From iSteve commenter Visionary:
Steve, I’m a physician in NYC, and want to know what your take is:
I’m wondering if people who are being hospitalized and who are likely to crump are much more likely to be those who have co-infections with COVID and influenza, *particularly if they have no comorbidities or are not old*.
I think that is what explains why the US has hospitalizations rates break down the way they do (which was confusing me), why Italy is getting hammered all at once, and why places like South Korea and Japan have ridiculously low hospitalization rates of people, particularly those who are young.
The US vaccinates the elderly the most (pretty well to be fair), the young (6-17) second most, 50-65 third most, and the lowest people from 18-49 (basically only 25-30 percent of people).
South Korea vaccinates an insane amount of the old, and pretty much is better than us even in the young. Japan isn’t so great BUT South Korea and Japan also tend to have low rates of influenza.
Italy sucks at vaccinating the elderly (they’re waaay worse than us) and pretty much everyone else too, plus they have high influenza incidence rates.
So you get a situation where South Korea and Japan have low hospitalization rates, with SK doing amazingly well, we have PREFERENTIALLY elevated hospitalization rates of our young between 18-49 (since we don’t vaccinate them very well), and Italy gets hospitalization of basically everyone, but the ones who get hammered are the elderly who are not vaccinated since they’re at highest risk a priori.
The tip off is multifold:
First: when you look at our non-overwhelmed death rate after hospitalization due to influenza and coronavirus, it’s almost literally the same (7-8 percent). This suggests to me that coronavirus just might be permitting opportunistic infections to nail you. Sort of like an HIV phenomenon. Meaning while it DOES have the capacity to kill you, that’s probably less likely than how you actually do die. And the opportunistic infection is influenza for the most part if you’re being hospitalized for “coronavirus”.
Second: you look at Italy and they’ve almost stopped counting influenza deaths, but somehow their numbers for coronavirus are kind of on scale of what you’d see for influenza, only it’s sped up (like everyone is presenting like they’re end-stage influenza). It’s also right at the peak of Italian flu deaths historically.
Third: There is a case report from China of co-infection shows that these patients get hammered as a proof of concept. It’s the only one I can find because they were intrepid in testing both. We are not testing both often at all.
Fourth: COVID19 host immune reponse: it causes lymphopenia which means some white blood cells actually go DOWN, which explains why it initially is asymptomatic. It ALSO explains why the secondary infections are so terrible (your immune system basically doesn’t stop the initial infection), and why the worst patients start mimicking apparently severe influenza. In particular, influenza often takes you out by causing severe ARDS. That’s pretty much what’s killing these patients with coronavirus. It also explains why anti-virals that “work” for influenza don’t work for coronavirus in severe cases — you need a decent immune system to clear out these viruses, and those things don’t work for coronavirus. So you get this double hit, but everyone just classifies it as “coronavirus”.
Fifth: Unless you’re a medical professional who is being exposed to massive viral loads every 5 minutes (so like ED or ICU), if you’re exposed and then go home, I’ve now heard of multiple medical professionals have relatively mild disease. Cough, some fever, but no hospitalization. I think part of that has to do with it being mandatory they get influenza vaccines. Indeed, I now suspect that if that weren’t the case, the rates of death for medical professionals would be insanely higher.
As an aside, the UK is great for older patients >65 in vaccinations (>70 percent) and TERRIBLE for younger patients (13 percent). They’re getting hit with younger patients being hospitalized.
France is even worse – they’re terrible for older patients (around 42 percent) and even more terrible for younger patients (13% but only for those who are “at risk”). Unsurprisingly, they’re getting hit with younger patients being hospitalized.
I can’t figure out Germany yet – they haven’t really released their hospitalization rate breakdowns, but they’re also terrible with young (and old) vaccinations.
Thanks for mentioning that Germany doesn’t fit the hypothesized pattern. It’s important right now when coming up with theories to point out the exceptions to your theory, as this commenter has done.
This clinically really matters because if the hypothesis is correct, one response is to try to get as many people vaccinated for influenza like right now, even if they’re 18-49 in the US, but in an orderly fashion. We actually HAVE that vaccine at present, and can potentially scale it up. Certainly we should think of doing so once we get through this to avoid the second round in like 5 months.
This hypothesis should be quickly testable using big data analytics on large databases of patients, such as Kaiser-Permanente’s. Are there any legal or regulatory roadblocks to doing this kind of thing fast?
One methodological problem might be that you can get a flu shot at a lot of places (I got mine at Costco last fall), not necessarily at your health care providers’s. I kept getting notices to come in to my health care provider and get a flu shot, but the Epic software didn’t seem to let me respond that I’d gotten one from a third party.
Anyway, this is speculation. Also, if this is valid, I’m not sure how much can be done about it right now.
As the commenter suggests, if this checks out, it would be a good to-do item for late summer: universal flu vaccination.
Commenter Dr. DoomNGloom points out:
Don’t be so sure about the power of Big Data. There is a long history of observational studies having opposite results to controlled studies. Only after carefully considering confounders and selection biases can the difference be resolved. Even then, often only after a considerable time has passed.
The key is thinking about the problem with a really deep understanding of the domain. Think “Bill James”.
Right, Bill James didn’t start out as a Big Data expert who then applied his skills to lowly baseball, he started out as an intense baseball fan who paid attention to what baseball people were already talking about — e.g., how smart is it to give a big 5 year contract to a 32 year old All-Star? — and then devised quantitative ways of getting fairly reliable answers.
Medicine is vastly more complicated than baseball.
One of the superstars in medical studies is Hernan at Harvard.
Examples, neither baby aspirin nor fish oil prevent heart attacks. OTOH, the big data can get you insights , for example doses of epotin,
In medicine, Big Data observational (i.e., non-experimental) studies might be better suited for preliminary falsification: if a hypothesis like this fails in a quick Big Data analysis, then no need to go on to controlled studies.