As global Covid-19 deaths top 1m, excess mortality gives a different picture – Whiteside

This week, the global Covid-19 death toll crossed one million and South Africa currently has the 10th highest official number of infections in the world. UN Secretary-General António Guterres called this an “agonising milestone”. Meanwhile, the World Health Organisation’s Mike Ryan says that the number of deaths is likely much higher than reported and that there is probably an “underestimate” of those individuals who have either contracted Covid-19 or died as a cause of it. In his latest contribution, Prof Alan Whiteside wants to revisit some fundamental issues relating to the way Covid-19 data is portrayed, focusing on events in the UK, where he is based. Currently, he says, there are far more new cases reported than there were in April at the height of the pandemic. Also, there is a sense of worry because the northern hemisphere is entering the winter, and that may well mean an inevitable increase in cases. At the same time, the number of deaths and hospitalisations has fallen dramatically and may well remain low. Confusing data indeed. – Lindiwe Molekoa

Covid-19 Watch: Confusion

By Prof Alan Whiteside* 


On Tuesday, the global death toll attributed to the coronavirus topped one million people. The largest share, by an order of magnitude, was reported from the USA. This bleak milestone has been extensively covered by the worlds’ media. However, in terms of the daily number of confirmed cases there seems to be a plateau, or at the very least, the numbers are not rising as rapidly. To put Covid-19 into perspective, in 2017 there were 620,000 deaths from malaria, 794,000 from suicide and 954,000 from HIV and AIDS.1 This is the greatest death toll from a pandemic for centuries.

In this blog I want to turn to, and revisit, some fundamental issues:

  • How many coronavirus cases have there been?
  • How many of the cases matter and how much?
  • What does excess mortality look like?
  • One major concern has been the link between HIV and Covid-19. It seems there is some clarity on this – and good news, as discussed in a special section.
  • Finally, in the conclusion, I ask what is the impact of the virus?
Professor Alan Whiteside

The reason for this revisit is because of the way data are portrayed. Each evening in the UK we are informed by newsreaders of the number of new cases and the number of deaths. One graph shows the new cases recorded since the epidemic began. At first sight is deeply concerning, there are far more new cases reported at present than there were in April at the height of the pandemic. On the 25th September there were 6,878 new cases, well above the previous peak of 5,505 on the 22nd April. It should be noted this is data for the United Kingdom, it can be disaggregated for the four nations: England, Scotland, Wales and Northern Ireland.

This pattern is seen in several other European countries. How concerned should we be? There is a sense of real worry because the northern hemisphere is entering the winter, and no one is quite sure what this means. Normally there will be many respiratory illnesses and indeed with schools having reopened and students returned to university, (where many students are now, unbelievably, locked in)2 there is a sense that there will be an inevitable increase in cases. At the same time, the number of deaths and hospitalisations has fallen dramatically and may well remain low.

The Numbers

The cumulative global number of Covid-19 cases has now exceeded 30 million. The USA has the most cases – over seven million, followed by India with over six million, Brazil with nearly five million and Russia with about 1.2 million. All other countries currently have fewer than a million cases. Columbia and Peru are fifth and sixth, respectively. Spain is in seventh place and here the shape of the graph is extremely concerning with significant increases since mid-August. The Spanish government is keen to make it clear that these are localised epidemics. The Mexican trend is downwards while the Argentinian is upward. South Africa is 10th place, while despite the significant increase in cases, the UK is in 14th position. Table 1 shows selected countries by cases. The rate of increase in most countries is slowing. Critically, the rates of hospitalisation and death are low.

