Author Archives: Joe Wein

About Joe Wein

Software developer and anti-spam activist

Covid-19 in Japan: Numbers vs. Testing

Posted on by
Reply

Every day Tokyoites are anxiously waiting for the latest daily count of new confirmed Covid cases: Is it less than 200 or more than 200? It was a big thing when it first exceeded 100. On April 11, it barely missed the next big number with 197 cases. But what does it really mean?

“Cases of Covid-19 confirmed with PCR tests” depends as much on the number of tests conducted as the actual number of people newly infected. Anyone not tested is by definition not included in the count. Until now, anyone testing positive has been sent to a hospital, which supposedly is legally required (only recently have light cases been moved out to specially rented hotel rooms). Therefore hospitals have been reserving tests to people with the most severe symptoms or with pre-existing conditions so that the most deserving patients will occupy the limited hospital beds they can take up.

It may seem counter-intuitive, but the primary purpose of testing has not been to track the growth of infections but to allocate hospital beds. The number does not reflect reality as it is artificially throttled. As long as this policy continues, new confirmed cases will basically only be allowed to grow at a rate at which hospital beds are found for them, which is becoming increasingly difficult, as more and more hospitals are turning away new patients to avoid healthcare associated infections (HAI). There have been a number of cases of infection hotspots in hospitals.

An article in the The Atlantic discusses positivity, the rate of confirmed infections found as a share of tests conducted in different states and countries (please don’t confuse this with prevalence, the rate of infected people as a percentage of the population).

Basically, the higher the percentage of tests that come out positive, the more likely a country or region is to be undercounting infections, for example because of limited lab capacities or limited access to hospitals that can do testing:

“[W]hile the U.S. has a 20 percent positivity rate, South Korea’s is only about 2 percent—a full order of magnitude smaller.

South Korea is not alone in bringing its positivity rate down: America’s figure dwarfs that of almost every other developed country. Canada, Germany and Denmark have positivity rates from 6 to 8 percent. Australia and New Zealand have 2 percent positivity rates. Even Italy—which faced one of the world’s most ravaging outbreaks—has a 15 percent rate. It has found nearly 160,000 cases and conducted more than a million tests. Virtually the only wealthy country with a larger positivity rate than the U.S. is the United Kingdom, where more than 30 percent of people tested for the virus have been positive.

Comparing American states to regions in other countries results in the same general pattern. In Lombardy, the hardest hit part of Italy, the positive rate today stands at about 28 percent. That’s comparable to the rate in Connecticut. But New York, so far the hardest hit state in the U.S., has an even higher rate of 41 percent. And in New Jersey, an astounding one in two people tested for the virus are found to have it.”

(The Atlantic, 2020-04-16)

So what this tells us is that a high positive rate in the tests conducted indicates a higher rate of uncounted cases that the system can’t keep up with.

What does that mean about the potential undercounting of explosive infection growth in Japan and specifically in Tokyo? According to statistics published by the Tokyo Metropolitan government on its website, by yesterday (April 16) Tokyo had 2,595 confirmed Covid-19 cases out of 7,244 cases tested. That’s a positivity of 36%, higher than the positivity for the UK (30%), Lombardy and Connecticut (28%), the US overall (20%), Italy (15%) and Canada, Germany or Denmark (6-8%). Only New York and New Jersey are worse.

Recently, Germany has been running 350,000 tests per week and is capable of running up to 500,000 tests per week. Tests are analyzed 7 days a week, 24 hours a day using shift work in the labs. Japan, with a population 52% larger, has averaged 22,000 tested individuals per week over the past two weeks (Apr 2-15).

Don’t take any comfort from numbers staying flat or growing moderately unless testing is also expanded exponentially to keep up with and exceed the growth rate of virus cases. Without greatly expanded testing, we’re like a pilot flying blind in a cloud without radar. Beware of the mountains ahead.

See also:

Covid-19 in Japan: Where are the experts?

Posted on by
Reply

America has Dr. Anthony Fauci and Dr. Deborah Birx, who provide scientific information at White House press conferences.

Germany has Prof. Dr. Lothar Wieler of the Robert Koch Institute, who speaks at biweekly press briefings. Dr Christian Drosten, chief virologist at the Charité hospital in Berlin, has provided frequent commentary to the public.

