![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
coronavirus testing; numbers, numbers, but what do they mean?
Sunday, June 14th, 2020 11:07 amcdc-and-states-are-misreporting-covid-19-test-data (article from 2020/05/21)
"We’ve learned that the CDC is making, at best, a debilitating mistake: combining test results that diagnose current coronavirus infections with test results that measure whether someone has ever had the virus."
I haven't been able to correlate the daily South Carolina numbers that are reported in different places. Maybe the above kind of thing is part of the reason why.
The way SC has reported COVID-19 numbers on its daily report page has changed a few times.
This website captures daily screenshots of the above page, so one can check the numbers from prior days (the site also captures data reported for other states):
https://covidtracking.com/data/state/south-carolina#historical
Since June 10, the SC page distinguishes between viral and antibody tests in the "COVID-19 Testing in South Carolina" section. Before then, it either didn't distinguish them or wasn't including numbers for the antibody tests.
Since June 10, that same section has included a timestamp, which shows that those numbers are as of midnight of the previous day.
By contrast, the top of the page has always listed a time (usually between 3pm and 6pm of the current day) when the page was last updated; I think the numbers shown in the top box are what's been reported as of that time.
From May 22 to June 9, the "COVID-19 Testing in South Carolina" section has included caveats:
"Total negative tests is the cumulative number of all tests without positive results. This total includes individuals who may have been tested once and individuals who have been tested multiple times.
Total positive tests is different than the total number of positive cases in the state. The number of cases is the number of individuals who have tested positive for COVID-19 at least once. Some of those individuals have been tested multiple times for various reasons, and the total positive tests captures all of the positive test results in the state. One positive test may have been tested multiple times, with potentially multiple positive results"
Since June 10, the section has a simpler caveat, which I think is saying the same thing:
"This table represents volume of tests received and not distinct individuals tested. Individuals could have multiple tests."
The daily news releases list the percent positive rate. For example, yesterday:
"The total number of individuals tested yesterday statewide was 5,794 (not including antibody tests) and the percent positive was 13.3%. When the percent positive is low, it may indicate that more widespread testing is being performed and the percent positive may more accurately reflect how much disease is present in the community."
I haven't figured out how that percent positive is being calculated, based on the numbers on the other page.
It seems to me that the percent positive rate should only be based on the viral test results of distinct individuals who have been tested. Or that it should at least not include repeat *positive* test results from the same individual who has previously tested positive. It probably also shouldn't include negative test results from individuals who previously tested positive (showing that they've fought the infection off.)
But the main page doesn't list those numbers. Maybe that is why I haven't been able to come up with a calculation that gives the same results.
The top box of the main page gives a "Total Positive Cases" number, which I *believe* (but am not sure) is the overall number of distinct people so far who have ever tested positive for either COVID-19 or the antibodies, not including repeat tests.
Comparing that number to the prior day's number would include not only the number of new people who have been diagnosed with COVID-19 since the prior day, but also people who have tested positive for antibodies, meaning they had COVID-19 in the past. So it wouldn't make sense to use that number to calculate a daily percent positive rate.
Ah, wait. There's another section of the main page, "Percent Positive Trends Among Reported COVID-19 Cases". It says: "The percent positive is the number of individual people that tested positive (770 as of June 12) divided by the number of individuals tested (5,794 as of June 12)".
Comparing the "COVID-19 Testing in South Carolina" sections for June 11 and 12 gives 957 positive viral tests out of 6019 tests. So the number of positive repeat tests must be (957 - 770 = 187). The total repeat tests must be (6019 - 5794 = 225). So (225 - 187 = 38) people must have previously tested positive and now tested negative... and/or maybe tested negative twice?
"We’ve learned that the CDC is making, at best, a debilitating mistake: combining test results that diagnose current coronavirus infections with test results that measure whether someone has ever had the virus."
I haven't been able to correlate the daily South Carolina numbers that are reported in different places. Maybe the above kind of thing is part of the reason why.
The way SC has reported COVID-19 numbers on its daily report page has changed a few times.
This website captures daily screenshots of the above page, so one can check the numbers from prior days (the site also captures data reported for other states):
https://covidtracking.com/data/state/south-carolina#historical
Since June 10, the SC page distinguishes between viral and antibody tests in the "COVID-19 Testing in South Carolina" section. Before then, it either didn't distinguish them or wasn't including numbers for the antibody tests.
Since June 10, that same section has included a timestamp, which shows that those numbers are as of midnight of the previous day.
By contrast, the top of the page has always listed a time (usually between 3pm and 6pm of the current day) when the page was last updated; I think the numbers shown in the top box are what's been reported as of that time.
From May 22 to June 9, the "COVID-19 Testing in South Carolina" section has included caveats:
"Total negative tests is the cumulative number of all tests without positive results. This total includes individuals who may have been tested once and individuals who have been tested multiple times.
Total positive tests is different than the total number of positive cases in the state. The number of cases is the number of individuals who have tested positive for COVID-19 at least once. Some of those individuals have been tested multiple times for various reasons, and the total positive tests captures all of the positive test results in the state. One positive test may have been tested multiple times, with potentially multiple positive results"
Since June 10, the section has a simpler caveat, which I think is saying the same thing:
"This table represents volume of tests received and not distinct individuals tested. Individuals could have multiple tests."
The daily news releases list the percent positive rate. For example, yesterday:
"The total number of individuals tested yesterday statewide was 5,794 (not including antibody tests) and the percent positive was 13.3%. When the percent positive is low, it may indicate that more widespread testing is being performed and the percent positive may more accurately reflect how much disease is present in the community."
I haven't figured out how that percent positive is being calculated, based on the numbers on the other page.
It seems to me that the percent positive rate should only be based on the viral test results of distinct individuals who have been tested. Or that it should at least not include repeat *positive* test results from the same individual who has previously tested positive. It probably also shouldn't include negative test results from individuals who previously tested positive (showing that they've fought the infection off.)
But the main page doesn't list those numbers. Maybe that is why I haven't been able to come up with a calculation that gives the same results.
The top box of the main page gives a "Total Positive Cases" number, which I *believe* (but am not sure) is the overall number of distinct people so far who have ever tested positive for either COVID-19 or the antibodies, not including repeat tests.
Comparing that number to the prior day's number would include not only the number of new people who have been diagnosed with COVID-19 since the prior day, but also people who have tested positive for antibodies, meaning they had COVID-19 in the past. So it wouldn't make sense to use that number to calculate a daily percent positive rate.
Ah, wait. There's another section of the main page, "Percent Positive Trends Among Reported COVID-19 Cases". It says: "The percent positive is the number of individual people that tested positive (770 as of June 12) divided by the number of individuals tested (5,794 as of June 12)".
Comparing the "COVID-19 Testing in South Carolina" sections for June 11 and 12 gives 957 positive viral tests out of 6019 tests. So the number of positive repeat tests must be (957 - 770 = 187). The total repeat tests must be (6019 - 5794 = 225). So (225 - 187 = 38) people must have previously tested positive and now tested negative... and/or maybe tested negative twice?