Background: Data from BreastScreen Australia Screening and Assessment Services (SAS) for 2002-2010 wereanalysed to determine whether some SAS characteristics were more conducive that others to high screeningperformance, as indicated by high priority performance indicators and standards. Materials and
Methods:Indicators investigated related to: numbers of benign open biopsies, screen-detected invasive cancers, and intervalcancers, and wait times between screening and assessment. Multivariate Poisson regression was undertaken usingas candidate predictors of performance, SAS size (screening volume), urban or rural location, year of screening,accreditation status, and percentages of clients from culturally and linguistically diverse backgrounds, ruraland remote areas, and socio-economically disadvantaged areas.
Results: Performance standards for benignbiopsies and invasive cancer detection were uniformly met irrespective of SAS location and size. The intervalcancer standard was also met, except in 2003 when the 95% confidence interval of the rate still incorporated thenational standard. Performance indicators improved over time for: benign open biopsy for second or subsequentscreening rounds; rates of invasive breast cancer detection for second or subsequent screening rounds; andrates of small cancer detection. No differences were found over time in interval cancer rates. Interval cancerrates did not differ between non-metropolitan and metropolitan SAS, although state-wide SAS had lower rates.The standard for wait time between screening and assessment (being assessed <28 days) was mostly unmet andthis applied in particular to SAS with high percentages of culturally and linguistically diverse women in theirscreening populations.
Conclusions: Gains in performance were observed, and all performance standards weremet irrespective of SAS characteristics, except wait times to assessment. Additional descriptive data should becollected on SAS characteristics, and their associations with favourable screening performance, as these maybe important when deciding on SAS design.