TECHNICAL NOTE 1 A MORE TIMELY ANNUAL COLLECTION: CHANGES TO ABS PROCESSES
Past mortality coding practices
Over the past 10 years, the ABS has released the annual Cause of Death, Australia dataset 15 months after the end of each reference period (i.e. data for the 2014 reference period was published in March 2016). The 2015 release of Causes of Death, Australia has been released 6 months earlier, representing a significant change in processing of the national mortality dataset.
The delayed publication of mortality data was implemented in 2007 to enhance the quality of data on coroner-referred deaths. Coroner-referred deaths include accidents, assaults and suicides, causes for which accurate statistics are particularly important from a public health perspective. The processing schedule allowed additional time for information on these deaths to be made available on the National Coronial Information System (NCIS), thereby enhancing the quality of initial counts of deaths, especially those from suicide.
To further mitigate issues associated with the timeliness of data availability, the ABS also implemented a revisions process for coroner-referred deaths. This revisions process enabled cases to be reassessed a further two years after initial publication, allowing further suicide deaths to be identified as cases on the NCIS were closed and additional documentation (including coronial findings) became available.
Reasons for changing processes
The ABS has carefully assessed changes to the processing and publication of mortality data. At the heart of that assessment was the balance between timely release of data to support policy, and maintaining high quality initial counts of coroner-certified deaths, including suicides. The decision to change the timing of initial processing and release of causes of death data was made based on a several factors:
- Coding practices have been enhanced to enable greater use of police, toxicology and autopsy reports on the NCIS. These changes have meant that more evidence can be sourced to code the cause of death regardless of whether a coroner has made a finding.
- Attachment rates for police, autopsy and toxicology reports on the NCIS have been improved, ensuring that the evidence required to code deaths is, in the main, available sooner.
- The cumulative effect of enhanced coding practices and faster report attachment rates on the NCIS has resulted in more accurate preliminary data, evidenced through smaller changes to data through revisions over time.
The majority of deaths are certified by a medical practitioner (approximately 87% of all deaths) using a Medical Certificate of Cause of Death. The details of doctor-certified deaths do not change over time. These deaths are currently coded shortly after they are received from the state and territory Registries of Births, Deaths and Marriages. Doctor-certified deaths are not impacted by changes to processing timelines.
Deaths which are referred to a coroner (approximately 13%) can be impacted by changes to the timing of coding. Extensive analysis of data relating to coroner-referred deaths in 2015 was conducted. Findings included:
- The number of deaths for which the cause is ill-defined or unknown (R99) increased from 1,215 in 2014 to 1,480 in 2015. Many of these are coroner-referred deaths. However, deaths coded as ill-defined or unknown have mainly been found to be from natural causes, and rarely have they later been found to be suicide deaths. Examination of those deaths coded to R99 in 2015 indicates that this continues to be the case.
- The number of deaths for which the mechanism of death is known but the intent (accident, assault, suicide) is unknown (Y10-Y34) was lower in 2015 (237) than in 2014 (280), and is in fact the lowest number for preliminary data since 2006. These deaths can be later found to be from one or other intent, but in most cases the intent can not be determined by a coroner and they remain in these categories.
- The number of deaths coded to intentional self-harm or suicide is the highest on record. There is a possibility that this number will increase more through the revisions process than counts have in recent years, but there is little evidence that this will be the case.
Analysis has shown that the data, while coded several months earlier than in previous years, is comparable with that from 2014 and prior. Further analysis of data comparability will be undertaken when data for the 2015 reference period has been revised.
Future revisions of ABS mortality data
The ABS remains committed to revisions of coroner-referred deaths, including the revision of the 2013 and 2014 reference periods. Releases of revised data for these periods will occur in early 2017*. The ABS also continues to monitor the timeliness of data availability on the NCIS, and will use this to further adapt both the initial release of data and the timing of future releases of revised data.
*NB Final data for 2013 and revised data for 2014 was released on 4 April, 2017, and is available in the 2013 and 2014 Revisions data cube of this publication. Any articles or materials published on or after 4 April, 2017 contain final 2013 and revised 2014 data. No previously-released data has been updated. As a result, small differences in published 2013 and 2014 data throughout this publication may be identified. See Causes of Death Revisions, 2013 Final Data (Technical Note) and History of Changes in this publication for further information.