Perinatal deaths

The scope of all ABS vital statistics datasets (births, deaths, marriages) is based on registered life events. This scope has also applied to stillbirths (fetal deaths). However, not all occurrences of stillbirths are registered in Australia.

To register a stillbirth, the state and territory Registries of Births, Deaths and Marriages (RBDMs) require a birth notification from the hospital and a birth registration statement completed by the family. RBDMs also require a completed Medical Certificate of Cause of Perinatal Death (MCCPD). There are differences in legislation and policy across jurisdictions as to how stillbirths are registered. For example, a stillbirth resulting from a termination of pregnancy is not required to be registered in some states and territories. This can alter what specific data the ABS receives from each jurisdiction.

Additionally, the registration may not be finalised. The RBDMs recognise this is a sensitive and challenging time for families.

Comparison with the AIHW National Perinatal Mortality Data Collection

An alternative official source of stillbirths data is the National Perinatal Mortality Data Collection (NPMDC) held by the Australian Institute of Health and Welfare (AIHW). The AIHW publishes this as part of the Australia’s Mothers and Babies report. Data is sourced from birth notification forms completed by midwives and other birth attendants with input from family and other informants, rather than registrations. 

Comparisons with NPMDC numbers confirm that the number of stillbirth registrations do not align with occurrences of stillbirths. The table below shows the level that registration data is lower than birth notification data at the national level. The level of difference varies by jurisdiction.

National and international reporting based on registration data can be potentially misleading due to registrations not accurately reflecting stillbirth occurrence.

Potential alternative data sources

The ABS has been working with the RBDMs to determine the availability and suitability of alternative data sources to supplement the ABS stillbirths collection. Two alternative sources are:

Birth notifications

State and territory RBDM Acts require that the Registrar be notified of a birth (which includes stillbirths) by a responsible person (e.g. the CEO of the hospital where the child was born or was brought to after birth, a medical practitioner or midwife responsible for the professional care of the mother at birth, or any other person in attendance at the birth if no medical practitioner was present). These notifications are required to be submitted within a specified period ranging from 48 hours to 1 month, depending on the state or territory. They provide documentation that a stillbirth has occurred regardless of whether a registration is completed. 

Medical certificates of cause of perinatal death

States and territories also require that a Medical Certificate of Cause of Perinatal Death (MCCPD) be completed following a stillbirth by a medical practitioner who was responsible for the professional care of the birth mother at the birth, or who examined the body of the stillborn child. Depending on the jurisdiction, the completed MCCPD must then be submitted to the Registrar, together with the birth notification, by the medical practitioner and/or to the funeral director (who then must provide it to the Registrar). These notifications are required to be submitted within a specified period ranging from 48 hours to 10 days, depending on the jurisdiction.

Both sources represent a more accurate count of stillbirths occurring in Australia as they are not dependent on registration. MCCPDs contain information on the cause of death that allow for mortality coding to the International Classification of Diseases (ICD). 

Decision to change scope of ABS stillbirths collection

To address the known quality concerns with the ABS stillbirths collection, the ABS has committed to making changes to ensure it represents a more accurate count of stillbirths that occur. To this end, the scope of the ABS stillbirths collection will be expanded to include alternative sources of data to supplement the birth and death registrations that constitute the current scope of the collection. 

The ABS will work with each state and territory RBDM to determine if alternative data sources are available, assess the quality of these sources and determine the best way to supplement the collection. 

The ABS will be releasing improvements to the collection iteratively. For the first phase, the ABS has worked with the Victorian RBDM, as Victoria is a jurisdiction with one of the largest undercounts of stillbirths when tabulated by registration. The following sections detail the outcomes of the first phase of improvements.

Change of data source for Victorian stillbirths data

The ABS worked with the Victorian RBDM to investigate alternative sources to improve reporting of stillbirths data for Victoria. The Victorian RBDM provided expert advice on relevant legislation and policy specific to how stillbirths are reported in Victoria. It also gave insight into how stillbirths data is received and processed at the RBDM prior to receipt by the ABS.

After initial analysis of all data sources provided by the Victorian RBDM, the decision was made to use the Medical Certificate of Cause of Perinatal Death (MCCPD) as the primary source of data for the ABS stillbirths collection for Victoria. 

Stillbirth registrations are still received by the ABS and will be used for reconciliation purposes and to supplement information provided on the corresponding MCCPDs.

Counts of stillbirths for Victoria will now be based on MCCPDs rather than registrations. Data has been revised back to 2020. This section provides a summary of the impact of moving to MCCPDs as the primary source of data for Victorian stillbirths.

 The number of stillbirths in Victoria increases substantially when the scope of the collection is based on MCCPDs. 

• Using MCCPDs produces higher counts than using registrations alone. Proportionally, this count is between 54.3% and 86.6% higher than using registrations alone for 2020-2024.
• The fetal death rate for Victoria is also higher when the scope of the collection is based on MCCPDs, with increases of between 2.9 and 4.4 deaths per 1,000 total births compared to registrations for 2020-2024.

