Gender, economics and culture: diversity and the international evolution of smoking prevalence.

Aims: To examine whether the observed diversity between national patterns of smoking prevalence could require modification of the World Health Organization (WHO) linear model for an international 'smoking pandemic' (a worldwide epidemic) to address data from non-western countries.

Method: We conducted secondary research using current measures in three publicly available databases: Globalink, the International Labour Organization and the World Bank (all Internet-accessible). The measures we used are the separate percentage data for men and women on: smoking and employment and national income per capita (US$) and percentage growth per annum.

Results: Regression analysis showed that women smokers were more frequent in countries with higher national income, but women were less likely to smoke in countries of rapid growth. Men were less likely to smoke in countries with higher national income, but more likely to smoke in countries of rapid growth. Two principle components together explained 62% of all the variance in the international data. The largest factor was positively correlated with the percentage of employed females, the percentage of female smokers and national income per capita, but negatively correlated with the percentage of male smokers and percentage annual. growth. The effect of female employment was not continuous, but above a threshold of 51%, was associated with a higher prevalence of female smoking. The smaller, second factor was only weakly correlated with any smoking variables.

Conclusions: In his 1994 model (subsequently adopted by the WHO) Lopez looked at historical trends in 'stages' of smoking prevalence. These have been associated with 'stages' of economic development. We extended this analysis to look at a dynamic change (% annual growth) and a social indicator (employment). Male and female smoking is affected differentially by economic change and by level of income. These are also strongly related to the percentage of women in employment. This has implications for workplace policies on smoking.