I would like to break down the fallacy of widespread COVID-19 testing in South Africa. These are my personal deductions rooted on my training as medical doctor, a specialist in anaesthesiology and during subspeciality training as an intensivist and cardiac Anaesthesiologist. I have corroborated my assertions with readily available data in the public domain.
As of 27 April 2020, South Africa had done 168 643 tests for COVID-19. These tests had yielded 4 546 positive tests. Over that period we had 87deaths. The cost of a COVID-19 test was R1 200, which means by then we had spent R202 371 600 conservatively on testing for COVID-19. By May 1st (1st Day of Level 4 lockdown) this number was at 217 552 tests at a cost of R261 062 400. For the additional 48 909 tests done, we got a yield of 1 405 (2.87%) positive cases and an increase of deaths by 29. With this update, on the 8th of May 2020, 292 153 tests have been done - an additional 74 601 tests. This was at an additional cost of R89 521 000 spent with a yield of 2 372 positive cases (3.17%). So we have spent in total R353 583 600 in doing widespread testing. Keep these figures in mind as you read the rest of this article and make up your mind on the validity of testing everyone.
Considerations when embarking on a test of any sort in the medical sphere:
1. The indication of the test - is there any value derived from testing? You can’t test for the sake of testing. What are you going to do with the result?
2. The accuracy of a test - it must have a high specificity and high sensitivity (as close to 100% as possible) and very low false negatives (where you wrongly think there is no problem when there actually is), low false positives (where people who do not have medical problem but wrongly assigned as such)
3. It must be meaningful for wide-spread use eg. if a cure exists and where knowing the status has impact on disease/population management then the test is useful
4. It must be cheap, easy to perform and interpret
5. Requires minimal expertise in the remote population settings
6. Less invasive eg pregnancy test
7. Short processing time to allow an appropriate intervention in the shortest time
The facts about the current available COVID-19 tests in South Africa:
1. 30-40% false negative results with the serological quick tests - that is, you can be falsely reassured that you are safe with someone who is in fact positive.
2. PCR testing increases accuracy but there are still up to 15% false negatives. It is more expensive and requires highly specialized training and takes longer to perform in the state sector in South Africa (1-7days)
3. They are not cheap for such an inaccurate test - R1 200
4. They are not readily available - only specialized testing centers at the moment
5. Special training is required to be able to meaningfully get accurate tests - this negatively affects accuracy and increases the false positive rates
6. Knowing that the status is positive does not change anything for the majority of patients who are not sick as the disease is self-limiting, but is useful in those presenting with moderate to severe disease
7. And because even if it’s negative there is a possibility that this could be wrong, does it mean you can drop your guard because the test is negative? NO! So you are better off assuming everyone is positive and testing those who are symptomatic for directed management
8. If 80% of the population has mild disease that does not require admission, what is the value of knowing that people are positive when they can’t be treated and in the face of high false negatives. It would be cheaper to assume that everyone is positive and continue practices aimed at limiting spread in the general population.
9. In fact, if there is anything we can learn from widespread COVID-19 tests it’s the following:
- the disease is highly contagious but less lethal. At the moment mainly symptomatic people are being testing. So the numbers we are deriving from these results mean there is probably a higher infection rate with even less mortality than we are projecting (something like closer to 0.1-0.2% true mortality in infected people)
- the wider you test amongst the asymptomatic population, the less will be the yield with a test that is inaccurate even in the setting of active disease. So there will be a lot of false negatives and false reassurance
- because there is no cure as yet, we can not intervene specifically anyway except for making sure we can isolate and protect other people and staff from positive “known” individuals. The false negatives still mean 15-40% of the time we’ll still get exposed.
Here is what we know about COVID-19 so far from the most severely affected countries and South Africa has shown the same profile:
1. 80% of infected people have mild disease - they do not need admission of any sort and the disease is self-limiting in this population
2. 15% of infected people have moderate to severe disease that requires admission for oxygen supplementation and maybe fluid therapy
3. There is NO CURE - so even if you are admitted, the hospital does not do anything specific for you at the moment anyway. All current therapies are either experimental with no proven efficacy or supportive
4. There is a 3-5% mortality in positive patients broken down as such:
- 72% of over 65years old,
- 23% Mortality rate in 45-64year olds (76% of these have co-morbidities),
- 4.5% between 18-44years old and
- 0.04% between 0-17years old.
- the true mortality of the net was even widened more is believed to be in the order of 0.1-0.2%
I therefore doubt the value of widespread testing in South Africa for a disease with the above profile and in a population with the following characteristics:
1. an average age of 26.4 vs Italy with an average age of 46.6 or Germany (47.4)
2. We have 6% people over the age of 65 in a country of just under 60million (3.6million) vs for an example Germany with 17.9% of over 65year olds in a country of 80million (14.32million)
3. Yes I know you’ll say there is a 20.4% HIV rate in SA (12million people) from the 2019 Stats SA numbers, with 32% of these people (3million) not on treatment. We are worried that if these people contract HIV they will die in numbers.
