2 April 2020 – Thursday – #18

Public transit is free in Barcelona! Now, if I could just use it.

The transit system is free until 9 April and, to avoid a fine, riders need proof that they are on their way to an essential job or a doctor.

Yesterday I mentioned that Spanish gasoline consumption is down 83%, which is easy to believe looking at how little traffic passes by the apartment. In the US, the drop in traffic is reflected in traffic accident statistics. For instance, after San Francisco’s lockdown order, weekly traffic accident reports dropped from an average of 260 to 110.

San Francisco traffic accident reports by day.

Given the fall off in automobile and truck traffic here, it’s no surprise that, in the second day of the new quarter, economic reports reflect the enormous impact of Covid-19 on Spain’s economy: 850,000 jobs lost and 300,000 new claims for unemployment. It’s easy to see on this chart the Covid-19 disruption.

Spanish unemployment claims jump at start of Covid-19 lockdown

Like the rest of the world, Spain’s economy was performing progressively better the past few years. In early March, the Spanish government anticipate Covid-19 with a €200 billion Covid-19 relief package including moratoriums on mortgage and utility payments, easier access to unemployment, and tax deferments. Still, today’s unemployment numbers are eye popping.

On 31 March, I mentioned a report that the official Italian Covid-19 death toll is low because it didn’t count people who died at home and weren’t tested. In Spain news is emerging that a quarter of the people who died of Covid-19 lived in retirement or nursing homes. Earlier reports suggest the facilities often were deserted by staff because of untenable conditions.

Some useful Covid-19 news. The anesthesia department at the University of Florida has invented a clever N95 mask substitute from materials many hospitals have in stock. Follow the link for instructions on constructing a substitute N95 mask.

The makeshift masks use Halyard H600 two-ply spun polypropylene that cannot be penetrated by water, bacteria, or particles. It blocks 99.9% of particulates, making the masks about 4% more effective at blocking particulate material than the N95 masks ….

Bruce Spiess, M.D., Professor of Anesthesiology, UF College of Medicine

I want to focus today on Covid-19 testing. Everyone needs to focus on testing as the next step. After lockdown, hand washing, N95 masks, and ventilators, nothing else matters in the next 90 days. Treatments and cures won’t arrive in time as most countries move through peak Covid-19 infection.

Until a Covid-19 treatment or vaccine gets to market in scale, testing is the only way to know at the end of the first wave of infection whether people need to quarantine or are safe to return to work.

Luckily , lots of test kits on the way, like this 2 minute antibody test just approved by the FDA. Alas, this test only works if you’ve been infected for a few days. This is similar to the days of AIDS when most widely available tests detected HIV only days or weeks after the initial infection.

Different Covid-19 tests will work differently from each other. Some will test for presence of the virus by using PCR to amplify any virus in the body. Others will detect antibodies, which show that the body has had an immune response to Covid-19. Other test attributes include sample acquisition type (swab, pin prick, blood draw), time for results (minutes to days), accuracy (% false positives and false negatives), and timeliness (how long a person needs to be infected to obtain a valid test result).

How much testing is enough? According to WHO, if 80% – 90% of tests come back positive, not nearly enough. 10% positive response indicates testing has reached enough of the population to manage most, if not all the Covid-19 infections. My guess based on the amount of testing South Korea needed to manage infections is that adequate Covid-19 testing requires on the order of 5 tests per thousand people per day. In the US, that works out to 1-1/2 million tests per day in round numbers. In Spain, about 1/4 million tests per day.

Countries need to have that kind of Covid-19 capacity in place in 1-3 months from now. That’s when they will be traversing the backside of the Covid-19 infection curve.

One issue with current Covid-19 testing is false-negatives. China believes its RNA tests may have up to 30% false negatives. US officials worry the US false-negative number are be worse. That means if you test negative, there’s still about a one in three chance you are actually Covid-19 positive. When calculating how many Covid-19 tests are necessary, the calculation has to assume that one person probably needs more than one test.

Okay, enough on testing. Everyone please build testing capacity asap.

We had two days of intense rain on Tuesday and Wednesday, but the sun is starting to come out today. The weather has a stronger influence on my mood during isolation. There was enough of a break in the rain yesterday to allow the neighborhood to cheer the healthcare workers at 8p local time. You can’t see many of my neighbors clapping, but you can hear them.

