Reports of a Decline in the Number of Positive Test Results for COVID-19 Are Based on a Mistake
May 22nd 2020
Lt. Cmdr. Danett Bishop, from Hawthorn Woods, Illinois, tests respiratory samples (Seaman Jonathan Berlier)
By Mark Sumner
As of Friday, the CDC reported that America has done over 13 million tests. On a per capita basis, that still puts the United States in the #39 position—behind the U.K., Italy, Germany, and most of Europe, way behind nations like Iceland and the U.A.E. who are determined that everyone is going to get tested. Still, 13 million tests is a lot of tests, even if it did take months to get there, and even if just one Chinese province intends to do 11 million in just 10 days.
Only it seems that those numbers may be a little less than they seem on the surface. Because on Thursday the CDC acknowledged that the number of tests being reported combines the PCR results looking for active cases, with antibody tests looking for people who have been exposed to the SARS-CoV-2 virus. By combining those two values, the CDC is including test results from tests whose accuracy is still in doubt. And it could also be badly distorting the nation’s ability to conduct additional tests.
On the surface, the confusion of the two tests, which was first reported by NPR, may not seem like a big concern. After all, several nations are only conducting PCR tests. Others are using antibody tests. Both can be valid methods of finding cases of COVID-19.
But the CDC action comes at a time when there is high concern over the ability of the United States to conduct a large number of cases in a short period. In order to support the reopening already underway in most states, it will require a very high level of testing ability—one that can be deployed again and again to spot active cases and support isolation and case tracing. By mixing together tests of all sorts, including those that were done to determine the percentage of people exposed across broad areas with no followup on individual cases, the CDC has made it seem as if the ability to test had greatly expanded, even though the value of all these tests was far from equal. It also appears that some of these tests were conducted across an extended period before they were added to the total, inflating the apparent increase in daily tests.
The mixing up of the two types of tests almost certainly contributed to the impression that the percentage of positive tests has rapidly declined over the last few days. That apparent decrease has been used both as an indication that the epidemic was somehow being brought under control, and that testing had expanded to the extent that it was broadly available. Neither of these impressions may be accurate, thanks to the mingling of PCR and antibody tests.
The CDC also seems to have made no effort to winnow out those tests whose results are suspect. Tests such as those used in two California counties generated national stories when they indicated that the number of people who had been exposed to COVID-19 may be far higher than expected. But it’s far more probable that these tests have a level of false positives that make their use in this kind of public survey questionable. These tests have still not been approved by the FDA.
Both antibody and PCR tests have value. But they are different tests for different purposes. Determining whether or not someone has been exposed to SARS-CoV-2 has critical value, especially as it seems increasingly likely that those with significant numbers of antibodies to the virus have obtained immunity to COVID-19 for a period of months or years. But determining who currently has COVID-19 and represents a potential vector of the disease is also critical. Putting both types of tests in one hat, and including the totals of tests that have not been approved along with tests that were conducted for purposes other than determining the status of an individual patient, inflates the number of valid tests that have been conducted and misrepresents the ability of the U.S. to conduct testing going forward.
Posted with permission