Hacker News new | ask | show | jobs
by Barrin92 2303 days ago
>To reiterate an important point: "Healthy" infected people go around unwittingly spreading it for several days to a week, so you have no idea who's really sick.

According to the newest WHO report (also posted on the frontpage of HN) this doesn't appear to be the case. There are apparently few asymptomatic cases and the disease mainly is spread by people who are symptomatic.

>You have to assume the number of infected is 10-30x the number

This sounds like pure speculation and is probably not a reasonable thing to assume.
2 comments
troydavis 2303 days ago
In Seattle, the best current estimate (https://bedford.io/blog/ncov-cryptic-transmission/) is 570 infections as of March 2:

> Knowing that transmission was initiated on Jan 15 allows us to estimate the total number of infections that exist in this cluster today. Our preliminary analysis puts this at 570 with an 90% uncertainty interval of between 80 and 1500 infections.

link
pmiller2 2303 days ago
That confidence interval is far too wide to contain any useful information.
link
evil-olive 2303 days ago
As of Monday there were 18 confirmed cases in WA [0]. So even the lower bound of that confidence interval supports the claim that the number of actual cases is significantly higher than the number of confirmed cases (though "only" 4-5 times higher, rather than 10-30 times).

0: https://www.washingtonpost.com/world/2020/03/02/coronavirus-...

link
pmiller2 2302 days ago
What should anyone do differently if there are 18 cases vs 80?
link
troydavis 2302 days ago
If you're going to pick a number to plan for, it's 570, not 80. Yes, the range of possible values includes 80, but it also includes 1500.

So, what would 570 (out of King + Snohomish County population of 3 million) mean? 22% of the population is under 18, so something like 125 minors would have had it (King County has ~600 schools).

The doubling period is 7 days, so ~275 of those cases would have been in week prior to March 2 and ~570 new cases could be expected between March 3-10. If 20% of those need hospitalization[1], that would be ~100 new beds.

With 570 patients (and 275 new), contact tracing is both impossible and ineffective. Let's optimistically assume that people notice their fever immediately and self-quarantine flawlessly. Even then, we don't know who these people encountered in the few days before symptoms began and can't notify them specifically. Either we accept whatever retransmission rate occurs during that period, or do what we can to reduce it by reducing the population-wide minutes spent within 6' of other people.

[1]: https://www.washingtonpost.com/world/europe/coronavirus-in-i... and others

link
evil-olive 2303 days ago
> This sounds like pure speculation and is probably not a reasonable thing to assume.

It is speculation, but not pure speculation - there is some data to back it up.

Trevor Bedford [0] is a researcher at Fred Hutch in Seattle and a professor at UW, focusing on genomic sequencing of viruses.

He estimated two days ago [1] that the number of cases was in the hundreds, with a large degree of uncertainty:

> This approach leads to a similar estimate of the number of current infections at 330 with a 90% uncertainty interval of between 20 and 1500 infections.

This is solely an estimate of the infections here in the Seattle area, based on gene sequencing.

0: https://bedford.io/team/trevor-bedford/

1: https://twitter.com/trvrb/status/1234589820946534401

link