DRAFT
which the authors attempted to overcome by using available demographic and socio-
economic variables and matching these to the Australian Household Expenditure Survey
to impute household income.
Taylor et al (2003), using a combination of survey and administrative data, separated
their analysis into women who had never screened versus women who were irregular
screeners. They found that (1) women living in high income households were more
likely to have never screened, whereas (2) those living in low income households were
more likely to be irregular screeners. However, these results may be confounded by the
fact that younger women in the cohort are more likely to have never used mammography
and may also have higher incomes due to higher labour force participation. Further, the
study only contained two income brackets; more or less than $40k annual household
income which, at the time of the study represented average household income (Australian
Bureau of Statistics, 2007).
This study extends previous work in this area. The aim of the paper was to firstly
determine which characteristics are associated with screening behaviour. It seeks to
compare the predictors of screening between those women who have never had a
screening mammogram (never screeners) and those who screen less than two-yearly
(irregular screeners) with women who screen every two years (regular screeners). In
doing so, we aim to inform potential policy responses directed at boosting breast
screening participation rates amongst women aged 50-69.
Methods
Study Data
Data from the 2002 and 2004 NSW Population Health Surveys were used for this
analysis. In NSW a Population Health Survey is conducted annually using computer-
assisted telephone interviewing (CATI) to study a random sample of NSW residents aged
16 years and over, living in households with private telephones. The survey randomly
selects one eligible person from a household. Questions covered a wide range of topics