Identifying and definitional attributes | |
Metadata item type: | Data Quality Statement |
---|---|
METEOR identifier: | 447953 |
Registration status: | Health, Superseded 04/12/2012 |
Data quality | |
Data quality statement summary: |
|
---|---|
Institutional environment: | BreastScreen Australia is a joint program of the Australian Government and State and Territory governments. The target age group is women aged 50–69 years. BreastScreen Australia program registers in each State and Territory are maintained by jurisdictional Program managers. Data from State and Territory registers are provided to the AIHW annually as unit record data. BreastScreen Australia is monitored annually. Results are compiled and reported at the national level by the AIHW in an annual BreastScreen Australia monitoring report. |
Timeliness: | Data available for the 2011 COAG Reform Council report is based on the two-year calendar period 1 January 2008 to 31 December 2009. Data are presented as a rate for the two-year period to reflect the recommended screening interval. |
Accessibility: | The BreastScreen Australia annual reports are available via the AIHW website where they can be downloaded free of charge. |
Interpretability: | While numbers of women screened are easy to interpret, calculation of age-standardised rates is more complex and the concept may be confusing to some users. Information on how and why age-standardised rates have been calculated and how to interpret them is available in all AIHW BreastScreen Australia monitoring reports, for example, BreastScreen Australia monitoring report 2006-2007 and 2007-2008. |
Relevance: | BreastScreen Australia registers collect information on all breast cancer screening undertaken as part of BreastScreen Australia. The use of ERP based on Census data for denominators provide the most comprehensive data coverage possible. While BreastScreen data are complete, some breast cancer screening may occur outside the program, and thus this is not a measure of all breast cancer screening in Australia. It is not possible to estimate the number of women screening outside BreastScreen Australia. The BreastScreen Australia data used to calculate this indicator are of high quality. For participation nationally, the numerator is the number of women aged 50–69 years screened in each State and Territory in 2008 and 2009, extracted from unit record data supplied by each State and Territory. The denominator is the average of the 2008 and 2009 ABS ERP for women aged 50–69 years. Caution is required when examining differences across states and territories of Australia due to the substantial differences in population, area, geographic structure, policies and other factors. For participation by Indigenous status, the numerator for Indigenous is the number of women aged 50–69 years screened in each State and Territory in 2008 and 2009 who self-reported that they were Indigenous at the time of their screen. Non-Indigenous is the number of women aged 50–69 years screened in each State and Territory in 2008 and 2009 who self reported that they were not Indigenous at the time of their screen. Women who choose not to identify as either Indigenous or non-Indigenous are classified as ‘not stated’ and are not included in either numerator. Caution is required when examining differences across Indigenous status, as some States and Territories do not allow for the ‘not stated’ category, and some Indigenous women may choose not to identify as such when presenting to a BreastScreen Australia service. Thus, some Indigenous women may be incorrectly assigned non-Indigenous status in the data presented. For participation by remoteness and socioeconomic status, the numerator is the number of women screened in 2008 and 2009 aged 50–69 years who reside in each of the remoteness and socioeconomic status categories. A postal area (POA) to statistical local area (SLA) to remoteness concordance and a POA to SLA to socioeconomic status concordance are used to allocate women screened to remoteness and socioeconomic status categories based on their postcode nationally. The denominator is the average of the 2008 and 2009 ABS ERP for women aged 50–69 years in each remoteness and socioeconomic status category, generated by applying an SLA to remoteness concordance and an SLA to socioeconomic status concordance to SLA ERP. Caution is required when examining differences across remoteness and socioeconomic status categories for several reasons. First, while the postcode of women screened is interpreted as postcode of residence, some women may supply an address other than where they reside, or their postcode may be invalid or missing. Second, because the concordances are based on the 2006 Census, the accuracy of both ASGC and SEIFA IRSD diminishes due to subsequent changes in demographics within some postcode boundaries, and some boundaries themselves may have changed over time. Third, many valid postcodes are omitted from the socioeconomic status concordance in particular, meaning that many screened women are unable to be allocated to a socioeconomic status category (the remoteness concordance contains a more comprehensive list of postcodes, but some women will still be missed). Breakdown of remoteness and socioeconomic status categories by State and Territory may introduce an additional source of inaccuracy, since screened women, once allocated a category, also need to be allocated to the State or Territory. Because some postcodes cross State and Territory boundaries, there is the potential for some women to be allocated to a State or Territory different to the one in which they reside. |
