Back to Africa Check
MARCO LONGARI / AFP

FACTSHEET: Excess deaths in South Africa explained

The number of excess deaths may be a more accurate and more concerning measure of the Covid pandemic than the daily death toll. But how are excess deaths calculated?

South Africans have become accustomed to daily updates from the department of health on the number of Covid-19 deaths recorded in the previous 24 hours. 

But countries around the world, including the United States and the United Kingdom, have also made efforts to record another indicator, referred to as “excess deaths”. 

The World Health Organization (WHO) says: “While 1,813,188 Covid-19 deaths were reported in 2020, recent WHO estimates suggest an excess mortality of at least 3,000,000.”

In South Africa, the South African Medical Research Council (SAMRC) has also been tracking these numbers. Its burden of disease research unit publishes a weekly report on excess deaths every Wednesday.

But what are excess deaths? How are they recorded? And why is there such a discrepancy between these excess deaths and reported deaths from Covid-19? This factsheet answers these and other frequently asked questions. 

What are excess deaths?

“In its broadest sense, excess deaths are defined as the number of people who die within a time period above that which would be expected based on historic death rates,” Dr Jonathan Clarke, a public health expert and biostatician from Imperial College London’s Centre for Mathematics of Precision Healthcare, told Africa Check. 

According to the WHO, many countries do not have the ability to accurately record vital data, such as deaths and causes of death.

Different countries also use different methods to test and report Covid-19 deaths. To overcome these challenges “many countries have turned to excess mortality as a more accurate measure of the true impact of the pandemic”.

The international health organisation says excess death data “accounts for both the total number of deaths directly attributed to the virus as well as the indirect impact, such as disruption to essential health services or travel disruptions”.

How are excess deaths recorded in South Africa?

Clarke said there is no single method to calculate excess deaths. Different methods are used around the world. 

“Generally speaking, methods try to predict how many people would be expected to die in a population within a time period based on previous data and compare this to the number of deaths observed in the time period.”

Prof Debbie Bradshaw is a chief specialist scientist at the SAMRC’s burden of disease research unit and a co-author of the weekly excess death report. 

She told Africa Check that the unit used death data from 2018 and 2019 to calculate expected deaths in 2020. In 2021 they are using an improved model based on additional data from 2014 to 2019. They are also taking population growth into account to predict the number of deaths that should occur. 

Data on natural deaths – resulting from aging or disease – and unnatural deaths – resulting from external causes, such as accidents or violence –  was then sourced from the country’s national population register.

It’s possible to calculate the number of excess deaths above what would be expected had the historical mortality trends prior to the Covid-19 pandemic continued.

“In addition, we have attempted to get a better idea of the impact of Covid-19 by tracking the number of excess deaths from natural causes against a base that was revised for the lack of influenza and lower rate of other communicable diseases last year,” she said. 

“While it is generally acknowledged that lockdown may reduce the number of injury deaths and that there was very little influenza in 2020, I am not aware of any other country that has used this approach.” 

How do recorded Covid-19 deaths compare to excess deaths?

The weekly death reports have revealed a huge discrepancy between the country’s confirmed deaths from Covid-19 and number of excess natural deaths. 

The latest reported number is 58, 702 Covid-19 deaths between 27 March 2020 and 21 June 2021. In comparison, there have been nearly 171,000 excess deaths. The unit started recording excess deaths on 3 May 2020.

It’s important to note that the excess deaths are recorded by date of the death while the confirmed Covid-19 deaths are recorded by the date of reporting by the national health department. This means there might have been delays between when the actual death occurred and when the department announced the deaths to the public.

According to Bradshaw, from May 2020 to June 2021 South Africa experienced two distinct waves of the pandemic during which the numbers of deaths were well above the numbers predicted from historical data. At the time of writing, some provinces were experiencing a third wave, which was beginning to show at the national level.

Deaths SA

Source: Bradshaw et al. Report on weekly deaths in South Africa 30 May-5 June 2021 (Week 21). SAMRC and UCT CARe, 2021.

The data shows that the number of deaths during April 2020 dropped below the predicted number of deaths. The stringent level five lockdown in April 2020 halved the number of deaths from injury, likely because of less traffic on the roads in combination with a ban on alcohol sales. The lockdown also reduced the number of deaths from natural causes. There was decreased influenza and other infections were possibly also reduced. 

Why is there a discrepancy between excess deaths and reported Covid-19 deaths?

Prof Tom Moultrie is the director of the Centre for Actuarial Research at the University of Cape Town and one of the co-writers of the weekly report. 

He said that reported Covid-19 deaths reflected people who died and were known to have been infected with the virus. (Read our factsheet on how deaths are recorded in South Africa.)

“The numbers come from provincial departments of health notifications to the national department of health. In addition to the requirement that these are known to be Covid-deaths, it follows that the vast majority of these deaths are those that occur in health facilities. It is much harder to identify ‘community’ or home-based deaths in this way.”

Bradshaw said people in long-term care facilities and older people living in the community “may have become ill very quickly, before considering seeking health care”. They would not have died in health care facilities or been tested for Covid-19. 

Moultrie said that the pattern of excess deaths “strongly follows the pattern of reported cases and infections”. 

A February 2021 joint report from SAMRC and the Centre for Actuarial Research estimated that “85-95% of the excess natural deaths are attributable to Covid-19”. The remaining deaths were thought to be due to “collateral causes” that may have occurred because people chose to avoid medical care due to concerns about the pandemic or due to hospitals being overwhelmed during the pandemic.

Could this discrepancy be rectified? Not readily, said Moultrie. It would require real-time death notifications including cause of death information and a transition from a paper-based system to an electronic one. He said South Africa was unlikely to fully understand the impact of Covid-19 for several years.

Do the figures have limitations? 

There are a number of limitations that need to be taken into account when reviewing the weekly excess death figures. 

SAMRC has to estimate the number of deaths that are not registered. They do this by comparing vital registration data with census data from the past 20 years. However, as more data becomes available, it is possible that the numbers would need to be revised. 

Bradshaw added that there wasn’t an accepted international standard for reporting and calculating excess deaths. This could hamper global comparisons. Data was also not available for many countries. 

She said the estimates ideally should be “age-standardised”, which would take into account the population age structure. This was important as in a younger population you would expect fewer Covid-19 deaths, as the virus affected older people more. However, there was insufficient global data to make it possible. 

The main limitation was that the medical information about the underlying cause of death was not available.

Further Reading

Add new comment

Restricted HTML

  • Allowed HTML tags: <a href hreflang> <em> <strong> <cite> <blockquote cite> <code> <ul type> <ol start type> <li> <dl> <dt> <dd> <h2 id> <h3 id> <h4 id> <h5 id> <h6 id>
  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.
limit: 600 characters