Where are these experts in Japan? The briefings I have watched in Japan primarily involved Prime Minister Abe Shinzo and Governer of Tokyo Koike Yuriko.

Yes, individuals scientists have been interviewed by the “wide shows” entertainment programs, but why are they not given a prominent official role in shaping public responses to the epidemic? Yes, the declaration of emergency is supposed to be based on the discussions of a circle of experts, but we don’t directly hear from them. It’s only the government that consults with them.

The impression this creates is that the decisions being made are ultimately driven by political and economic considerations, which take priority over any medical interpretation of the situation.

As Son Masayoshi of Softbank pointed out in a recent tweet, it is strange that in Japan Covid-19 policy is spearheaded by the Minister of Economic Revitalization instead of a health expert such as Dr. Fauci in the US. That tells you all you need to know about the priorities of the Abe government.

Covid-19: Fleeing to the countryside

Posted on by
Reply

Following the “not a lockdown” in Tokyo and 6 other prefectures, the summer resort town of Karuizawa has reportedly experienced an influx of cars with Tokyo license plates as Tokyo residents with second homes (bessō) in the area are flocking to the town in Nagano that is not yet subject to the restrictions.

I would expect a similar pattern to unfold in resort areas around Yamanakako and Izu (in Yamanashi and Shizuoka, also not among the 7 prefectures).

Not only does this exodus from the city risk the spread of infections to prefectures that currently still have a lower incidence of Covid-19, it could also have infected bessō owners take up rural hospital beds then no longer available for the local population after they also get infected.

Other countries have clear policies that discourage people from dashing to second homes, which earned Prince Charles some criticism for self-quarantining in Scotland and cost Scotland’s chief medical officer her job. Don’t get me wrong, I understand why they’re doing it, but it’s selfish and against the spirit of sheltering in place. Why should others stay put in their tiny Tokyo apartments when some can drive across the country?

While it can be said that refugees from the city will spread the load on the health care system between urban and rural hospitals, that is something the national government should be thinking about, not something only those rich enough to own two or more homes to take advantage of.

Covid-19: Japan to declare state of emergency

Posted on by
Reply

Quoting a government official, the local media (both English and Japanese-language) are predicting the government wil declare a state of emergency from as early as Tue, April 7 in big cities such as Tokyo and Osaka, to initially last until May 6 (Golden Week).

“The move would give governors in hard-hit regions legal authority to ask people to stay home and businesses to close, but not to impose the kind of lockdowns seen in other countries. In most cases, there are no penalties for ignoring requests, although public compliance would likely increase with an emergency declaration.” (Japan Times)

The number of confirmed infections with the SARS-CoV-2 virus have doubled in a week or less in Tokyo, Osaka, Kanagawa, Saitama and Chiba.

Here are the numbers on April 1 and April 6 and the corresponding doubling time in days:

  • Tokyo: 527 => 1033 (5.1d)
  • Kanagawa: 120 => 265 (4.4d)
  • Saitama: 98 => 185 (5.5d)
  • Chiba: 164 => 260 (7.5d)
  • Osaka: 245 => 408 (6.8d)
  • Hyogo: 147 => 203 (10.7d)
  • Aichi: 176 => 228 (13.4d)
  • Hokkaido: 177 => 194 (38d)

The growth rate in Tokyo and its neighbour prefectures Kanagawa and Saitama is similar now to growth in the US over the past couple of days (12-14% daily increase of total confirmed cases). The US expects daily deaths to peak at a rate of about double the current date rate in about 10 days if everybody follows social distancing rules.

Italy experienced similar growth levels around March 17, i.e. 20 days ago after it had been in shutdown for the first week. Over the next two weeks of total shutdown, Italy’s growth rate gradually dropped from 12-14% to 4% where it’s been stable for about a week now. It was only this weekend that the number of ICU beds in use slightly dropped, the first time since the beginning of the crisis in Italy.

I can see how the government would want to avoid hurting the economy in prefectures whose official case numbers are still low and therefore restricts the shutdown to only the worst-hit urban centers, but with no closed political borders it remains to be seen how effective a partial shutdown will be. Rural Japan has a much bigger elderly population than the cities. Will families in cities refrain from visiting their rural relatives? And will the population in the rest of Japan understand the severity of the situation when their prefecture is excluded from the shutdown?

