Holding the Gains

Over the past month the US lockdown has succeeded in sustaining a steady rate of contagion and death. However, projected decreases have yet to materialize, and it’s becoming evident that they never will. Some countries have done it, have brought the epidemic to a standstill  — South Korea and Australia offer prime examples. Some, like Germany and Czech, are in the process of doing it.

The US hasn’t and isn’t.

Americans have been encouraged to adopt social distancing measures as a temporary measure against a short-lived crisis, rather than as a long-term adaptation to an altered environment. So too with economic measures: increased compensation for the newly unemployed and underemployed is scheduled to sunset at the end of July, while actual payout of those benefits has been shamefully slow in coming. Right-wing agitators are getting militant as national and local politicians assure the public that the worst is over, despite an abundance of evidence to the contrary. Ready or not, things are opening up again.

Not ready.

It’s becoming increasingly apparent that no systematic widespread coordinated efforts will be made to contain the epidemic once shelter-in-place restrictions are lifted. There aren’t enough diagnostic tests to identify and quarantine contagious cases until they’re so sick they’re already home-bound. Findings from most antibody tests are so inaccurate that false positives exceed actual positives. Most locales don’t have nearly enough case trackers to isolate contagious individuals and to identify others with whom they’ve been in contact.

Full church pews and Fourth of July parades and miracle cures and the worst is behind us — public clarions of optimism are beginning to sound shrill and hollow even to true believers. Herd immunity, conspiracy theory, the intrinsic ADHD of the American character — rationales are being trotted out as post-hoc excuses for imminent failure.

It seems foolish to hope that the status quo, bad as it is, can persist for long before taking a sharp turn for the worse. At the present rate, maybe 85 million Americans — a quarter of the population — will have been infected by the end of the year, accompanied by a death toll of maybe half a million. That’s still a long way from herd immunity. But maintaining the steady rate of infection keeps the medical system from being overwhelmed any more than it already is. And if the news from Oxford U. and AstraZeneca comes from reliable sources, then a vaccine could be widely available in early 2021. So there’s motivation to hold a steady course.

Surprisingly, maybe the best hope for avoiding a spike relies less on governmental  mandate than on a decentralized array of everyday interpersonal transactions. Schools and meatpacking plants and nursing homes don’t want to fling their doors open if it means exposing their workers and customers to infection. They’d lose revenues, and they’d incur higher operating costs — a lose-lose situation for return on investment. Put infrared thermometers at all the entrances; wear masks; maintain social distancing; send sick workers home with pay. Diagnostic testing and contact tracing? They’re essential if the goal is to bring infections to a halt, but it’s become clear that the supply will never catch up with the demand imposed by the virus itself. If instead the goal is the more modest one of sustaining the plateau until the vaccine shows up, then any and every useful resource should thrown into the fray.

 

 

 

Short Takes on Social Distancing and Diagnostic Testing

Lapsed social distancing. Mobile phone telemetry shows that Americans are starting to move around more. That’s especially true in the southeast and central areas of the country.

A month ago, communiqués from US government and media stressed the importance of sheltering at home during the first two weeks of April, identified as the turning point in the pandemic. That sort of crisis rhetoric likely mislead large swaths of the public into believing that, if they made it past the peak unscathed, they’d be relatively safe going forward — like making it through a hurricane. That’s not the case. The hope at the time was that the extended regime of commercial shutdown and social distancing would finally begin to pay dividends in the form of a marked downward trend in new infections, leading toward an end of the first wave by the end of April. That hasn’t happened: US covid infection rates — and deaths — have plateaued, but have not yet shown a marked and persistent downward trend.

Three times over the past week the IHCE has increased its body count projections. No doubt increased public mobility will accelerate as businesses start reopening. It’s likely that infection rates will go up as well.

Diagnostic testing protocols seem inappropriate. Who should get tested? Here’s what the CDC has to say:

Most people have mild illness and can recover at home without medical care. They may not need to be tested. At this time, there is no treatment specifically approved for people who have COVID-19.

It usually takes a few days to get the test results back from the lab. Surely doctors aren’t waiting that long to treat a patient who’s already very ill. Since there are no covid-specific treatments to offer, test results won’t alter the course of medical care, which must focus on alleviating symptoms rather than curing the disease. Most people who meet the criteria for testing have already been running a fever, coughing, feeling short of breath and achy — too sick to get out of the chair, let alone get out of the house infecting others.

Why then test at such an advanced stage of disease progression, if making the differential diagnosis has no impact on medical treatment or disease transmission?

 

A Dream

I was seated in the bleachers at a high school football field. I was there by myself. Over my left shoulder I exchanged a few remarks with the guy seated about three empty rows behind me. From the P.A. system a metallic man’s voice announced that it was time for the school fight song. The sparse crowd fell silent. No one stood; no one sang. No band played. I sang:

Hail to thee the blue gold and white
Warriors loyal we’ll cheer you ever onward
Fight team win team vict’ry proclaim
V-I-C-T-O-R-Y
Maine West Warriors that’s our cry
Yay rah for Maine West High

I sang the song all the way through, remaining in my seat, facing the empty playing field, loudly enough to be heard, more or less in tune but without much inflection. No one sang the school’s own song; no one booed, or cheered me on, or made wisecracks. When I’d finished, the group of high school boys seated just to my right wondered how long the school was going to let this go on before doing something about it. I turned toward them; they didn’t turn toward me. I faced the field again, waiting for something to happen.

