Table 1

 Rate ratios of 25–77 year old all cause mortality for current smokers compared with never smokers, adjusted for: (1) baseline of age and ethnicity; (2) baseline plus one socioeconomic factor (deprivation, income, or education); (3) baseline plus deprivation, income, and education simultaneously; and (4) model 3 plus car access, housing tenure, and labour force status

Poisson regression modelRate ratios for current compared to never smokers (95% confidence intervals)Percentage reduction of excess rate ratio (that is, rate ratio−1) compared with baseline model
The New Zealand index of small area socioeconomic deprivation (NZDep96) was used to measure neighbourhood deprivation (see Methods). Equivalised household income (five level categorical variable) was calculated by summing the incomes of all people in the household, and equivalising for economies of scale (number of adults and children in household) using the New Zealand specific Jensen index. Highest educational qualification was obtained directly from census data (nil, school, post-school). *In addition to the variables listed in the table, the fully adjusted model also includes census data on: household car access (nil, 1, ⩾2), housing tenure (freehold, rental and other), and labour force status (employed, unemployed, non-active).
Tobacco smoking: current compared with never
(1) Baseline model—adjusting for just age and ethnicity2.10 (2.01 to 2.20)2.06 (1.95 to 2.17)
(2) Adjusting additionally for just one socioeconomic factor:
    (2a) Model 1, plus neighbourhood socioeconomic deprivation1.97 (1.88 to 2.06)1.95 (1.85 to 2.06)1210
    (2b) Model 1, plus equivalised household income1.98 (1.89 to 2.07)1.98 (1.87 to 2.09)118
    (2c) Model 1, plus highest educational qualification2.05 (1.96 to 2.15)2.01 (1.90 to 2.12)55
(3) Adjusting simultaneously for deprivation, income, and education1.88 (1.80 to 1.97)1.89 (1.79 to 2.00)2016
(4) Adjusting fully for all socioeconomic factors*1.81 (1.72 to 1.89)1.86 (1.76 to 1.97)2619