I could be open to a conditional approach to mortality rates, as you suggest, for infections diseases. Indeed, cancer survival rates can be reported that way, i.e., conditional on diagnosis.
The problem chiefly with your suggested design is that of data collection. We would need a credible authority that collects and cleans the data needed of infected people across all infections. That collection effort would also have to be systematic and in a consistent manner, for your suggested design to work.
To be fair, that kind of systematic and consistent data collection effort on infected people exists for Ebola. But problems still exist. As this article notes,
It can be difficult to clinically distinguish EVD from other infectious diseases such as malaria, typhoid fever and meningitis. Thus, we may not be getting good diagnosis data.
But let's assume that the Ebola collection effort is accurate and robust. Thus, using your suggested approach regarding Ebola, the latest data show 1,743 deaths out of 2,592 cases or a 67% death rate.
We now have a problem of comparability. Let's take influenza. There isn't an Ebola-like infected person data collection structure for influenza. Indeed, as this article notes, reported influenza infection and even death data are subject to several limitations.
Thus, your suggested metric of deaths/infected people for 67% for Ebola (or if it's going to be calculable for coronavirus) , isn't comparable to the "irrelevant" deaths/population metric of 0.0019% for Congo. But, that "irrelevant" metric IS comparable to the 34,157 influenza deaths in 2018-2019 to 330 million people in the USA metric or 0.01%.
Wait...you mean the uncertainty nditional US death rate from influenza is higher than the unconditional Congo death rate for Ebola? It sure is. Why? It seems that it's not that easy to become INFECTED with Ebola, but if you do, your chances of surviving aren't that great. In contrast, the estimated conditional influenza death rate in America is low, but it seems to be way easy to become infected with influenza.
Thus, net/net/net, freaking out over Ebola seems to be as mature as freaking out over guns. And I bet the current freak-out over coronavirus will prove to be as equally-mature as gun and Ebola hysteria. Parenthetically, the "irrelevant" metric of deaths per 100,000 is the industry standard.
Troll.
Um, no. But thanks for the dialogue.
And, yes, I agree nailing down the rate of propagation is easier said than done.
In this case we have
a fluid situation
people dying from *likely* sequelae (pneumonia) but either with no testing kits, or no pathognomonic symptoms (e.g. petechiae, for some hemorrhagic fever or other)
a government with both a propensity to lie and reason to downplay.
So, you take your guess...
But the other thing we can do, is to observe the *actions* of the government.
And in this case, they're not lying, sweeping it under the rug, pretending nothing is wrong, and hoping everyone else believes the lie.
They're running around with their hair on fire, trying to keep people from talking about it, and hoping it gets better on its own.
And that part, is what convinces me, to consider higher numbers for the lethality and R0 values.
Thanks for fleshing out your thinking; you do have a systematically sound approach...but as you point out, the collection has to be systematic for that kind of thing to work.
Right now we're still in the Wild, Wild West phase of things...