Wat an effect of the horribleness that was last March and April in NY has been that even though cases have been going up again this winter and closures are back in effect, it doesn't feel anywhere near as bad as what we had early last Spring.  The infection rate is approaching 10% again but the hospitals aren't filled beyond capacity like they were before.  This must be that curve flattening in action.  What do you think the mortality rate is in NY taking into account that 1/2 to 2/3 of true cases being missed?    Maybe more like 2-3%?  That's very high for such a high infection rate!  Also, latest projections indicate that within weeks the majority of new infections will be coming from that new highly infectious strain- is this evolution in action?  We already have detected cases from the new strain in NYC and also on Long Island where I live.  The saddest thing is reading or hearing about entire families dying.....it reminds me of Ebola in Africa   I think the US may be near one million deaths by the time this thing is finally under control (hopefully by summer).  That new J&J vaccine is a single dose vaccine which is a big plus!

Do you think professional sports should be stopped again?  Not having the benefit of the bubble, the NBA has seen a multiplication of new cases.

Also as an aside, the Post Office has experienced severe delays with over 1000 of their employees out sick.  I had a package shipped to me from California (three prime lenses for my camera) on December 4th and I didn't get it until January 4th, exactly a month later!  The shipper told me he didn't know that shipping first class mail meant it was shipped on the back of a turtle!

 What do you think the mortality rate is in NY taking into account that 1/2 to 2/3 of true cases being missed?

I think it is probably between 0.5% and 1% right now, due to better treatments and health care capacity. (And, of course, it is much higher than 1% in the elderly, and much less for the young.)

Let's compare the official cases and deaths curves across the state of New York. I pull the figures from covidactnow.org. 
Last spring, new cases across New York peaked at about 9800 per day, while deaths peaked at about 980 per day. That is a case fatality rate (CFR) of 10% -- staggeringly high. But most of that is because true cases were missed, by probably a factor of 10. We were only catching about 1 in every 10 cases, which is evident by the extremely high positivity rate.

Today (or better yet, 3 weeks ago to account for the lag), the rate of new cases was 11,000 per day, and deaths now are averaging 180 a day. That's a CFR of 1.6%. So if the actual mortality rate is still 1%, then we're catching about 6 out of every 10 true cases. More likely, hospital capacity and care have reduced the mortality rate below 1%, and we're probably catching fewer true cases, maybe 1/2 to 1/3, as you say. If we're catching 1/3, then the infection mortality rate now is about 0.5%.


A good way to predict rate of death in 2 to 3 weeks is to take the rate of new cases today and multiply by the CFR, which you get by dividing rate of death today by the rate of new cases 2 to 3 weeks ago.

I was reading that perhaps the number of deaths have also been underestimated and we'll find out later that many more people died from covid complications.  I wonder if there is some truth to that.  Numbers in NYC have been adjusted upwards when the cause of death of thousands of cases was changed to covid, makes me wonder if we might end up with a million or so deaths with it in the US by the time this ends.

I was reading that perhaps the number of deaths have also been underestimated and we'll find out later that many more people died from covid complications.  I wonder if there is some truth to that. 

Yes, official deaths estimates generally get adjusted upwards over time as more data come in. This is true for most pandemics. This is also why we like to follow "deaths above normal" data (i.e. reported deaths from any cause, and subtract the number of deaths that normally to happen by that time of year) to get a better idea of the true impact of the pandemic, though that has its limitations, too.

I wonder how flu seasons (before 2020) in Norway, Sweden, and Denmark compare with other European countries? Maybe it can help gain some insights about normal social interactions in these countries vs. their responses to COVID-19.

I searched a bit and found this.  It's just for 2018, though, but indicates that there the Nordic countries aren't much different from other European countries.  The severity might reflect the availability of vaccines more than the social distancing.  Personally, I get the yearly vaccine through my employer.  I did get a bad flu early in the 2018 season, though.  I was in China in late september 2017, and though I went to the airport in Beijing three hours before my flight, all that time was spent in endless lines, and I recall whilst spending an hour in the packed emigration line, the lady before me in the line was clearly seriously ill, and sure enough, a few days after I got home, I got ill, and this was before vaccines had been distributed.  I joked that it was probably me who brought the flu to Norway that season (actually, nobody else in my family got ill).

I was reading that perhaps the number of deaths have also been underestimated and we'll find out later that many more people died from covid complications.  I wonder if there is some truth to that. 

Yes, official deaths estimates generally get adjusted upwards over time as more data come in. This is true for most pandemics. This is also why we like to follow "deaths above normal" data (i.e. reported deaths from any cause, and subtract the number of deaths that normally to happen by that time of year) to get a better idea of the true impact of the pandemic, though that has its limitations, too.

That's very logical, I think this is also how the number of deaths was upped in NY, as many of the extra deaths were originally categorized as "unexplained" and only later were grouped in with the covid deaths.  I wonder if they distinguish between covid-related and covid-caused?
One good thing that has resulted from the pandemic (aside from the 10% reduction in fossil fuel emissions), is that in NYC we've seen an explosion in community gardens.  Bronx and Queens both have high rates of obesity, hypertension and diabetes type 2 and it's attributed to lacking access to healthy fresh food.  But because of the pandemic more people are trying to move away from unhealthy food (potato chips, soda, etc.) and now we have community gardens where fresh vegetables are being grown.

