Najman's commentary  concerning the Lindström et al. paper on economic stress and tobacco smoking  presents valuable insights regarding present-day tobacco prevention. The main conclusion is that broad and population-based tobacco control policies can reduce smoking consumption rates substantially in the general population, while risk factor-based interventions, for example towards socially defined population subgroups at increased risk, may be relevant only to a small minority of the remaining smokers in countries with low rates of tobacco consumption. Population-based preventive approaches are thus preferable compared to risk factor-orientated approaches in such countries, according to Najman .
Although Najman's conclusion concerning countries with low rates of tobacco smoking such as Sweden is correct in principle, other aspects of prevention and policy need to be considered. One main public health objective in Sweden and world-wide is to reduce the overall disease incidence and mortality in the population in order to improve population health. However, a complementary public health policy objective is to reduce socio-economic differences in health in the population . Socio-economic status (SES) differences in tobacco smoking have increased since tobacco smoking rates started to decline in Sweden among men in the 1970s and among younger women in the 1980s. The socio-economic gradient in tobacco smoking in the 2008 study  ranges from 9.8% of higher non-manual employees to 26.3% of unskilled manual workers among men and from 12.3% of higher non-manual employees to 28.7% of unskilled manual workers among women. Groups outside the workforce, such as the early retired, the unemployed and those on long-term sick leave, mainly have even higher rates . Despite the fact that all the tobacco control policy measures mentioned by Najman  have been implemented in Sweden, together with additional policies such as a total ban on tobacco smoking in restaurants in 2005, substantial SES differences persist and currently constitute a substantial contributor to, for example, SES differences in life expectancy in Sweden . This is one reason why economic and social risk factors in early life as well as in adult life should be considered.
The second reason for considering risk factors such as economic stress and SES pertains to the fact that if tobacco smoking has become relatively more concentrated in risk groups defined according to risk factors such as low SES or economic stress, then the tobacco smokers within these risk groups will no longer be a ‘small’ minority of all tobacco smokers. This is illustrated by the fact that the unskilled manual workers, the early retired, the unemployed and those on long-term sick leave constitute 18.9% of the total male population but 33.0% of the subpopulation of male tobacco smokers (from data in reference ).
In conclusion, risk factor-based interventions and population-based smoking policy initiatives should not be seen as mutually opposing but instead as complementary. The Swedish experience suggests that population-based smoking policy initiatives will substantially decrease rates of tobacco smoking in the general population. However, it also suggests that risk factor-based interventions are needed to decrease SES differences.
Declaration of interests