Stories I have heard from friends of friends make me wonder.

One is invited to a wedding and the couple isn’t sure whether to go ahead, as there would be a huge cancellation fee to the hotel they booked. If the the government took action, this would no longer be a question.

Another was sick with a fever and breathing problems (i.e. symptoms matching Covid-19), yet the employer insisted he still come to work. I think that’s insane and outrageous.

A Japanese acquaintance I recently happened to meet still wanted to shake my hand when he saw me.

People who have been quarantined in Europe and North America have changed their thinking. In Japan this change of attitude was delayed by attempts to keep the Tokyo 2020 on course and now it will take time to turn the ship around.

See:

Covid-19 mortality in Germany

Posted on by
Reply

There have been many comments on the relatively low mortality in Germany relative to the number of confirmed cases. The median age of infected persons in Germany is 47 compared to 64 years in Italy. Many Germans were infected while skiing in Austria or northern Italy (the same is true for Norway and Iceland, which also imported the problem this way) and these younger, healthier patients have a lower risk of severe outcomes.

Germany has relied heavily on PCR testing from an early stage. Because of this the total number of confirmed cases includes many light cases that might otherwise have remained undetected. Last week about 350,000 PCR tests were performed in Germany, an average of 50,000 a day. This compares to 28,000 people in total who have been tested in Japan so far. Japan’s population is about 50% larger than Germany’s.

By tracing these mostly mild cases with large scale PCR testing, the authorities have prevented the spread to other more at risk groups of the population.

The youngest person in Germany to die from Covid-19 so far was 28 years old and had a pre-existing condition.

There is a distinct gender bias in fatality. 84% of the 31 persons under 60 to die were male (26m, 5f). 73% of the 44 persons between 60-69 were male (32m, 12f). In the 70-79 bracket, 78% of the 130 who died were male (102m, 28f). For 80-89, the largest age group with 305 deaths, the ratio is 61% male (185m, 120f) even though far fewer men reach this age than women. Even in the 90 and above bracket (71 deaths), male deaths still outnumber female deaths (38m, 33f).

On March 30, the number of new infections in Germany dropped close to the number of fresh recoveries (4,450 vs. 4,289), which recently increased. Combined with 104 deaths this meant that the number of active cases (total confirmed infections – deaths – recoveries) only climbed by 57 or 0.1%. This is the lowest rate of increase since the start of the epidemic, which if it continues gives reason to be hopeful.

COVID-19 Growth Rate Trends

Posted on by
Reply

The number of confirmed COVID-19 cases has increased by this much daily (most recent three days average) in the following countries:

1) Korea: 1%
2) Japan: 5%
3) Italy: 14.2%
4) Spain: 22.8%
5) Germany: 26.8%
6) USA: 29.6%

Korea
Korea seems to be getting close to getting the epidemic under control. Patients who have officially recovered now outnumber new infections day by day. That is very encouraging.

Japan
Due to geographic proximity and economic relations with China, Japan was also one of the countries with early cases, many of them visitors to China. It crossed the 100 confirmed case threshold on February 22, one day before Italy, yet the outcome could not have been more different. Since then the case number doubled three times in Japan (about every 8 days on average) versus 8 times in Italy (about every 3 days). Confirmed case numbers can not always be taken at face value as they can be quite dependent on the amount of testing and most mild cases will most likely never be counted. There could easily be an order of magnitude more cases than listed in the official statistics. However, 28 deaths in Japan vs 2,503 in Italy (i.e. about 90 times more) suggests that there were actually hugely different outcomes in these two countries. Japanese infection case numbers have increased about 9 times since they were below 100 whereas in Italy they increased about 400-fold since that threshold. One would expect higher mortality in Italy once the medical system was stressed to limit and beyond. Quite plausibly Japan and Italy have been counting a similar percentage of actual cases, but case numbers have been growing much slower in Japan (9% daily average over 24 days) than in Italy (28% daily average over 23 days) and with less stress to the medical system, outcomes have been less lethal on top of numbers being smaller. With intensifying efforts at social distancing, Japan may be next at halting the spread for now. It will still be a difficult road.

