Smartphone-based colorimetric determination of GHB and GBL in beverages | Hacker News Reader

Smartphone-based colorimetric determination of GHB and GBL in beverages

onlinelibrary.wiley.com

1. There is not only GBH, in our lands clopfelin is much more common for that goal.

2. Color analysis? Good luck on doing that. I really surprised it worked at least in lab conditions.

I think they used a "clean" drink to normalize their values. I don't see a reasonable path to real-world use. As they drily point out - "As with all colorimetric methods, the natural color of some beverages would prove to be a challenge".

Not to mention the varying light conditions in different venues and variations in smartphone cameras, this seems impossible

I guess that can be somewhat calibrated based on the rest of the image, but it seems to be basically impossible to know what color a "clean" cocktail should be.

I didn't have high hopes for this, but the second word of the abstract incorrectly refers to Gamma-hydroxybutyrate as 'GBH'. Many other grammatical errors in the abstract alone. Maybe it's unfair to judge a preprint for this? I don't know.

Interesting research, but a dangerous idea in practice. Better to just not consume any drink that falls under suspicion rather than presume to rely on an app. It is assured that any such app will regularly produce false negatives even under the most generous circumstances, but even one failure is unacceptable for this use-case. Further, there's also the risk of false positives, which could result in serious and sometimes irreversible consequences for the falsely accused.

This looks more like an app for law enforcement than something someone would take to a bar. The alcohol sample has to be processed as follows (see Fig 1, materials and methods):

> "Scheme II shows an overview of the reaction mechanism to give the purple-colored complex. A 0.3 ml aliquot of sample was taken and added to a 30 ml Universal glass vial with a 24 mm polypropylene screw cap. The sample was then adjusted to pH 1 by the addition of H2SO4, to convert the GHB (I) present to the lactone, GBL (II) via intramolecular esterification. A 1.0 ml aliquot of a 1.0 M hydroxylamine HCl solution and 0.2 ml of 12 M sodium hydroxide solution was then added. The resulting mixture was then mixed gently by hand to form the hydroxamate derivative (III). Upon the addition of 0.2 ml of ferric chloride, a purple-colored complex (IV) is then formed instantly."

You'd have to carry around a little kit with all these pre-made solutions (plus some pH paper to get step 1 right). The 'app' part is just using your cell phone as a crude spectrophotometer to get a rough estimate. Also, colored drinks would be a problem. The idea here seems to be more related to 'in the field' analysis for forensic law enforcement without the need to transport samples to a lab.

Incidentally alcohol is low-quality drug with nasty side effects, even though its use is socially normalized. I group it with morphine derivatives, as alcohol triggers several opiate-response pathways, accounting for some of its addictive properties. The yearly death toll of the 'alcohol epidemic' is also comparable to the yearly death toll of the 'opiate epidemic'.

Oh god I hope law enforcement doesn't decide to jump in on this as a cost-cutting measure (that they don't need) that would result in evidentiary and potentially constitutional issues. Most LE departments have arrangements that give them access to far more accurate results from, say, GC/MS testing, and performed by those who does testing and analysis for a living. The police in America at least has a pretty poor record on using pseudoscientific methods to build a case while leaving the easy and proven part of scientific testing - sending rape kits and such DNA collected to labs - on the backburner. And of course none of this is going to actually help the police since reporting rates and clearance rates for rape and sexual assault remain very low and by the time a report comes in, the evidence may no longer exist.

The focus on GHB/GBL seems... lacking in coverage, considering that there's a whole range of water/alcohol soluble substances that is easier to carry (in powder or pill form, instead of liquid) and doesn't have the taste problem GHB/GBL has when mixed drinks tend to drown out the chemical taste of Midazolam or, if push comes to shove, research chemicals based off existing formulations such as Nifoxipam represent just the tip of the iceberg in the analogues of just one category of potential "date rape drugs", quotes because without question the likes of Midazolam have contributed much more to the health and well-being of patients than as the moral panic du jour once again that overshadows its utility.

And I would posit that alcohol is not viewed as an epidemic in part because those who drink tend to know what they're drinking, while there's virtually no street demand for fentanyl and its analogues and the supply-side demand is generated entirely by the failed war on drugs that targets the supply side aggressively enough so that the golden rule of prohibition kicked in and the most potent and easiest to carry form of the prohibited substance becomes what the supply side focuses on. Regular, non-opiate-naive users were overdosing and dying, which is very much an aberration when it comes to overdoses generally. And we've utterly failed at implementing anything more than the bare minimum in terms of harm reduction while politicians go out of their way to blame everyone but themselves. It's effectively a government sanctioned epidemic. If you want to kick off your week on a shitty note, read Inspector Generals' reports on DOJ and the DEA specifically. However impractical, this is at least an attempt at putting a harm reduction product on the market, and should be applauded as such. It's more than what congress have managed to do for god knows how long.