Many people I know in Barcelona are originally from Latin America. I’m starting to hear news about Covid-19 through them. Ecuador seems to be a disasters right now. My friend said, “In my city there has been a lot of unofficial deaths and they can’t even do anything with the bodies. People are forced to stay with their loved ones dead at home.”

One Mexican friend was able to keep her apartment here so she could delay a months-long visit to Mexico. Another Mexican friend who’s having trouble working out immigration to be with his boyfriend was asking on Facebook two days ago if there were any flights from Europe to Mexico.

A friend in Buenos Aires reports that the lockdown started there more than a week ago when there fewer than 50 Covid-19 casualties. Friends in Brazil are anxiously watching rise in Covid-19 cases. Some countries are responding, many don’t have adequate resources even if they do. Death and dislocation in Latin America is following the US and Canada.

I don’t have a happy ending today. The US just flew past the 5,000 death mark. The number of Covid-19 cases reported in the world will fly by one million today.

There’s hope. Many Asian countries have Covid-19 under control. Children in South Korea are going back to school. Also, in Europe all the slopes are turning the right direction after a month of lockdown. Countries that are about to experience a Covid-19 epidemic need to lockdown right now.

The US is going to be terrible for the next two months. The White House estimate is 100,000 – 240,000 casualties. When I look at this chart, I think my estimate of 350,000 – 500,000 US casualties may be optimistic. It makes me profoundly sad that, in spite of the resources America has, its people will suffer because poor leadership discounted Covid-19 for months. Under different leadership, America would be helping the rest of the world, not begging it for N95 masks and ventilators.

This chart doesn’t even register the rest of the world.

I know it’s selfish, but I’m glad I stayed in Barcelona. Now, if I could just get on one of those free buses and go somewhere.

1 April 2020 – Wednesday – #17

Let’s start today with maps. Kinsa, a company that has sold a million smart thermometers, sees evidence that social distancing in the US is paying off. The company provides a live map of data from is US customers. Map viewers can drill down to results in their zip code. Here’s a map from two days ago.

Kinsa smart thermometer data at 30 March 2020, atypical readings in the US from 162,000 thermometers

The company’s tracking algorithms typically pick up flu trends 1-2 weeks before government agencies that rely on reporting from healthcare providers. Where Covid-19 lockdowns took effect, the company measured drops in both flu and Covid-19 infections. I expect the CDC will report decreases in other infectious and sexually transmitted diseases as well.

So, now that we have measurements showing that social distancing works, how is the US doing at social distancing? Even more important, how are your neighbors doing at social distancing? Unacast uses cell phone data to determine how people have changed the movement patterns.

Unacast cellular-based social distancing by state from 30 March 2020

Other than the chart legend characterizing the results as “Change in Average Mobility,” I wasn’t able to find an explanation of how Unacast produces its scores, but the models appear to be good enough to drill down on your neighbors’ social distancing by state and county. Unacast also provides Covid-19 infection data with its social distancing data. For instance, Manhattan (New York County, New York) has earned an A grade for social distancing:

Unacast cellular-based social distancing data for Manhattan as of 30 March 2020

The Canadians have maps, too! Here’s a Covid-19 heat map from Flatten. Shrey Jain, a University of Toronto student, started the site after his university sent students home. Jain simply asks people to provide data about their health and then combines this data with location data into a heat map. Because the data set includes more users from Toronto, the high level analysis is skewed towards Covid-19 events in that city.

I’m looking for European Covid-19 map data. If you know of any please leave a comment or send me a note.

FlightAware isn’t a Covid-19 tracking site on purpose, but take a look at this morning’s data.

Even the airplanes in Western Europe are practicing good social distancing. Three aircraft over all of Spain! There are small clusters around Heathrow and Amsterdam, but does anyone know what the curve of aircraft on the right is about? It looks like Russian airspace to me, but I’m wondering what the route is.

In US Covid-19 news, many of the three million newly unemployed workers are losing healthcare benefits during a pandemic. The Trump White House won’t open up Obamacare enrollments to help out. Luckily, though, since it’s April Fools day, US residents don’t have to pay their April rent or mortgage.

April Fools!

Spanish petroleum distribution numbers suggest energy use is much lower than previously thought during the Covid-19 pandemic.