What’s Up — Two Covid Anomalies Briefly Explored

What’s up with New York? New York State accounts for 6 percent of the US population but 30 percent of confirmed covid diagnoses and 40 percent of covid-related deaths. They got hit early and hard, but the other states are catching up: last week, 21% of newly diagnosed cases and 21% of new deaths came from NY.

What’s up with testing?

Hypothesis 1:  State A and State B have equal populations. State A tests everyone with a fever, whereas State B tests everyone with a fever and a dry cough. All else equal, State A would test more people overall, while State B would have a higher hit rate for the smaller number of tests it conducts.

Hypothesis 2: Again, assume two states with equal populations. State A’s hospitals report a surge in patients testing positive for covid, whereas State B experiences no such surge. All else equal, State A would test more people, and have a higher hit rate, than would State B.

Which hypothesis is better supported by the data? Turns out it’s Hypothesis 2. Using Worldometer’s data, I correlated the states’ testing rate per million of population with their case rate per million. The correlation coefficient is 0.65 — a pretty strong positive relationship between the two variables. I.e., the more a state tests, the higher the rate of test-positive rate. That’s not what I expected.

 

Coming Down from the Plateau

The covid pandemic has plateaued in the US, but the lockdown that’s slowed the spread will soon be relaxed. What’s to keep contations from spiking again? Maybe even more importantly, what would it take to come down the other side of the plateau? The accepted answer: widespread testing, quarantining, social contact tracing.

Those who test positive get quarantined, and in many locales their social contacts get traced. However, diagnostic tests are administered primarily to those who’ve become very ill, which means they’ve already been infected and contagious for a week or more. And nine of ten infected experience few or no symptoms, so they never get tested, all the while being fully capable of transmitting the virus to others.

How do you identify, isolate, and track social contacts of the untested and the asymptomatic? More testing: that’s the usual answer. But who should get tested? The US tests 150K seriously symptomatic people per day, of whom around 20% test positive — that’s 30K new diagnoses per day. Meanwhile the actual new case count is ten times that, or 300K per day. Even if ten times as many daily tests are administered, it’s just a crap shoot if you’re trying to find the asymptomatic and presymptomatic cases.

Suppose the country were to focus its case-tracking program exclusively on individuals presently experiencing symptoms, at all levels of severity. That would limit potential impact, at least initially, to maybe half of those who are infected and contagious. Individuals’ symptoms can be monitored, continually and inexpensively and rapidly, with self-report apps and infrared temperature checks. Expand lab testing to 750K per day — five times current capacity — and those symptomatic from corona can be identified and quarantined. That would prove a challenge, technologically and logistically and financially, but it’s probably within reach.

Next comes contact tracing. Yes, there are apps based on GPS proximity of mobile phones to each other, but let’s assume personal phone contact is the preferred tracing strategy. Ignore the random brief encounters in the shop or at the park: tell me about those with whom you’ve shared personal space for at least ten minutes during the past week. One of those contacts is the likely source of your infection; the rest probably weren’t infected, at least not yet. I’ll track those people down and test them, regardless of whether or not they feel ill or are running a fever. Assume an average of 7 close extended recent social contacts for each newly diagnosed case — this approach would require another sevenfold expansion of testing capacity. But if it’s successful, tracing can work back upstream, capturing an increasing percentage of the asymptomatic carriers of the disease and further reducing contagion.

Here’s a potentially viable program for bringing the contagion rate low enough not just to flatten the wave but to extinguish it altogether. Maybe 50 to 100 thousand FTEs would be needed to do the job thoroughly: that can achieved by hiring and training a small fraction of those who have lost their jobs due to the pandemic shutdown. And maybe 150K x 5 x7 = 5 million diagnostic tests per day would need to be performed, with results being obtained not in days but in hours in order to keep up with the virus.

This vast extension of testing capacity, it’s widely agreed, is the bottleneck. Suppose testing can’t be ramped up fast enough — are there alternatives?

Let’s say that the contact tracer gets in touch with someone who’s had recent close contact with a newly diagnosed case, but who isn’t experiencing symptoms. The tracer can encourage this social contact to self-isolate for two weeks, just in case. It might be difficult for this person to collect two weeks’ paid sick leave without feeling sick. Can the contact tracer “write a note to the boss” — an official notification that can be presented to the contact’s employer? Maybe, but it’s likely to require quarantining tens of millions of Americans at any given time, most of whom aren’t even infected. Probably not feasible.

