How many people have died of Covid-19? The answer, one might imagine, can be found at the ‘global dashboard’ updated hourly by Johns Hopkins University using publicly available data. Government departments and agencies supply much of the data. Most publish daily updates, such as the UK summary, and cumulative trackers such as that maintained by the US Centers for Disease Control and Prevention (CDC). The numbers are striking because they are large and presented out of their natural context. Although we are now familiar with the notion of excess deaths, how many of us know the usual annual number of deaths in our own countries?

The correlation between the figures for excess deaths and death certificates mentioning Covid is arresting, and also misleading

The UK’s Office for National Statistics (ONS) recorded 530,841 deaths in 2019, a 2% decrease from the 2018 figure of 541,589. This was a significant blip but explained largely by the fact that there were 1.6% more deaths in 2018 than in 2017. 2018 was the ‘worst’ year since 1999 for registered deaths. The provisional data for 2020 suggests the 2018 figure will be eclipsed significantly. According to the latest ONS bulletin, “the number of deaths up to 25 December 2020 was 604,029, which is 72,900 more than the five-year average”. This is a 13% increase from 2019. The same bulletin observes that, “of the deaths registered by 25 December 2020, 78,467 mentioned Covid-19 on the death certificate. This is 13% of all deaths in England and Wales”. The correlation between the figures for excess deaths and death certificates mentioning Covid is arresting, and also misleading. We should – but for reasons discussed below we frequently do not or cannot – distinguish between a death caused by Covid and a death certificate that mentions Covid.

The situation is similar in the USA, if the numbers are rather larger. As of 8 January 2021, provisional data compiled by the CDC suggest that more than 3.2 million people died in the USA in 2020. This number is higher than the two previous years by around 350,000, or about 13% (the same as in the UK). The same chart includes a column, “All deaths involving Covid-19”, which on 8 January offered the number 318,786. The number of excess deaths and the number of deaths involving Covid-19 are similar. It is tempting to regard them as the same, but we should not.

Death counts are actually collations of death certificates. In the UK, the ONS provides weekly updates for deaths registered, interpreted in a series of publications. During 2020, most of these publications have related to Covid. Similarly, Public Health England (PHE) maintains a running tally of excess deaths with charts that begin with the first recorded Covid deaths in March. The rationale offered is that, “monitoring excess mortality provides understanding of the impact of Covid-19 during the course of the pandemic and beyond”.

In the USA, provisional death counts generally run at a delay of six months, to allow reporting of deaths to reach an approximation of 100%, and confirmed data is not available for more than a year. However, like the ONS and PHE, the CDC has generated a wealth of additional data tables and analyses relating to Covid that are updated daily and weekly.

It is worth noting that more than a quarter of US death certificates are incomplete, and more than a third of deaths take longer than ten days to report to the CDC. Covid-related deaths have been among those most likely to be delayed, by about a week, although they are processed far more quickly than many certificates that specify injury or drug overdose as a cause of death. In the UK, death reporting overall has actually got faster during Covid, although deaths requiring examination and certification by a coroner have been greatly delayed.

When certifying every death, doctors must look for an immediate and underlying cause in a patient’s medical record. The CDC issues a full guide to medical certification of death in the USA, with a Covid related supplement. Similarly, in the UK, a revised guide has been published to assist in death certification during “the emergency period”. It includes an explicit instruction to “avoid ‘old age alone’” as the sole cause of death and to “never use ‘natural causes’ alone”. Important voices have stressed the importance of recording all instances where Covid may have contributed to a death, even if no test was carried out. In the USA, there are also long-standing professional guidelines against including “old age’”, “senescence”, “infirmity”, and other such terms on a death certificate.

In death certification in both the UK and the USA, if nothing else is apparent or obvious, a chronic condition or recent infection may perforce become a cause of death. For example, even if a heart attack cannot be identified as an immediate cause of death, the “underlying cause” of a death may be identified as coronary heart disease. How many of us had an elderly relative whose death certificate specifies they died of “congestive heart failure” rather than “unidentified natural causes after a full life well lived”? We shall never know how many elderly people have been determined to have died of Covid rather than the regular go-to “underlying causes”.

A death certificate would, under normal circumstances, only specify Covid-19 infection as an underlying cause of death if it led to pneumonia or acute respiratory distress syndrome (wet lung) as a direct cause of death. However, as a new pandemic disease, World Health Organization (WHO) guidelines specify that “Covid-19 should be recorded on the medical certificate of cause of death for ALL decedents where the disease caused, or is assumed to have caused, or contributed to death”. Assumed to have caused or merely contributed to a death is not a very high bar. Furthermore, the guidelines continue: “A death due to Covid-19 may not be attributed to another disease (e.g. cancer) and should be counted independently of pre-existing conditions that are suspected of triggering a severe course of Covid-19”. When Covid is detected, therefore, WHO guidelines appear to demand that it be included on a death certificate, if it cannot be ruled out as having contributed to a death.

The bar to recording a Covid death is, therefore, extremely low

Fair enough, one might say: If a post-mortem test demonstrates that a person was infected with Covid at the time of death, then both official guidance and the imperfect state of knowledge about a novel disease and its consequences together demand that Covid-19 be included as an “underlying cause of death”. However, a post-mortem test to confirm Covid is not considered necessary under UK or US guidelines, which allow doctors to “use professional judgment to determine if a decedent had signs and symptoms compatible with Covid-19 during life and whether postmortem testing is necessary”.

The bar to recording a Covid death is, therefore, extremely low. And where Covid appears on a death certificate, in the vast majority of cases it is listed as an “underlying cause of death.” In the UK, in the week ending 25 December, “of the 2,912 deaths involving Covid-19, 2,497 had this recorded as the underlying cause of death (85.7%)”. An independent study published in September suggests an average in the UK may be closer to 70%. This is high, but far exceeded in the USA where the CDC reports that where Covid-19 is reported on the death certificate in approximately 95% of cases Covid is selected as the underlying cause of death. This would only make sense if it were the percentage of people who were hospitalised with severe Covid and subsequently died. However, for reasons that we have set out above, it includes deaths in every other setting, even where there is no confirmed case of Covid.

Clearly, we shall never know how many people died of Covid, only how many doctors felt compelled to record Covid on a death certificate. Nor will we ever have accurate data for the number of deaths caused directly and indirectly by policies introduced in response to the pandemic. The manner in which data has been compiled and presented to the public has served to obscure rather than clarify the true impact of the novel coronavirus on public health.