ProductYear-over-Year Consumption Decrease
Gasoline83%
Diesel61%
Aviation Fuel85%
YoY decrease in Spanish petroleum product consumption during Covid-19 pandemic

It looks like truck usage is down by more than half while automobiles and airplanes are essentially parked. Assuming Spanish energy use is similar to the rest of Europe, the Covid-19 pandemic will provide climate change models useful data about reduction in human CO2 generation.

While the cars and trucks are taking a vacation, the Internet is on fire. Nielsen estimates that US streaming video viewing is up 85%.

I’ve had a couple of good writing days and even allowed myself to start thinking about life after isolation. I like my cooking, but a great meal out would be such a treat right now. Unfortunately, I’m not sure what will be open in May because I read what’s happening in California. The California Restaurant Association wrote Governor Newsom that it expects 30,000 California restaurants, or 30% of California’s restaurants, to close in the wake of the state lockdown.

Here in Barcelona, a food blogger provides updates on the cuisine scene during the lockdown. Chefs are providing online tips and the city’s Mercats are open reduced hours.

The calçot is a regional favorite. A calçot is skinnier version of a leak that’s typically roasted on a grill outside and served with Romesco sauce. It’s a mess and total fun. The season ended just before the pandemic took off.

Calçots on display in Gracia

Ana, a friend and local food critic, took Brad and me to one of the specialty calçot restaurants. What a feast! If there’s a silver lining for the restaurant business around Barcelona, it’s that these calçot restaurants only operate during the calçot season, so they’re normally closed from March until November. I’m counting on more calçots later this year.

Last but not least, if you have a fever, make sure it’s not cabin fever before you call your healthcare provider. This video may help with your cabin fever diagnosis.

31 March 2020 – Tuesday – #16

Will the US Covid-19 response be worse than Italy’s? There are new data points since I estimated US Covid-19 mortality a few days ago. It’s not good news.

The Chinese Covid-19 death toll may be ten times higher than officially reported based on cremation activity in the area. Reports that Wuhan residence were paid for their silence suggest the government is covering up actual Covid-19 mortality there. I want to point out that even though the mortality rate may be much higher than reported, the test-and-quarantine method the Chinese used to quell the epidemic worked.

Based on overall death rates in several northern municipalities, the Italian Covid-19 death toll may be four to ten times higher in that region than officials report, apparently because people who died at home weren’t counted as having Covid-19.

It is extremely reasonable to think that these excess deaths are largely elderly or frail people who died at home or in residential facilities, without being hospitalized and without being swabbed to verify that they have actually become infected with Covid-19.

Corriere del Sera

As a result of this, I’m revisiting a chart I made two days ago. Here’s the original chart:



Estimated Total
Covid-19
Deaths
Covid-19
Deaths per
Capita
Implied US
Covid-19

Deaths
Spain13,000 – 19,0000.3 – 0.4 per 1k96,000 – 128,000
Italy20,000 – 30,0000.3 – 0.5 per 1k96,000 – 160,000
Original Chart from 29 March 2020

I was a little surprised when I made the original chart that the overall Italian mortality was so close to Spain’s because, while the cities of Madrid and Igualada were Covid-19 hot spots, Spain hasn’t experienced an entire region turning into a hot spot the way Lombardia and its surrounds did.

There are two assumptions I’ll make in revising the chart. One is that Italian under-reporting was only in the hardest hit cities. In areas that weren’t hit hard, I’ll assume there were not scores of people who died at home and went unreported. The second assumption is that 1/3 of the deaths happened in hot spot cities that under-reported. Based on the 4x – 10x under-reporting from the article, that means that countrywide Italian mortality is something like 1.3x – 3.3x higher than reported. Putting all that together, here’s the revision.

Estimated Total
Covid-19
Deaths
Covid-19
Deaths per
Capita
Implied US
Covid-19
Deaths
Spain13,000 – 19,0000.3 – 0.4 per 1k96,000 – 128, 000
Italy26,000 – 99,0000.4 – 1.7 per 1k128,000 – 544,000
Estimate US Covid-19 mortality based on Spanish and Italian Covid-19 mortality rates

This makes more sense to me. Higher mortality reflects the larger healthcare system overload Italy experienced. It says healthcare overload probably increases countrywide mortality by something like 2x – 3x.

