🇳🇱
Ya know how some of the less intelligent people on social media keep griping about how the death toll of COVID-19 is no worse than a normal flu... well next time show them this graph.
@freemo This does not take into account inflated death counts, though. People dying from accidents are being labelled as #Coronavirus deaths. My own great aunt who probably died from a nose bleed (but there's some suspicious stuff there, and a chance foul play was involved) has been labelled as a #Coronavirus death.
But this shouldn't surprise anyone, since you apparently don't have to be tested for #COVID to be labelled as a #COVID death. 🙄
@realcaseyrollins No the way we count deaths is the same as corona as it is for the flu and other things.. Its not so much accidents people point out but things like someone with a heart issue that is provoked by covid and dies.. but this is no different than what we would do with the numbers for the Flu. In general the reporting has been pretty consistent in this regard.
@freemo I thought you have to test positive for the flu in order to be listed as having dying from it? Or am I wrong?
@realcaseyrollins No, most people dont even get tested for the flu.. a doctor just needs to diagnose you as having the flu, which can be done without needing to test.
@freemo Interesting. I didn't know it was so subjective.
What if the diagnosis is wrong? Esp. with something like #Coronavirus where the symptoms are extremely similar to the flu?
@realcaseyrollins when it comes to data science have infallability isnt really needed to having good numbers.. think about it like this...
Some people get diagnosed with the flu (or coronabirus or whatever) without being tested. Tests are, in some cases, performed later on either as part of a study or for other reasons. We then can look at all the diagnosis across the country for a particular disease and get a sense for how often those diagnosis are true in the cases of where they were later tested. From this we can conclude the false-positive and false-negative rates of diagnosis and from this determine how reliable a diagnosis without a test is, from this number we can adjust the death rate figures to get accurate numbers when someone hasnt been actually tested.
Now when we talk specifically about deaths however its a lot easier to diagnose COVID-19 then it is for other diseases. If someone dies of the disease they are late-stage and late-stage COVId is extraordinarily easy to diagnose vs the flu or a cold as it is the only one that will cause lower respiratory tract infection and scaring.
It is hard to diagnose COVID more generally but thats only because it is highly asymptomatic. An asymptomatic person is impossible to diagnose with COVId without a test. But these people are also not attributing to the mortality rate in any way.
@freemo I generally agree with this; however, it assumes that doctors are diagnosing patients with #Coronavirus in good will.
The rehab facility where my great aunt was living literally lied about her temperature so they could put her in isolation. She didn't have symptoms. Some of my family members were video calling her mere hours before she died, and she was doing just fine, save for the nosebleed (which I don't think is a symptom of the #coronavirus, it was actually an existing condition that she had before this whole thing).
@realcaseyrollins Well no not at all, it would just as easily catch it if they werent doing it out of good will.
Ultimately we can check the success rate of diagnosis int he way i explained in my last post. If doctors are not exhibiting good will and diagnosing people with COVID when they know they dont have COVID then you'd see this in the numbers that their false-positive rate would be much much higher. But that isnt what we are seeing t all. In fact what we are seeing is the opposite, that we arent actually catching the vast majority of COVID cases because it is being underdiagnosed not overdiagnosed due to the asymptomatic cases being overwhelming right now (at least 25% of cases are entirely asymptomatic, perhaps much more).
@realcaseyrollins Thankfully the quarantine has helped a lot for sure.
@freemo That's a good point! I wonder if assuming folks would break quarantine contributed to the inflated predictions?
@realcaseyrollins Well depends on the prediction.. Any model that would assume 100% lock down without any social interaction would also show now growth of the virus at all. this of course isnt reality and any model will assume some breaking of quarintine either by choice or due to being an "essential worker". So there is a lot of room for variations in the model on that alone.
Then on top of that there is the fact that we dont completely know the R0 value of the virus either (how easily it will spread from person to person when no measures of any kind are taken to stop it). So... models are going to have a huge error rate on this sort of thing.
@freemo At the end of the day, inflated or not, the #Coronavirus deaths are way below the model predictions (at least in the #USA), so that's encouraging, at least. 🙂