The last time I commented on why recovered numbers are incorrect. I will try to explain why case numbers are incorrect.
#1 Until Feb. there was a hope to stop spreading like we did against SARS/MERS. The only way was a well organized process of detecting, isolating and discharging.
To achieve it, WHO recommended increasing rtPCR capacity and medical capacity, and recommended keeping strict discharge rule. Only these measures can assure viruses will not spread in our community.
Definition of Confirmed Cases is for concept of public health. I am sorry to say that 'Confirmed Cases' is about who has the virus NOW. It is not for already recovered person. (this is related to Ab test survey performed by Spanish government)
Confirmed cases means only rtPCR positive. But 2 major exceptions were emerged during outbreak.
#1 Until Feb. 11, China's accumulative confirmed case was only 44K.
In Feb. 12, China announced 14K daily cases including symptomatic patients without rtPCR testing. It was just after new big hospital was built.
Surprisingly the number of 14K without PCR testing results was recorded in WHO situation reports as confirmed cases.
Considering the economic indicators, we can guess the current situation of China. I think the completion of new hospital was the turning point.
I remember 14K cases gave me the first difficulty because I only tried to show confirmed cases using rtPCR.
#3 In Apr. 14, US CDC announced new terminology "Total cases/deaths".
Total Cases/Deaths included "Confirmed ~" and "Probable ~". This change was very similar to what China did on Feb. 12.
Some states of US began to report only "total cases/deaths", thus a discrimination between confirmed cases and probable cases has been impossible.
WorldOMeter described in detail about this change. www.worldometers.info/coronavirus/us-data/ Because of this, WorldOMeter changed the names of columns from "confirmed ~" to "total ~" according to the new US CDC guideline. But WHO situation reports, and JHU dashboard and WHO dashboard don't use terminology of "total ~". I think WHO accepted US total cases as confirmed cases. WHO keeps using the term of "confirmed cases".
#4 New Zealand government has been separately announced confirmed and probable, but recovered numbers exceeded confirmed cases. Thus every dashboards are showing confirmed + probable numbers.
#5 In April, antibody test(Ab test) was a big issue because it is cheaper and faster than rtPCR test. (Some said that the Ab test will be a game changer)
One big misunderstanding about the Ab test is that it can be used for multi-purpose.
But the Ab test should be carefully applied to symptomatic patients in emergency, for symptomatic patients only positive results have meaning.
Because our body makes antibodies usually 1 weeks after infection, negative results can't assure that a patient doesn't have a virus.
#6 I remember US medical staff urgently requested approval of Ab test because they had to face a bad situation and had to check the patient's status.
It wasn't because Ab test can replace the rtPCR test.
#7 Confirmed cases have the importance that someone has a virus that can be spread now.
Active cases imply the number monitored by health care workers.
We must know who has the virus now.
Interpreting the results of the Ab test is difficult.
For an asymptomatic person, both positive and negative results guarantee nothing.
Positive doesn't mean a person has a virus now. Negative doesn't mean a person doesn't have a virus now.
To know whether a person has the virus, rtPCR should be performed again.
So I want to say that Ab testing on asymptomatic people has no benefit to the community.
I heard a rumor that some US clinics want to receive negative results of Ab test from visitors.
Ab test can't be used as certification of virus-free for normal people.
I doubt that the negative result of rtPCR can be used as certification because of false negative.
(In previous post, I commented that the definition of recovered requires 2day consecutive negative results due to the possibility of false negative)
#8 All testing methods can be evaluated using Sensitivity/Specificity/Accuracy.
To maximize diagnostic ability(accuracy), almost all testing methods should adjust cut-off value.
As adjusting cut-off value, the number of false-positive(Sensitivity) or false-negative(Specificity) can be reduced.
But unfortunately, an adjusting value to reduce false positives causes increasement of false-negatives, vice versa.
Usually the cut-off was determined to minimize total false results.
Until now several guidelines indicate PCR test has 95% of accuracy, but we don't know how many false positives and false negatives were during evaluation.
#9 Early Pandemic, WHO urged to increase accuracy of PCR test kits.
And the quality of PCR test kits was issued in many countries.
But now no one cares about the accuracy of the PCR test kit now.
We are facing the 2nd wave. I think we should check testing methods again.
Roylab Stats
The last time I commented on why recovered numbers are incorrect.
