Researchers – SA’s Covid-19 death toll could be double reported 20,000

New analysis of the country’s excess deaths by a preeminent scientist has added weight to evidence that South African authorities have could be significantly under-reported the Covid-19 death toll.

If a significant proportion of the excess deaths found by researchers are in fact linked to Covid-19, it would push South Africa from relatively low in the global rankings of Covid-19 mortality per one million people of the population, to almost 700 deaths per million which would place it among the hardest-hit countries in the world.

The analysis by Professor Alex van den Heever, the chair of the University of the Witwatersrand School of Governance’s social security systems and management studies, found a significant proportion of excess deaths reported can be linked to Covid-19, based on the timing of the deaths compared with confirmed cases.

Between 6 May and 3 November, researchers found an estimated 49 251 excess deaths of which 19 539 were reported Covid-19 deaths. The official causes of death for the remaining 29 712 deaths are unclear.

As of 11 November, the reported number of Covid-19 deaths was 20 011.

The information that holds the key to providing answers to questions over the possibility that a significant proportion of the 29 000 excess deaths for which no definitive cause is publicly known, can be attributed to Covid-19 or not sits with the Department of Home Affairs and Stats SA, who are custodians of official death notifications.

Van den Heever has found that between 80% and 97% of excess deaths can be linked to the Covid-19 local epidemic.

More conservative estimates, however, place this proportion at 70%, allowing more leeway for the causes being collateral such as ill patients being unable or too scared to access healthcare during Covid-19 surges.

“The close correlation between new infection trends and excess mortality strongly suggests that virtually all excess deaths are Covid-19 deaths,” a presentation by Van den Heever on his analysis, dated 5 November, reads.

He made the presentation virtually during a South African Medical Association webinar earlier this month.

If the conservative 70% estimate is taken, it would push the actual death toll from Covid-19 to just over 40 000. This would take South Africa from low on the list of countries in terms of deaths per million people from its current 350 to roughly 700 deaths per million people.

News24 previously undertook a similar analysis and confirmed that a strong correlation existed between the timing and location of excess deaths and Covid-19 surges – nationally and in each of the provinces.

By mapping confirmed cases, delayed by two weeks to account for incubation periods and delays in reporting, with the excess deaths which are recorded by date of death, this clear correlation emerges not only nationally, but for each province.

As Covid-19 cases surge, excess deaths surge, and as Covid-19 cases decline, excess deaths return to expected bounds.

On the following graph, the zero value on the right-hand axis represents the expected baseline for weekly deaths, which is a rough average of 8 200 natural deaths per week.

If the graph reaches 7 000 weekly deaths, as it does in mid-July, it means roughly 15 000 people died that week of natural causes.

Van den Heever’s analysis was undertaken independently of the team of scientists from both the University of Cape Town’s Centre for Actuarial Research and the South African Medical Research Council’s Burden of Disease Research Unit that has produced weekly mortality reports estimating excess deaths since April.

Epidemiologists have described the tracking of excess deaths – deaths that occur above the expected number of deaths which is based on data from previous years – as the most accurate method of tracking the true impact of an epidemic.

Professor Debbie Bradshaw of the BDRU and co-author of the weekly reports in July said the “timing and geographic pattern leaves no room to question whether [the excess deaths are] associated with the Covid-19 epidemic”.

“The weekly death reports have revealed a huge discrepancy between the country’s confirmed Covid-19 deaths and the number of excess natural deaths,” Bradshaw said at the time.

That “huge discrepancy” is most apparent when weekly, reported Covid-19 deaths are compared with weekly excess deaths from natural causes.

The health department previously refused to acknowledge the official death toll is under-reported, instead of pointing out that simply because the strong correlation existed, it did not mean it was true that a high proportion of excess deaths were Covid-19 deaths that were “missed”.

The department told News24 last month that studies would need to be done to determine the proportion with any certainty. The department was asked whether its position remains the same, but had not responded at the time of writing.

This graphic is a representation of the weekly natural deaths the BDRU/CARe team has reported. The black line, which peaks in July, represents the excess deaths, or deaths over and above expected deaths.

Bradshaw and Professor Glenda Gray, the SAMRC’s chief executive, have for months made repeated calls for the departments of health and home affairs and Stats SA to fast-track the processing of death notifications, official letters that record a person’s date and cause of death.

Stats SA is currently busy with 2018 death notifications.

Researchers have explained there could be multiple causes for these “unexplained” 29 000 deaths, including what they term “collateral deaths” – which are deaths as a result of patients not being able to access healthcare due to health systems being repurposed to handle Covid-19 patients.

While it is possible to show the correlation between excess deaths and confirmed cases, the confirmed cases are an unreliable metric to track the spread of Covid-19, due to changes in the availability and strategy of testing as well as the delays between a person being infected, obtaining a test, receiving the result and for that positive result to be first reported to the national health department and the time it takes the department to report those cases.

A more reliable tracking method, according to multiple scientists interviewed by News24, is the proportion of tests that return with a positive result. South Africa has implemented a prioritised testing strategy, which means symptomatic patients who arrived at hospitals for care were prioritised.

But testing numbers have remained relatively stable, at a daily average of between 17 000 to 20 000 tests on a week-by-week basis for several months.

The following graph tracks the proportion of positive tests, as reported by the National Institute for Communicable Diseases (NICD) against the natural excess deaths found by researchers.

It is clear as more people tested positive – a strong indicator of a surging epidemic – the higher number of excess deaths would follow in the subsequent days.

In the latest mortality report, published on Wednesday, researchers pointed out that excess deaths had started increasing again in the Eastern Cape city of Nelson Mandela Bay.

The estimated excess deaths in the metro, which is seeing a sharp spike in cases, rose by 200 deaths for the week ending 3 November compared with the previous week.

For more news your way, download The Citizen’s app for iOS and Android.