So the official numbers are going to be increasingly unreliable.

[LAS14]

This makes sense to me, given the inadequate numbers of kits, but I wonder if they'll come up with some alternate strategy for counting those they're presuming to have the disease?


"The Los Angeles County Department of Public Health has advised doctors to only test patients for the coronavirus if a positive result would actually impact their treatment. The department said in an official letter on Thursday that it was “shifting from a strategy of case containment to slowing disease transmission and averting excess morbidity and mortality.” The recommendation signaled a realization that there is little hope of containing the outbreak and came as a result of a growing number of patients and a limited number of tests. The decision could make it difficult to ever accurately determine the number of coronavirus cases in L.A. County, the country’s second-largest municipal health system."

https://www.thedailybeast.com/skip-coronavirus-testing-of-some-patients-los-angeles-county-department-of-public-health-tells-doctors

[hlthe2b]

1. After the first wave has slowed down, we will have to use serology (antibody testing in various

communities around the country to develop prevalence and extrapolate to the number of cases.

Such a test is routinely used in emerging community infection outbreaks but unfortunately was not developed here simultaneously to RT-PCR antigen testing for acute infection. One is being used on a research basis in Telluride, so we will likely learn a lot in a few weeks.

[Igel]

2. They've been increasingly unreliable day by day.

Testing has not kept up with increase in infections (and probably never would have).

Testing has not been distributed randomly across cohorts in an area.

Testing has not been distributed evenly across areas.

Assume that death numbers are fairly accurate--people that are very ill and dying of respiratory disease are likely to be tested and have at most a 20-day lag from infection (with some deaths occurring within 4-5 days).

Where I live has about 65 confirmed cases reported and one death, of an elderly man. Those numbers are compatible. If there were 1000 reported I'd expect a spike in deaths; if there were 20 deaths and 65 reported cases, I'd strongly suspect something was wrong. Of course, the details matter: If 20 dead/65 reported were from a large hospice facility for the aged, it would still seem reasonable.

There'll have to be studies done based on antibody assays after the fact. Even random PCR testing now would miss cases more than a few days after recovery.