Italy and Spain
In Italy the absolute number of daily new infections has been more or less flat for the last 4 days. As a percentage of existing cases it is now half of what it was about week ago, but as total numbers still tripled, the new cases are still at a peak, even if steady. In Spain too the daily increase as a percentage has slowed but like in Italy it has some way to go before it comes close to zero.

Germany and US
Neither in Germany nor in the US is there a clear drop in the daily case growth percentage yet. They are still in exponential growth. Hopefully self-isolation measures will bite soon.

Other numbers
Germany has 5 times as many respirators as France (25,000 vs 5,000).

Italy now has about half as many infections per million inhabitants (~500) than Hubei province, where Wuhan is located, had at the peak of the local outbreak (1,100 per million).

Sources:

COVID-19 Growth Rate in USA and Italy

Posted on by
Reply

“And again, when you have 15 people, and the 15 within a couple of days is going to be down to close to zero, that’s a pretty good job we’ve done.”
(Donald Trump, Feb 26, 2020)

Looking at charts for the number of total COVID-19 cases in the US and in Italy, it looks like in the 11 days from March 1 to March 12, case numbers increased by an average of 33% per day in the US (75 to 1697).

A week earlier (February 22) Italy already had a similar number of cases and numbers have also been growing exponentially (79 to 1701 in 8 days, an average daily increase of 48%). The infection growth rate seems to have slowed to around 24% after quarantine in parts of Northern Italy and then a countrywide lock down on March 9 (1701 to 15113 cases in 11 days). Nevertheless, in some of the worst affected areas in Italy the health care system is already stressed to the breaking point.

Assuming the US case numbers keep increasing at 33% a day, it would exceed a million cases around April 4. At 24% it will happen a week later, on April 11. But regardless of the timing, if 5% of a million infected Americans were to need intensive care, that’s more than there are ICU beds in the entire US (and most of those are already in use, of course).

This is a very worrying picture. Obesity, high blood pressure and diabetes are known risk factors for severe outcomes of COVID-19 and these are highly common in the US. 45% of the US population over 45 are either obese or severely obese (both men and women). On top of that, some 25 million people in the US are not covered by any health insurance.

Worldometers Coronavirus data:
USA Coronavirus
Italy Coronavirus

How Taiwan handled the COVID-19 epidemic

Posted on by
Reply

As one country whose economy is closely entwined with mainland China, Taiwan was expected to take a major hit from COVID-19, but it appears the Taiwanese authorities’ response has been exemplary, resulting in a small number of infections (45 as of March 6) and only a single death so far.

Almost a million Taiwanese live in mainland China and close to 3 million Chinese a year visit Taiwan. China accounts for 23.9% of trade with Taiwan.

This report makes for fascinating reading on how quickly and efficiently the government dealt with the emergency:
Response to COVID-19 in Taiwan: Big Data Analytics, New Technology, and Proactive Testing

It helps to have competent people in charge:

In addition to daily press briefings by the minister of health and welfare the CECC, the vice president of Taiwan, a prominent epidemiologist, gave regular public service announcements broadcast from the office of the president and made available via the internet. These announcements included when and where to wear a mask, the importance of handwashing, and the danger of hoarding masks to prevent them from becoming unavailable to frontline health workers. The CECC also made plans to assist schools, businesses, and furloughed workers.

Many other countries could learn from Taiwan’s common-sense approach, yet the WHO does not even provide Taiwan with any information on the worldwide epidemic or allow its representatives to attend its conferences. That’s because the People’s Republic of China, a UN and WHO member, claims Taiwan as one of its provinces. The WHO expects Taiwan to receive all information from Beijing and report its findings that way too.

In the March 6, 2020 WHO Situation Report (PDF), the WHO lists Taiwan under “Taipei and environs” in a table on “Confirmed and suspected cases of COVID-19 acute respiratory disease reported by provinces, regions and cities in China”. Interestingly, all but three of the provinces listed there have far higher number of infections and those three have but a fraction of the population of Taiwan (e.g. Macao SAR has only 3% of the population but 22% of the number of infections of Taiwan; Hongkong SAR has 231% of the infections of Taiwan but only 32% of the population). Though it probably helps that Taiwan doesn’t have a land border with mainland China, it also has fewer infections than far away countries such as Norway or Spain that are 8 time zones away from China.