What about more intensified social distancing of social contacts that doesn’t go all the way into quarantine? Avoid extended physical proximity with others. Wear a mask whenever social distancing can’t be maintained. We’ll check in with you daily to see how it’s going. If you start experiencing symptoms, we’ll get you tested. Granted, this is the preventive armamentarium that’s already widely recommended to all. But if you know that you’ve been in recent close contact with someone who’s tested covid-positive, wouldn’t you be more motivated to adhere to the protocol, protecting both yourself and others? And might not the ongoing support and encouragement of a health professional keep you on track during your two-week stint of relative self-isolation?

 

 

Breathing a Little Easier?

As the number of diagnosed covid cases in this country approaches one million, I have to remind myself: not everyone who gets sick stays sick.

The total number of American diagnosed with the virus increases by about 30K every day. Based on the NY immunity survey, multiply that number tenfold to estimate the actual incidence: make it 300K new infections daily, or 2M new infections weekly.

How long does someone who contracts the virus stay contagious? Since the verdict is still out, I’ll be cautious: 3 weeks.

So there are about 2M x 3 weeks = 6 million Americans who are presently infected and contagious.

For the past 3 weeks the rate of new infections has plateaued. The 2M new cases added to the roster every week are counterbalanced by 1.9M cases who got sick 3 weeks ago and who have recovered, plus 0.1M who have died. Either way, the 3-week-old cases are no longer contagious.

About 330 million people live in the US. 6M/330M = 1.8% — that’s the percentage of people out there who are presently infected and contagious. As long as the new infection rate remains stable, so will the 1.8% who are contagious at any given time.

When I go for a walk, each random person I encounter has about a 2% likelihood of being presently covid-positive. Not bad.

Say I were to go shopping at a grocery store occupied by 100 people: odds are that 2 of them would be covid-positive. That wouldn’t be a bad risk if everyone in the store were to stay in one place: — then the likelihood of my being positioned within 6 feet of either of those two disease carriers would be small. But everyone just keeps moving up and down the aisles, passing me and coming up from behind, blocking my progress in temporary clumps and jamming up the checkout lanes…

There are regional differences. Using these same calculations, fewer than 1 percent of my fellow North Carolinians are currently infected.

The downside: as long as the infection rates remain plateaued, the law of large numbers will prevail, with nearly everyone getting the chance to spend their own personal three weeks as a member of that select minority.

The NY Immunity Study: What It Could Mean

Preliminary results from a statewide random sample indicate that 14 percent of New Yorkers are immune to covid. Assuming the results prove valid and generalizable, what interpretations can be drawn?

Prevalence is high. The running tally of New Yorkers testing positive for the virus stands at 268 thousand. Extrapolating from the immunity study, the actual count would be around 2.7 million — ten times the official rate.

Prevalence isn’t as high as some thought it might be. Two immunity studies recently conducted in California found that the rate of infection might be as much as fifty times as high as the official diagnostic counts. If the CA results had been replicated in NY, the actual statewide prevalence of covid would be 13.4 million, or 70 percent of the state population.

The mortality rate is higher than projected. Early in the pandemic, mortality rate estimates  ranged as high as 2 percent; more recent estimates have dropped to around 0.2 percent. To date nearly 50 thousand US deaths have been attributed to covid. If, per the NY study, 9 million Americans have been infected, the mortality rate would be 0.6%.

****

Some societal implications:

The US is a long way from herd immunity. Coronavirus is highly contagious — on average, newly infected individuals pass the virus on to 2.5 others.  Herd immunity can’t be achieved until the contagion rate drops below 1.0, which won’t happen until the virus has infected at least 60 percent of the population. Based on the NY findings, only around 3 percent of the US population has been infected.

The death toll for achieving herd immunity would be sizable. 57% of a total  population of 330 million: that’s another 190 million people who need to be infected in order to get the rate of contagion below 1.0. At a mortality rate of 0.6 percent, another 1.15 million would need to die en route to herd immunity. To put that in perspective, the viral body count to date is only 0.05 million people — that’s rounding  error.

Individual case tracking is more manageable than previously expected. Eventually an effective vaccine will convey acquired immunity on the herd. That’s at least a year off. Until then, the best alternative strategy for bringing the contagion rate below 1.0 is to isolate newly infected people, as well as their social contacts, from the herd until they recover. For the past three weeks the incidence of newly diagnosed cases has plateaued at around 30K per day: multiply by 10 and you get around 300K new cases daily. That’s a lot of identifying and quarantining and contact tracing; still, it’s far less than the California immunology findings implied. As noted in yesterday’s post, the State of Washington is hiring a brigade of case trackers, estimating that each tracker can keep tabs on 7 new diagnoses and 21 new social contacts per day. At that rate the country would need around 43 thousand case trackers, or a little over 800 per state. 26 million Americans have lost their jobs due to the pandemic; fully staffing a nationwide case-tracker brigade would provide paid and meaningful work to 2% of the newly unemployed.

****

Some public policy inferences:

The worst is yet to come.

Herd immunity is a worst-case scenario.

Individual case tracking should be diligently implemented.

Prioritize and fast-track the vaccine research.