to the introduction of the PoS display ban, the immediate step level change in smoking prevalence, and the change in the smoking trend after the PoS ban was introduced compared with the preban period trend

Trend pre-PoS display ban
Step level change
Change in trend
Percentage change 95% CI p Value Percentage change 95% CI p Value Percentage change 95% CI p Value
Log binomial models
 Model 0: unadjusted 0.15 −1.04 to 1.63 0.846 −2.30 −6.19 to 1.76 0.263 0.53 0.73 to0.33 <0.001
 Model 1: +sociodemographics* −0.93 −2.31 to 0.47 0.190 −2.25 −5.96 to 1.61 0.249 0.21 0.40 to0.02 0.029
 Model 2: +e-cigarettes 1.68 3.19 to0.14 0.032 4.54 8.66 to0.22 0.039 0.41 0.67 to0.15 0.002
 Model 3: +seasonality 2.07 3.59 to0.52 0.009 −3.69 −7.94 to 0.75 0.102 0.46 0.72 to0.20 0.001
Poisson models
 Model 0: unadjusted 0.14 −2.04 to 2.37 0.902 −1.94 −7.77 to 4.26 0.531 0.53 0.82 to0.24 <0.001
 Model 1: adjusted for seasonality −0.21 −2.34 to 1.96 0.849 −0.49 −6.35 to 5.73 0.874 0.56 0.84 to0.27 <0.001
 Model 2: +autocorrelation −0.23 −2.21 to 1.80 0.826 −0.46 −5.95 to 5.34 0.873 0.56 0.82 to0.29 <0.001

Significant associations are indicated in bold (p<0.05).

*Age, gender and social grade.

PoS, point of sale.