Government ministers are wrong to say that the impact of coronavirus on health in the UK cannot yet be compared with other countries, according to leading academics whose data shows the peak of deaths in England was higher than that in Italy.
International comparison charts were shown at the Downing Street daily briefing until England’s death rate from Covid-19 hit a trajectory set to take it above other European nations. Ministers have said comparisons are not accurate because countries collect their data in different ways.
At a Downing Street briefing on 5 May, Dominic Raab repeated previous warnings by ministers and experts against making international comparisons, describing it as “speculation” to say the UK had the highest death toll in Europe.
“I don’t think we’ll get a real verdict on how well countries have done until the pandemic is over and particularly until we’ve got comprehensive international data on all cause of mortality,” the foreign secretary said.
“We now publish data that includes all deaths in all settings and not all countries do that so I’m not sure that the international comparison works unless you reliably know that all countries are measuring in the same way.”
But a team from the London School of Hygiene and Tropical Medicine (LSHTM) say fair and accurate comparisons can be made by looking at the total deaths each week since Covid-19 took off and comparing them with total deaths from the same week in the years before it existed.
This total excess mortality expressed as a percentage, they say, gives a true picture of the public health impact. It captures not only the Covid-19 deaths, whether or not they were certified as such, but also the extra deaths caused by cancer, heart attacks, stroke or any other condition that was neglected because of the pandemic.
Their data shows Italy’s peak came on 13 March, with just under 15,000 total deaths that week, which was 103% higher than average mortality compared with the same week in previous years. England hit the peak on 15 April, when there were nearly 20,000 deaths in a week, which was 109% higher than previous years.
All countries collect total mortality data. Gaming the data is hard, because the total number of deaths cannot be hidden. Michel Coleman, a professor of epidemiology at LSHTM, who has analysed the data with colleagues Veronica Di Carlo,Melissa Matz and Claudia Allemani, said many of the extra deaths remained unaccounted for.
“Even if all the persons whose death certificate mentioned Covid-19 are counted, a quarter of the excess mortality in England and Wales is not explained,” said Coleman.
“This occurs because some deaths caused by coronavirus occur among people who were not tested. Other deaths occur among people with pre-existing cardiac or respiratory conditions that were made worse by coronavirus, and some deaths from unrelated conditions may occur because the health system was overwhelmed. All these deaths form part of the overall public health impact of the epidemic, but they will not be revealed by restricting reports to deaths among people who were tested for Covid-19.
“So, in a fast-moving pandemic, the cause of death on the death certificate is not a good way to assess the overall public health impact of the disease.”
The team’s data also shows how hard-hit certain areas and cities have been. The epidemic in Italy was disproportionately lethal in Lombardy in the north, where excess mortality hit 300% by 27 March, four times what would be expected. But this hides extremely high excess deaths in three cities: Bergamo (754%), Lodi (489%) and Brescia (344%). In Milan, it reached 143%.
Similarly in the US, the total excess mortality hides grim figures in some cities. On 11 April, total excess mortality for 15 states and New York City, where the data from the National Center for Health Statistics was at least 90% complete, was below 10% in the five weeks to 20 March, but it then rose rapidly to 60% in the three weeks to 11 April. The area includes 138 million people, representing 42% of the US population.
But excess mortality reached 68% by 11 April in Massachusetts and 209% in New Jersey, more than three times the expected number of deaths. In New York City alone, with 11.8 million people, excess mortality rose steeply to reach 580% by 11 April, almost seven times the expected mortality.
Coleman is well versed in international data comparisons. He and his team were responsible for the cancer survival statistics that tracked the UK lagging behind much of Europe and persuaded the government to invest in cancer strategies to improve the NHS’s performance.
The importance of excess mortality data rather than registered Covid-19 deaths has become increasingly recognised as the pandemic has unfolded. It was only when the figures for all deaths in care homes were released by the Office for National Statistics that the real impact in England and Wales was understood. But the excess death toll is more than hospitals and care home excess deaths combined, suggesting some people are dying at home who would in previous years have gone to hospital – and possibly recovered.
David Spiegelhalter, the Winton professor of the public understanding of risk at the University of Cambridge, told the independent organisation Full Fact that all-cause mortality was the best measure.
“I feel the only unbiased comparison you could make between different countries is by looking at all-cause mortality,” he said. “There are so many questions about the rise we have seen in deaths that have not got Covid on the death certificate, yet we feel are inevitably linked in some way to this epidemic. Those are the figures that I would prefer to look at.”
The deputy chief medical officer for England, Jenny Harries, has also suggested all-cause mortality would be “probably the most useful statistic” for comparisons between countries.