Current public data don’t distinguish between diagnostic tests and screening tests for Covid-19, and screening programs are very likely resulting in a vast majority of false positive results because they are by definition testing asymptomatic people; Hawaii data portals and public officials should distinguish these key data points and act upon them
UK government official emails were leaked in April that warned of up to 98% false positive rates from widespread twice-weekly Covid-19 testing with rapid antigen tests. This kind of widespread testing is known as “screening” because it tests everyone even if they’re not showing symptoms.
The official wrote: “As of today, someone who gets a positive [antigen] result in (say) London has at best a 25% chance of it being a true positive … or as low as 2% (on a more pessimistic assumption).”
It’s well-known in epidemiology that a big risk with screening testing is large numbers of false positives, even using accurate tests. This occurs because, when the background disease rates are low, false positives will swamp the true positive counts.
This is not an “FYI” kind of risk — it’s the kind of risk where the large majority of positive test results can be false positives.
The Food and Drug Administration warned in a letter last November, based on exactly the same reasoning, of up to 96% false positives with screening programs.
Harvard medical School professor and epidemiologist Westyn Branch-Elliman and her coauthors warned in a recent article that up to 71 out of 72 (98.6%) Covid-19 screening programs in schools could be false positives.
What does this have to do with Hawaii?
Well, everything. Hawaii’s testing rates and screening programs have skyrocketed in the last two months. At the same time, our case rate has skyrocketed even more.
Until June of this year the Big Island required rapid antigen tests upon arrival on the Big Island. Public data reported by the Tribune-Herald in October of last year revealed that fully 93% of the positive test results up to that date (14 out of 15) were found to be false positives when re-tested with PCR tests (generally considered to be more accurate than rapid antigen tests).
The Big Island’s travel screening program, while revealing just how inaccurate antigen tests can be — as the various articles mentioned above have warned about in the context of screening programs like screening all arriving travelers, regardless of symptoms — did at least include the highly important secondary test to verify the first result.
Because of the verification test, no harm was done by the massive false positive rate of the antigen tests in the Big Island travel screening program. People who tested positive at the airport were free to leave and go about their business unless the PCR verification test came back positive a day or two later, in which case they were required to quarantine from that time forward.
But here’s the problem: many of Hawaii’s other screening programs do not include the crucial second test to verify initial results. As an example, we’re seeing in Hawaii that pediatric “cases” in July and August comprised 20–25% of all new cases (and about 22% nationwide), while pediatric hospitalizations remained extremely low. Why? It’s likely that the large majority of these “cases” are coming from screening programs for kids already in school or heading back to school, or from testing programs that test whole classes and whole families after possible exposure to another case.
The federal Health and Human Services Department (HHS) has been spending $12 billion in funds that Congress approved last March to expand screening programs around the country, including in schools and many other settings. But these programs don’t require verification tests of initial positive results. HHS’s March press release stated it will allocate “$10 billion to states to support COVID-19 screening testing for teachers, staff and students to assist schools in reopening safely for in-person instruction.”
Clearly, Hawaii has a problem with our latest surge in cases and hospitalizations. The surge in cases is certainly not all false positives. But until policymakers, the media and the public start to distinguish between the inevitable false positives, resulting from vastly expanded screening programs, and the actual positives resulting mostly from diagnostic test results (usually PCR tests), we’ll continue to see unnecessary panic and confusion about the actual scale of the problem.
And if the large majority of “cases” are actually false positives we risk being stuck in a vicious cycle of false positives breeding fear and thus more testing and then more false positives — with increased stress on emergency rooms and hospitals from people getting false positive test results and also from others seeking to be tested because they’re healthy but worried that they might have been exposed to the virus by being around someone who was sick or had a positive test result.
Here’s my policy suggestion: all (yes, all) screening programs must include a second verification test with either PCR, live culture calibration (more accurate than PCR for identifying active infections), or genetic sequencing of test results (more accurate still because genetic sequencing looks for long genetic sequences that are truly unique to the virus).
If all screening test results are verified with a second test, only the second test results should be publicly reported. This means that if, as will almost certainly be the case, the large majority of second test results are negative, we won’t see an artificial inflation in the case rate from false positive first test results.
A second suggestion: Hawaii’s public health officials and websites should distinguish screening test results from diagnostic test results, antigen test results from PCR test results, and symptomatic versus asymptomatic “cases.” These are all crucial data points that are simply not available currently. Only with these data in hand can we really start to understand what’s going on — and get out of this neverending pandemic.