AAP FactCheck Investigation:
Is the fatality rate for COVID-19 only marginally worse than the seasonal flu?
“The Ioannidis study shows the death rate (for COVID-19) is only very marginally worse than the standard flu viruses that kill hundreds of sick and elderly New Zealanders every year.”
Simon Thornley, senior lecturer in epidemiology, University of Auckland. September 6, 2020.
A group calling for New Zealand to relax restrictions to prevent the spread of COVID-19 claims the virus is only marginally worse than the seasonal flu.
The assertion was made by COVID Plan B, a group of six academics including three health science lecturers. The group has called for an end to COVID-19 lockdowns, arguing the risks of the virus have been overstated and lockdowns are more harmful than the virus itself.
During an interview on TVNZ’s Q+A with Jack Tame on September 6, group representative and University of Auckland epidemiology lecturer Simon Thornley said New Zealand’s COVID-19 response was driven by fear and panic and was unwarranted and irrational. (Video mark 1min).
“The Ioannidis study shows the death rate is only very marginally worse than the standard flu viruses that kill hundreds of sick and elderly New Zealanders every year,” Dr Thornley said.
“The response should be a measured one, like we’ve planned for lower mortality pandemics, not lockdowns.
“We’ve sacrificed our humanity, our society and our economy for the wrong virus.”
AAP FactCheck examined Dr Thornley’s statement that the death rate for COVID-19 is only marginally worse than the seasonal flu.
Dr Thornley said his statement was based on a July 14, 2020 report by Stanford University epidemiology professor John P.A Ioannidis.
The non-peer reviewed report provides a meta analysis of 36 studies which looked at the prevalence of COVID-19 antibodies in populations to determine the portion that had been infected by the virus.
This can be used to more accurately calculate the proportion of people who die after catching the virus, which is called the infection fatality ratio (IFR).
Dr Ioannidis found the studies had an estimated IFR ranging from 0 per cent to 1.63 per cent, and a median of 0.27 per cent. (Page 2)
Dr Thornley also directed AAP Factcheck to a study which looked at the prevalence of COVID-19 in the state of Indiana and estimated the IFR was 0.26 per cent. However, that study excluded deaths in nursing home residents which accounted for 54.9 per cent of COVID-19 deaths in the state.
Dr Thornley said the IFR for seasonal flu varied, “but are usually quoted at less than 0.5 per cent” and pointed to a 2009 study that said 0.5 per cent was “similar to the upper range” for the fatality rate for the flu. The report cites a webpage from the US Centers for Disease Control and Prevention (CDC) on influenza as the source of this figure, however this webpage does not currently include flu fatality rates, and the page archives for 2009 also do not show the flu IFR.
A flu IFR of 0.5 would be similar to the upper range for COVID-19. However, other research has put the IFR for COVID-19 significantly higher, and the IFR for flu significantly lower.
Another non-peer reviewed meta-analysis of COVID-19 infection fatality ratios looked at 24 studies and estimated the IFR at 0.68 per cent – nearly triple that found by Ioannidis. (Page 2)
The authors noted a high level of variation between different studies and said their IFR estimate was likely an underestimate. (Page 3).
The World Health Organisation (WHO) has found estimates for the IFR for COVID-19 “converging at approximately 0.5 – 1 per cent” based on three studies – two in Switzerland and one in Stockholm, Sweden.
By contrast, the WHO says “mortality is usually well below 0.1 per cent” for seasonal influenza, which suggests the fatality rate of COVID-19 is five to 10 times higher than for the flu.
Between July 10 and September 9, website archives show the CDC officially estimated the IFR for COVID-19 at 0.0065 deaths per infection, or 0.65 per cent (Table 1). These figures were replaced on September 10 with a breakdown of IFR estimates by age range.
Overall CDC figures for the number of flu deaths each year suggest about five times more people have already died from COVID-19 in the US than die from flu in an average year.
Between 2010/11 and 2018/19, the number of people in the US who died from the flu each year ranged from 12,000 to 61,000 with an average of 37,462 flu deaths, while the number who had flu symptoms ranged from 9.3m to 45m, for an average of 28.6m (table 1).
Using the average figures, which do not include asymptomatic cases, this would put a crude IFR at 0.13 per cent. However one meta-analysis found between 4 and 85 per cent of people with flu are asymptomatic.
By contrast, as of September 10, the US had already recorded 188,608 COVID-19 deaths.
University of Otago public health professor Nick Wilson has published multiple papers on pandemic influenza in New Zealand and said the IFR for COVID-19 was significantly higher than for seasonal flu.
He said while he was not aware of studies which had directly calculated the IFR for the flu, a 2017 study found on average 0.0135 per cent of the population dies from flu each year.
As an estimated 10 to 20 per cent of New Zealanders catch seasonal influenza each year, this suggests an IFR between 0.0675 per cent and 0.135 per cent.
University of Otago epidemiologist Professor Michael Baker agreed, saying Dr Thornley’s statement is “not supported by the evidence generated in New Zealand”.
He said a comprehensive study of influenza rates in the country found about 35 per cent were infected by the virus each year, which would put the IFR for the flu even lower, at about 0.04 per cent.
This would mean the fatality rate for COVID-19 was seven times higher than seasonal flu based on Ioannidis’ figures, or nearly 18 times higher based on the 0.68 per cent rate from another meta-analysis, or 16 times higher based on the 0.65 per cent rate cited by the CDC.
University of Auckland epidemiologist Professor Rod Jackson was critical of the findings of the Ioannidis study. He said the fatality ratio he used was derived by using the median figure from studies which had vastly divergent results and so was unreliable.
“That is a complete no-no for evidence reviews when there are very different estimates from different studies,” Prof Jackson said.
He said the studies Ioannidis included in his meta-analysis were also based on specific sub-populations – cities or regions – and tests were conducted over a relatively short period of time, which could also give inaccurate results.
Prof Jackson said his own unpublished research found the IFR for COVID-19 was closer to one per cent.
He also pointed to other research which looked at the number of COVID-19 deaths in a single week in the US and compared that to the peak number of flu deaths a week in an average year and found COVID-19 was between 9.5 to 44.1 times more deadly.
AAP FactCheck found the statement that the fatality rate for COVID-19 is only marginally worse than the seasonal flu to be mostly false.
While a meta-analysis by Stanford University epidemiology professor John P.A Ioannidis did indeed find COVID-19 had an infection fatality ratio of about 0.27 per cent, the study has been criticised as flawed, and figures from the WHO and CDC suggest COVID-19 is significantly more deadly than seasonal flu.
The number of people in the US who have died from COVID-19 in the first nine months of the outbreak is already five times greater than the number of flu deaths in a typical year.
Three leading epidemiologists from New Zealand who have written extensively on influenza believed the fatality ratio for COVID-19 is significantly higher than seasonal flu based on a number of studies.
Mostly False – The claim is mostly false with one minor element of truth.
* AAP FactCheck is accredited by the Poynter Institute’s International Fact-Checking Network, which promotes best practice through a stringent and transparent Code of Principles. https://aap.com.au/