Socio-economic factors and mortality among 25-64 year olds followed from 1991 to 1994: the New Zealand Census-Mortality Study.

AIM: To measure the association of income, education, occupational class, small area socio-economic deprivation, car access and labour force status with mortality among 25-64 year old males and females using the 1991 census-cohort of the New Zealand Census-Mortality Study. METHODS: Mortality records for 1991-94 were anonymously and probabilistically linked to 1991 census records, thereby creating a cohort study of all New Zealand census respondents. Odds ratios of mortality comparing categories of each socio-economic factor were calculated using logistic regression. For income, education and deprivation (NZDep91) a modified relative index of inequality (RII(10:90)) was calculated. The RII(10:90) estimates the relative risk of mortality for low socio-economic people (10th percentile rank) compared to high socio-economic people (90 percentile rank) allowing direct comparisons across socio-economic factors. RESULTS: The relative risk of all-cause mortality for 25-64 year old males with an equivalised household income less than $20,000, compared to greater than $50,000, was 2.16 (95% confidence interval 1.99 to 2.34). For females, this relative risk was 1.68 (1.52 to 1.86). Using the RII(10:90) all-cause mortality was 2.22, 1.94 and 1.58 times greater among low compared to high socio-economic males for income, NZDep91 and education, respectively. For females, these RII(10:90) estimates were 1.77, 1.69 and 1.57, respectively. By cause of death, the strongest gradients were observed for respiratory diseases, followed by lung cancer, cardiovascular disease and unintentional injury. For suicide deaths, unemployed males and females had 2.70 (1.84 to 3.95) and 2.86 (1.19 to 6.85) greater rates than the employed. CONCLUSIONS: There are strong socio-economic gradients for all-cause mortality and most specific causes of mortality among both males and female adults in New Zealand, regardless of the choice of socio-economic factor. The gradients were strongest for income, followed by small area deprivation and education, and strongest for 'preventable' causes of death.