20th Australian Population Association Conference: Opening Address

A tale of two pandemics in Australia: 1918-19 Influenza and COVID-19*

Print
Dr David Gruen AO
Australian Statistician
Wednesday 23 November 2022

Introduction

It is my pleasure to be invited to give the opening address at the Australian Population Association’s 2022 Conference.

While I am not a demographer, I have a keen interest in population and demography and today I want to talk about Australia's experience with two pandemics: the 1918-19 Influenza and COVID-19.

Influenza deaths in the 1918-19 panademic

  • Once the 1918-19 influenza pandemic arrived in Australia, excess deaths were almost immediate and in large numbers. Excess deaths occurred over a short period of time – mostly within the first 10 months of the flu strain arriving in Australia.
  • Comparing the number of flu deaths to the number of deaths in years preceding the flu pandemic is a simple way to estimate excess deaths.
  • The expected average number of flu deaths, based on 1911-1917 outcomes, is very low.
  • There were at least 12,000 Australian deaths from the 1918-19 flu pandemic, out of a population at the time of 5.3 million.

Weekly all cause deaths 2019-2022

  1. Dates for key events are indicative only and may differ to other sources. 
  2. Data is provisional and will change as additional death registrations are received.
  3. 2016 is the starting point for the regression when modelling the projected number of deaths. 
  4. Data includes all doctor certified deaths occurring by the end of February 2022 and registered and received by the ABS until 30 April 2022. 
  5. People who died with COVID-19 as an associated cause of death (as opposed to the underlying cause of death) are included in the without COVID-19 series.
  6. Data is for January and February is expected to be approximately 98% complete. See section on Timeliness and completeness of data for more information.

Source: Measuring Australia's excess mortality during the COVID-19 pandemic (doctor-certified deaths)

  • Focus on the episodes when the number of deaths falls outside the confidence interval.
  • This is vastly different to the ‘excess’ deaths pattern for the influenza epidemic of 1918-19. With COVID-19, deaths have occurred over a longer period with the number of deaths picking up almost two years into the pandemic.  By contrast, many flu deaths occurred three months into the 1918-19 pandemic.
  • The higher number of deaths in 2022 reflects a policy shift towards living with the pandemic, post mass vaccination of the population.
  • This is in contrast with the 1918-19 pandemic when there was no effective vaccination developed. There was even an unresolved debate about whether the pandemic was viral or bacterial in origin.
  • During the first wave lockdowns in mid-2020, deaths dropped below the expected rate. This is due to factors such as fewer accidents, and less deaths from the flu as a result of lockdowns suppressing transmission, as well as changed behaviour of the population. This is based on new data from the ABS, released on a weekly basis since June 2020, which supports understanding COVID-19 mortality.

Weekly all cause deaths, August 2021 - July 2022

  1. Data is by occurrence.
  2. Data is provisional and subject to change.
  3. Weeks are defined as seven-day periods which start on a Monday as per the ISO week date system. Refer to 'Weekly comparisons' on the methodology page of this publication for more information regarding the data in this graph.
  4. The baseline includes deaths from 2015-19 (for 2021) and from 2017-19 and 2021 (for 2022).
  5. Change in baseline was introduced

Source: Provisional Mortality Statistics, Jan - Jul 2022

  • There were 16,375 estimated excess deaths from January to July 2022. However, this is a crude estimate with no adjustments for population growth or improvements in disease treatment. The number of excess deaths for January to July 2022 is expected to be approximately 14,000-14,500 when taking these factors into account. 
  • This is to be compared with estimated flu deaths in 1918-19 of at least 12,000.

Additional information

  • The Australian population at the time of the 1918-19 flu pandemic was about 5.3 million people, while the current population is nearly 5 times larger at close to 26 million people.

Age-standardised all cause death rates

Source: ABS Deaths collection and ABS Provisional Mortality collection

  • A broader look at all causes of mortality over the past century or so shows the different impact of the two pandemics.
  • The spikes at the time of the 1918-19 influenza epidemic are noticeable.
  • In comparison, there is a small up-tick in 2021.
  • There is likely to be a larger uptick in 2022, but that is strongly influenced by age distribution of COVID-19 deaths.

Additional information

  • Age standardisation is based on the 2001 Australian population distribution.

