Hacker News new | past | comments | ask | show | jobs | submit login

Especially when you know that all the electronics actually do is interpret the result of a standard paper based test.



I was surprised to discover that the tests as used without expert supervision had a 25% misread rate, vs 1% with the display.

I think that’s a dire indictment of the education system, but we need to make stuff for the people we have, and a quarter of them can’t read the test.


Check out https://www.reddit.com/r/TFABLinePorn/

A subreddit where users post positive or probably positive pregnancy tests.

Pregnancy tests generally measure hCG concentration in urine, early in the pregnancy concentrations are low and the test line will be very faint if visible at all. It's only later on in the pregnancy when the line will become more visible.

Pregnancy also involves a lot of emotion, hope & fear, in combination it's not that surprising that a portion of people misread pregnancy tests.


Wow I did not expect that! Many of them are really hard to read actually!

A quick example of two positive tests side by side https://i.redd.it/llw9fmn6dxk51.jpg

Even worse, sometimes the reaction is so strong that there is a "dye stealer" effect, where the test line steals the ink from the control line. See for example here: https://i.redd.it/skdynwwi6ll51.jpg (Notice how the line is on the "wrong" side)

No wonder the misread rate is so high indeed...


Those aren't misreads, those are ambiguous tests. The computer would get false negatives on those.


Strictly speaking if the control line disappears that's an invalid test.

https://www.early-pregnancy-tests.com/inpregtesstr


It eliminates a wide variety of sources of error, including user ignorance but also possible issues with colorblindness, poor lighting conditions, or a positive result that's not as easy to discern with the naked eye.


> a positive result that's not as easy to discern with the naked eye.

Those are false negatives in the machine.


> 1% with the display

Source on that? I've only seen 1% reported for lab testing of paper strip.


>1% with the display

Which is, high as it is, another surprise! I wonder how that happens.


The bottom 2% of the population is illiterate or severely mentally handicapped and do no better than chance at reading pregnant vs. not pregnant.


There was a good comment about this on reddit: https://www.reddit.com/r/nextfuckinglevel/comments/ini2aw/do...

Essentially - it's easy to mis-read the paper tests and there's a small portion of users who may decided to use an electronic device as a double-check for the standard paper test especially if they've been mis-led in the past.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: