What’s the smart way to use spare Covid testing capacity?

I have a question for you and/or your MR readers: what’s the smart way to use spare Covid testing capacity?

With the virus (currently) receding in many places fewer and fewer people are getting symptoms and seeking tests.

Even without a second wave in the next few months, we’ll need testing capacity again for the next flu season, when we’ll need to distinguish between flu patients and Covid patients.

How should we use spare testing capacity in the meantime? Increase random testing? Weekly tests for everyone in a single city? Weekly tests for everyone in particular economic sectors?

I would be grateful for your thoughts on this.

That is from O.L.  My intuition (and I stress this is not a scientific answer in any way) is to test people who take elevators every day, to get a better sense of how risky elevators are.  And then test systematically in other situations and professions to learn more about transmission mechanisms, for instance the subway when relevant, supermarket clerks, and so on.  Test to generate better risk data.  What do you all think?

Comments

Mobile testing teams at the direction of contact tracers. The point being so that when a contact tracer is speaking with a person who's at high-risk of being infected but hasn't gotten tested due to inconvenience or similar, the contact tracer can tell that high-risk person, "I can have a mobile testing team at your door within two hours."

And I can tell the contact tracer to Fuck Off right before I hang up on them!!

That is one of the likely responses yes, but if you make testing both free AND highly-convenient for people they're much more likely to get tested.

Let me know when they develop a reliable test that does not attempt to probe my brain at the same time.

Respond

Add Comment

What exactly do you think testing accomplishes at this point?

Other than pushing the death rate down even further, naturally.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Testing is in fact a public service. Some folks should be paid to get tested.

Respond

Add Comment

Respond

Add Comment

Can we get things set up for rapid (<2 hours) results in airports?

IMHO, the primary goal should be to maintain a "ready reserve" for the next hot-spot. This means that the capacity shouldn't be integrated into any business process.

Respond

Add Comment

Respond

Add Comment

I would say, that if it's possible to convert some of the surplus Covid19 testing to antibody testing that would be useful. A large, sample of randomized anti-body testing over a sustained period would give the medical establishment a much better understanding of the current situation.

We still only have a vague idea as to what the percentage of the population that has had Covid19 actually is.

Those are very different tests, with little overlap in material or capacity. I don't think you can convert from the one to the other. Virus testing is PCR on saliva, i.e. DNA replication machines. Antibody testing is an enzyme-linked immunosorbent assay on blood, using different machines and different collection tools.

Yes, I realize to a certain extent that's true. But also to an extent when you are talking about billions of dollars per month, then the money is fungible.

The antibody testing would be useful, like (honest) polls. I would shutter the infection testing, but make sure it can be restated quickly. It will be misused now.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Slightly off-topic:

"Association of country-wide coronavirus mortality with demographics, testing, lockdowns, and public wearing of masks"

"In countries with cultural norms or government policies supporting public mask-wearing, per-capita coronavirus mortality increased on average by just 8.0% each week, as compared with 54% each week in remaining countries"

https://www.researchgate.net/publication/342198360_Association_of_country-wide_coronavirus_mortality_with_demographics_testing_lockdowns_and_public_wearing_of_masks_Update_June_15_2020

Respond

Add Comment

Respond

Add Comment

Users of elevators and public transit seems like a good idea. I think there are now cameras in most or at least many of those places? Can researchers get access to the video and (hopefully using AI/ML so a computer does the work) count what percent of riders are wearing masks? And if those data can be shared nationally we can get an idea how well mask wearing can mitigate the risk of those transportation modes.

I would say continuously testing students in K12 schools. Keeping a handle on cases in schools will ensure that schools can actually remain open instead of closing down at the first site of ‘rona.

It would be over my dead body that I would allow the school test my child. There is zero chance they will do the nasal swab... None...
Clearly this is put forward by someone without children.

Dude chill out. I have two kids. I just want schools to open up. Are you also an anti-vaxxer, "ain't no gummit gonna tell me to stick a needle in my kid" kind of person? If my kid had to get a nasal swab a few times a month for schools to stay open, I'm fine with that.

Respond

Add Comment

Why? What possible harm can testing do to a child?

