DON’T GET TESTY — PART 1. In testing for for SARS-CoV-2, “an abundance of caution” might ironically cause net harm

[Also appearing on Linkedin]

PROCEED WITH CAUTION

COVID-19 counts are going up. Now is not the time to panic. Panic leads to bad policy. Bad policy harms people. Let’s pause. Breathe. Think.

In a U.S. leadership vacuum, testing for SARS-CoV-2 was never prioritized. States were left to fend for themselves. Organizations had to figure things out on their own. Among such organizations were a multinational food company and a local school. I have been advising for both.

Understandably, neither the company nor the school has in-house expertise in pandemic response. The company has its food scientists. The school has its dedicated nurses; it also has volunteer group of committed clinician-parents. But for both the company and the school, knowledge about viral testing (or about screening tests more generally) is lacking.

Having relevant training and expertise to lend perspective, I recognized in both the company’s and the school’s responses two problematic assumptions:

1. If others are doing it, it must be a good idea
2. Out of “an abundance of caution,” more is better

Regarding the first, just as we wouldn’t have our kids jump off a bridge because others are doing it, we shouldn’t jump ourselves without carefully considering the repercussions. That other companies/schools have taken an approach doesn’t necessarily make it a good idea. Context matters. Different situations are different. Also, with everyone rushing to do something (anything), it is likely not all ramifications have been considered.

Regarding the second, more is not necessarily better. Action taken out of “an abundance of caution” is not necessarily cautious. It can be the opposite. A law in public health: even the best-intended policies have unintended consequences. Attempts at COVID-19 mitigation should not cause harm.

The remainder of this piece considers the harms of attempted COVID-19 mitigation — specifically, harms through a strategy of repeat SARS-CoV-2 testing. The experience for the multinational company is currently being prepared for peer-review. Here I will focus on the case for a school.

The arguments are detailed. Different readers will find different sections helpful. I have put discussion/conclusions first. Supporting evidence follows thereafter (in Part 2). Here’s what you’ll find where:

• Section A: bottom line — weekly testing is ill-advised (discussion and reasoned conclusions)

PART 2.

• Section B: testing basics — an illustrative example for non-experts (ok to skip if you are experienced in epidemiology)
• Section C: calculation inputs — actual numbers for the current SARS-CoV-2 situation (information to justify the numbers)
• Section D: categorical harms — enumerating the ways testing goes badly (this is the crux)
• Section E: gray areas — dealing with uncertainty (additional nuance in the weeds)

A.) FOLLOW THE LEADER? WAIT. STOP. DON’T JUMP!

Many schools are doing weekly, school-wide, SARS-CoV-2 screening. I suspect most are struggling. Whether or not they recognize the consequence (e.g., acting on “false positives” as if they were “true positives”), they must be realizing the consequences (e.g., unnecessarily excluding students and potentially closing classes, cohorts, and campuses).

In considering SARS-CoV-2 testing strategies, guidance for schools has been sparse. The CDC has posted indicators of risk. But indicators are mostly based on problematic metrics — e.g., percent positivity or cases per 100K (Section C). Moreover, CDC guidance is framed entirely around viral transmission; there is a myopic focus on SARS-CoV-2 to the exclusion of all other aspects of life.

Similarly, a recent modeling study views the word exclusively through a SARS-CoV-2 lens. The study presents helpful mathematical projections. But analyses focus solely on interrupting viral spread. If the only consideration was viral spread, wouldn’t the safest thing be to live in a a bubble, with filtered air, and zero contact with others? Obviously that is no way to live. What about non-COVID-19 issues? What about harms?

Giving some consideration to harms is a Duke-Hopkins collaborative report. The report focuses on risk assessment and testing protocols. Referenced in the report is the afore-mentioned CDC guidance. Acknowledged are the risks of false positive results — although the harms of false results are not explored in any detail. Indeed, report authors concede that “equity, educational attainment, and family well-being” are “out of scope for this document” (pretty important considerations to leave out). Additionally, the authors recognize the following:

“Most tests available now have not been evaluated specifically for performance in children or people who do not have symptoms. It is possible that some tests are not as accurate at detecting infection in school-age populations and in asymptomatic individuals.”

Got that? So here are some questions for schools doing weekly SARS-CoV-2 testing:

• When parents understand that screening is an unsupported experiment, using untested technology (in their children), without clear benefit (Section D), and with inevitable harm (Section D), is the response going to be favorable?
• If case counts at a school wind up being higher than those in the general population (as they very well could if there is little asymptomatic community testing) … or higher than those at other schools (who might only be testing in cases of symptoms) … will parents/teachers have confidence in the school or its protective measures?
• When untrue results close a class/cohort/campus and everything goes remote (perhaps in perpetuity), will that be a good thing for anyone? (students, teachers, families … society?)

