Hacker News Reader

Comment by leoedin | Hacker News Reader

It's not hard to write "Faint lines count" in the test instructions. Any line at all is a positive test.

You have to take into account the effect the extra cost has as well - if it's between doing one automated bluetooth test that costs $25 or 5 manual tests which cost $5 each - the 5 tests is clearly a better outcome for society.

View on HN

> Any line at all is a positive test.

It's not obvious what counts as a "faint line" -- is this too faint? Is it not really there? Am I misreading natural variations in the colour of the test strip?

It's not a matter of reading instructions -- even intelligent adults who have followed the directions 100% correctly can still struggle with knowing what counts.

Anyone who has ever taken a pregnancy test (or seen someone else take a pregnancy test), can tell you about the major amounts of stress possible attempting to read a "simple faint line" -- see https://www.reddit.com/r/TFABLinePorn/ for one example.

I'm not saying that every test strip needs $20 of Bluetooth gizmos and e-waste. Just that normal, intelligent, fully-instruction-following adults can reasonably be expected to not always know how to read "any faint lines" accurately.

On the contrary, it is extremely clear what counts as a faint line. If its different from the white background, in any way or shape, the test is positive. Simple as that. It couldn't actually be more simple, but someone had to over-engineer and forget about security during the process.

Example here [1] (as linked elsewhere throughout this thread)

[1] https://twitter.com/zeynep/status/1471493781073530882

Not to mention motivated reasoning/perception. Is this positive test going to mean you’re going to cancel your holiday plans and lose thousands of dollars?

> It's not obvious what counts as a "faint line" -- is this too faint? Is it not really there? Am I misreading natural variations in the colour of the test strip?

If you think of how the test works, then it is obvious that if you can spot the line it is positive.

I have literally no idea how the test works; I'm a software engineer.

They're supposed to be similar to pregnancy tests right? But those detect hormones afaik, not viruses (or antibodies).

The test contains antibodies for the virus, so if they react in any way (assuming you weren't tampering with it[1]) then you have the virus. Those tests are considered unreliable, because the false negative is high (i.e. not enough of the virus to register).

I don't know enough about pregnancy test, but I assume the hormone is always present, it's just certain level of it signals pregnancy.

[1] there were videos where people where testing juices and soft drinks and that was messing up the test.

> the 5 tests is clearly a better outcome for society.

I don't think that's obviously true. Misread pregnancy tests are fantastically common, and someone mistakenly believing they don't have covid is going to be a significant negative for society.

Instructions simply don't work, a large number of people don't read instructions. Heck, I might not read the instructions for this if it looked understandable at first glance. Decades of pointless barely-understandable instructions on cheap devices have trained me not to bother.

It's pretty tricky to self-administer one of these lateral flow covid tests without reading the instructions. There's a whole sequence of steps - swabbing, mixing the swab in the buffer solution, dropping a set number of drops into the sample window. This device doesn't appear to get rid of any of that complexity - instead adding significant complexity and to the simplest part of the test - look for a line.

Indeed, there's also considerable variability in the different lateral flow tests. If you don't read the instructions then you may not work out the differences when the next set come along.

In the UK we have, amongst others I'm sure, the NHS branded boxes, Flowflex, and Orient Gene.

Some require swabbing nose and throat, some are just nasal swabs. Some require 2 drops onto the device, some require 4. Some require a certain amount of time to be left in the buffer solution, others just a set of steps of pushing/turning/squeezing the swab in the buffer. Some require waiting 30 minutes (minimum), some are 15 minutes (minimum) and 30 minutes maximum. etc.

I somehow managed to miss the critical fact that the line can be very faint for quite a long time.

I'm a developer. I suck at reading documentation.

Problem is, I can often see the faint line right out of the package, before any reagent or sample is applied. I get really OCD about that.

The tests I have here say if you see any lines that are not the control line the test is not good. You might want to double check and contact the manufacturer.

I may be imagining the lines. But there’s no difference between imagining one and seeing one when it comes to your confidence level and thus the value of the test. Google “imagining faint line” and see what a common problem it is.

There was an interesting thread last week about different abilities to visualize images in your head - some people describe visualizing things in a faint sort of way, not a full on visual hallucination. Take that ability, mix it with the anxiety of fearing a test result and the tendency to second guess yourself, and I think you can see the value of electronics making the determination for you.

What I have been wondering about is a positive line showing up after 3 hours. I've had this happen consistently, and found no indication of what to do. The official docs say any results after 30 minutes are void. But why? Does the false positive rate go up to high? What could cause false positives.

I have had a few PCRs that all said negative, but the quicktest really worried me.

>if it's between doing one automated bluetooth test that costs $25 or 5 manual tests which cost $5 each - the 5 tests is clearly a better outcome for society.

are we talking about doing 5 tests on the same women, or doing 5 tests on different women?

Reply on news.ycombinator.com