Table 1: Global and National Cumulative Numbers of Confirmed Covid-19 Cases (alphabetical order every 2 weeks)3, 4, 5
DateGlobal casesBrazilChile∞India∞Mexico∞Peru∞RussiaSouth AfricaUKUSA
15 Feb69,000020943
4 Mar93,00043086149
18 Mar201,5003721471162,6007,800
1 Apr861,0006,8362,7771,40029,900213,400
15 Apr1,982,55228,28024,4902,41594,845609,422
29 Apr3,117,75679,68593,3994,996162,3501,012,583
13 May4,262,799180,000*232,24311,350227,7411,369,964
27 May5,594,175291,222362,34224,264241,408β1,681,418
10 Jun7,250,909739,503493,02352,991262,0981,979,893
24 Jun9,264,5691,145,906264,689456,183196,847264,689598,878106,108277,2002,347,022
8 Jul11,830,8851,668,539312,911742,417261,750312,911693,215215,885286,9792,996,098
29 Jul16,741,0492,483,191349,8001,531,669402,697395,005827,509459,761302,2954,352,083
17 Aug22,145,6433,407,354388,8552,767,253531,239549,321930,276592,144322,1775,482,602
3 Sep25,761,4303,950,931413,1453,769,523606,036657,129997,972628,259339,4156,075,652
16 Sep29,576,1914,382,263437,9835,020,359676,487738,0201,069,873651,521376,6706,606,293
30 Sep33,646,8134,777,522461,3006,225,763738,163808,7141,162,428663,282448,7297,191,062
Table 2: Covid-19 Deaths and Cases per million (alphabetical order)6
BelgiumFranceItalyRussiaSouth KoreaSouth AfricaSpainUKUSA
Deaths (19 May)421.07529.6418.845.135.26593.04*523.33275.8
Total cases (20 May)2,1893,7361,9912162774,9533,6294,557
Deaths (3 June)429.83533.9333.565.2713.35580.58587.24320.93
Total cases (2 or 3 June)2,3203,8562,9052255795,1254,0705,472
Deaths (17 June)438.73568.7649.015.3827.14580.78627.71354.46
Total cases (16 or 17 June)2,4103,9243,6812371,2395,2214,3726,386
Deaths (1 July)44457463543606655385
Total cases (30 June or 1 July)2,5163,9764,3932492,4325,1404,5957,826
Deaths (8 July)44457564546606657388
Total cases (7 or 8 July)2,759*3,9994,7132573,3175,4004,2098,877
Deaths (22 July)449581935.7122608688451
Total cases (21 July)2,8044,0735,6062777,6305,6934,42012,961
Deaths (12 August)4515821045.9186611699501
Total cases (13 August)3,1274,1556,1502879,5457,2144,60715,532
Deaths (26 August)8654535861136227617623540
Total cases (25 August)7,0753,7514,3056,58835010,3098,2574,81117,344
Deaths (8 September)8674565881236259631624576
Total cases (8 September)7,6515,0404,6107,06241810,78011,240°5,15719,035
Deaths (23 September)9664605891297270642626597
Total cases (23 September)9,0717,1704,9767,64545211,18314,5925,94420,834
Deaths (30 September)8744715931408287672631624
Total cases (30 September)9,9458,3135,1497,94546211,32416,0046,46621,595

*misread these data °data from the previous one or two days

How many cases have there been? How much does it matter?

The Economist is producing some of the best and most readable analysis of the pandemic. The article on 26th September ‘One million and counting: The COVID-19 pandemic is worse than official figures show’ is particularly interesting.7 At present the global case fatality rate is about 3% of the 32 million recorded cases.

As has been extensively discussed the 32 million underestimates the number that have been infected. There are those who are asymptomatic and may not know they had been infected. Many people will have minor symptoms and will not be seen or recorded by the health system. There have been suggestions that some people will deal with the infection even without antibodies being activated. In this case even more people may have had symptomless infection than we realised.

Related articles:

In earlier communiques the difference between antigen and antibody testing was discussed. Antigen is the testing which health services push: “test, know if you have coronavirus, then self-islolate”, an instruction fewer than 20% of British people are currently following. Antibody tests tell if someone has been exposed, and hence antibodies have developed, I predict an antibody passport will come.

In HIV a key source of data are sero-surveys that establish how many people are infected in a population. There are an increasing number of Covid-19 sero-surveys being undertaken by scientists and public-health officials around the world. We know from 40 years of experience with HIV that these surveys are something of a blunt instrument.

The Economist takes note of the many caveats and states:

“But by constructing an empirical relationship between death rates, case rates, average income—a reasonable proxy for intensity of testing—and seropositivity it is possible to impute rates for countries where data are not available and thus estimate a global total.”

