Does the Southern Hemisphere Provide a Leading Flu Indicator for the North?

With the world largely returning to normal and Covid-19 precautions being eased, the flu, which has been largely silent for the last two years, is making a comeback.

Although influenza can be transmitted at any time of the year, it typically peaks in winter. This means that peak flu season occurs alternately in the Southern Hemisphere (where June, July and August are the winter months) and the Northern Hemisphere (where December, January and February are the winter months).

This hemispheric shift has recently led some experts to believe that the severity of influenza season in the Southern Hemisphere could be a leading indicator of what to expect in the Northern Hemisphere.

Melinda Wenner Moyer, a highly respected science writer, was quoted in the New York Times:

“We can look at what’s been happening in the southern hemisphere – their flu season usually ends when ours begins. And it’s been a pretty bad flu season down there, which doesn’t bode well for us.”

Katherine Wu, a PhD microbiologist, takes this line of reasoning even further in a recent article The Atlantic. The article, entitled “The Strongest Sign Americans Should Be Worried About Flu This Winter,” reports that the number of flu cases in Australia has made this season “one of the country’s worst in several years”. Which, she argues, “does not bode well for us in the North,” since “the same viruses that cause outbreaks in the South are also those that cause epidemics here as the seasons make their annual turn.”

I’m not so sure, and here’s why.

The argument contains two claims, both of which are disputed. The first claim is that the southern hemisphere actually experienced a “pretty bad flu season” in 2022, i.e. “the worst in the country in several years”. The second claim is that there is indeed a connection between the hemispheres – what we might call the Leading Indicator Hypothesis. In fact, the evidence for both claims is pretty thin.

Data from the World Health Organization’s Global Influenza Program is freely available to everyone. So we can take a look around.

It’s true that the number of laboratory-diagnosed cases in Australia in 2022 was higher than ever before.

However, more flu is only one possible explanation for this fact. An alternative explanation is that more testing was simply done. This means that the more you look for something, the more of it you will find. Influenza is a nationally notifiable disease in Australia. This means that every time a test is performed, the result must be recorded in a register. As a form of passive monitoring, this data suffers from variability in testing frequency, making it difficult to interpret. With the presence of Covid-19, it stands to reason that Australian healthcare providers may have been testing for flu more aggressively than ever in 2022.

Is there evidence for this more testing hypothesis? Yes. The red line in the figure above is the percentage of positive tests and, with the exception of the two years of Covid-19 (when there was almost no flu), this ‘positivity rate’ in 2022 was actually significantly lower than what is normal for Australia . The combination of a low positivity rate and a large number of positive tests means that the total number of tests must have been higher than usual, perhaps much higher.

A potentially more accurate measure of influenza intensity comes from the Australian Influenza-Like Illness Surveillance (also known as ILI). The figure below shows the number of ILI cases (in green) found in two ‘sentinel’ networks of GPs known as the Australian Sentinel Practices Research Network (ASPREN) and the Victorian Sentinel Practice Influenza Network (VicSPIN).

While the 2022 season was definitely tougher than 2021 and 2020, that’s not surprising. The reason is the Covid-19 pandemic. But other than that, Australia’s 2022 flu season is hardly notable and actually appears to have been mild compared to years like 2012, 2014, 2015 and 2017. I see no evidence to support the claim that Australia’s 2022 flu season is “pretty bad”.

We can use the same data to consider the second claim, ie that the number of cases in the southern hemisphere is a leading indicator of cases in the northern hemisphere.

The US also reports ILI data.

If Australia were used as a leading indicator of ILI in the Northern Hemisphere, then we should expect the number of cases to be strongly positively correlated at the peak of each season. The juxtaposition should generally be along an upward sloping diagonal. The diagram below shows that this is not the case. (In fact, the correlation is even slightly negative!)

So there is no evidence here that Australia is a leading indicator of the flu in the US. That’s not to say the US will have a mild flu season. I think we just don’t know yet.