The advent of Covid-19 following on so closely from the centenary of Spanish influenza has led to a renewed interest in that last great pandemic. Yet, more than 100 years after the event, there is still a wide discrepancy in the estimates of how many it killed.
Globally, estimates range from 17.4 to 100 million, even in research papers published in the last 20 years. In Australia, estimates for the number of Australians who died of Spanish flu, also vary widely, even among reputable sources.
The website of the Commonwealth Department of Health advises, ‘By the end of 1919 … around 10,000 Australians … had died of influenza’. The State Archives and Records Authority of New South Wales disagrees: ‘It is certain … that over 12,000 Australians died’.
According to historian Humphrey McQueen, ‘the figure was certainly between twelve and thirteen thousand’. The State Library of Victoria is more specific: ‘Around 12,500 … died from the highly infectious virus’.
Historian Frank Bongiorno contends, ‘Australian losses were probably about 12,000-15,000 deaths’. While yet another historian, Peter Hobbins, has written, ‘the flu left nearly 15,000 dead in under a year’, an approximation with which the National Museum of Australia and the ABC agree.
Although the range 10,000 to 15,000 is not as great in relative terms as that for global estimates, it is still very wide. Is it possible to be more specific about the number of Australians who died from Spanish flu? The answer to this question, like many in life, is both yes and no.
There are a number of reasons that calculating the true number of deaths is difficult. One problem is classification. At a federal-state conference in November 1918, the Commonwealth and state governments designated ‘pneumonic influenza’ as the name of the disease commonly referred to as Spanish flu. Yet each State collected its mortality statistics under different headings. For instance, the NSW Government Statistician tallied deaths from three categories of disease: pneumonic influenza, influenza with pneumonia, and influenza.
A related problem was diagnosis. Many of those classified as victims of pneumonic influenza had signs of other disease, especially heart disease, leaving the primary cause of death debatable. Conversely, some whose death was not attributed to pneumonic influenza nevertheless died as an indirect effect of contracting the disease.
Another problem is the time period used for calculating the numbers. Some sources, such as government yearbooks, use the calendar year 1919. Other official sources use January to September 1919, the period when the outbreak was most prevalent. While others include some months in 1918 on the basis that there was a high incidence of influenza in late 1918 that may have been pneumonic influenza.
Even if agreement could be reached today on classifications, time periods and diagnoses, the raw data to enable a precise calculation is no longer available. Some patient records might have survived in some jurisdictions, but certainly not across the whole country. And even then, how are undocumented or inadequately documented cases to be brought to account?
Statisticians use various methods to calculate the number of deaths due to an epidemic. A common method is to compare the number of deaths in a particular year with the average number in, say, the preceding five years. But that method is not fool proof. There might be other factors that elevate the number of deaths in the subject year which might be difficult to identify and eliminate. Looked at from the other direction, the comparative period might not be typical over the long term. This is a particular problem in the case of the Spanish flu. The preceding five years included the First World War when hundreds of thousands of young males were overseas. What effect did that have on the rate of death in Australia during that time? Results will differ depending on the assumptions made.
If it is impossible to calculate the precise number of deaths, is there a way to reconcile the various estimates? The answer is yes.
Helpfully, the means to do so are contained in a single contemporary source, the Official Year Book of the Commonwealth of Australia 1920 (CYB).
On p. 192 the CYB states, ‘In 1919 Australia experienced the full effect of the, world-wide epidemic, the number of deaths reaching the unprecedented figure of 11,552.’ On the next page it gives separate figures for each state showing the number of deaths due to ‘Ordinary influenza’ and ‘Pneumonic influenza’. The total for the former is 10,263 and the latter is 11,552.
On p. 1129 of the CYB, in a section entitled, ‘The Influenza Epidemic of 1918-19’, a table provides a breakdown of figures for each state showing ‘deaths and death rates from influenza in 1919’ with the total shown as 11,989. For reasons that are unclear, the individual state figures and their total differ from the tally on p. 192.
No distinction is made in that table between ordinary influenza and pneumonic influenza. However, on p. 1130 another table gives monthly totals for 1919 showing that pneumonic influenza claimed the lives of 10,493 while 1496 died from ordinary influenza. The sum of those numbers is 11,989.
But, it gets more complicated. On p. 1132 the statistician provides a table calculating excess deaths in 1918-1919 compared to previous years for each of influenza, pneumonia and heart disease. He explains that he included 1918 because the number of deaths from influenza became greater than normal from July 1918 and that he included pneumonia and heart disease because they too showed an increase which he states ‘may possibly have been the indirect effect of the influenza’. The table on p. 1132 shows the total excess deaths to be 14,528 of which 590 were in 1918 and 13,938 in 1919. He concludes:
This number was the death-tribute for the two years owing directly and indirectly to the epidemic of influenza, on the supposition indicated, viz., that the increase in deaths from pneumonia and heart disease were associated more or less directly therewith.
Thus, so far as the CYB is concerned, we have the following numbers for the death toll from Spanish flu in Australia:
· 10,263 (being deaths in 1919 from pneumonic influenza as per pp. 192-193)
· 10,493 (being deaths in 1919 from pneumonic influenza as per p. 1130)
· 11,552 (being deaths in 1919 from ordinary influenza and pneumonic influenza as per pp. 192-193)
· 11,989 (being deaths in 1919 from ordinary influenza and pneumonic influenza as per pp. 1129-1130)
· 13,938 (being excess deaths in 1919 from influenza, pneumonia and heart disease as per p. 1132)
· 14,528 (being excess deaths in 1918-1919 from influenza, pneumonia and heart disease as per p. 1132)
Using the standard convention for rounding to the nearest thousand, the CYB thus provides a range of 10,000 to 15,000 for deaths from Spanish flu, thus accommodating the commonly quoted figures exemplified by the sources cited above.
Although state-based yearbooks and statistical compilations provide different numbers to those which the CYB attributes to each state, they do not alter the outcome. The state-based figures set out in those sources total just over 12,200, yielding a rounded total of 12,000. Because each state used different classifications, that figure is no more authoritative than any of those based on the CYB.
But, having come this far in reconciling the various estimates, it should be noted that they are nevertheless an underestimate. While they may include some indigenous Australians who died of Spanish flu, they do not include all. Those living in the most remote communities avoided infection. But those who were in contact with whites but were outside the health system were particularly vulnerable. Some communities had a mortality rate approaching 50 per cent. Historian Tom Gara has calculated that ‘at least 1000 Aboriginal people throughout Australia died from influenza in 1919-1920’.
In the end, reconciling the various estimates of the number of Australians who died from Spanish flu is about comparing apples and pears, and in that regard there is a vast orchard from which you can take your pick.