They produce a graphic based on 279 sero-surveys in 19 countries which

“suggests that infections were already running at over 1m a day by the end of January—when the world at large was only just beginning to hear of the virus’s existence. In May the worldwide rate appears to have been more than 5m a day. The uncertainties in the estimate are large, … but all told it finds that somewhere between 500m and 730m people worldwide have been infected—from 6.4% to 9.3% of the world’s population.”8

This is not far off the estimates the WHO is set to release, they set an upper bound at 10% of the global population.

The Economist goes on to conclude:

“If the disease is far more widespread than it appears, is it proportionately less deadly than official statistics, mainly gathered in rich countries, have made it look? Almost certainly. On the basis of British figures David Spiegelhalter, who studies the public understanding of risk at Cambridge University, has calculated that the risk of death from COVID increases by about 13% for every year of age, which means a 65-year-old is 100 times more likely to die than a 25-year-old. And 65-year-olds are not evenly distributed around the world. Last year 20.5% of the EU’s population was over 65, as opposed to just 3% of sub-Saharan Africa’s.”9

Let us return to increase in daily cases in the UK. According to the JHU data the daily cases fell in May and then rose from early September. There is, despite the political battles and blame being apportioned, considerably more testing being done. The level of testing March and April was very much lower and so if the results are adjusted the picture seems less concerning (see Table 2’s cases per million). If we look at cases as a rate rather than an absolute number, then the picture is quite different, and this indicator has been falling since early August. The Economist notes:

“During the first wave little testing was being done, and so many infections were being missed. Now lots of testing is being done, and vastly more infections are being picked up. Correct for this distortion and you see that the first wave was far larger than what is being seen today, which makes today’s lower death rate much less surprising”.10

While the cases are underestimated and are not as good an indicator of what is going on as we would like, the excess mortality gives a different picture. The graphic below is taken from Public Health England.11 Note this shows the number of deaths over those that were anticipated.

Weekly excess deaths by date of registration, England

This type of graphic is widely available for most countries and gives a clear indication of the deaths that can be attributed to Covid-19. In addition, there are excess deaths from other causes. This is because other diseases are not being treated, largely because patients are not presenting to health facilities. According to the Health Foundation:

“A&E visits across all unit types dropped by 57% in April, but the percentage fall was larger in minor A&E units (71%) than in major A&E units (48%). This is consistent with people with less severe conditions being directed to other NHS services or choosing to avoid (emphasis mine) seeking medical care. Public Health England … suggest(s) that some of the falls are the result of reductions in the prevalence of certain infectious diseases during the lockdown”.12

A similar pattern was recorded in South Africa and was ascribed, in part, to the ban on the sale of alcohol. The concern is deaths from other causes are rising, especially in resource poor countries. The socio-economic consequences of the pandemic: unemployment and poverty mean that people cannot afford food and health care, there is a real possibility of widespread starvation. The future looks bleak for most of the world’s population.

HIV and Covid-19

This section is largely based on an article by academics at the University of the Western Cape.13 It is further informed by my work on HIV over the past 35 years. I wrote my first paper on HIV in 1987 on migrant mineworkers and presented it at the ‘First Global Impact of AIDS Conference’14 at the Barbican Centre in London. I was working on foreign labour migration to South Africa and was concerned by vulnerability of the hundreds of thousands of men who travelled to apartheid South Africa to work in the mines and industries and on farms. The conditions were generally appalling, and in the mines, dangerous. The migrants were mostly housed in single sex hostels and could expect to see their wives and children just once a year. This, and the resulting growth in a sex industry was well documented by Cathy Campbell.15 HIV rates skyrocketed among these men and, of course, their families in rural Southern Africa, and the local women around the mines. The legacy of the HIV epidemic is that a significant proportion of the Southern and Eastern African population are living with HIV. In South Africa, an estimated 7 million people are HIV positive. This used to be a death sentence. Now HIV is treatable and provided people are adherent drugs they can lead normal lives.16

Those of us who knew understood HIV shuddered when we learnt of the arrival of Covid. We thought Covid numbers would be overwhelming. In Africa public health facilities are often poor. The interventions of handwashing, face masks and social distancing are hard to observe in crowded living conditions and slums. A huge concern was that the tens of millions living with HIV and tuberculosis would be disproportionately affected by Covid-19. This has not happened. Why?

