Hacker News Reader

Comment by drtillberg | Hacker News Reader

The Boston Globe this morning ran an op-ed advocating testing a random sample of residents to get a better sense of community-wide infection rates.[1] It's amazing that this needed to be stated in a major market newspaper four weeks after most business and school was stopped due to the virus. Goverent is doing some interesting things, random sampling apparently isn't one of them, for reasons it's hard to understand.

[1]. https://www.bostonglobe.com/2020/04/15/opinion/how-get-bette...

View on HN

Everybody knows this, but the US healthcare system was so unprepared for this that there simply weren't (aren't) enough test kits around to do a proper study. Even people who have the symptoms struggle to find a place that will test them in some cities.

I like Step 0 of Paul Romer's proposal: "Stop testing people with symptoms. Presume positive and isolate."

https://twitter.com/paulmromer/status/1248712889705410560

It's probably low prevalence in the population based on our understanding of when covid-19 first appeared.

The UK has some randomized seriological surveillance and it's about 13% prevalence. That matches with the pregnancy study of 15% prevalence (which may be enriched due to hospital visits).

There's already a change in the public narrative.

New signs suggest coronavirus was in California far earlier than anyone knew: https://www.latimes.com/california/story/2020-04-11/bay-area...

Whoops.

That's not inconsistent with the Seattle timeline. But for sure genetic analysis of the viruses would reveal the source.

"Stanford’s virology lab, looking retroactively at some 2,800 patient samples collected since January, did not find the first COVID-19 cases until late February — from two patients who were tested Feb. 21 and Feb. 23"

Again not inconsistent with the Seattle timeline.

Random testing was performed in Johnson County(KS) over the weekend.

https://www.kansascity.com/news/local/article242010356.html

The idea that nobody has thought to do them seems absurd. I'm getting more and more suspicious the longer we see no results from randomized studies in US cities - it's getting more and more likely that either the tests are turning out to be inaccurate, or the results are turning out to be problematic.

But either way folks are probably being very careful because the results will have significant impact.

In the developing countries without universal health care, there are government health workers, who are into vaccination, prevention, etc. In situations like this, these workers are mobilized for random testing. However, in the states, I don't see such a workforce. It is mostly done by volunteers, such as that Stanford drive in testing.

Reply on news.ycombinator.com