Hacker News Reader

Comment by prox | Hacker News Reader

It’s good to see solutions for testing are coming.

This test by Singapore a*star nanobiolab can test in 5 minutes : https://www.straitstimes.com/singapore/health/singapore-scie...

View on HN

Besides comparing times, I guess the accuracies or FP/FN rates are also important. The linked article states 95% accuracy for the Bosch test. Are there any numbers for the 5-minute test?

It only states high sensitivity and accuracy, you might find actual data on the a*star nanobiolab website.

This article [0] suggests that false negative results will not be as problematic if we manage to start screening on a population level.

Inaccuracy of up to 15% could still keep the viral R0 under 1.0 (the critical value for eliminating an epidemic).

From the article:

> "False negative rates up to 15% could be tolerated if 80% comply with testing, and false positives can be almost arbitrarily high when a high fraction of the population is already effectively quarantined"

https://medium.com/@sten.linnarsson/to-stop-covid-19-test-ev...

Shouldn't we optimize for a higher number of false positives? In the end, sending someone in quarantine for 2 weeks without them being infected is probably less dangerous than a false negative. So making tests overly sensitive would seem more applicable to me here.

No, we should optimize for a balance between the two.

For example a 150,000 batch of rapid tests China recently sent to Czechia had 80% error rates. You don't want to be mass quarantining people because of such terrible testing supplies.

There's nothing inherently bad with mass quarantines.

If you're quarantining a large number of medics then that's potentially serious and possibly more dangerous overall if your numbers are off.

You could administer two tests for those, I believe that's what's already done.

Reply on news.ycombinator.com