Does it detect alcohol?

EDIT: As the father of three daughters, I'm much more concerned with my daughters consuming too much alcohol, which puts them into a compromised state of decision-making. I know those drugs exist, and have been used, but alcohol is much more prevalent, much more socially acceptable, and just as effective at reducing decision-making ability, and ultimately, the physical ability to resist unwanted advances.

Have you helped them get an understanding of how alcohol effects them? In many jurisdictions, even if it is illegal for a minor to purchase alcohol, it is legal for them to drink at home under the supervision of their parent or legal guardian. Lots of parents are horrified by the idea of letting their teen drink at home but it's an opportunity to help them learn their limits, understand how it affects their behavior, and removes the mystery and taboo from it. I'm not talking about having them do keg stands. Wine with dinner once a week and occasionally other types of drinks goes a long way teaching them how to deal with alcohol in a responsible manner. Obviously you want to control this because even adults sometime have trouble finding and abiding by their limits but you want them to discover this while still at home rather than during their first semester away at college.

Anecdotally, and to reinforce your point, people in my family and a lot of my friends have been drinking since childhood and all of us have a normal relationship to alcohol.

That being said I'm pretty sure every teenager will get way too hammered at one point or another, it's a quintessential teenage experience.

A lot of people "learn how to drink" in college, often through trial and fire. Many of us have a few stories of getting too drunk in our early 20's and learning those consequences the "hard way". Considering how likely it is for any given young college student to engage in drinking/partying after leaving home, getting that knowledge in the safety of their parents home, ideally through experiences less harmful then maybe expected through learning the "hard way", guided by the more-experienced parents, what you suggest makes a lot of sense to me. For example, some people find it nearly impossible to stop drinking when they start, or they may have an alcohol intolerance leading to inebriation with much smaller amounts of alcohol - great information to know before going into an environment like college.

I second that. It is pretty inevitable that teenagers get to experiment with alcohol, it is best to do that in a safe environment. My mother told me back then, that if I have to get drunk (and sooner or later one will get, especially when inexperienced and getting carried away), please do that in safety at home.

While one shouldn't exactly encourage drinking, responsible drinking is part of our culture so allowing that for teens at the appropriate age (in Germany 16 and above) at home is the best way to give them the necessary experience and also being able to give advice where needed, as this then isn't a "taboo" topic.

If alcohol was tasteless, and a few milliliters could incapacitate someone to the point they risk being date raped, you'd have a point.

GP has a very valid point. We've all seen college students too drunk. Too much alcohol could easily be consumed at, say, a party, leading to someone sexually assaulting the drunk individual. Yes, GHB will "knock out" someone in a way different from alcohol, but considering how ubiquitous alcohol is and how hard it can be for many people to cut themselves off before getting too drunk, I too would personally be more afraid of a college student getting too drunk over getting administered GHB or similar. Not to downplay date-rape drugs, but I just don't think they're nearly as common as alcohol.

Yes, but if the issue is detection, one can detect that they're consuming alcohol long before it would incapacitate them to that degree. They can't do that for GHB.

I am aware they wrote that facetiously, but the issue really is about detecting things you don't want in your drink, not changing people's choices about what they want to consume.

Spoken like a man who hasn't paid $5 for access to the jungle juice at a house party hosted by a friend of a friend.

It is much harder to unknowingly consume it in the amounts needed for that, though.

The solution for that is proper education about its effects and about knowing their own limits.

"When to leave the party" is knowledge everybody should have.

You don't need an app to detect alcohol because you can taste it, unlike GBH.

GHB is not tasteless; my understanding is it is incredibly salty.

Serious question: Have you ever had an alcoholic beverage made with hard liquor? Alcohol has a very strong flavor/burn. Even half an ounce of cheap vodka, the "flavor-neutral spirit", will throw off the taste of a glass of soda. Grain spirits with even more ethanol like Everclear are even harsher. There are cocktails that can mask the taste of alcohol well, but not entirely.

You're looking at this from the "is it tasteable?" standpoint, and I am looking at this from a probability-of-risk standpoint.