Accuracy: | This indicator is calculated on data that have been supplied to the AIHW by individual State and Territory registers. Prior to publication, the results of analyses are referred back to States and Territories for checking and clearance. Any errors found by states and territories are corrected once confirmed. Thus participation by State and Territory, based on the State or Territory in which the woman was screened, is both robust and readily verified. However, States and Territories are unable to check or verify participation by Indigenous status, participation by remoteness or participation by socioeconomic status, since their data, once supplied to the AIHW, are nationalised and thereby lose their State or Territory identity. Further, while previous reports can be used to verify these data at the national level, participation by remoteness and socioeconomic status categories has never before been disaggregated by State and Territory across all of Australia, and would thus be unable to be verified. Further, due to the very small numbers involved, disaggregation of participation by Indigenous status by State and Territory is not robust, and leads to issues around confidentiality and comparability. The number of women who choose not to identify as either Indigenous or non-Indigenous, and the number of Indigenous women who choose not to identify as Indigenous are sources of inaccuracy in the data. While the latter cannot be quantified, the former can for those States and Territories that use the ‘not stated’ category; in 2008-2009, 8556 women did not identify as either Indigenous or non-Indigenous nationally. The need to apply concordances to numerators and denominators introduces a level of inaccuracy. As COAG denominator populations for disaggregation by socioeconomic status and remoteness are based on SLA, the numerator counts for this indicator also need to be based on SLA. Since the allocation of women screened is based on the woman’s postcode, this required the generation of a POA to SLA to remoteness concordance and a POA to SLA to socioeconomic status concordance. However, postcode data are limiting. These concordances are based on 2006 boundaries and classifications, while the current data for this indicator are for 2008-2009. Overall, many postcodes may not have valid SLA-based socioeconomic status or remoteness data available, and many may have changed classification group since 2006 and be giving inaccurate information now. There are several sources of missing values. First, the state or territory may not have a postcode for all women screened, or the postcode supplied may not be valid. For those women that do have a valid postcode, many cannot be allocated to a remoteness or socioeconomic category, as their postcode may not be included in the concordances — this is a greater issue for socioeconomic status, since this concordance contains fewer postcodes than does the remoteness concordance. Further, this may affect some remoteness and socioeconomic categories more than others. Remoteness: 1079 women excluded (NSW: 221 excluded; Vic: 6 excluded; Qld: 304 excluded; WA: 462 excluded; SA: 0 excluded; Tas: 1 excluded; ACT: 24 excluded; NT: 1 excluded). No adjustments have been made to account for excluded women in the data. Women are counted only once in the two-year period 1 January 2008 to 31 December 2009, even if they were screened more than once during this period. All women screened in each State and Territory are included in order to present the most accurate national picture of breast cancer screening. No cell suppression was required for the data presented. The Estimated Resident Population data are provided by the ABS. |
Coherence: | Some of these data are published annually in Program monitoring reports prepared by the AIHW. The most recent of these reports is BreastScreen Australia monitoring report 2006-2007 and 2007-2008, published in 2010. This report includes participation by State and Territory, participation by Indigenous status, and participation by remoteness and socioeconomic status categories nationally for the two-year period 2007-2008. However, there were differences in the concordances used and data included compared to this report. Data for 2008-2009 will not be published until 2011, and while also covering participation by State and Territory, participation by Indigenous status, and participation by remoteness and socioeconomic status categories nationally, will differ from the current report due to slightly different methods (adjustments to account for missing postcodes for remoteness and socioeconomic status) and categories (socioeconomic status quintiles rather than deciles) that are considered optimal for national monitoring. Further, State and Territory participation in BreastScreen Australia monitoring reports is based on State or Territory of screen, rather than State or Territory of residence, since this is more appropriate for program monitoring. |
Source and reference attributes | |
Submitting organisation: | Australian Institute of Health and Welfare |
Relational attributes | |
Related metadata references: | Supersedes National Healthcare Agreement: P10-Breast cancer screening rates, 2010 QS Health, Superseded 08/06/2011 Has been superseded by National Healthcare Agreement: PI 10-Breast cancer screening rates, 2012 QS Health, Retired 14/01/2015 |
Indicators linked to this Data Quality statement: | National Healthcare Agreement: PI 10-Breast cancer screening rates, 2011 Health, Superseded 30/10/2011 |