Balancing absolute and relative risk reduction in tobacco control policy: the example of antenatal smoking in Victoria, Australia


Correspondence to:
Nathan Grills, The Nossal Institute of the University of Melbourne and Department of Health, 187 Grattan Street, Carlton, Victoria 3010. Fax: (03) 8344 9130; e-mail:


Objective: This descriptive epidemiological analysis aims to explore the benefits, risks and policy balance between a whole-of-population and high-risk reduction approach to reducing antenatal smoking prevalence.

Methods: Using Victorian hospital antenatal statistics the rate-ratio for smoking in each hypothesised high prevalence group was calculated and combined with the absolute number of births in each high-risk group. The effect on smoking prevalence of whole-of-population reductions and high-risk group reductions was then modelled.

Results: In Victoria, there were higher rates of antenatal smoking among single [RR = 4.67 (3.46–4.42)], teenage women [RR (95%CI) = 3.26 (3.00–3.54)] of indigenous ethnicity [RR = 4.39 (3.94, 4.88)] with low income [RR = 4.67 (4.17–5.22)] and low education attainment [RR = 3.89 (3.47–4.36)] who lived in less accessible areas [RR = 2.14 (1.92–2.39)]. However, as each of these high-risk groups represents a relatively small proportion of mothers, most antenatal smokers are aged 25–34, educated, city-based, non-Indigenous and non-impoverished.

Conclusions: The majority of Victorian women who smoke in pregnancy do not belong to traditional high-risk groups.

Implications: Absolute reductions in smoking prevalence in high-risk groups can potentially be achieved by whole-of-population prevalence reductions, despite a potential continuance in high relative risk among these groups. Conversely, an exclusive focus on smoking reduction in high-risk groups may fail to reduce the whole-of-population antenatal smoking prevalence.