Virus update

Posted on by

Jeff Harris, who did some research claiming that the NY subway system affected the outbreak there, is still ringing that bell.

That leaves us with the public transportation system, particularly New York City’s public subway system. We continue to stress the word system, because we should think of the subways not as a loose aggregate of individual stations docked in individual neighborhoods, but as a whole, as a mechanism for efficiently pooling millions of individuals into one large mixing basin.

New York City’s unique subway system had the capability in late February and early March to rapidly disperse SARS-CoV-2 throughout the city’s boroughs

Elsewhere, Harris notes that the falling ratio of deaths to reported cases suggests that treatment is getting better. That goes along with my prediction that treatment is more likely than a vaccine to be the solution.

Tyler Cowen notes that testing with rapid results could change the game, also, and he put some grant money where his mouth is. Read the whole post.

Suppose that with current best treatment practices (aided by rapid-results testing?), out of 10,000 otherwise healthy people who get the virus, fewer than 5 suffer adverse long-term consequences. Are we still supposed to structure our lives around fear of the virus?

Of course, health experts cannot or will not give us an estimate of how many out of 10,000 otherwise health people will suffer adverse long-term consequences if they get the virus. Because we are in the dark, every outbreak of cases becomes a justification for allowing our lives to be directed by health experts.

I note that this paper looks at India’s performance with respect to the virus by comparing age-specific case fatality rates across countries. The U.S. is not one of the comparison countries. Probably because we do not have the data broken down that way?

24 thoughts on “Virus update”

  1. There is some US age cohort data around, and the results are similar: young people face very low risk, often less than normal seasonal colds and flus. The result here says that if you correct for India being so much younger on average, the rate is higher, not lower, than that in other countries, suggesting treatment impact. (Score minus one for Hansonian medicine?)

    This is also the answer to why Africa and I’m sure Afghanistan are doing ok. The population is mostly very young people, and not as many folks in the tougher spots make it past 60.

      • That is a good point. One site puts Japan’s Covid deaths per million population at only 8.8.

        The India paper has certainly changed my impression of India’s response.

        My first reaction to it was to look at the age structure in Sweden and Norway, since the difference between death rates in the two countries, it is so often argued, is dispositive in condemning the Swedish approach.

        Norway: 65 years and over: 16.94% (male 420,178 /female 489,759) (2018 est.).
        49.11 Covid deaths per million.

        Sweden: 65 years and over: 20.37% (male 946,170 /female 1,098,986) (2018 est.). 573 Covid deaths per million.

        So, age structure might explain some of the death rate difference. Sex, another possible factor given higher male death rates, is about 46% male in each country, so scratch that.

        Size of major cities is also a likely factor. Stockholm has 1.56 million, Oslo only 693,000, only about 100,000 more than Sweden’s second largest city, Gothenburg with 590, 580. So perhaps city size is a possible explanatory factor?

        Still, even given those adjustments, it doesn’t look like the Swedish approach can be exonerated.

        Looking at Japan, much has been made of different strains of the virus being more or less lethal than others, and some say that is what explains. Nevertheless, it may be useful to consider Japan’s governmental as well as individual citizen responses to the virus.

        First, although Abe was visible, it has been noted that it was prefectural governors who actually managed the concrete measures in each region. Japan called for a voluntary lockdown in February and closed schools until March. This did do a lot of economic damage so it appears private citizens reduced activities even if self-isolation was minimal among people with symptoms, per one study, and compliance with many recommended safety behaviors minimal.
        See article entitled “Original article
        Adoption of personal protective measures by ordinary citizens during the COVID-19 outbreak in Japan .”

        Japan limited testing to high risk populations. They still have only done about 9,385 tests per million, compared to the test-crazed USA which is at 63,285 tests per million. Japan does seem to have been innovative in its tracing efforts by focusing on clusters, rather than individuals. It also seems to have been wiser than the USA by focusing on warning citizens about high-risk environments, in particular the Three Cs: closed spaces, crowded places, and close-contact settings. The preoccupation in the USA with closing parks and beaches seems unduly puritanical.

        Lots of other possible explanatory factors.
        I have not found much yet on housing ventilation, or relative use of ventilators in hospitals, or nursing home measures.

