Apples to Apples II

More than one person has suggested that this entire endeavor is quixotic at best, distracting from real solutions at worst. Yes, I hear you, but am ignoring you. One of my usual lines when people make pronouncements is “Compared to what? Or whom?” When the accusation goes out about the evils of American interaction with oh, South America, I like to say “Compared to who? Portugal? The USSR? More recently, China? What do you mean, exactly?” So when we are looking at whether America as a whole is doing the right things about C19, it is quite natural to me to ask “Compared to who?” Especially in an election year we tend to fall into Sweden yea vs Sweden nay, or “Look how stupid Trump/Biden/Cuomo/Abbot/lockdown/opening is.” I am first trying to get out of that and see if there are large tendencies related to population density, international contact, and primary strategies.

You are free to find that a ridiculous approach. I still like it.


Iceland and Switzerland are outliers, both with a lot of cases but few deaths. Switzerland’s cases are largely along its border with Italy and they have just had a resurgence, so the conclusions people were drawing two months ago are now looking premature. Iceland’s story is that they have done a level of aggressive contact tracing that likely only Scandinavians would put up with. ( Maybe Benelux as well.) While most people in all countries are happy to help with contact tracing, when it comes to doing so against their will, some will very much dislike that on principle. When Black Lives Matter or QAnon or drug dealers or pedophiles or CNN reporters refuse to give up their phones, it wouldn’t just be American libertarians who say “You can’t make them. It’s a terrible precedent. Once that seal is broken who knows what it will be used for?” There are lots of Canadians and Scots and Dutchmen who would say the same.

Still, the tradeoff on that was that Iceland did not lock down and believes they’ve got everything normal and under control at present, so maybe I’m wrong about what folks will put up with. Privacy vs having a job? Tough call. Also, it’s an island, so even if we thought their approach was the bee’s knees we might not be able to do it.

The other Scandinavian countries are not particularly good comps, but maybe a little something. They had much less international contact during the leadup and aren’t clamoring for more since; they have no megacities and their few large cities do not have especially dense centers; they are traditionally very cooperative with government directives. Sweden is so cooperative that they voluntarily did a lot of what everyone else had to be ordered to do. Lots of places closed on their own, but they did not close schools and government offices. Their death rate has been very high for Scandinavia, on a par with France, Spain, UK, etc. As their cases and deaths have gone very low recently (but not to zero, as is sometimes claimed), some are claiming their method worked. The hope was increased herd immunity, so maybe it will work in the long run. But at present the other Scandi countries also have low rates, without having lots of people die at the beginning, so I would say they are still in the hole. All of central Sweden has had problems, not just Stockholm.

The UK, with very high deaths per million* has a megacity and a series of cities across the Midlands, all of which have been hit hard. London has one of the highest rates of international contact in the world. France likewise has high DPM and a megacity, and international contact ,but it also has two other hotspots for other reasons: its border with Northern Italy and the area around Strasbourg, where there was an early superspreader event of 2500 people indoors for an annual religious gathering. Italy has high case rates and death rates, mostly in the northern section most similar to the rest of Europe, as opposed to southern Italy which has a more Mediterranean culture and economy. While there is some concentration of cases near those northern Italian cities, as we all remember from the early accounts, villages were also hit, and their medical services overwhelmed. Italy has large cities, but no megacity. While they technically did not have a superspreader event to kick things off, their contacts with China cannot be regarded as equivalent to more general “international contact.” A relatively small amount of contact was devastating.

Sarah Hoyt over at Instapundit assured us that nothing like Italy and Spain would happen in America because our hygiene practices were far superior to even prosperous European nations. She spoke with authority because she is Portuguese, but it doesn’t seem to have turned out that way. This is one of the reasons I skip over her posts now. Raising it as a possibility to be alert for going forward is quite different from large assertions like that. She may be right about the hygiene. I did find European bathrooms rather appalling even in England. General cultural hygiene doesn’t seem to have been the key, though.

Madrid and Barcelona do not qualify as megacities, but they are large, and Spain’s alarming number of cases do concentrate there, even on a percentage basis.

No megacities in the Netherlands, Belgium, Portugal, Germany, but they have lots of urban areas, lots of international contact, and their rates of cases and deaths are all high. Canada is very regional. Quebec has a high case rate and death rate, comparable to all but the very worst US states and the worst of western Europe, and Ontario has some moderate problems, especially around Toronto, but the rest of Canada has very few cases. Quebec’s problem is mostly, though not entirely a Montreal problem. I don’t know what’s happening there. The Canadian Maritimes mirror the northern New England states**, and Western Canada mirrors the low rates of Montana, North Dakota, etc.

What few things can we learn from the general demography, then?

1. Megacities are a large vulnerability for their entire regions, and this does not seem to be simply linear. It may be impossible to control the movement of that many people for long. Dependence on public transportation is an added vulnerability and going underground is worse.

2. Lack of population is protective, though I can’t say whether that is also nonlinear. It sure looks it. From continent maps to city maps, denser areas have more cases not only in absolute terms but per capita. The followup worry is the availability of medical care if your rural area gets the bad superspreader d20 roll of 1 and now you’re screwed. While everyone else is fine and yelling “What’s the problem? It’s just the flu!”