In the end the smartphone is quite irrelevant - the detection happens through a color test that should also be visually clear as a purple color. Also, it is a bit impractical because a color test will ruin the drink (hydroxylamine is poisonous and has a strong bad taste), so the drink would have to be sampled and the test done in a second container. The authors note esters can interfere, which might effectively be a problem in fruity mixed drinks with fruit esters and fermentation+distillation related esters. All in all, it is probably easier to throw away your drinks if you suspect they are spiked. If there is a serious suspicion, the drinks can also be given to the police for reliable investigation.

Very interesting. I wonder if it's robust enough to be used in real life. For example, if distortions/variance from different phones/cameras, laser lights or other club lighting, or specific drink ingredients could cause false results.

The people who are commenting that this is not something that could be used in a real-world situation are absolutely right. However, that's not what it's for - it's a study to determine whether the basic principles might be workable at all.

There would be much more work required before it could be applied in practice.

One confusion is about how this might be used. It's about the possibility of a very cheap and readily-available means of basic testing (not a substitute for laboratory testing) in criminal investigation.

The main problem I see with anything like this are the social and legal consequences.

All tests are imperfect, so you have to choose whether to be biased toward false positives or false negatives. Clearly a false negative would be a disaster for the drink recipient. So the test would be biased heavily toward false positives. Normally that’s fine, but in this case it would falsely implicate the drink provider with disastrous legal and social consequences. That’s no good!

There was a bunch of PR stuff around a nail polish that did something similar a few years ago [0] in South Africa but I don't think a product ever came out of it.

[0] https://www.news24.com/health24/Lifestyle/Woman/Your-body/St...

This reminds me of the "poison snooper" as described in Dune.

A bad test is worse than no test.

This is a very bad test. Cute paper, but I hope anyone foolish enough to try and deploy such an app gets a strongly-worded letter from FDA.

I disagree, a bad test can lead to a false sense of security which can translate to more harm than before in cases when the test produces a false negative and the person has lowered their guard due to the presence of a test.

I think y'all are saying the same thing.

It sounds like you're agreeing with OP, who said a bad test is worse than no test. Isn't that what you're saying too?

"Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something."

https://news.ycombinator.com/newsguidelines.html

Alright.

A test like will have very low sensitivity, being unable to detect most adulterants added to drinks, and low specificity as well, given that colorimetry is a bad way to detect a substance which differs more by specific gravity than color, and given that drinks are seldom clear to begin with.

A false positive could have unfortunate social outcomes for everyone involved, a false negative would create unwarranted confidence, spiked drinks are uncommon enough that it's a safe bet that false one or the other would be the normal outcome, and the damned thing is, the cup is known to already contain the most common date rape drug in existence, ethanol.

Seems like just throwing the drink out and getting a new one is an easier solution?

But wouldn’t a better solution be to get out of the place altogether, if you suspect that a drink might have been spiked?

If someone is targeting to drug and rape you, what does leaving accomplish? They can just follow.

In fact, leaving is likely what they want you to do. Drug kicks in, they pose as the Good Samaritan or friend taking the clearly drank-too-much person home.

A bartender could spike it and you'd be none the wiser. Always keeping your drink close and under your eyes doesn't always work.

A lot of bartender pure "date-rape" alcohol into the drinks under the eyes of the customers.

The bartender is just about the only person who leaves a paper trail if they do this. I really don't think they're the likeliest of threats.

If you're worried your bartender is spiking your drink, the only logical thing to do is leave, not pull out your phone to run a test.

Jesus.

https://youtu.be/Chrd0GjBq9o?t=15

I wouldn't have suspected a thing. (15 seconds in)

Better approach:

    1. put smartphone in stand
    2. point camera towards drink
    3. open app
    4. go dancing / wait until someone puts drug in drink
    5. app detects it, using deep learning / computer vision
    6. call police / throw away drink

Sounds like a good way to discover people steal unattended phones.

The app can use the accellerometer, and sound an alarm or contact wrist watch.

Muggers literally take phones physically from people; despite full knowledge the phone is being taken, and physical proximity to the person, they still get away. A watch notification that your phone is getting stolen is just gonna give you a nice view of them disappearing out the door of the bar.

The solution to date rape is not setting up your phone on a tripod in a bar.

I don't see how that would help in a typical busy, noisy club - the thief is simply going to grab the phone and disappear into the crowd. Plus, they can just turn the phone off.

But here's the clever bit: you have a second phone guarding the first phone!

You don’t have to leave drinks unattended for someone to spike it

It can happen very fast, while you are holding it

Reply on news.ycombinator.com