- This is unproven, and we have the luxury of watching what has happened to the East Asian countries with an HIV burden of 4,734million during their winter where the exposure is maximum. There is no proof that HIV positive patients are dying in higher numbers there, well at least no published data. China is leading Asia with 680 000 people living with HIV.
4. It is widely accepted that the predictions guiding what we are doing in South Africa at the moment were based on a worst case scenario which does not exist anywhere in the world. We know that the assumed Ro for COVID-19 is 2.5 at the moment, that is the number of people a positive person can infect on average. Even with the 10% of the population used to predict infection rates in South Africa testing everyone makes no sense:
- This would mean 6million would have to be assumed infected
- They could therefore spread the disease to 15million people, meaning we need to test more than 25million people
- 25million tests would cost us 30billion rand
- This for a disease which is self-limiting in 80-85% of patients (20-21.25 million). That means we can prove a positive result at a cost of R24Billion where the disease will burn itself off.
- and all of this is for a disease with no cure as yet anyway
The mounting and growing concern now is the collateral damage that is caused by delays experienced from waiting for a COVID-19 test results in conditions like sepsis, bacterial pneumonia, stroke, heart attacks and other medical emergencies, that still occur even during the pandemic. All level of staff are gripped with fear from what they see on social and mainstream media about the lethality of COVID-19. There is significant stigma associated with the disease and victimization of those who are deemed suspects or PUI’s (persons under investigation) by their colleagues and superiors is rampant. This feeds into the collateral damage from people reluctant to come early to hospital for medical interventions, to staff members being unwilling to help patients until they know the COVID-19 status. We are unfortunately not counting these numbers currently and nobody knows the amount of harm, but we should tally these against COVID-19 deaths if we are to measure our efforts against the harm caused.
I hope I have made a reasonable case for the futility of widespread testing in South Africa, at least at this particular juncture. This will be gross wasteful expenditure with no real return on that investment as it will not aid the country in making any meaningful interventions nor will it aid us in limiting disease spread. It produces false security while causing maximum panic. There are way cheaper and sensible ways of dealing with this pandemic than knee jerk reactions borne out of fear. What we know from basic sciences and common sense still has a place even in a 2020 pandemic.
Dr. Nathi Mdladla
Associate Professor and Head of ICU at Sefako Makgatho University and Dr George Mukhari Academic Hospital
Cardiac Anaesthetist in Private Practice
As of 27 April 2020, South Africa had done 168 643 tests for COVID-19. These tests had yielded 4 546 positive tests. Over that period we had 87deaths. The cost of a COVID-19 test was R1 200, which means by then we had spent R202 371 600 conservatively on testing for COVID-19. By May 1st (1st Day of Level 4 lockdown) this number was at 217 552 tests at a cost of R261 062 400. For the additional 48 909 tests done, we got a yield of 1 405 (2.87%) positive cases and an increase of deaths by 29. With this update, on the 8th of May 2020, 292 153 tests have been done - an additional 74 601 tests. This was at an additional cost of R89 521 000 spent with a yield of 2 372 positive cases (3.17%). So we have spent in total R353 583 600 in doing widespread testing. Keep these figures in mind as you read the rest of this article and make up your mind on the validity of testing everyone.
Considerations when embarking on a test of any sort in the medical sphere:
1. The indication of the test - is there any value derived from testing? You can’t test for the sake of testing. What are you going to do with the result?
2. The accuracy of a test - it must have a high specificity and high sensitivity (as close to 100% as possible) and very low false negatives (where you wrongly think there is no problem when there actually is), low false positives (where people who do not have medical problem but wrongly assigned as such)
3. It must be meaningful for wide-spread use eg. if a cure exists and where knowing the status has impact on disease/population management then the test is useful
4. It must be cheap, easy to perform and interpret
5. Requires minimal expertise in the remote population settings
6. Less invasive eg pregnancy test
7. Short processing time to allow an appropriate intervention in the shortest time
The facts about the current available COVID-19 tests in South Africa:
1. 30-40% false negative results with the serological quick tests - that is, you can be falsely reassured that you are safe with someone who is in fact positive.
2. PCR testing increases accuracy but there are still up to 15% false negatives. It is more expensive and requires highly specialized training and takes longer to perform in the state sector in South Africa (1-7days)
3. They are not cheap for such an inaccurate test - R1 200
4. They are not readily available - only specialized testing centers at the moment
5. Special training is required to be able to meaningfully get accurate tests - this negatively affects accuracy and increases the false positive rates
6. Knowing that the status is positive does not change anything for the majority of patients who are not sick as the disease is self-limiting, but is useful in those presenting with moderate to severe disease
7. And because even if it’s negative there is a possibility that this could be wrong, does it mean you can drop your guard because the test is negative? NO! So you are better off assuming everyone is positive and testing those who are symptomatic for directed management
8. If 80% of the population has mild disease that does not require admission, what is the value of knowing that people are positive when they can’t be treated and in the face of high false negatives. It would be cheaper to assume that everyone is positive and continue practices aimed at limiting spread in the general population.