3. Borders matter, but not so much as governments hope, especially in dense population areas. People must be leaking over, whatever the authorities claim. So one rule going forward would be “Don’t have a border with Italy” and “residents of big cities get out anyway.”

Additional – not really part of what I’ve looked at here but part of what I’ve picked up:

4. I thought ventilation was going to be big, but a couple of people here have put forth solid arguments that this is unlikely.

5. There are lots of co-morbidities that are a problem, and many seem to be part of Metabolic Syndrome, but obesity seems to be rising out of the pack. I thought COPD was going to be enormous, but it seems to be only moderate or even only mildly a problem. Huh. Type II Diabetes is a problem, but I don’t think well-separated from the often co-occurring obesity. I have to think morbid obesity is going to even more of an issue. The others are middling.

* Deaths per million is the most useful statistic, but only in a context of other similarity. As I noted before, Andorra, Peru, New Zealand, and India’s overall numbers do not tell us anything about the US. Whatever lessons we learn from those would have to come from unexpected factors that turned up. The larger nations of Western Europe and Canada are the only real comps, and even those have limitations.

**Except for two NH counties that are partly Boston suburbs

11 thoughts on “Apples to Apples II”

  1. Sarah Hoyt over at Instapundit assured us that nothing like Italy and Spain would happen in America because our hygiene practices were far superior to even prosperous European nations. She spoke with authority because she is Portuguese, but it doesn’t seem to have turned out that way.

    586 deaths/million
    617 deaths/million
    547 deaths/million.

    That’s with with several US states putting known infected people into nursing homes when it was not required, and known issues such as “guy who was shot to death tested positive for COVID19 and was reported to match.” Remove those (with death rates in the thousands per million) and the number drops like a rock.

    It’s also with the US counting things as a COVID19 fatality if it could have contributed to the death, while Italy reported only those known by testing to be infected, and per some sources where it was the primary cause of death. One of their vital statistics guys was pointing this out this spring. It looks like Belgium uses something like the US’ standard, and I can’t tell if the UK gives World of Meters their daily rate or their weekly one. (weekly includes suspected, daily doesn’t)

    Looking at population density with deaths, Italy is between Maryland (611 deaths per million) and Delaware (620). Delaware banned requiring a negative COVID19 test before admitting a patient into their nursing homes, can’t find data on Maryland.

    Spain matches California. (310 deaths per million)

  2. If the cultures and governments aren’t similar enough to have roughly identical reporting of vital statistics, then there can be no apples to apples comparison.

    Multiparty parliamentary systems, especially without ongoing election campaigns, have a potential difference where internal fraud is concerned. Three or more parties dilutes influence, making it less likely that a single party could tamper with statistics for partisan gain.

    We have factions potentially making partisan hay from “the lockdown is reasonable and serious” and “the lockdown is fakery purely intended to hurt Trump”. Both explanations cannot be correct. At least one of them may be fraud, with partisan actors chiming in to support.

    The quest for the apple might require a) factional control of provincial vital statistics reporting b) a heated political fight between factions.

    If you have to make comparisons, it might be good to wait a couple of election cycles in case the advantage in preserving fraud expires.

  3. I agree that differences in reporting will affect and even skew results. Johns Hopkins has been trying hard to fight through that, I believe, and I have significant trust in them.

    One could also say “drop those few Chinese flights to Italy and its numbers drop,” or “take out that superspreader event in France and its numbers drop.” My thought is that there are vulnerabilities that make those events take hold, where they might not elsewhere. The common assertion that America is overcounting, padding its numbers by being less strict about the criteria is not solid. I have seen it claimed often, but evidenced only by small-scale events which the claimant believes is more general. Skepticism of anyone’s numbers may be warranted, but that is not evidence in itself.

    Also, we do not apply that standard to lost business asking “Oh, did they go bankrupt solely because of Covid or did they have a pre-existing condition like poor location or being stupid? They probably were going to go out of business anyway, then.” When you try and make that switch you see the weakness of the argument.

  4. We know about how many businesses go bankrupt in a given year. Well, I don’t but someone does. I don’t think it’s too unreasonable to estimate that the number of excess bankruptcies is approximately the number caused by Covid. You could argue whether they are caused by the lockdown, or by changes in people’s behavior, or you could say it isn’t even a meaningful distinction. But the number of if bankruptcies this year is vastly higher than in a typical year, which I think it is, that means something.

    We could say the same thing about excess deaths. It may impossible to say whether a particular individual who died with Covid died of Covid, but looking at the number of people in a given demographic group die in a typical year gives you at least an initial estimate of how many would have died anyway. Of course, you have indirect effects e.g. the lockdown probably significantly affected the number of car accidents and suicides, although I think one is going down and the other is going up. And people will die of cancer next year because they missed a screening this year.

  5. Not promising news:
    SEOUL, South Korea (AP) — South Korea is closing schools and returning to remote learning in the capital region as the country counted its 12th straight day of triple-digit daily increases in coronavirus cases.

