DRAFT
household income was not a significant predictor amongst the never screener group,
except for women in the $20,000 to $40,000 income category whose odds of never
screening fell by 42% relative to a woman living in a household with an annual income of
less than $10,000 (p < 0.05).
The relative risk ratio for hours worked in the last week were statistically different from
unity (p < 0.05) in the irregular screeners’ group; that is, an extra hour worked in the last
week increased a woman’s odds of screening irregularly by 1%. Hours worked was not
significant in the never screened group.
None of the ARIA variables played a significant role in predicting screening behaviour.
Nevertheless, there was evidence of regional differences depending on which AHS a
woman resided in. Women living in the outer-metropolitan regions of Wentworth and
Illawarra were more likely to have never screened compared to their Northern Sydney
counterparts. Women living in the outer metropolitan regions of Wentworth, Sydney
South West and the Central Coast tended to have a higher likelihood of being an
irregularly screener.
Discussion
Participation rates based on administrative data are typically found to be lower than those
in self-reported surveys. In our case, self reported data from the NSW Health survey
indicates an overall participation rate of 73.7% compared to the 50.1% rate indicated by
administrative data. There may be several explanations for this apparent discrepancy.
First, survey data may be capturing mammograms provided under the Medicare program
in addition to those provided by Breast Screen Australia. In 2003 and 2004, Medicare
subsidised over 650,000 mammograms and it is feasible that a proportion of these may be
reported in the survey data as screening mammograms. Second, women’s recall of
whether their last screening mammogram occurred in the last two years may be imprecise
and third, respondents may be over-stating their use of mammography for screening. It is
of course possible that the administrative data are also imprecise.