Increased testing -> increased detection rate -> increased recorded cases. Unless you put changed testing rates/ population up against measured cases, you end up with false appearance of increased prevalence as testing goes up.
Example:
The fictional town of Kekistan has 1000 people, of whom 10 have the Poopoo Flu
If you test 10% of the population in April, you'll detect 1 of 10 who are positive.
Resultant prevalence rate: 1/1000, or .1% prevalence.
If you test 50% of the population in November, you'll detect 5 of 10 who are positive.
Resultant prevalence rate: 5/1000, or .5% prevalence. In fear-mongering language, that's a 500% spike in Poopoo Flu cases!!! Run for the hills! Grab your wife! Grab your kids! Give your money to China and give up your guns!
Only if you take the new prevalence rate (.5%), then divide by the change in testing : population (in this case, 5x increased testing), do you come up with an accurate change in prevalence. In Kekistan's example, .5% / 5.00 = .1%, identical to the first testing case.
Caveat: cold weather DOES increase actual cases of this entire class, including the Chinese Virus, so reported numbers are some combination of increased testing + increased cases. I'd lean toward the majority of cases being a factor of increased testing access and normalization, with some jobs requiring biweekly testing just to continue coming to work. But the latter IS a variable.
Increased testing -> increased detection rate -> increased recorded cases. Unless you put changed testing rates/ population up against measured cases, you end up with false appearance of increased prevalence as testing goes up.
Example: The fictional town of Kekistan has 1000 people, of whom 10 have the Poopoo Flu
If you test 10% of the population in April, you'll detect 1 of 10 who are positive.
If you test 50% of the population in November, you'll detect 5 of 10 who are positive.
Only if you take the new prevalence rate (.5%), then divide by the change in testing : population (in this case, 5x increased testing), do you come up with an accurate change in prevalence. In Kekistan's example, .5% / 5.00 = .1%, identical to the first testing case.
Caveat: cold weather DOES increase actual cases of this entire class, including the Chinese Virus, so reported numbers are some combination of increased testing + increased cases. I'd lean toward the majority of cases being a factor of increased testing access and normalization, with some jobs requiring biweekly testing just to continue coming to work. But the latter IS a variable.