On March 1, the US Center for Disease Control (CDC) removed Taiwan from its list of countries with community spread of COVID-19.

What about COVID-19 in Turkey?

In stark contrast with the openness of the government in Taiwan, there has so far not been a single officially confirmed case in Turkey, which beggars belief considering that the WHO lists 15 countries in the Eastern Mediterranean region with COVID-19, including over 3500 confirmed cases and over 100 deaths in next-door Iran.

Turkey claims to have tested 940 individuals with symptoms by March 3 but that every one of them tested negative. The same day, a passenger on a Turkish Airlines flight to Singapore tested positive for COVID-19. It was a French Citizen transiting through Istanbul from Europe. By March 5, Turkey claims to have tested 1,363 persons, still all negative. Turkey claims to have developed its own corona virus test.

Closing the land borders to Iran and Iraq as happened on February 23 could be like shutting the barn door after the horse has bolted.

Also, almost 300,000 Chinese tourists visited Turkey in the first eight months of 2019 alone. And therein may lie the rub: The Turkish economy is highly dependent on tourism. If Turkey were to share Italy’s fate with hundreds of counted cases, it would deal a heavy blow to an important source of foreign currency and employment, which would weaken President Erdoğan’s grip on power.

But if infections have been spreading in Turkey, which I suspect is highly likely, it will not be possible to hush it up forever. Viruses don’t tend to respect authoritarian politicians’ sensibilities.

Test-driving a Tesla Model 3 in Tokyo

Posted on by
Reply

Recently my son Shintaro and I went to the Tesla showroom in Aoyama, Tokyo to take a Tesla Model 3 for a test drive. I wanted to see for myself how this electric vehicle compared to my almost 12 year old Prius hybrid and to be able to compare it to future EVs from other brands that we may eventually consider.

I’d noticed an increasing number of Teslas around Tokyo, though they’re still far rarer than around the San Francisco bay area. Given that much of Japan is densely populated, range anxiety (an often cited reason for slow electrification) should be less of an issue here compared to the US, particularly with cars that already have over 400 km of range.

I love the practicality of the rear hatch of my Prius that allows me to carry two road bikes without disassembly by simply folding the rear seats. The Tesla Model 3 has a much less accessible trunk, which pretty much rules it out for me. The Model Y will be more practical, but is also even bigger. Apparently it won’t be available in Japan until a year or two after it starts shipping in the US this month (March 2020).

Tesla’s models are quite large by Japanese standards, with implications for parking and for driving on narrow back streets. For example, these are the dimensions of the Tesla Model 3 vs. the current generation Toyota Prius (XW50):

Length: 4690 / 4570 (+120 mm)
Width: 1850 / 1760 (+90 mm)
Height: 1440 / 1470 (-30 mm)

Exact numbers for the Model Y aren’t available yet, but it’s expected to be about the same width but about 1600 mm tall (160 mm taller than the Prius).

The test drive was an unusual experience by Japanese standards. Somebody had mentioned that the dealer experience with Tesla is more like visiting an Apple store than a traditional dealer showroom. I’d say the difference was even greater.

Customer service expectations in Japan are incredibly high and that is probably one factor for Tesla’s relatively sluggish sales here, see a recent Japan Times article.

Shintaro had tried to make the reservation online and was promised a callback within 48 hours, but that never happened so he had to call again to fix up an appointment.

Even when I take my Prius to an oil change at a local gas station, I’ll be served a cup of coffee while I wait. By contrast, when we visited the Tesla showroom to evaluate a JPY 5,100,000 (USD 48,000) car, all we received was a business card of the sales person. They don’t even give you paper brochures. You can look it all up on the website, right?

Before the test drive they took photo copies of our drivers licenses. We were instructed not to take any pictures and to follow the rules of the road. We would be liable for any incidental damage to the car during the test drive. Then we got into the car parked by the roadside outside the showroom, first as passengers, then later taking turns driving it around Akasaka.

I liked the seats, which were nice and firm. The acceleration when you put your foot down is amazing. It feels like a big car but with enough power for its weight. Getting back into the Prius later, it felt quite light by comparison, by which I don’t mean acceleration but it simply feels like a lot less metal being moved around. It tips the scales at about 280 kg less than the base Model 3 (1335 kg vs. 1612 kg).