Keeping schools open is vital. While there remains vigorous debate about how societies should best handle the pandemic, there is emerging consensus that schools should be open. Even within apparently opposing camps — unfairly characterized (or caricaturized) as “open it up” and “lock it down” — there is general agreement: in-person school is a priority.

Towards having schools open, strategies like weekly SARS-CoV-2 testing are often framed a as “an extra layer of protection.” In reality, they may be the opposite: They may guarantee schools close.

Nonetheless, some people have argued that science or not, offering testing may provide a psychological benefit for teachers and parents. But does providing a false impression of safety justify squandering limited resources — while ensuring avoidable harm? I hope we can agree that daily hydroxychloroquine would not likewise be a reasonable approach. Or having students gargle bleach …

Relatedly, the timing of test initiation at the school for which I’ve been advising was particularly surprising. The decision was made just as the White House was becoming a hot zone of COVID-19 … in spite of daily (and sometimes twice daily) testing!

Actually the White House, in spite of all of its blundering, provides a compelling case for what does work when testing does not: i.e., the masking and distancing not practiced there. Schools should focus on these measures. The afore-mentioned CDC guidance explicitly lists five mitigation strategies:

• Consistent and correct use of masks
• Social distancing to the largest extent possible
• Hand hygiene and respiratory etiquette
• Cleaning and disinfection
• Contact tracing in collaboration with local health department

To these, I would add copious ventilation and, if feasible, air filtration. If the solution to pollution is dilution, the inflection for infection may be convection. Stated less ridiculously, air flow matters! Air exchange = good. Stagnation = bad. Opening some windows might provide more protection against COVID-19 than weekly testing.

In fact, with mitigation measures in place, the Duke-Hopkins collaborative guidance does not recommend weekly testing at all. Rather it recommends:

“Continued mitigation measures in school; students and staff are offered diagnostic testing as needed; if a confirmed positive case is found, all individuals in that pod and any other close contacts are quarantined and tested”

(In other words, symptom-based and exposure-based testing)

Exactly! This is the point. It is not testing that is bad. It is weekly screening in asymptomatic individuals that’s the problem.

Before doing any test (in medicine or otherwise), you must have a reason. [And that reason should not be that everyone else is doing it.] The condition (SARS-CoV-2 or other) has to be reasonably likely. Otherwise (as demonstrated with the 50% →1% prevalence example in the APPENDIX) you can expect mostly false results.

Keep in mind, as an extreme example, men have precisely zero chance of being pregnant. The prevalence of pregnancy in men is zero. Nonetheless a man can test positive for pregnancy. A man’s positive pregnancy test is obviously a false positive. The solution is not expanded prenatal care. The solution is not to test men in the first place.

Granted, the risk of SARS-CoV-2 in asymptomatic school students is not zero. But even in the setting of rising case counts and relatively high community prevalence, the chance of false results will generally outweigh the chance of true results (Section D). Conversely, the chance of true harm is all but guaranteed.

Providing some offset for harms is the possibility that actual test specificity, in asymptomatic individuals, is better than assumed (Section C). But we don’t actually know one way or the other. The tests have not been tested. If specificity is higher, there will be slightly fewer false positives. That would be good. But there would still be false positives; even a lesser degree of harm remains unacceptable when there might be negligible benefits.

Regarding benefits, yes screening will detect true positives too. However, unhelpfully (as explained Section C), “true” positives might be inconsequential cases of dead virus … or cases live virus already in the school for more than a week. Also, as case counts go up, the number of false negatives will explode. Testing or not, virus is going to enter the school. Period. At any open school in the U.S. right now, SARS-CoV-2 is already there. The goal cannot be elimination. That is unrealistic. The goal must be mitigation. That is fully achievable.

Reassuringly, schools are already safe places. International data to that effect is mounting. So from a place of relative advantage, we need to keep schools secure. We need to get and keep schools open. We need to keep students, teachers, and families safe. We do that by masking, and distancing, and disinfecting. We do that, by promoting outdoor time and by opening windows when able. We do that by cohorting, and handwashing, and screening daily for symptoms. And we do that by testing …

…NOT testing every week, school-wide, in asymptomatic individuals. Rather, we test only when pre-test probability (APPENDIX) is high: when there are symptoms, in cases of exposures, or after high-risk events (e.g., following holiday travel and group celebrations).

Weekly SARS-CoV-2 testing is not to a school’s benefit. Rather, it is to substantial harm. The case that follows in Part 2 is just the facts. Please don’t get testy.

Originally published at https://www.linkedin.com.