Related articles:

South Africa provides a case study. It has half of Africa’s Covid-19 cases. On 29th September, according to the JHU database there were 671,669 cases and 16,586 deaths. The pattern of the epidemic is informative. South Africa went into a strict lockdown on 23rd March, this was eased in stages from the 1st May, but the number of cases climbed rapidly. They peaked 24th July with 13,944 cases and began to fall almost as quickly. On 26th September just 1,268 new cases were reported.

This is far fewer than predicted. “These comparatively “low” numbers have come as a surprise. Many possible explanations for this modest-by-comparison outbreak are being floated.”17 There are seven human coronaviruses, and four are common triggering about 10%-30% of all colds globally.18 The three that cause concern spread zoonotically to humans since 2000. They are the severe acute respiratory syndrome, or SARS, which emerged in 2002, caused a global panic, infected 8,096 people and killed 774 or 9.6%. The second was the Middle East Respiratory Syndrome (MERS) of 2012 which infected 2,494 people and killed 34% or 858 people. The Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2), is the third and is the virus that causes Covid-19.

Makoti and Fielding note:

“Of particular interest in South Africa is the risk of COVID-19 and HIV co-infections. … Before the COVID-19 pandemic, there was a lack of published academic work on HIV and coronavirus co-infections. This is, in part, why the spectre of COVID-19 running amok in countries with large numbers of people living with HIV – like South Africa – raised anxiety levels. … there were few known associations between HIV and coronaviruses. … The COVID-19 pandemic has propelled the study of HIV-coronavirus co-infections.”19

Their laboratory looked at Covid-19 infections among more than 11,000 HIV-positive individuals. The estimated Covid-19 prevalence does not suggest increased rates of hospitalisation or mortality among HIV-positive patients, indeed clinical characteristics and disease outcomes were comparable to the general population with Covid-19.20 The hypotheses to explain this include:

  • Antiretroviral treatments have anti-CoV properties and offer some protection against Covid-19 infection.
  • The weakened immune system of people with HIV stops “overreaction” to the presence of the coronavirus, avoiding the inflammation linked with Covid-19. Linked to this is the possibility that, because helper T-cells are deactivated in HIV+ patients, the response of the immune system is tempered, reducing excessive inflammation.
  • “A more intriguing hypothesis suggests that the original HIV infection changes the host cells, so they no longer offer a favourable environment for other viruses. This phenomenon is called “viral interference”.” It is suggested the potential 2009 European flu epidemic did not occur because of common cold strains circulating then.21 This idea of ecological space is well understood by people in natural sciences, as well as all gardeners and farmers.

The message is that being HIV positive does not put patients, at least those taking and adherent to treatment, at any greater risk from Covid-19.


This week’s blog has focussed on events in the UK. On the day it is published, there is a strong possibility the Boris Johnson’s government will be defeated in the House of Commons over their desire to push through regulation without parliamentary oversight. This was foreshadowed in the letters from the two groups of scientists referred to last week.22 If anything the divisions are deepening. The Observer notes the country has reached a crossroads, needing to choose between returning to the lockdown or “turn their back on ‘an authoritarian nightmare” that is preventing the nation from getting on with the business of living”.23 It is not just here that these stark choices are faced but that the clowns in charge of the UK are completely and utterly unable to make a decision and stick to it. I will appropriate Alan Paton’s words and say: ‘Cry, The Beloved Country’.24

  • Thank you for reading, reposting and providing comments. What I write is public domain so please share, forward and disseminate. My contact is: [email protected]
  • Professor Alan Whiteside, OBE, Chair of Global Health Policy, BSIA, Waterloo, Canada & Professor Emeritus, University of KwaZulu-Natal

I have been asked by a reader to add a resource to the blog. Given the real need to address mental health I do so. See You do not have to buy anything!

(Visited 1,024 times, 120 visits today)