OaktownBear said:
wifeisafurd said:
golden sloth said:
sycasey said:
golden sloth said:
Unfortunately, yesterday and today have seen a pronounced spike in the numbers of new cases, to the extent that they are about 500 above what the previous record for new cases in the state was. That can't be good, and suggests a longer lockdown. Also, today is the second deadliest day on record in the state, with about 4 hours more to go.
I think this is due to a big testing spike in CA, not so much increased spread of the virus.
Hopefully that is the explanation. And hopefully the increased deaths is just a result of it being a delayed reaction to new cases.
I did find it interesting that the la times is now tracking hospitalizations related to covid and whether they were ICU or ACU patients. That number seems to be pretty steady.
I still argue you need to look at trends on a longer term basis because of reporting and testing biases. Again, I would encourage people to look at the state curves which you can see on the internet with a google search. It is not a smooth curve, but peaks and valleys. You need to a few weeks to see the curve shapes in general. Silver and others are just wrong on relying on short period changes.
Looking at cases is pointless especially given there are clearly places that are actively trying to keep their numbers down by not testing as much. Hospitalizations would be the best number, but it is hard to find that. Death is a lagging indicator, but that is the best number to look at.
Also, you have to look at where places are on their curve. South Dakota looks great on deaths right now, but that big spike in cases from the meat processing plant has the potential to blow up their deaths per capita, for instance.
it is hard to look at a small population of cases like in South Dakota, for any statistical understanding of trends. Statistics 101.
That said, case reporting, and death reporting, despite problems, at least gives you a reportable number, where someone actually is confirming cases. That is the best you have right now for trend analysis. And if you are looking for trends, as opposed to daily or even weekly results, these numbers, where statistically significant, probably are accurate enough.
If you start using the school studies like John Hopkins or Stanford that extrapolate (i.e., make up stuff) you not only get wildly inconsistent results, you get people accused of pushing an agenda with bad data applications (how many articles are there now that say the Stanford prognostication re: Santa Clara infection rates was simply fiction?). Death totals have the similiar problems as case sturdies. Not all cases are reported or reported on time. For starters, when people die, they don't aways examine for COVID, especially where there may be multiple causes of death. My friend works at UCLA COVID critical care. In numerous cases, somebody is brought down in peril from another hospital area, with just diagnosed COVID or suspected COVID in addition to other fatal conditions, only to have them code out. Is that a COVID death? The determination is at the discretion of the attending, and not always consistent. But you figure a lot of those decisions just even out and use either cases or deaths, since they do have actual results. Maybe you adjust the numbers when there is more testing (most states have way higher per capita testing then California, but at least the base statistic are people who actually are determined to have COVID, and you then look for a trend line, not daily or even weekly data that morons on Twitter use in order to find ways to support their political agenda.
Newsom has ordered a relook at early deaths and cases, with the idea of demonstrating if there were more deaths or cases early, Cali really is on the flat or lowering side of the curve, so he can justify dropping certain restrictions. That may sound like a politician under pressure, but that may also be the right way to do this. Get the most accurate reportable results you can get. And tell places like Stanford or John Hopkins they do more harm than good with these hypothetical studies that are blown-up all over the media. Who really benefits when you have a patently false college study that says there are 80 times more cases?