The Department of Health has gotten snippy with Scott Johnson at the Powerline Blog in answering his questions. In the most recent round they basically gave non-answers except to a question about the distribution of PCR test results at the state lab. Results from the last 3 months (why not the whole epidemic, something fishy right there) indicate that out of over 2200 tests 59.5% had a cycle number of 25 or less, so pretty clearly positives (meaning reflecting presence of actual viable virus), 29.4% were a cycle number between 26 and 35, so a mix of very likely positives, ambiguous results and very likely not true positives, and 11.1% were cycle numbers over 35, very unlikely to be true positives.
Now, that wasn’t so hard to share, was it? Unfortunately, you can see that the state lab ran an incredibly small percentage of tests over the last 3 months. The answer said the tests done there are largely for long-term care outbreaks, a population which one would suspect is more likely to have high viral loads. Across a more representative population, the distribution would almost certainly be much more heavily weighted to high cycle number positives. The state lab doesn’t have the incentive that hospital labs do to call every possible test positive so they get more revenue. So it is almost certain that other labs are running even a greater percentage of non-positive positives. But just based on what we are told, I would estimate that 20% of the positives represent non-viable positives, i.e., people who aren’t infectious, maybe aren’t even infected. If it is 20% in the state lab, I would risk a bet that it is at least 30% in the non-state ones. Wonder why DOH doesn’t do an actual study?
Then, of course, we get a couple of paragraphs of gibberish about why we shouldn’t pay too much attention to cycle number. If not, why is the state using it for defining a positive test. Once again, we see pure incompetence and ignorance of the current research, as they say: “Ct values aren’t a direct indicator of more virus (how sick someone is). There are so many variables to why a Ct value is what it is that a physician can’t make a clinical decision on it.” I have lost track of how many studies I have posted on which show a very direct correlation between cycle number and amount of virus and that very specifically recommend that physicians be given Ct numbers because they have clinical utility.
Someone who helps me regularly with data analysis and who was getting fairly regular answers to questions from DOH also was recently told that such help would no longer be available, that DOH didn’t have time for it. They are clearly shutting down and trying to control the messages even more tightly.
Join the discussion 5 Comments
Today, KSTP 5 Eyewitness news reported on a new CDC study that “raises concern regarding COVID-19 immunity after infection. In the report they stated that “28% had such a large drop in their antibody levels by 60 days, it would be highly unlikely that they would be protected from getting reinfected.” It also stated that “The study also found differences in antibody levels based on the severity of the illness. Those with a mild case of the virus tended to have more rapid declines in antibodies.” Do you think the way we conduct PCR testing could account for this and if perhaps the “mild cases” were in fact false positives? The report also said that MDH has investigated over 130 cases of individuals with repeat positive tests… and so far have not been able to confirm reinfection.”, which would seem consistent with this theory. Is this important information? Is this something we should be concerned about? How come we are getting such conflicting information regarding immunity?
Excellent post, Madeline. Ivor Cummings (@fatemperor on Twitter) recently cited this article as more proof that immunity is long term.
As a lay person, I feel the antibodies fade but the B-cells and T-cells will give the long lasting immunity common for most viruses. Of course, how can anyone “prove” long term immunity for a new virus?
The lack of information on cycle thresholds (Ct) for PCR tests is troubling at best and leads to numerous problems in analyzing the pandemic. How it relates to reinfection is one of them: was the first infection actually an infection? Was the second infection actually an infection?
The Minnesota Department of Health doesn’t seem to feel the Ct is an important metric. This article from late summer hints at how important this could be.
Florida supposedly started collecting Ct data from its labs in mid December; however, I have not been able to find if this information is being released. I have found other articles on the internet (dangerous, I know) of people in other states trying to find Ct data. If as Mr. Roche suggests even 20% of the “positive” PCR tests are in fact false positives, how does that alter our view of the pandemic? What if the false positives are 40%? The cascade effect on cases, hospitalizations and deaths could be material. It is mind boggling to me that this seemingly basic piece of information is not being used in the public discussion of how society should respond to this virus. People can draw their own conclusions why.
So even though the FDA recommends 40 cycles for PCR, we are to believe the MDH numbers of 25 of for most tests? The same MDH that ‘follows the science’ and the scientist recomendations?
Here is the Florida Dept. of Health Requirement that labs submit Ct values, but I don’t know if/when the information received from the labs is being made public.
Minnesota could do this – require all PCR tests from private (non-state) labs to report Ct values – if it wanted to…