Respond

Add Comment

Respond

Add Comment

Currently there is nearly no evidence that k though ~6 really spread the virus to older people. And since they are at a very minimal risk themselves, this is probably not a good use of the tests.

The issue is more about perception. There is less risk from kids spreading but if people think Covid is spreading in the school, schools shut down for a prolonged period. If you can stem any spread, schools can stay open.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

I think we learn the most when we focus tightly on a few groups, the results from which we can then extrapolate to the population. Take a neighborhood with willing civic leaders, try to test everyone in that area (maybe 5000 people), then test them again every week for 20 weeks. Ideally you'd want antibody testing for the same group, although not as often (beginning, middle and end of the period). The repeated testing will help to take some of the testing inaccuracy out of the data. Do that in 10 places around the country, and we'll learn more about how the disease spreads.

I'm not sure how elevator or subway testing is going to work, or lead to useful data.

Respond

Add Comment

I’m alarmed by the apparent perception of this person that cases are generally going down. The US had its highest ever report of new cases on June 19th and I think many areas where people are socializing more without masks are about to see increasing spikes. That being said, the question is a good one, although it depends on the goal of the test deployer. If the goal is to continue to contain cases and we’re assuming regional deployment of tests, I’d use it to create mandatory weekly testing of all frontline workers. Combined with mandatory masking inside establishments, that would have a significant impact on further spread.

Typical dimocrap. The Pelosi hoax virus is fake news.

Hey, what happened to the CMG? Hope she's ok.

Respond

Add Comment

Could you expand on exactly which reported facts you consider to be mistaken?

Respond

Add Comment

Respond

Add Comment

"I’m alarmed by the apparent perception of this person that cases are generally going down."

That's because they are. Take a look at the data.

" The US had its highest ever report of new cases on June 19th "

There were 35K cases reported on June 19th. The highest count was 39K cases on April 24th. There were at least 7 other days with higher counts in March and April than June 19th.

source: World Infometers (US / 7 day rolling average clicked)

But to the important part of the picture, the Deaths per day has been trending steadily downward since the middle of April.

April 21st had a 7 day rolling average of 2,214. June 22nd had a 7 day rolling average of 619.

Daily new US cases are up by over one quarter from a couple of weeks ago.

Yesterday's figure for new US infections is the third highest it's ever been. This doesn't mean actual new infections is the third highest it's been, but it's a terrible sign. Looking at the 7 day average it's the highest its been since the original peak in April. If this keeps up it won't be long before the US will be contracting COVID-19 at a faster rate than ever before, even allowing for a fair amount of inaccuracy in the figures.

Respond

Add Comment

Respond

Add Comment

7-day rolling average of cases is trending up could be milder cases, but the fact remains cases aren’t going down; they are going up.

Respond

Add Comment

Well, yesterday reported higher cases than 5 of the last 7 days, but JWatts insists on portraying this as cases going down. Sure, we can make a reasonable case that things are improving - although the reverse case is not without some merit as well. But I'm growing tired of idealogues insisting that the picture is clear, citing misleading cherry-picked data to support their opinions. This is how we become stupider every day. Learn to tolerate some ambiguity.

"who cares about facts?"

The person that actually posts the data and isn't trying to craft a narrative. I made some simple statements and backed them up with actual numbers.

"But I'm growing tired of idealogues insisting that the picture is clear,"

That's not what I wrote. Re-read it. Try to get better at reading comprehension and don't just assume that the writer means things that they didn't say.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

I was looking at this today using data from https://covidtracking.com/data/us-daily

As of yesterday, 27.6 million tests had been conducted with 2.3 million positives, for a rate of 8.3% of total tests. This number has been declining
from a maximum of >30% back in March when testing was less than 1000 per day. The daily positives are around 5.5% but bounce around more.

Somehow the news reports that breathlessly talk about surges don't tell you this. I suspect that the demographics of the tests are changing as fewer people have symptoms. Protests, perhaps?

So, testing only sick people should report fewer positive results than testing mostly healthy people?

Testing today is still very limited because there is no "excess capacity", but workers reporting to work with no symptoms are being tested heavily in certain sectors, eg, first responders, nursing home and health care workers, sports players and training staff.

