>Something weird happened in California, where 20% of their tests were coming back positive, then all the sudden 04/04/2020 they started seeing <10% positive tests:

This is not surprising. When tests were scarce, tests would be rationed to those most suspected of infection.

The cases reported per country, or state are only a narrow window into what is actually happening. The most advanced medical testing company in the world, Roche, is Swiss-German so Germany's testing capability relative to most countries is off the charts. Combined with a highly distributed, yet also highly co-ordinated health care system they have been ahead of the curve in testing on every conceivable metric. As a result they have reported relatively high case totals, because they have been doing mass testing for a long time, but very low death rates because this enabled them to effectively isolate cases quickly.

As countries increase their testing capacity it is quite possible some countries (or states) will end up eventually reporting accelerating numbers of diagnoses even if the actual new case rate in the population is falling.

One way to mitigate that is to report only cases requiring hospital admission. This isn't ideal by any stretch, but is a reasonable proxy for the spread of the virus and at least gives a good indication of the impact on the health care system.

> One way to mitigate that is to report only cases requiring hospital admission. This isn't ideal by any stretch,

It doesn't count all the frail people living in care or nursing homes who are put onto a palliative pathway and allowed to die, for example.

This could be about half of all Covid-19 deaths.

https://ltccovid.org/2020/04/12/mortality-associated-with-co...

> Data from 3 epidemiological studies in the United States show that as many as half of people with COVID-19 infections in care homes were asymptomatic (or pre-symptomatic) at the time of testing

> Data from 5 European countries suggest that care home residents have so far accounted for between 42% and 57% of all deaths related to COVID-19.

But hospital admission cases depends on perceived hospital resources too... if people think going to the hospital is worse than staying at home, or impossible, that will distort conclusions also.

That's true, but is mainly a problem when comparing one country or region with another. Within a region, assuming a roughly stable general perception of hospital capability, it should till give a reasonable indication of the relative rise and fall in numbers of severe cases. I don't know, it's a really hard problem.

Tests are still scarce.

Maybe but it’s all relative. Even a little less scarcity can change those numbers in significant ways.

I can't figure out what you mean, can I go someplace to get tested and put my mind at ease that I'm not about to die or kill my familymembers? That's the only "relative" that matters here.

Are they scarce everywhere? I was under the impression that LA county, at least, could test everyone that was symptomatic.

No, they have been available to only “seniors and high risk individuals”.

Literally just today Mayor Garcetti announced testing for everyone is available.

That's great for LA county but there are 320 million of us who don't live there. That is: should they be scarce anywhere?

Remember, this is the USA.

This whole thread is about California's roadmap. LA County is something like 25% of the state. Why stop at the USA? With people moving between places, outbreaks anywhere will matter until there's a reasonable vaccine/treatment.

Maybe or may not be true.

I just got an email from local urgent care that they have the fast test equipment that can offer a result within 15 minutes if you are negative.

The situation is changing quickly

My friend in SF was told he wasn't severe enough to get tested a few weeks ago, but he's convinced he had it.

Assuming that's the Abbott 15 minute test, there's only going to be 50,000 of those a day for the near future (and it seems less than that in the immediate future).

That's a big chunk on top of current testing, but probably not enough to make it easy to move forward with a testing based strategy.

And the stories I've read say their accuracy is questionable.

When testing is scarce it is used as a clinical tool to help make decisions about treatment. For example: patient has pneumonia. If its bacterial pneumonia a course of the right kind of antibiotics will be very helpful. If its COVID, those antibiotics won't do anything, and the patient may not get the therapies they really need.

Once testing becomes more available, doctors use it to rule out COVID even on patients who aren't very sick, to help make sure they those patients aren't spreading it, instead of just making treatment decisions.

Ideally once testing capacity gets high enough we could start testing every contact of an infected person, or just randomly sample people.

Those charts look very different from the Worldometer data, which shows the April 4th high water mark in new cases still holding 10 days later.

Their graph is linear growth in new cases per day until 4/4, followed by a gradual but spiky decline, even as tests/day keep increasing.

That’s because worldometer data is community reported. This data comes from the states (mostly):

https://covidtracking.com/