This increase in stillbirth numbers also means that ABS stillbirths data for Victoria is closer to the NPMDC figures when the scope of the collection is expanded to include MCCPDs. 

The table below shows:

• The difference between ABS and NPMDC numbers is significantly lower when MCCPDs are included in the scope, compared to using registrations alone.
• Between 2020 and 2023, ABS figures using stillbirth registrations were between 39.8% and 49.7% lower than NPMDC stillbirth numbers.
• ABS figures using MCCPDs were only 0.5% to 1.7% lower than NPMDC stillbirth numbers.
• Some small differences still exist between the two datasets. This is expected as they are compiled from different sources.  

Impact of change of scope on national stillbirths data

The change of scope for compiling Victorian stillbirths also has a noticeable effect on the Australia level stillbirth counts and associated fetal death rates.

• The inclusion of MCCPDs increases the count of stillbirths occurring in Australia. Proportionally this increase is less than for Victoria alone, with an increase of between 12.4% and 18.5% for 2020-2024.
• Fetal death rates for Australia are also higher when MCCPDs are included in the scope, with increases of between 0.7 and 1.2 deaths per 1,000 births across the same period.

While expanding the scope of the ABS stillbirths collection to include MCCPDs has improved counts of stillbirth occurrences, there are other impacts on quality to take into consideration. As the MCCPDs are a new dataset the ABS undertook data quality checks (in addition to usual processes) to ensure the information would address the existing issues in the registration dataset. Details are outlined in the following sections. 

Processing MCCPDs

In addition to the quality checks performed by the ABS on registration data, the following were applied to the MCCPDs:

• Scoping rules were applied to ensure only records within the gestation and birth weight limits were included in the dataset.
• Duplicate checks to ensure all MCCPDs were unique and not double counted.
• Consistency checks to ensure the data items within each record were coherent with one another and resolved any contradictory information.

Characteristics of unregistered stillbirths

As part of the move to basing stillbirth counts on MCCPDs, the ABS undertook analysis of the unregistered stillbirths. This is valuable in understanding the characteristics of records that previously would not have been a part of ABS reporting but are now included in the collection due to the change of scope. 

For the period 2020-2024, around 43% of stillbirths that occurred in Victoria have not been registered. Compared to stillbirths that have been registered, unregistered stillbirths were more likely to:

• Be of a lower gestational age. Unregistered stillbirths had higher proportions in the lower gestation period categories compared to registered stillbirths.
• Unregistered stillbirths had a higher proportion where termination of pregnancy was mentioned in the cause of death information provided on the MCCPD.

Other quality considerations 

• Timeliness. MCCPDs are required to be completed and submitted to the Registrar within a short period after a stillbirth, ranging from 48 hours to 10 days, depending on the jurisdiction. Therefore their receipt by the RBDM is more timely than the time it takes to complete a registration.
• Sex. MCCPDs for stillbirths are more likely to not have a sex specified compared to stillbirths that have been registered. The move to using MCCPD data as the source of information for Victorian stillbirths has impacted on the number of stillbirths where the sex has not been specified as male or female at the Victorian and Australian levels. The reason for the high levels of unspecified sex is not well understood but is likely influenced by administrative factors. The ABS will continue to work with the Victorian Registry to better understand the quality of the sex information on the MCCPDs.
• Aboriginal and Torres Strait Islander origin. Stillbirths data by Aboriginal and Torres Strait Islander origin has not previously been published for Victoria due to quality concerns. This data has been published for the first time for the period 2020-2024 (see Table 7 in the data cube). The derivation of this data is based on the status of the Mother and Father as provided on the MCCPD only.

Stillbirths data presented in this publication for Victoria is based on MCCPDs rather than registrations. This change in scope represents a break in time series for stillbirths and perinatal deaths data at the Victorian and Australian levels. It is recommended that previously published data for Victoria and Australia be used with caution, and not used for comparison purposes.

Other things to note when interpreting data in this publication:

• Stillbirths data for all jurisdictions other than Victoria is still based on registrations.
• Where data is presented by year of registration, for MCCPD data this corresponds to the year in which the MCCPD was received at the Victorian RBDM.
• Data presented in the Perinatal deaths data cube available in Data Downloads is presented by state and territory of usual residence. There is a small number of stillbirths where the MCCPD was received at the Victorian RBDM and the mother's usual residence was a state or territory other than Victoria, most notably New South Wales.
• Cause of death data presented for Victorian stillbirths in Table 5 in the Perinatal deaths data cube is based on cause of death information received in the stillbirth registrations for the period 2020-2024. Cause of death information available in the additional stillbirths received via the MCCPDs for that period have not been coded.  

With the first phase of improvements now implemented, the ABS will continue to investigate other changes to further improve the collection. Specifically for Victoria, future improvements will include a system of linking MCCPDs to registrations. This will allow for the most complete and up-to-date information to be available for demographic data items like Aboriginal and Torres Strait Islander origin and sex, as well as geographic data. 

Longer term, the ABS will continue working with other jurisdictions where the number of stillbirth registrations does not align closely with stillbirth occurrences to determine if suitable alternative data sources exist to supplement the ABS collection and improve reporting of stillbirths data.