The UK's 'coronavirus dashboard' may be under-reporting deaths significantly

The UK's daily death toll updates for COVID-19 could be underreporting the true number by a significant margin, according to government numbers published Tuesday.

In its daily tally of deaths, UK government counts only people who tested positive and then died of the virus in hospital. But that doesn't match the number of UK death certificates that list COVID-19 as a factor.

On Tuesday, figures released by the UK's Office of National Statistics (ONS), showed that the 1,568 people had COVID-19 listed on their death certificates between March 5, when the first death was reported, and March 27.

In contrast, death tolls released by Public Health England over the same period show only 926 confirmed deaths due to the virus in England.

There are several factors for the difference. The two tallies are counting different things -- suspected deaths versus confirmed cases that died in the hospital. There may also be a lag in reporting, as ONS have more time to record the deaths.

UK Prime Minister Boris Johnson's spokesperson cautioned on Tuesday that the ONS numbers can't be used to draw trends about final death tolls from the disease. "You can't project one onto the other in a statistically safe way," they said.

But "it's important to have both numbers," said UK chief scientific adviser, Patrick Vallance on Tuesday. "The ONS data, which are important, look at overall deaths on death certificates where coronavirus is mentioned, so that they are not confirmed deaths necessarily," he explained.

The UK is using the "international reporting standard for deaths," he added, which he described as "hospitalized deaths confirmed."

The UK is hardly the only country with data discrepancies and fears of undercounting deaths. China's official data on coronavirus-related death and infection rate has drawn scrutiny from outside. And some of the hardest-hit European countries, like France, Italy, and Spain, count more than just hospital deaths in their daily statistics, but could nevertheless be under-reporting for other reasons.