Age distribution of COVID-19 deaths

  • There have been four COVID-19 waves in Australia. In all four waves, people aged 80-89 years had both the largest number and proportion of deaths due to COVID-19.  
  • In the second wave, almost 80% of deaths were of people aged 80 and over.
  • The Delta wave was the only wave where more than half of the deaths (53%) occurred in people younger than 80 years old. During the Omicron wave, 71% of deaths were of people aged 80 and over. This proportion has varied considerably during this wave, ranging from 60% in March 2021 to 77% in June 2021.
  • The age distribution of total deaths closely resembles the age distribution during the Omicron wave because a large majority of COVID-19 deaths occurred during this wave.
  • Death rates for Australian born were 27.8 per 100,000 people but much higher for some other groups – for example, those born in Tonga (178.9), Samoa (178.2), Syria (118.3), Romania (97.1) and Iraq (91.0).
  • Across all waves, there were more deaths among the disadvantaged. Overall, there were three times as many deaths in the lowest income quintile than in the highest income quintile.

Additional information

  • Wave 1: occurred between March and May 2020. The predominant variant during Wave 1 was the original virus strain. 
  • Wave 2: occurred between June and November 2020. Wave 2 predominantly occurred in Victoria. The variant during Wave 2 was the original virus strain. 
  • Delta wave: occurred between July and December 2021. 
  • Omicron wave: occurring during 2022 (until the end of September 2022). 

Age-specific death rates

Source: AIHW Grim Books and ABS Provisional Mortality Statistics collection

  • There are strikingly different age distributions for the two pandemics. For COVID-19, death rates are only high enough to be visible in the graph from the age of 50 onwards, with rates for those 85+ being much higher than for younger age groups.
  • Death rates are much higher during the influenza pandemic overall and across all ages other than the very oldest. In the 1918-19 pandemic, 80% of flu deaths were people aged under 50, a stark contrast with COVID-19, for which less than 2% of deaths were of people aged under 50 years. Looking at the spread of disease from the 1918-19 pandemic shows that males were at high risk. A possible explanation is that males were more likely in the working population which had a higher risk of exposure to the flu. Social conventions also led to males being in other high-risk places such as sporting events and hotels.
  • With COVID -19, most deaths were for those aged 80 years and over. The spread of disease occurred within vulnerable populations, particularly aged care residents and/or those with pre-existing conditions and disability.
  • The two pandemics had a similar sex ratio of deaths. In 1919, the sex ratio of deaths was 1.4 males for every female, compared with 1.3 males for every female during the COVID-19 pandemic.

Additional information

  • Data for age-specific death rates is updated to September 2022.
  • Death rates are as a proportion of the whole age cohort in each case, not proportions of those who caught the flu or COVID-19.

Proportion of deaths from and with COVID-19 during the Omicron wave

During the Omicron wave: 

  • The proportion of people dying with rather than of COVID-19 of all COVID-related deaths has increased over the course of 2022. Vaccination has played a part in this as those vaccinated are less like to develop COVID pneumonia.  So, death is now more likely for those who are old or vulnerable who get COVID-19. For example, the most common cause of death for those who die with COVID-19 is cancer.
  • From March 2022, over 20% of COVID-19 associated mortality are deaths where the person has died with COVID-19. 
  • Close to 30% of COVID-19 associated deaths in August were of people dying with COVID-19. 

Life expectancy: International comparison of top 10 countries, United Nations

  • Australia was one of the few countries that showed an increase in life expectancy in the first two years of the COVID-19 pandemic and has the third highest life expectancy in the world according to the United Nations’ estimates.
  • Other countries that did not see declines in life expectancy in 2020 were Norway, Denmark, Finland (for females only), New Zealand and Korea.
  • In 2022, higher death rates due to COVID-19 may lower Australia’s life expectancy.
  • Apart from 2020, the graph shows the top 10 countries for life expectancy.
  • Australia’s life expectancy at birth increased slightly – by 0.1 years for both males and females – from the average life expectancy in the years, 2018 to 2020, to the average in the years, 2019 to 2021.

Crude birth rate

Source: ABS, Historical Populations

  • The crude birth rate shows a slight increase in 2021, following the lowest rate on record in 2020, the first year of the pandemic.
  • The birth rate fell in 1919 during the influenza outbreak. This was likely partly a consequence of the number of pregnant women who died, as there is some evidence that pregnant women had high death rates from flu. There were also reports of higher numbers of stillbirths for the year for those infected with the flu.
  • In 2021, the crude birth rate was 12.1 per 1,000 women, up from 11.5 per 1,000 in 2020.
  • In 2021, the total fertility rate was 1.7 babies per woman.

