Smoking and mortality from 1981 to 2011 in New Zealand: effect modification and changing contribution to inequalities over time


Tony is an epidemiologist. His research covers a range of topic areas intersecting with methodological advancements. He has appointments at both the Department of Public Health, University of Otago, Wellington & School of Population and Global Health, University of Melbourne

Since 2010, he has directed the HRC-funded Burden of Disease Epidemiology, Equity and Cost-Effectiveness Programme (BODE³). This programme builds infrastructure (e.g. linked routine datasets) and capacity (e.g. epidemiological and economic decision modelling) to rapidly assess the health impact and cost effectiveness of a range of preventative and cancer control interventions – and examine their equity impacts.

Tony initiated and implemented the New Zealand Census-Mortality Study (NZCMS) in the late 1990s, a pioneering study linking the national censuses with mortality data to allow monitoring and research on ethnic and socio-economic inequalities and the contribution of smoking to mortality (the NZ census periodically includes smoking).

Cutting across all Tony’s research is a strong focus and interest in epidemiological and quantitative research methodologies (e.g. causal mediation analysis, simulation modelling).


There is now 30 years of linked census-mortality data for cohort analyses in New Zealand, with smoking on many of the censuses.  In this presentation the large variations in rate ratios of the smoking-mortality association by ethnic group and over time will be presented.  This ‘effect modification’ is of interest in and of itself, but it also has impacts on how one estimates changing ‘mediation’ of ethnic-mortality gradients over time.  Causal mediation analyses, using a potential outcomes approach and weighted multiple mediators, will conclude the presentation, with estimates of both: how much of the ethnic-mortality association is additionally mediated by smoking over and above socio-economic position (‘natural effects’); and how much ethnic inequalities might have reduce had smoking been – counterfactually – removed from New Zealand (‘controlled effects’).  This presentation will be of interest to those interested in health inequalities, smoking and epidemiological methods

Hosted by the Director's Seminar Series & NCEPH Seminar