Some of the controls took some getting used to, such as the lever action of the indicator stalk (which is on the left unlike in Japanese cars) or putting the car into park or into drive with the right stalk. Much of the demonstration involved showing the use of the center screen and its user interface. Many of the functions of the car, such as the electrically assisted steering or the regenerative breaking can be tweaked there, to change the feel of the car.

Headroom in the Tesla was good but personally I don’t much care for the glass roof. In a roll-over accident I would feel safer with a steel roof, but maybe those are not so likely with the low center of gravity afforded by the floor-based battery. The car interior felt overheated when we got into it and no fan was blowing, but I only asked about fan control towards the end of my driving portion. In any other car I would have easily figured it out on my own.

Checking out the trunk and the “frunk” (front trunk) after we got out of the car, the limited access for bulky luggage from the rear was quite a contrast to our Prius, in which we regularly move large items from a DIY center or bicycles for cycling tours far from Tokyo. The Model Y will address that, but it’s also 160 mm taller than the Prius on top of being 90 mm wider like the Model 3. That’s more air resistance and more kWh used to overcome it. That’s one thing I love about the Prius, it offers all this interior space despite being compact and efficient on the outside. 🙂

The width would already make a Model 3 or Model Y a very tight fit in our driveway. We would also have to figure out if there’s enough clearance around the car to plug in the charging cable for overnight charging.

In summary, Tesla’s range of cars is not an easy sell for me as a Japanese customer. While they have great technology, some of the design choices are not a good fit for Japan and the customer experience when dealing with the company (especially given the price range) will not match a lot of cultural expectations.

UPDATE (2020-03-19):

Size information has finally been released for the Model Y. These are the exterior dimensions compared to my current Prius:

Length: 4751 / 4570 (+181 mm)
Width: 1921 / 1760 (+161 mm)
Height: 1624 / 1470 (+154 mm)

Given the width and height it looks like it has roughly 20% more frontal area than the Prius which will impact its air resistance and hence energy usage at freeway speeds.

The train to Galápagos

Posted on by
Reply

If you have read my previous article about the route of the Japanese Chuo Shinkansen, you will know that I am both interested in and sceptical of the high speed Maglev train route now under construction between Tokyo and Nagoya, later to be extended until Osaka.

I recently came across an insightful article by German train expert Sven Andersen in which he highlighted some serious drawbacks of the superconducting magnetic levitation technology to be used for the project. As he pointed out, despite its exorbitant cost the new line will only add about 25% of the capacity of the existing Nozomi trains and 19% of the capacity of the Kodama and Hikari trains (the former only stop at major stations while the latter also stop at intermediate stations).

The mix of different train classes on either rail and wheel based or maglev routes means that (without two separate tracks in either direction) faster trains will have to pass slower trains while those are stopped at stations, with a need for switches to direct them to alternative track sections around each station. A new type of switch had to be developed for maglev trains which did not even exist yet when JR Central already decided on maglev as the technology for the new line: A lengthy piece of concrete track with embedded magnets is push sideways using hydraulic pistons, which takes considerable time. This slow operation of switches limits how closely different classes of trains can follow each other or how closely spaced they can arrive at different platforms at a head station (Tokyo Shinagawa and Nagoya or Osaka).

The two main reasons given for building the new Chuo Shinkansen line were 1) to provide more passenger capacity as the Tokaido Shinkansen line is running at capacity and 2) to provide an alternative route in case the Tokaido Shinkansen is hit by a natural disaster.

If the new maglev line can only provide a quarter or a fifth of the capacity of the existing line then it will not really be able to live up to either objective.

Andersen therefore strongly favours rethinking the plans by going for a rail and wheel-based approach on the new line. Though it would limit top speeds to 350 km/h instead of 500 km/h, it would allow many more trains to be run per hour, which dramatically increase capacity. It would also make it possible for trains operating on the new line to interconnect with the existing rail network, e.g. on to Hiroshima, Kyoto and other parts of Japan, instead of passengers having to get off one train with their luggage, ride 10 floors of escalators and then board a different train to get to their ultimate destination.

Like the German Transrapid Maglev system that was a commercial flop, the Chuo Shinkansen without wheels and rails will be an island within the rail network of an island – an amazing technical feat, but not really a solution for the needs of passengers in Japan and elsewhere.