Respond

Add Comment

The lack of random testing from the beginning and even now is still astounding.

+1, we need far more random / spot testing

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

There is surplus testing capacity in places where the positivity rate continues to grow?

"While most states are increasing testing, the number of tests coming back positive is rising. At least four states are averaging double-digit rates in the percentage of tests that are positive for the virus: Arizona at 20%, Florida and Utah both at 11%, and Texas at 10%.

By contrast, New York, formerly the centre of the US outbreak, has been reporting positive test rates of around 1%.

The World Health Organization considers positivity rates above 5% to be especially concerning."

Cite please.

Likely came from this article - www.reuters.com/article/us-health-coronavirus-usa/as-us-coronavirus-cases-surge-texas-arizona-and-nevada-hit-new-records-idUSKBN23U2QT

It is a bit scattershot, but then, so is what is going in the U.S.

+1, thanks for the post. I was hoping for some analysis on the data. But oh well, it is what it is.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

The false positive rate on PCR tests is probably around 1%. In many places where pandemic is declining, the positivity rate is now also around 1%. That implies that false positives are becoming an important factor.

This makes me wonder about the utility of these kinds of testing experiments. It makes sense, to test people who have high risk jobs, work with high risk people, or who have been in contact with those who have positive tests. But random testing is going to lead to a lot of wasted contact tracing efforts going to chasing false positive contacts who will themselves generate false positive contacts.

A lot of companies resuming work are requiring that everyone be tested. Many of those people and their contacts will be affected by false positives.

Targeted studies like elevators, subways, buses, etc... seem more useful. At least for scientific purposes the uncertainties can be quantified and one gains some knowledge.

"The false positive rate on PCR tests is probably around 1%." How do we know that? What's the false negative rate (and how do we know that)? More generally, how are the tests tested?

Does a PCR test identify infectious virus or does it identify viral RNA? Does the difference matter (and how do we know that)?

Here they give a 4% false positive rate (That seems too high, if it were that high you couldn't consistently have a 1% positivity rate. I guess it's really in the range of 0.1-1%):
https://theconversation.com/coronavirus-tests-are-pretty-accurate-but-far-from-perfect-136671
Here is a preprint discussing the problem:
https://www.medrxiv.org/content/10.1101/2020.04.26.20080911v2

The usual test they give to see if you are infected does not distinguish. That is thought to be the reason for all the studies that showed people having positive tests long after they recovered, which got a lot of attention a few months back. Whether you count those as false positives or not depends on your perspective.

Thank you, Steven. the medrxiv authors say "we estimate a conservative false positive rate from external quality assessments of similar viral assays".

Jolly good but that just displaces the question to how are those assays tested.

In the physical sciences I suppose it's fair to say that you learn to trust measurements if two measurement techniques working on distinctively different physical principles come up with the same results. And you can often cheaply test for reproducibility and cheaply run calibration experiments.

I take it that establishing the virtues of virus testing is typically a harder nut to crack.

Respond

Add Comment

Respond

Add Comment

The false positive is probably close to 1% , which gives it zero reliability if that's the prevalence in the population you're testing. At any time in the general population not a large proportion is infected as they make their way over time from susceptible -> infected -> recovered.
The test detects viral DNA so a positive may not be infectious . You have to actually culture the virus to see if it's infectious.
I would just test people who are likely to have it ( e.g they show symptoms) or health care workers or people with high exposure.

This paper estimated from similar RT-PCR tests the FPR at ~ 1%
https://www.medrxiv.org/content/10.1101/2020.04.26.20080911v2

Thanks, Cat.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Close and continuous contact seems to be the most likely transmitter of the virus, so I would identify those encounters for testing. Elevators is close but not continuous; but since it's how thousands (millions) get to and from work, it makes sense to test at elevators. Restaurants, bars, and churches (an odd assortment) seem to be the most effective transmitters. But we already know that and nonetheless it doesn't stop people from going to restaurants, bars, and churches. Maybe instead of testing for coronavirus we should test IQ. Alas, we already know about IQ in America, and there's nothing we can do about it.

The continuous element is what contributes to viral load, which produces the worst symptoms. That's not news for readers of this blog, but sometimes it's a good idea to state the obvious.