Using Spanish and Italian Covid-19 mortality to estimate US mortality is a stretch because of differences in culture, healthcare systems, politics, etc. But I’ll go with my chart as a reasonable first-order estimate. I acknowledge that Dr. Fauci’s current estimate of 100,000 – 200,000 US casualties is a tighter estimate than mine. I also think Fauci may be low for a number of reasons I’ll discuss below. Retrospective studies will be able to determine what factors most influenced each country’s final outcome.

[1 APRIL 2020 UPDATE: I realized the shape of the prediction curves are not symmetric. The tails are always longer than the head. That suggests that I use the higher end of my estimates, which assume we’re 1/3 of the way through Covid-19 mortality curve when we reach peak, rather than the low end, which assumes we’re 1/2 of the way through mortality. Yesterday’s revised White House estimate is 100,000 – 240,000 US deaths. My estimate is the higher end of my range, so 350,000 – 500,000.]

There are many reasons I think the US response will turn out worse than Italy’s. Before Covid-19 hits peak, we’re seeing overwhelmed healthcare systems in New York City, Atlanta, and New Orleans. So there’s evidence that the US already is experiencing the kind of healthcare system overload Italy experienced.

One key issue is the US federal government’s inability to prioritize its efforts effectively. Let’s use the administration’s current focus on the malaria drug hydroxychloroquine to treat Covid-19 as an example.

The French have given provisional permission to doctors to use hydroxychloroquine as a Covid-19 treatment after Dr. Didier Raoult issued a pre-print of his second “observational” study with 80 patients. His second study is larger and more flawed than his first. The problems with Dr. Raoult’s new study include the lack of a control group and of a presentation of per-patient data. The UK has banned the use of hydroxychloroquine for treatment of Covid-19 and a few other countries like Iran are testing it. No country has adopted it as a cure. On the other hand, the US FDA, has accelerated the acquisition of hydroxychloroquine, presumably at the behest of the president who has stated it is the magic bullet.

It’s unclear why the Trump administration seems to be putting all its Covid-19 eggs in its hydroxychloroquine Easter basket. There is one New York doctor claiming to have cured hundreds of patients with hydroxychloroquine. He has produced many videos claiming he has a cure, often providing his phone number, but not a study. Other American doctors are trying it, but none have declared it a cure.

The reality is that, like all the other candidate Covid-19 treatments, and there are many of them, hydroxychloroquine needs to go through a human testing process much more rigorous than Dr. Raoult’s. Luckily, the Gates Foundation is funding the Therapeutics Accelerator to assess and prioritize candidate treatments, something I would have expected the FDA or CDC to manage. So there is at least one organization coordinating and prioritizing Covid-19 treatment analysis. Six months is an optimistic time frame for any candidate, including hydroxychloroquine, to pass the human trials needed to show efficacy.

I’m going to harp on testing. Again. Testing is the real short-term answer, and short-term is what we’re talking about. Most of Europe is hitting peak Covid-19 infection now and US will hit peak Covid-19 infection in April or early May. Investing time and money into hydroxychloroquine is a huge gamble, especially without a study that shows it works. The administration is shifting resources from testing, which is known to work, and ventilators, which are know to save lives, to an unproven treatment.

The White House is placing a risky bet when safer bets are available.

Testing is easy, it’s approved, it works, and it is scaling. The US is too far along to flatten its Covid-19 infection curve the way South Korea, China, and Vietnam did with their test-and-quarantine programs. But testing will mitigate healthcare overloading in US hot spots, keep currently unaffected areas, especially rural areas with inadequate healthcare infrastructure, from turning into hot spots, and save lives of healthcare providers everywhere. In other words, even though testing won’t give the US as good an outcome as South Korea at this point, it can keep the US outcome from becoming a lot worse than Italy’s.

In my opinion, either Trump isn’t smart enough to understand this or he wants a cure for political reasons, to look like a hero. Or he’s good at bad bets. He put his casinos into bankruptcy, after all, proving that the house doesn’t always win. But no one has to be a hero to reduce Covid-19 mortality. What the US needs to do in the next week is increase testing and build ventilators.