I was reading that perhaps the number of deaths have also been underestimated and we'll find out later that many more people died from covid complications.  I wonder if there is some truth to that.  Numbers in NYC have been adjusted upwards when the cause of death of thousands of cases was changed to covid, makes me wonder if we might end up with a million or so deaths with it in the US by the time this ends.

Yes it is possible. After all, complications do not disappear and the cause of death of a person is still the transferred COVID. Maybe this is an incorrect comparison, but after all, a person who is being treated for oncology survives, recovers, but his life expectancy decreases. Why? Because such serious conditions and such intensive treatment

Global infection fatality rate is approximately 0.15% with 1.5‐2.0 billion infections as of February 2021.

The fatality rate is in the very lower end of previous estimates, and the estimate of total infections seems very high.  These two estimates influence each other, of course.

The author is known for his paper Why Most Published Research Findings Are False in which he makes some good points.  And early in the pandemic he criticised the lockdowns on the basis that the data were unreliable in the strict scientific sense, which may be a reasonable argument.  However, early in the pandemic he also estimated the fatality rate to be around 0.15%, which begs the question: what did he base that on?  Now he defends that estimate again.

I'm not saying he must be wrong, but I think this illustrates how difficult science is when it becomes political, and did this author step into just about all the traps he himself warned about?

Statistics Norway (the official statistics) just released their data for 2020.  The 2020 number for bed-days due to diseases of the respiratory system is interesting.  I wonder if people will dare to discuss it in light of the covid restrictions. A 15.6% drop in bed-days from 2019 to 2020!

How about bed-days regardless of cause? A 3.3% drop.  If we exclude respiratory diseases the drop is still 2.4%.  This raises some questions.  Did we suddenly get healthier, or has it become more difficult to get hospital treatment?  We should expect a drop in exotic (imported) diseases due to travel restrictions, but that seems to only explain a small part of the drop.

The number of total deaths per 1000 population was also the lowest on record (7.5).  That number obviously also depends on migration and changes in the age distribution.

The fatality rate is in the very lower end of previous estimates, and the estimate of total infections seems very high. 

Yes, I think that estimate for the IFR is too small. I am still confident it is between 0.5% and 1% (but decreasing thanks to preferential vaccination of the elderly and more vulnerable). Speaking of which, I just became eligible and hope to get my first dose of Moderna's vaccine later today.

Even if we believe this study's estimate of the number of infections, there is a problem with the number of deaths. They estimate between 1.5 and 2 billion infections. Let's take 1.75 billion. Then a 0.15% IFR comes to 2.62 million deaths, which was roughly the official figure for global deaths to that date. See the problem? We know the true number of deaths is much higher, because we have been following deaths above normal data.

NYtimes last updated their deaths above normal figures early February, estimating an additional half million deaths globally on top of the official number of COVID-19 deaths. But remember these data are highly incomplete for the most recent several weeks. They also only include about 35 countries, and those countries account for 25% of the global population outside of China.¹ So we can anticipate the true number of deaths globally is not just half a million more than the official figure, but potentially 1 or even 2 million more. We might have already passed 5 million deaths globally. This would not be very surprising because it is normal to find estimates of deaths from a pandemic to be revised upward afterward. It is very difficult to track them accurately in real time.

1 I'm excluding China because I roughly believe their figures that they have mostly kept the pandemic under control there. But even if China is hiding a massive wave of the pandemic, then this shouldn't significantly change the math for IFR because we would expect to add roughly the same proportion of infections to deaths.

We know the true number of deaths is much higher, because we have been following deaths above normal data.

It's complicated.  Using the latest official numbers in Norway, I get a death rate of 0.37% (681 deaths of 182,241 estimated cases - 99,366 confirmed, so below 0.7% for sure), but the median age for covid deaths is 84, above the general life expectancy, so yes, the "above normal" figure needs to be considered.  But in Norway there has been a below normal situation for the past year.  The other killer for the 80+ group is the seasonal flu, which has been non-existent for the last year.  The simple conclusion to explain the below normal deaths would be that covid has lower mortality than the typical flu it replaced, but we know the true numbers of flu cases much less than covid cases.


A single number seems pretty meaningless.  It needs to be broken down based on age and existing health condition.

I just became eligible and hope to get my first dose of Moderna's vaccine later today.

Congratulations.  Here the age group 70-75 is next.  I'm getting an antibody test this Thursday, as part of a research project, which is less useful.  It will be negative, or incorrect, or positive (in that order of likelihood, I think).  Was also tested last summer.

Vaccines for all age groups 16+ across most of the US

Vaccines for all age groups 16+ across most of the US

Apparently, Trump did a much better job than the politicians in European countries regarding vaccines.  Here, the vaccination of the 85+ group started in January, and now three months later >85% of the 75+ group have at least one shot and vaccination of the 70-75 group is starting.  This will take some time...

The mRNA vaccine is in me. I like xkcd's explanation of how it works. Also "definitely feeling a little sore" is right. An ibuprofen helps.

Here the vaccination started with AZ for younger people (health care personnel) and other vaccines for old people.  AZ was halted in early March, and, though not officially announced yet, I think it will be scrapped, as well as J&J, as the immune response from these vaccines seems likely to kill more young, healthy people than covid itself if continued.  Herd immunity would be reached earlier, saving lives in the older population, but that argument obviously has serious ethical issues.  This will slow things significantly.