Updates on factors that will ease living with a coronavirus pandemic

Three weeks ago, I posted about things that will help us loosen restrictions as we move forward through the coronavirus pandemic.  In this post, I'm following up on a few of those in the form of a progress tracker.


PCR viral tests

March 23, 2020:  58,500 tests/day
Apr 14, 2020:  145,000 tests/day.

Source:  https://covidtracking.com/data/us-daily
Taken as the average of a 5 day window centered around the date of concern.
SARS-CoV-2 tests per day

As with all things, test availability is subject to multiple bottlenecks.  It takes swabs, PCR reagents, PPE for the people doing the tests, machines on which to run tests, trained personnel, etc.  The manufacturing and supply of all of these can take time to spin up. We're not yet to the point where we can test all likely cases, but we're running about 2.4x as many tests as we were three weeks ago.

This is not enough.  Estimates vary widely, but one Harvard study suggests we may need in the millions of tests per day range.  Others put the numbers in the high hundreds of thousands per day, but no matter what, we need at least to go 5x where we are now.

Serum antibody tests

Serum antibody tests are starting to roll out.  The NIH has started a study to quantify the true prevalence of cases.  Several commercial test providers have created tests and they are available (at high cost) in some areas in limited quantities.  They're not yet playing a big role but I expect that they'll begin to within a month.  As an example, UW Virology in Seattle is hoping to begin 1000 tests/day next week.  Stanford is testing 500 per day now.

As a warning, of course, the false positive and false negative rates on these tests are too high for them to be used alone for purposes such as "certify that I am immune to covid-19".  For example, the Cellex test is somewhere in the 95% range, which means that for every 100 people who have never had coronavirus, the test will tell 5 of them it has.  That's higher than the expected fraction of people in the US who have been infected thus far, which means that if the test says you've had it you ... might have had it.  So for the coming month, at least, the value of these tests is primarily to help understand the spread (and infection fatality rate) of the disease better, but that understanding is critical to crafting policy on how to respond to it.  But for individual benefit, more improvements and more data sources are needed.


This is hard to quantify;  data is limited and messy.  Some anecdotes but not overall numbers:
Melt-blown poly continues to be one of the key bottlenecks, which as I noted a few weeks ago, may take several months to ramp up supply.  N95s are not the only masks we need, and other manufacturers are bottlenecked by other components, such as elastic, when making ASTM level 1 and 2 masks (standard procedure masks).

Relations between the US and China are an important factor in mask availability, which may hopefully be improving for medical supply shipments.

This is all a long-winded way of saying we're better, we're not better enough, and it's going to be at least several months before we're up to speed on this one, but it's helped by...

Progress in reusing masks

There have been several independent efforts to start reusing masks, and these are entering real deployment.  For more information here, see https://www.n95decon.org/.  As one example, see Battelle's 80k masks/day (theoretical) decontamination system.  There are a few of these in deployment now, so we might expect them to reduce N95 mask demand by a million or two masks per month.  As they ramp up production (and these are fairly easy to build), they could conceivably reduce demand by something like 50M masks per month by the end of May.  That's equivalent to doubling our current mask production, which is a great start.  Other hospital systems are deploying UV or heat-based mask sanitization.  But we need all of these and more to meet demand, and it won't happen overnight.

(Calculation note:  I've de-rated the decontamination system throughput to 25k masks/day sustained for this calculation to be conservative about the effective duty cycle and the fact that it's likely hard to keep the system perfectly balanced as far as masks available to clean at all times.  In theory the 60 machines ordered for HHS could top 100M mask reuses/month).

But in total, this means that over the course of 3 months, we'll have increased "effective mask production" fourfold, with more increases possible in the subsequent 3 months.


Three weeks is basically no time in terms of drug development and testing with humans, so very little new to say here.  Testing in humans takes time, and even with a lot of emergency leeway, it still takes time.  There are more (preliminary and small) studies coming in about hydroxychloroquine combinations that make it clear that it's no wonder drug;  it's still up in the air if it helps, but every week that passes we get more data.  Data about Remdesivir is still coming in, but is patchy and many of the studies lack a control group, which makes it very hard to say much with certainty.  There are more than a dozen drugs in various stages of development as either vaccines or therapies.  A lot of activity, which is promising for future results, but I don't pretend to have a crystal ball.

Societal Changes that Reduce Contact

This section should probably be called:  Go go human adaptability!  There have been lots of promising advances here.

No-contact shopping:  Retailers and delivery services continue to try to optimize for more and more store-packed or delivered grocery service.  As with everything, improvements here require process optimizations, infrastructure changes, and hiring more people, all of which will take time.  Instacart reports that orders have quadrupled from a year ago, and says it has doubled its workforce from 200k to 350k in the last two weeks.  Amazon is reported to have hired 100,000 workers and is trying to hire another 75,000.  Many of the platforms are improving their process:  Instacart has added a "fast and flexible" delivery option to help smooth demand and keep drivers busy.  Amazon is reducing in-person store hours at some locations to facilitate more delivery.

Even in-person shopping is getting on the ball.  Some grocery stores are installing plexiglass barriers to reduce contact between customers and (high contact fan-out) cashiers.  Many are upping cleaning schedules, providing sanitizer, and requiring masks.  (Of course, this also requires an increase in the manufacturing of sanitizer, but that's happening too).

Masks for all:  There's a growing consensus that widespread mask wearing can help reduce the spread of coronavirus.  (There's also a lot of fighting and nitpicking about it, but this seems to be the median opinion).  The CDC has updated its guidelines to suggest that everyone wear a mask when out in public.  New York now has a mandatory mask law "when social distancing is not possible."  These changes are just coming online, so we can't see their effects yet, but this seems likely to start making progress by a few weeks from now.

Doin' it online:  More states are beefing up their vote-by-mail efforts.  Some are not.  Doctors have been getting increasingly online with providing telemedicine services, and for some visits, this appears to be working very well.  This may have a long-lasting effect in reducing the amount we all go to the doctor's office to share germs with each other.  We do need primary care visits to resume, however, because data also shows that vaccination prescriptions are down right now as people can't get physically to their doctors.

Tl;dr:  We've made measurable progress for three weeks, but there's a lot more to go before we're near getting things back to normal.  I'll try to update this again in a few weeks unless some better site comes online for tracking these things.


Popular posts from this blog

Reflecting on CS Graduate Admissions

Chili Crisp Showdown: Laoganma and Flybyjing

Two examples from the computer science review and publication process