Respond

Add Comment

Respond

Add Comment

Why not focus on dense neighborhoods? LA County has been doing weekly random testing for months, but the data gained from a sampling of 10 million across 4,000 square miles are useless if the hope is to limit the spread.

Some neighborhoods are much more affected, and their residents presumably would like information they could use -- locations of hot spots, or early notification if they are carriers.

Also, why not test people who traveled on commercial flights two or three weeks ago? Some will prefer not to participate, but many others would be willing. This might be unpopular with airlines, but as one who needs to travel soon, I'd like to know.

Or test every employee of every school that is planning a re-opening, and then re-test the same people every two weeks, for their safety and that of the students in their care.

Respond

Add Comment

Regular test dentist and hygienist, so we could get teeth cleaned without worrying would by my choice

+1. I would add anyone who works in a nursing home, salon, or massage parlor.

Respond

Add Comment

Respond

Add Comment

Is anyone testing circulating currency notes?

Respond

Add Comment

Test samples which have already been otherwise collected and centralized, for example blood/plasma banked samples.

https://www.sciencemag.org/news/2020/06/could-global-observatory-blood-help-stop-next-pandemic

Test sewage samples. At the building-level, if possible. That would identify elevator problems, too.

https://www.nature.com/articles/d41586-020-00973-x

Respond

Add Comment

If the sewage testing is accurate to see if cases are building in an area then i'd test everyone when the sewage is showing the virus has an increasing presence.

heavy testing in elderly populations too

Respond

Add Comment

Attention, an illegal, anti-Brazilian coup took over the Igreja Universal do Reino de Deus in Angola. The putschists are not the real leaders of the Church in Angola. They are usurpers. Their orders must not be obeyed. I repeat, they are not the Church's legitimate leaders. Their orders must not be obeyed.

Thiago,
40 peoplegot shot in chicago on monday
seems like a lot for a weekday
if biden was a houseplant what species of houseplant do you
think biden would be?

I am Mr. Salles, a Brazilian farmer.

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Biweekly screening of everyone living or working in a nursing home.

Respond

Add Comment

What percentage of americans take an elevator every day?

Respond

Add Comment

Stockpile for fall/winter.

I don't think you can stockpile to any significant degree. The bottleneck at this point is the lab time to process the samples. Though you could make preparations to expand the lab capacity for fall/winter.

Respond

Add Comment

Respond

Add Comment

I would use some of that spare capacity to test kids at summer camps and pre-schools, to hopefully get a better sense of the extent, if any, opening up schools will cause spread of the virus to accelerate.

Great idea. I'd love to get the faculty of my university tested in early August.
As a daycare parent, I'd also be willing to participate in a study of my daycare and associated families.
I was told two months ago that tests were to be reserved for very ill people. Even though I follow this news in an amateur way, I don't have any sense of whether my local health system has excess capacity now.

Respond

Add Comment

Respond

Add Comment

Children public school random sample testing.
->Every month, sample in person public schools, ie 15% of student population.
-> Test all public school teachers/aides/workers.

Yes

Respond

Add Comment

Respond

Add Comment

Send some excess capacity to countries that need it. The virus also need to be defeated abroad to go back to normal life.

Respond

Add Comment

Test young people who are at very busy bars on the weekends.

These are the same people who have likely been ignoring calls for social distancing for months, and are (or were) probably super-spreaders.

I would venture to guess that the % positive rate would be lower than expected, because many of these super-spreaders *already* contracted the virus weeks/months ago, and now would only test positive for an antibody test.

I can say that personally, there's almost no chance I haven't contracted the virus at some point over the past several months. I think boomers fail to understand the amount of young people that statement applies to.

Respond

Add Comment

Just so you know, it apparently costs about $3.50 US in Australia to run a SARS-CoV-2 PCR test. That doesn't include the cost of collecting the sample, but considering how much practice the people doing it have had by now, they should be pretty efficient. So the total marginal cost should be pretty low by now.

Respond

Add Comment

Uh, nursing home workers. Healthcare workers. Anyone else around old people.

How can you guys not realize this now? This virus is but an annoyance for most everyone who isn’t old. Case count doesn’t matter. Stopping the spread is not a sane goal.

