Yes, I believe he is.
But, I'll need to see the numbers behind those numbers to get on board.
Perhaps my error is doing my evaluation of the US but trying to lay his comments about the UK on top of it. As I said earlier, if the testing is targeted only to one group - then it could skew my analysis.
Let's look at the UK.
The UK currently has 12.1k deaths and 94k positive tests - that is a 12% case fatality rate. And given that delay between a positive test and death that I talked about before, it is really more like 12.1k deaths vs. 55k cases, or 22%. That suggests to me that the UK may only be testing their hospitalized cases. That is an incredibly high CFR for a country that isn't saturating its medical system. So, IF they only count deaths from CV+ individuals and these are the numbers, I'd say that's a very high rate of death. So high that it makes me think the only people that are getting CV tests are those that are being hospitalized. Surely the odds of dying aren't higher than 1 in 5 when you enter the hospital with CV? I've seen numbers like 30% of ICU admissions end in death. So 22% of hospitalizations would still seem high.
My point is that if their positive cases only reflect the people struggling the most with the disease (likely older and already ill, not just at-risk) and you can only be a CV death with a CV+ test, then it becomes more possible that a large fraction of your deaths would have happened within the year.
So, how at risk of dying would this group have to be? Using 55k cases and 12.1k deaths as the baseline. If 2/3 of those deaths would happen this year anyway and using that same 6 week window over which deaths have been recorded, let's do the math:
12.1k * .667 = 8.1k deaths over the course of this year (we started counting deaths around March 1, so we'll call that 44 weeks left in the year) would be 8,100/44*52 = 9572 deaths on an annual basis. And that is from a sampled population of 55k. So the inherent risk of this population (on a 100,000 population basis which is how these statistics are usually recorded) would be 9572 / 55000 = 17,400 deaths / 100,000 population. So what group has a mortality rate that is that high? The mortality rate of 85+ individuals from all causes is 13,600 in the US. So of they have only tested really old people with other risk factors then maybe I can get there.
Maybe I'm missing something in the way I'm looking at it?
I did some more digging. If you look at the 65-84 age demographic, US in 2010 had populations of:
65-74 = 21.7 MM (slightly skewed female)
75-84 = 13 MM (60% female)
85+ = 5.5 MM (66% female)
Let's focus on major cardiovascular disease:
Let's look at mortality rates within these age populations from cardiov disease (this does not include the subset of the age group that has cardiov disease - it is the total age group, we'll have to correct for that later). I put the number per year we would expect after in parentheses based on the populations above:
65-74: 500 / 100k (109k deaths / year)
75-84: 1400 / 100k (182k deaths / year)
85+: 5250 / 100k (289k deaths / year)
Now if we look at major heart disease incidence rates in these populations:
65-74: 27/1000 people (weighted evenly for men and women)
75-84: 48/1000 people (weighted at 60% female)
85+: 68/1000 people (weighted at 66% female)
So we can expect the population with heart disease to be (using the statistics above):
65-74 = 586k
75-84: 624k
85+ = 374k
So, now we can calculate age + underlying condition - specific mortality rates:
65-74 = 109k / 586k * 100k = 18,600 deaths / 100,000 of specific population
75-84: 182k / 624k * 100k = 29,167 deaths / 100,000 of specific population
85+ = 289k / 374k * 100k = 77,272 deaths / 100,000 of specific population
In my previous work, I calculated that for him to be right, the people tested would have to have an inherent mortality rate of about 17,400 deaths/year. Looking at the number above, if you are highly limiting TESTING to 65+ individuals with a major underlying condition such as cardiovascular disease AND you only count a CV death as a death that occurred following a positive CV test, then I could see that your final recorded deaths could be 2/3-comprised of individuals that would have died in that year.
But you kind of have to manufacture that statistic by ensuring the CV deaths only come from a very high-risk group and then ignore the other CV deaths as just pneumonia deaths if a test was never given. However, if those deaths never seem to occur from CV and it is just 65+ with major cardiovascular disease, then I guess the testing aspect doesn't matter. I have no idea if that is true.
I should also add that major cardiovascular disease is by far the biggest risk factor for death in the elderly. Diabetes, for example is considerably lower. Do we know if the UK denotes how may of their deaths have come from individuals with heart disease? The US tracks underlying conditions - but I don't know if they publish specific conditions.