        Perhaps there really is a lot more that we don’t know than would justify much judgment making.

        • Other interesting Japan tidbits:

          “Of about 2,600 people hospitalized at around 230 facilities [in Japan] between March and June, 7.5 percent have died, compared with 28 percent in China, 26 percent in Britain, and 21 to 24 percent in the U.S. state of New York, according to the National Center for Global Health and Medicine.” Less use of ventilators?

          And, besides subway cars that are well ventilated and in which talking is minimized, Japan is not just sanitizing subways, but treating with antiviral surfacing: https://www.straitstimes.com/asia/east-asia/is-that-hand-strap-clean-tokyo-metro-sprays-silver-to-fend-off-coronavirus

          • Japan asked for voluntary distancing in early April, not February.

            Also, Japan’s unemployment rate has risen from 2.4% to 2.8%, although hours were often cut.

  2. According to a friend in the business, COVID treatments have not gotten that much better. It’s just that healthier people are getting the virus now, and they’re not dying as frequently.

  3. As the number of people tested increases one should expect the CFR (case fatality ratio) to decrease. In March,April and May almost all the tests were given to those showing symptoms as testing increased more asymptomatic infections received a test. Some data shows that asymptomatic infected people die less frequently. So, the CFR will decline with increased tests. Here’s a paper that makes the same argument.
    https://towardsdatascience.com/why-testing-completely-skews-coronavirus-case-fatality-rates-c7cbf53ac4c8

    • The link is to an absolute must-read on the whole HCQ controversy. Highly recommended.

    • It was the subways.

      1. Lots of places have cars, but NYC got hit especially hard, especially fast.
      2. The scatter plot in figure 1 in that linked article looks like a mess, not a clear line at all.
      3. We know a lot about how contagious respiratory diseases viruses spread around in general, and about how cv19 does in particular, because we have a good catalog of track-and-trace successes and super-spreader events. It’s when a bunch of people are cooped together tightly and exchange the same air without most people wearing masks, and exacerbated when many people are oriented randomly and facing each other as opposed to when they are all facing the same direction, as in a theater or cabs or some buses. You are going to get a lot of rapid spreading around the population if people are doing this several times a day, and every time they do it, you are shaking the social kaleidoscope and mixing them up with a bunch of random strangers. That’s the perfect storm of what happened on the NYC subway. That’s just not what happens with cars. Remember back when people were actually thinking we could build a “track and trace regime” in the US? Should one laugh or cry? But at the time people said that the all subway mixing would make this completely impossible, even in the theoretically perfect scenario of 100% mandatory smartphone surveillance, because one’s odds of being within contagious distance of almost any the millions of other transit users increase very quickly with number of rides, and that number was high in the “potentially contagious” period of infection. Again, that’s not how cars work.
      4. If we’ve learned anything during this pandemic, it’s that international comparisons aren’t very useful, because different places are really, really different in lots of ways. For instance, while it’s true that lots of dense Asian cities are also very dependent on public transit, they also tended to go all-in on all-masks-all-the-time with almost universal compliance, which one could see on the social media pictures. Meanwhile in NYC, compliance was not good and was not enforced, for a long time, and wasn’t even that good after being enforced. The same thing happened in DC, with metro being reluctant to insist on a mandatory mask policy for a long while, and then not doing anything about it even after they did (saw it with my own eyes). Also NYC is special because this disease tended to spread between high status, globally jet-setting, super-socializes like Tom Hanks first, then trickle down to everyone else. That means that for the US, NYC was going to get it first and get hit hard and fast. The local peak in any other place was going to be proportional to the social-status-distance from those people and places.
      5. Fundamentally we are dealing with a data quality problem of GIGO. In NYC, we know that when we try to figure out how many people died because they caught it in nursing homes, if the individual got so bad that they needed ICU care, got transported to the hospital and died there, the health department doesn’t record it as a “nursing home” case. What we want is to compare data about serious cases, corrected for age and comorbidities and living conditions, against extremely granular details of personal lifestyle and behavior “how many hours were you on the subway on average during these critical weeks?” The use of rough, indirect proxies that embed assumptions about equal characteristics for different areas is not at all an adequate methodology when it runs directly counter to the common sense understanding of how viruses like this spread.