Net overseas migration

  1. Estimates from 1972 are year ending June, prior to this they are year ending December. Estimates for 2020-21 are preliminary. See revision status on the methodology page.
  • Many events have shaped overseas migration in Australia.
  • Again, the two pandemics have shown a completely different pattern.
  • Recently, we have seen what we call the “COVID-19 cliff”, whereas the opposite happened during the influenza of 1919. The flu pandemic hit at the end of WWI after a prolonged period of low migration and coinciding with a massive migration influx at the end of the war. This was not stopped by the flu pandemic. Instead, quarantine was used to stop spread of disease in the community.
  • In contrast, COVID-19 hit at a time of sustained high levels of overseas migration which plummeted when stringent international border restrictions were introduced. 2021 saw negative net overseas migration for the first time since 1947.

Components of population change by capital city

Components of population change by capital city 2018-19

Components of population change by capital city
The chart displays the components of population change for each Australian capital city in 2018-19 compared to 2020-21. Data for this chart is located below under the 'Data for the components of population change by capital city' heading.
  • Popular reporting suggests there was a mass exodus from capital cities to the regions. This is partly true but is not the whole story. In 2020-21 during the height of the COVID-19 lockdowns, the smaller capitals grew while the larger ones, Sydney and Melbourne, contracted. 
  • Lockdowns and cross-border restrictions did play a role in the pattern of movements.
  • Now that Australia has moved into the next phase of “living with COVID” it will be interesting to see how the internal migration patterns play out.

Additional information

  • People living in the capitals increased by 2.5 million (17%) between 2011 and 2021.
  • Regional Australia grew by 832,000 (11%) over the same period.
  • A comparison with 1919 cannot be made as component-based regional population estimates were introduced only in 2016.

Data for the components of population change by capital city

Components of population change by capital city 2018-19
 Natural IncreaseInternal MigrationOverseas MigrationPopulation change
Sydney38,701-25,55573,91987,065
Melbourne33,8592,25277,369113,480
Brisbane17,88415,91418,78952,587
Adelaide4,914-3,91412,90013,900
Perth14,837-1,40213,97027,405
Hobart7339861,7263,445
Darwin1,530-3,213542-1,141
Canberra3,409-1943,1106,325
Components of population change by capital city 2020-21
 Natural IncreaseInternal MigrationOverseas MigrationPopulation change
Sydney36,943-34,849-7,245-5,151
Melbourne27,363-33,501-54,367-60,505
Brisbane17,39715,001-10,52821,870
Adelaide5,252-226-3,1311,895
Perth15,0836,468-5,38216,169
Hobart863-828-369-334
Darwin1,656-1,602-381-327
Canberra3,318357-3,277398

Components of population change by rest of state

Components of population change by rest of state

Components of population change by rest of state
The chart displays the components of population change by rest of state in 2018-19 compared to 2020-21. Data for this chart is located below under the 'Data for the components of population change by rest of state' heading.

This is the corollary of the previous graph – we see the heightened importance of internal migration (red section of the stacked bar) to the regions in the three largest states

Data for the components of population change by rest of state

Components of population change by rest of state 2018-19
 Natural IncreaseInternal MigrationOverseas MigrationPopulation change
NSW6,5783,49212,14022,210
Vic.3,2859,9467,22020,451
Qld11,9456,91713,38932,251
SA343-441,2411,540
WA3,209-5,0491,93191
Tas.3871,0221,3352,744
NT997-1,15822766
Components of population change by rest of state 2020-21
 Natural IncreaseInternal MigrationOverseas MigrationPopulation change
NSW7,44618,0271,30526,778
Vic.2,32615,154-1,76715,713
Qld11,95516,178-4,01524,118
SA2881,121-1851,224
WA3,387-1,117-3941,876
Tas.475719-811,113
NT989-902-861

Conclusion: Data insights and opportunities

This presentation dipped into the range of population data we have which can provide information on the impacts of the pandemic.

The ABS introduced a range of new measures during the pandemic – not just on population. Some examples are weekly payroll jobs & wages, and rapid surveys of households and businesses.

For population statistics, examples include up to date and rich analysis though provisional deaths, early indicators through provisional overseas arrival and departures data, more frequent Net Overseas Migration revisions and adjustments to the preliminary model to account for changed traveller behaviour, more accessible information on international students, and treatments to handle the impact of vaccination programs on internal migration data sourced from Medicare.

ABS also harnessed integrated assets to support policy responses to vaccination with timely regularly updated data. You’ll hear more about this on Friday from Dr Phillip Gould in his plenary. Integrated data was also used to support Census and help assure its quality, and to support population rebasing.

There are huge gains to be made through wider use of administrative data, and these continue to be realised, often as a by-product of the increasing aspects of our lives now being intermediated through digital platforms.

Thank you.
 

*  The graphs shown are those I presented at the conference. The accompanying dot points, while not a transcript of my presentation, provide a summary of the points I made at the conference.