I will try to explain why case numbers are incorrect.
#1 Until Feb. there was a hope to stop spreading like we did against SARS/MERS.
The only way was a well organized process of detecting, isolating and discharging.
To achieve it, WHO recommended increasing rtPCR capacity and medical capacity, and recommended keeping strict discharge rule.
Only these measures can assure viruses will not spread in our community.
Definition of Confirmed Cases is for concept of public health.
I am sorry to say that 'Confirmed Cases' is about who has the virus NOW.
It is not for already recovered person. (this is related to Ab test survey performed by Spanish government)
Confirmed cases means only rtPCR positive.
But 2 major exceptions were emerged during outbreak.
#1
Until Feb. 11, China's accumulative confirmed case was only 44K.
In Feb. 12, China announced 14K daily cases including symptomatic patients without rtPCR testing. It was just after new big hospital was built.
Surprisingly the number of 14K without PCR testing results was recorded in WHO situation reports as confirmed cases.
Considering the economic indicators, we can guess the current situation of China. I think the completion of new hospital was the turning point.
I remember 14K cases gave me the first difficulty because I only tried to show confirmed cases using rtPCR.
#3 In Apr. 14, US CDC announced new terminology "Total cases/deaths".
Total Cases/Deaths included "Confirmed ~" and "Probable ~". This change was very similar to what China did on Feb. 12.
Some states of US began to report only "total cases/deaths", thus a discrimination between confirmed cases and probable cases has been impossible.
WorldOMeter described in detail about this change.
www.worldometers.info/coronavirus/us-data/
Because of this, WorldOMeter changed the names of columns from "confirmed ~" to "total ~" according to the new US CDC guideline.
But WHO situation reports, and JHU dashboard and WHO dashboard don't use terminology of "total ~".
I think WHO accepted US total cases as confirmed cases. WHO keeps using the term of "confirmed cases".
#4 New Zealand government has been separately announced confirmed and probable, but recovered numbers exceeded confirmed cases.
Thus every dashboards are showing confirmed + probable numbers.
#5 In April, antibody test(Ab test) was a big issue because it is cheaper and faster than rtPCR test. (Some said that the Ab test will be a game changer)
One big misunderstanding about the Ab test is that it can be used for multi-purpose.
But the Ab test should be carefully applied to symptomatic patients in emergency,
for symptomatic patients only positive results have meaning.
Because our body makes antibodies usually 1 weeks after infection, negative results can't assure that a patient doesn't have a virus.
#6 I remember US medical staff urgently requested approval of Ab test because they had to face a bad situation and had to check the patient's status.
It wasn't because Ab test can replace the rtPCR test.
#7 Confirmed cases have the importance that someone has a virus that can be spread now.
Active cases imply the number monitored by health care workers.
We must know who has the virus now.
Interpreting the results of the Ab test is difficult.
For an asymptomatic person, both positive and negative results guarantee nothing.
Positive doesn't mean a person has a virus now. Negative doesn't mean a person doesn't have a virus now.
To know whether a person has the virus, rtPCR should be performed again.
So I want to say that Ab testing on asymptomatic people has no benefit to the community.
I heard a rumor that some US clinics want to receive negative results of Ab test from visitors.
Ab test can't be used as certification of virus-free for normal people.
I doubt that the negative result of rtPCR can be used as certification because of false negative.
(In previous post, I commented that the definition of recovered requires 2day consecutive negative results due to the possibility of false negative)
#8 All testing methods can be evaluated using Sensitivity/Specificity/Accuracy.
To maximize diagnostic ability(accuracy), almost all testing methods should adjust cut-off value.
As adjusting cut-off value, the number of false-positive(Sensitivity) or false-negative(Specificity) can be reduced.
But unfortunately, an adjusting value to reduce false positives causes increasement of false-negatives, vice versa.
Usually the cut-off was determined to minimize total false results.
Until now several guidelines indicate PCR test has 95% of accuracy, but we don't know how many false positives and false negatives were during evaluation.
#9 Early Pandemic, WHO urged to increase accuracy of PCR test kits.
And the quality of PCR test kits was issued in many countries.
But now no one cares about the accuracy of the PCR test kit now.
We are facing the 2nd wave. I think we should check testing methods again.
4 years ago (edited) | [YT] | 960