      • NYC didn’t really have any kind of Plan B for what to do when the subway became dangerous. It’s the inherent downside of having America’s best public transportation asset: people build their lives around it.

        • This is either the best or worst year ever to joke, “Hindsight is 20/20”, but NYC / MTA officials probably could have made things better had they:

          1. Spread riders out. One mistake NYC made was to cut service to the bone, which kept essential workers without plan-Bs crammed close together. An alternative could have been to have kept as many cars running as possible, but reduced ridership by at least 50%, and also reduced crowding on platforms waiting for infrequent trains. Yes, ythis would have been expensive, but they did got billions of dollars of bailout already specifically to cover losses like that, and the subway has a lot of reasonable room to raise prices to profitable levels. It’s a lot cheaper than DC’s metro for example. If your job is ‘essential’, you should either be willing to pay a few bucks more for a ride, or your employer should help chip in. If that’s going to break the bank, the value you are generating per hour is probably not enough to justify exacerbating an epidemic.
          2. They could have required face-coverings and implemented rigorous enforcement. Yes, there was a total fiasco about commercially available masks, shortages, bad advices, etc. early on. But even scarves or bandanas or other improvised coverings would have helped somewhat. As it happened, they wasted not just critical days but well over a month before anyone started to think seriously about mass masking while East Asia ramped up their domestic production two orders of magnitude.

      • “late February and early March” => he seems to imply the virus spread very fast, and there must be some special factor that explains why very fast, hence the subway. What if the virus did not spread very fast:
        https://duckduckgo.com/?t=ffsb&q=sewage+sample+covid&ia=web

        Can we agree that if the virus was in New York in say december, not february, then it didn’t spread very fast and no special explanation is needed?

  6. In July, my family got sick and had several of COVID’s legion of symptoms. We were consistently not treated because of potential COVID by our normal physicians because we didn’t develop dire symptoms. For example, my 18 month daughter would need to have her 103 fever for 72 hours for her pediatrician’s office to see her. When our area had no tests at the usual mass testing centers, I was encouraged to go to an ER to get a test while less sick than my daughter. The average turnaround for our tests was 7 days. We were all negative.

    From my cluster of personal friends in the wild, the medical system is fraught with confusion over protocol and what Gigerenzer calls defensive decision making. I’m skeptical that mortality is falling due to improved quality of care.

  7. More COVID content! I find your coverage of COVID more informative and interesting than just about any other source on the internet

  8. –“Elsewhere, Harris notes that the falling ratio of deaths to reported cases suggests that treatment is getting better. That goes along with my prediction that treatment is more likely than a vaccine to be the solution.”–

    I don’t necessarily disagree with the second sentence, but do we have enough information to make that determination? We don’t have a complete view of cases by age and health status.

    Earlier on, there was a huge problem of COVID circulating through nursing homes, which led to tens of thousands of deaths. Is that still going on with the current wave? If it is, I’m not hearing about it. If nursing homes aren’t being hit as hard this time around, that alone would lead us to expect a lower ratio of deaths to cases from the current wave.

  9. University of North Carolina (Tarheels) just shifted all classes to on-line. All the students had arrived for the start of classes . More than 130 students tested positive.

    https://www.npr.org/sections/coronavirus-live-updates/2020/08/16/903071127/less-than-a-week-after-starting-classes-unc-chapel-hill-reports-4-covid-19-clust

    Plenty of other schools went to on-line, but this is the first that i heard of that changed over after the start of the semister.

    This doesn’t bode well for colleges this fall.

  11. Healthcare has to get away from the idea that someone has to go somewhere and queue with a lot of other sick people before getting medicine. That just spreads disease and makes more work for them.
    Covid19 is giving telemedicine an enormous boost, much to the disappointment of doctors who like to control people.
    Medicines that people can have in their cabinets and apply at the first symptoms stop disease spread. Only go somewhere if this does not work.

  12. We only a little bit less running around with our hair on fire.

    Like I wrote months ago, we are no longer a very serious people.

Comments are closed.