Yeah, it's only a tragedy if you happen to love an old person. And who would be stupid enough to do that?

No one said that?

We are saying protect the elderly...

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

There is no excess testing capacity.

A big problem is the Trump anti regulation procurement is buying supplies that can not be used.

Examples, swabs in bulk so testers will contaminat swabs grabbing swabs one after another from the box. The individual sterile pack allows a swab to be used with only possibly contaminated hands touching only the discarded stick.

Test swabs without the scoring on the stick just below the swab. The sample is taken with a long stick swab, the sample inserted in the tube of medium and the stick bent so it snaps off cleanly to allow capping the tube.

Sample tubes filled with a medium that contains ammonia or similar that produces chlorine gas if exposed to residual Clorox used to clean the machines.

Soda bottle tubes that are not sterile and don't fit the automated test machines. (These are the raw material for soda bottle machines which blow up the bottle as filling. Simpler than the blow mold factory on one side of the wall feeding bottles through the wall to be filled with soda. Allows smaller bottling plants.)

Standards/regulations are required when the end user is not directly connected to the supplier. The medical industry does have standards for these products, but the Trump administration has not adopted them for use in procurement contracts which seem not to be competitively bid. In competitive bidding, losing suppliers will protest substandard products being supplied by the winning bidder. Instead the Trump administration ships them to mostly State logistics centers who them have no use for them.

I did see a report on a pizza shop using its pizza over to make protective shields, heating up the lexan and then forming on a mold, employing two pizza cooks. But they supply these directly to hospitals, nursing homes, dentists, so they get direct feedback on problems with what they produce.

Respond

Add Comment

Yes, design tests to generate better risk data for high frequency exposure scenarios and alternative strategies for reducing risk such as maintaining  distance, mask wearing and common sense precautions. Are people who test positive more exposed than people who don't or less cautious?

Respond

Add Comment

And my intuition is that TC has no particular expertise in deciding optimal public health priorities and policies. Elevators? Is he serious? Seems to me we need to continue to improve testing methodology, understand the transmission of the virus, and monitor nursing homes closely. As far as super-spreader events: coupons for free testing? Coupons and a nasal swab kit/mailer? (What could go wrong, there? LOL). I doubt we'll see another lock-down like what happened in March. It was too economically damaging. As they say, politics is the art of the possible, so what's possible? Ohio continues to open up - I think the last closed venues (large amusement parks, etc.) are set to open up in July. (That's assuming the current alarming increase is due to Memorial Day spread and is only a blip.) Seems to me public health needs to be given as much information as possible to 1) convince politicians that they have an optimal plan and 2) that they will have the information come Fall to trigger the various stages of that plan. What needs to be done now to get to there isn't clear to me. Given limited resources, it isn't clear to me that spending money on random general public testing now is the best way forward. We now need to know what the number (and distribution) of actual cases are, so if virus testing takes away from antibody testing, then it may be worse than useless.

Respond

Add Comment

Best option seems to be mass random testing twice before and twice after changes lockdown policy. Learn the true impact of NFI.

Respond

Add Comment

Perhaps one strategy would be to use testing to create green zones where there is no community spread—these could start small even at the level of a neighborhood and then gradually expand. People entering a green zone would be tested and isolated if they test positive, whether they are coming from China or Myrtle Beach. The US (and really the whole world) otherwise seems like it might have a rolling epidemic where outbreaks bounce around from place to place.

Respond

Add Comment

Testing when there is no cure is only helpful to reduce the probability of spread.
What is needed is a government funded "Manhattan Project" to develop an effective medicine that people can keep in their homes and take at the first onset of symptoms.
This is against the thinking of healthcare professionals, as it takes away their control of patients.
It is also against the thinking of pharmaceutical companies as they lose their vast market of seasonal cold remedies.
A medicine that works against RSV, rhinovirus and coronavirus could be economic suicide to many businesses.

Respond

Add Comment

Daily stratified random (perhaps pooled) testing with cohorts big enough to yield geographic and demographic results.

Respond

Add Comment

Test all airline travelers. Randomly test automobile travelers between states.

Respond

Add Comment

Respond

Add Comment