    South Korea is a country usually pointed to as having done things the “right” way…

  6. its pretty simple. If your population widely complies with sensible methods to alleviate a pandemic, you will have a better outcome that one does not. If the population is so non compliant, as to require serious regulation, then the outcome will be worse again. if you tell your population there is nothing to worry about, you are an idiot.

    We in Canada are contemplating more regulation, as our population is becoming rather too brave, and we have spikes all over. As schools go back there will be more spikes all over the place. Kids are good at spreading viruses. Luckily this one mostly is easy on kids, but there are many they will infect, who will die.

  7. When this started, we were told that it might continue until 80% of the population developed resistance either by being infected or by immunization. The explicit reasoning behind this was that this was thought to be a completely novel virus with no existing resistance in the population. At that point this was the worst case scenario and actual knowledge was negligible.

    The basic equation is simple. A susceptible individual will come down with it if he absorbs a sufficient number of virus particles. There are two variables there. What constitutes a sufficient dose varies from virus to virus and from person to person. If they have established what constitutes a infectious dose of wuflu, I haven’t heard other than more is worse. In any event, enough people have come down with this to prove that it can be spread fairly easily.

    The other variable is susceptibility. The initial assumption was that 100% of the population was susceptible. The early ship board examples however, topped out at around 20% infected. I believe the exposure rate was probably 100% there and went on for weeks. Since then nobody seems curious why the other 80% didn’t get sick.

    I think part of the answer is here:

    On Page 15:
    “Of note, we detected cross-reactive T cell responses against spike or membrane in 28% of the unexposed healthy blood donors, consistent with a high degree of preexisting immune responses potentially induced by other coronaviruses (Braun et al.,2020; Grifoni et al., 2020; Le Bert et al., 2020)”

    My interpretation is that there was a preexisting reservoir of resistance, probably from other Corona viruses in circulation.

    So a lot of the patterns we are seeing may depend on whether particular viruses may have circulated in an area. I don’t know if there’s anyone willing to give even a rough order of magnitude of the number of viruses circulating. The ones that cause something noticeable like a cold are in the thousands, some portion are Corona viruses. This is a lot more plausible than low death rates as a manifestation of virtue.

    I thought at the time that an infection rate of 80% was unlikely just because it doesn’t seem to have happened before. Most of the really bad epidemics in history seem to top out at around 30%. Most places this time around seem to get rapidly better when the rate of antibodies gets to around 20%.

    20% still gives plenty of people to be on the unfortunate side of the balance of recovery, thankfully less that 80%.

    So the answer to AVI’s question might be pure chance. Beyond the trivialities that with something like this, young is better than old, healthy is better than sick, good care beats bad care, it may simply come down to whether or not someone was exposed to a virus that they may have never noticed, or whether that virus (or those) ever made it to your town or country.

  8. via Taylor Cowen:
    Deaths in the frontline: Occupation-specific COVID-19 mortality risks in Sweden
    We use data the Swedish authorities organized as an early release of all recorded COVID-19 deaths in Sweden up to May 7, 2020, which we link to administrative registers and occupational measures of exposure. Taxi and bus drivers had a higher risk of dying from COVID-19 than other workers, as did older individuals living with service workers. Our findings suggest however that these frontline workers and older individuals they live with are not at higher risk of dying from COVID-19 when adjusting the relationship for other individual characteristics. We also did not find evidence that being a frontline worker in terms of occupational exposure was linked to higher COVID-19 mortality. Our findings indicate no strong inequalities according to these occupational differences in Sweden and potentially other contexts that use a similar approach to managing COVID-19.

  9. I’ve recently realized that everyone is talking about viral spread.

    I understand you often see bacterial and viral respiratory infections together.

    You could say “that doesn’t matter, the bacteria on you already will not make you sick.” Staph would seem to be a counter argument for that.

    If you switch enough previously unmasked people to a heavily masked lifestyle, you should get an increase in bacterial respiratory infections, whether or not there is any viral spread.

    In particular, right now universities are of interest. If some kid is coughing up green stuff, is the university going to call it COVID? The university staff are politically motivated Democrats model says yes. The university staff are motivated by economic interest model (charging undergraduates residential fees and bringing international graduate students to do research) says yes, because they can pretend to let students move in, and qualify for visas, before just doing online courses. Flip the first model along partisan lines, and we predict that a Republican will say that a death showing the three stages of symptoms, with no response to antibiotics, is bacterial. Flip the second model along slightly different lines, why isn’t restricting visas proportionate and in line with restrictions imposed on domestic workers?

    In theory, you could discuss effect of masks on self sourced bacterial infections in terms of OSHA, etc. air quality requirements for HVACs and so forth. In practice, I don’t have a graduate degree in industrial hygiene, don’t have time for the coursework or self study before the crisis is likely to pass, and don’t know any industrial hygienists I am willing to trust to that degree.

    Beyond that, isn’t their enough political hay to be made to make trusting any recent study a little fraught?

Comments are closed.