Hacker News Reader

Comment by giantg2 | Hacker News Reader

There are many things that doctors do not recommend because they don't believe the average person can use it correctly.

For example, a doctor prescribed two medications (A, B) to take twice per day. Yet there is an alternate medication (C) for one of them (B) that can be given 3 times per day, which reduces blood level fluctuations resulting in better symptom management. She told me that they don't usually prescribe that one because they don't trust the parents of their patients to correctly administer one drug only twice per day (A) when giving the other drug (C) 3 times (they are worried about overdoses of the twice daily medication A).

View on HN

> There are many things that doctors do not recommend because they don't believe the average person can use it correctly.

> ...She told me that they don't usually prescribe that one because they don't trust the parents of their patients to correctly administer one drug only twice per day (A) when giving the other drug (C) 3 times (they are worried about overdoses of the twice daily medication A).

I wonder if they're not even worrying about "the average person", but a kind of "lowest common denominator" person. Depending on how severe an "overdose" is, that might be totally the right call. Even if only 1 in 100 patients would mess up and OD, prescribing the less-effective combination would actually end up helping a lot of patients.

IMHO, software engineers also end up infantilizing their users with similar "lowest common denominator" thinking, but usually with less justification.

Have you ever dealt with support tickets or read "normal" user forums for a consumer product that you work on? PEBKAC is often times a kind way to describe the problem.

Now engineers can (& do) get the analysis wrong (the famous "I can't send e-mail more than 500 miles). I don't think it happens with any less regularity than doctors. Thankfully our problem domain allows us to adjust the guard rails for a problem domain, or build better guard rails.

> Have you ever dealt with support tickets or read "normal" user forums for a consumer product that you work on? PEBKAC is often times a kind way to describe the problem.

I totally understand that, but I think it's far less justifiable to design non-safety-critical systems for a brainless lowest common denominator user than it is for things that are arguably safety-critical (like the dosing mentioned in the GGP), because then you end up with something like a world where the only books available (even to adults) are children's books.

Users also want free support. The more knobs you add the more complex your support becomes (both front line and automated tests) and your support costs skyrocket (now 1% of your users are turning on knobs you haven’t tested and those are 30% of your support costs). There’s a balance to be struck for sure and OSS can some times try to ignore the support cost because it’s just “spare time”. There’s still a cost though.

There’s also something to be said for sensible abstractions. Even if the power users don’t have the control they want, the solution works for 90% of all other users. For an example of this look to Apple. They regularly sacrifice the power users for the security and convenience of the general customer. From a profit, support and just pure “% of satisfied users” that’s the right balance. Don’t get me wrong. They don’t get it right for every single little thing, but in aggregate when there’s a conflict and your market is the “general” user, trying to stay appealing to the power users is still likely the wrong choice.

> There’s also something to be said for sensible abstractions. Even if the power users don’t have the control they want, the solution works for 90% of all other users.

There's something to be said for writing some things in some kind of Simple English [1], but it's quite different to take the position that most things should be written in it. Too often modern system designers seem to take something analogous to the latter position.

[1] https://en.wikipedia.org/wiki/Simple_English

IMO this is infantilizing is completely justified in many fields, but not all.

Ignorant people are like children, infantiles. And we are ignorant in different things, by in many things. There's a reason you infantilise infantiles, do you not?

Exactly. Anytime you are giving guidance to the public, you need to factor in that the average punter is capable of incredible feats of stupidity or craziness.

If you've had issues with your ear, you'll know that many treatments sound alot like the problems describe. People with lots of issues will have a procedure where they cut the eardrum and insert a tube. Or have a vaccuum or waterjet pull out earwax.

Are there people who will compulsively clean their ears and cause a buildup of wax? Probably. The only difference between that and people who damage their teeth or nose by picking is that they are using an implement to do so.

I believe an example is the use of alcohol to delay the onset of frostbite, at the cost of hastening the onset of hypothermia. Given the later is more immediately lethal and is accompanied by a loss of rationality, it would be irresponsible for doctors to admit that the vasodilative effects of alcohol will stave off frostbite by increase bloodflow to extremities (the same effect that hastens the onset of hypothermia.)

For prescriptions that are pills, I'd like to see a smartphone/tablet app to help with things like that.

At the simplest, it would just remind me when it is time to take pills and tell you which pills to take. That could help with the confusion of having pill A on a twice a day schedule and pill C on a thrice a day schedule.

Better would be if you could also show the app your pills using your device's camera when you first start a new prescription. Then when you take your scheduled doses you could show the app what pills you are going to take and it could check that you have the right number of pills and they are the right colors and shapes and sizes.

Weekly pill organiser boxes mostly solve this problem already, e.g. https://www.bing.com/images/search?q=weekly+pill+organiser&f...

You can already get technology-supported boxes but it seems to me that each extra 'check' step is likely to contribute to more failure cases.

In this case there's no upside to doing it "correctly". Or at least that's what many doctors think. So have no reason to care about overstating risks.

Reply on news.ycombinator.com