9. In fact, if there is anything we can learn from widespread COVID-19 tests it’s the following:
- the disease is highly contagious but less lethal. At the moment mainly symptomatic people are being testing. So the numbers we are deriving from these results mean there is probably a higher infection rate with even less mortality than we are projecting (something like closer to 0.1-0.2% true mortality in infected people)
- the wider you test amongst the asymptomatic population, the less will be the yield with a test that is inaccurate even in the setting of active disease. So there will be a lot of false negatives and false reassurance
- because there is no cure as yet, we can not intervene specifically anyway except for making sure we can isolate and protect other people and staff from positive “known” individuals. The false negatives still mean 15-40% of the time we’ll still get exposed.
Here is what we know about COVID-19 so far from the most severely affected countries and South Africa has shown the same profile:
1. 80% of infected people have mild disease - they do not need admission of any sort and the disease is self-limiting in this population
2. 15% of infected people have moderate to severe disease that requires admission for oxygen supplementation and maybe fluid therapy
3. There is NO CURE - so even if you are admitted, the hospital does not do anything specific for you at the moment anyway. All current therapies are either experimental with no proven efficacy or supportive
4. There is a 3-5% mortality in positive patients broken down as such:
- 72% of over 65years old,
- 23% Mortality rate in 45-64year olds (76% of these have co-morbidities),
- 4.5% between 18-44years old and
- 0.04% between 0-17years old.
- the true mortality of the net was even widened more is believed to be in the order of 0.1-0.2%
I therefore doubt the value of widespread testing in South Africa for a disease with the above profile and in a population with the following characteristics:
1. an average age of 26.4 vs Italy with an average age of 46.6 or Germany (47.4)
2. We have 6% people over the age of 65 in a country of just under 60million (3.6million) vs for an example Germany with 17.9% of over 65year olds in a country of 80million (14.32million)
3. Yes I know you’ll say there is a 20.4% HIV rate in SA (12million people) from the 2019 Stats SA numbers, with 32% of these people (3million) not on treatment. We are worried that if these people contract HIV they will die in numbers.
- This is unproven, and we have the luxury of watching what has happened to the East Asian countries with an HIV burden of 4,734million during their winter where the exposure is maximum. There is no proof that HIV positive patients are dying in higher numbers there, well at least no published data. China is leading Asia with 680 000 people living with HIV.
4. It is widely accepted that the predictions guiding what we are doing in South Africa at the moment were based on a worst case scenario which does not exist anywhere in the world. We know that the assumed Ro for COVID-19 is 2.5 at the moment, that is the number of people a positive person can infect on average. Even with the 10% of the population used to predict infection rates in South Africa testing everyone makes no sense:
- This would mean 6million would have to be assumed infected
- They could therefore spread the disease to 15million people, meaning we need to test more than 25million people
- 25million tests would cost us 30billion rand
- This for a disease which is self-limiting in 80-85% of patients (20-21.25 million). That means we can prove a positive result at a cost of R24Billion where the disease will burn itself off.
- and all of this is for a disease with no cure as yet anyway
The mounting and growing concern now is the collateral damage that is caused by delays experienced from waiting for a COVID-19 test results in conditions like sepsis, bacterial pneumonia, stroke, heart attacks and other medical emergencies, that still occur even during the pandemic. All level of staff are gripped with fear from what they see on social and mainstream media about the lethality of COVID-19. There is significant stigma associated with the disease and victimization of those who are deemed suspects or PUI’s (persons under investigation) by their colleagues and superiors is rampant. This feeds into the collateral damage from people reluctant to come early to hospital for medical interventions, to staff members being unwilling to help patients until they know the COVID-19 status. We are unfortunately not counting these numbers currently and nobody knows the amount of harm, but we should tally these against COVID-19 deaths if we are to measure our efforts against the harm caused.
I hope I have made a reasonable case for the futility of widespread testing in South Africa, at least at this particular juncture. This will be gross wasteful expenditure with no real return on that investment as it will not aid the country in making any meaningful interventions nor will it aid us in limiting disease spread. It produces false security while causing maximum panic. There are way cheaper and sensible ways of dealing with this pandemic than knee jerk reactions borne out of fear. What we know from basic sciences and common sense still has a place even in a 2020 pandemic.
Dr. Nathi Mdladla
Associate Professor and Head of ICU at Sefako Makgatho University and Dr George Mukhari Academic Hospital
Cardiac Anaesthetist in Private Practice