Unfortunately what I believe we’re seeing with Dr. Raoult and his two incomplete hydroxychloroquine studies is a strange repeat of the Paris AIDS treatment that Rock Hudson begged for and Nancy Reagan refused. It wouldn’t have mattered if Reagan had granted Hudson his wish, though, because the Paris AIDS treatment didn’t work as advertised. Like dozens of other AIDS “treatments” in that desperate time, the Paris AIDS treatment might have worked a couple times, but without a proper published human trial, it’s difficult to understand what “worked” means. Without solid published results, I doubt Dr. Raoult’s Covid-19 remedy works, either. If he had the goods, he’d publish them. Even if hydroxychloroquine does work, testing is still the preferred short-term approach because it can get to market in the next week or two.

I’m pretty sure Trump doesn’t understand this. I’m going to step out on a limb and predict that at the state level, the governors who want all the federal support they can get, like Florida Gov DeSantis, will shove hydroxychloroquine pills down the throats of unwitting Covid-19 patients.

Sorry, I’m blowing off some steam. Such a brain dead federal response.

Another reason the US response probably turns out worse than Italy’s is that you can’t take two steps in any directions in the US without ending up in a crowded religious service. Whether you’re in Lousiana, Florida, Washington, or Virginia, you can expose yourself to the Lord and to Covid-19 at a nearby religious institution.

Theaters have closed, concert halls have closed, sports stadiums have closed. Maybe pastors think their relationship with God will protect their flocks. Maybe Trump’s Evangelical base believes the president when he says Covid-19 is just like the flu. It’s hard to know how certain segments of the US population are rationalizing their resistance to isolation, but they are.

Actually, in the case of Jerry Falwell, Jr. and his Liberty University, it’s not hard to understand the rationalization. Falwell’s power stems from his relationship with Trump so, when Trump says Covid-19 is no more harmful than the flu, Falwell has no choice but to put his 46,000 member institution at risk. Well, actually, it’s probably money, too. Liberty U will refund just $1,000 of fees to students who opt to return home. So far, eleven students have Covid-19 symptoms.

Lockdowns work if everyone does them. If the religious right decides on its own that it gets a pass, the US response will be worse than Italy’s.

To the president’s credit, he has extended US social distancing through the end of April. Also to his credit, the president now is acknowledging that there may be as many as 100,000 Covid-19 casualties in the US. Which brings me to the main reason I think the US response turn out worse than Italy: Trump took too long to acknowledge the severity of the Covid-19 pandemic.

Worried Covid-19 would derail his reelection, Trump claimed that Covid-19 would be no worse than the flu and would go away as Spring brought warmer weather. Now we know a paper called “Coronavirus Perspective” by Richard A. Epstein of the Hoover Institute provided the Trump administration with its rationalization to continue downplaying the severity of the outbreak through the end of this month, even as the numbers climbed worldwide, as most of Europe went into lockdown, and as WHO provided clear warnings.

When Trump repeated throughout mid-March that he wanted the US to reopen for business on 12 April, he relied on Epstein’s analysis. But Epstein is as much of a quack as Raoult, it seems. In a recent New Yorker interview, Epstein is unable to distinguish between the way viruses and bacteria work and claims incorrectly that the strength a virus attenuates over the course of an epidemic. The irony is that Epstein bases his erroneous claims on his misunderstanding of Darwin, the same Darwin whose theories are an anathema to conservatives and Evangelicals.

While it may not seem like a big deal that Trump dillydallied for another week or two while he espoused Epstein’s views, consider this chart.

San Francisco issued its lockdown order on the same day as Epstein’s paper came out. New York’s isolation order came one week after San Francisco’s. This isn’t a completely fair comparison. If San Francisco waited a week, it wouldn’t necessarily have looked like New York, and vice versa.

However, that one week is a large part of the difference between looking like Vietnam’s response in San Francisco’s case and looking worse than Italy’s response in New York’s case. Why? Because Covid-19 is an exponential math problem. Every day the president ignored Covid-19 meant more dead Americans. Exponentially more dead Americans.

Time to wrap it up for today. Good luck to the state officials who are doing the real work in the American response to Covid-19. They are the ones who leading the American Covid-19 response.

In the unexpected consequences department, sales of shirts at Walmart are zooming. Not because people are buying any of the zillion t-shirts supporting Covid-19 causes, but because people are wearing the new stuff in front of the camera and the old stuff under the table while they video.

In the Spanish department, because I try to provide a little local color in each post, how about some bagpipe music. Wait, what? I’d always assumed the best use for bagpipes during isolation was to scare people back inside. However, this rendition of the Spanish anthem seems to have good intentions to keep us safe and happy inside.