Why can't we prevent or cure hangovers?

A case study in market failure

There is a nice theory in economics about why capitalism is better than socialism. It goes like this. Under socialism, central planners try to estimate what the public wants and they try to decide which products are good. They aren't able to estimate this as well as the market can. The way the market does this is to allow people to start companies to sell products they think people want. People buy the stuff they want, making those companies flourish, and they avoid the stuff they don't want, so those companies stop selling those products by choice or bankruptcy. Over time, this leads to better products being produced under capitalism than under socialism. I agree this kind of theory is right, yet capitalism has some conspicuous failings in this regard and we don't discuss them enough.

Alcohol is the most popular drug on the planet. It's been with humanity for something like 15,000 years. Some ethnic groups, and humans in general, literally evolved to adapt to this pleasant poison. But still, alcohol consumption leads to a delayed unpleasant state that we call a hangover. The causes of hangovers are actually not that well understood, but in general it involves some mix of dehydration, poor sleep quality, and acetaldehyde poisoning. Alright, you think, given that humans have been dealing with this alcohol consumption into hangover problem for 15,000 years, surely, someone has figured out how to combat this effectively. I mean, under capitalism, the market for avoiding hangovers must be enormous. The average Dane drinks about 9.5 liters of pure alcohol a year (that's about 25 liters of vodka at 40%). That's a lot of alcohol, and thus a lot of hangovers that could would want to avoid. Now, if one searches around for avoiding hangovers, one will get a never ending list of articles with good advice. Most of these primarily ascribe hangovers to dehydration, so their solution is drinking more water. If one asks people avoid hangovers, they will also give exciting anecdote based advice on what alcohols give the most hangovers (red wine, vodka, beer??) and which mixes they can be given in to avoid the hangover (don't mix X and Y!). In supermarkets, one can buy an assortment of herbal remedies to cure or prevent hangovers. However, it is not really possible from unsystematic personal observations to figure out which alcohols cause hangovers more than others, or which mixes are worse than others. The noise in the signal is too strong. The same kind of anecdotal evidence people supply for these claims could be just as well used to argue for nonsense like homeopathy. Any evidence that can be used to argue for homeopathy is suspect. It's the control group for medicine. We can do better. What does science say? There must be some proper studies on this right. Well:

• Pittler, M. H., Verster, J. C., & Ernst, E. (2005). Interventions for preventing or treating alcohol hangover: systematic review of randomised controlled trials. Bmj, 331(7531), 1515-1518.

Objective To assess the clinical evidence on the effectiveness of any medical intervention for preventing or treating alcohol hangover.

Data sources Systematic searches on Medline, Embase, Amed, Cochrane Central, the National Research Register (UK), and ClincalTrials.gov (USA); hand searches of conference proceedings and bibliographies; contact with experts and manufacturers of commercial preparations. Language of publication was not restricted.

Study selection and data extraction All randomised controlled trials of any medical intervention for preventing or treating alcohol hangover were included. Trials were considered if they were placebo controlled or controlled against a comparator intervention. Titles and abstracts of identified articles were read and hard copies were obtained. The selection of studies, data extraction, and validation were done independently by two reviewers. The Jadad score was used to evaluate methodological quality.

Results Fifteen potentially relevant trials were identified. Seven publications failed to meet all inclusion criteria. Eight randomised controlled trials assessing eight different interventions were reviewed. The agents tested were propranolol, tropisetron, tolfenamic acid, fructose or glucose, and the dietary supplements Borago officinalis (borage), Cynara scolymus (artichoke), Opuntia ficus-indica (prickly pear), and a yeast based preparation. All studies were double blind. Significant intergroup differences for overall symptom scores and individual symptoms were reported only for tolfenamic acid, γ linolenic acid from B officinalis, and a yeast based preparation.

Conclusion No compelling evidence exists to suggest that any conventional or complementary intervention is effective for preventing or treating alcohol hangover. The most effective way to avoid the symptoms of alcohol induced hangover is to practise abstinence or moderation.

Read those last lines again: No compelling evidence exists to suggest that any conventional or complementary intervention is effective for preventing or treating alcohol hangover. The most effective way to avoid the symptoms of alcohol induced hangover is to practise abstinence or moderation.

This should make you surprised given the extreme demand for treatments and capitalism to supply this. Why don't we have effective treatments? You might say "It's impossible! We've tried literally everything we could think of and nothing works." Not really. This systematic review found only 15 randomized trials that looked into this matter. It's from 2005, so obviously there can be newer studies, but consider that a 2010 meta-analysis found 83 studies published on ego depletion, something that doesn't actually seem to exist exist, and we only have 15 studies on potential cures for the largest drug side effect we have, that's... curious and insane. How come? Why hasn't capitalism delivered us a solution to this problem already? Why have our scientists not discovered the causes so we can intervene? Why haven't private sector scientists figured out how to reduce or cure this? It's not like the social costs are small, here's the introduction from the paper above:

The alcohol hangover has substantial economic and health consequences.1,2 In Britain, the associated problems have been estimated to account for about £2 billion in lost wages each year, mostly due to sickness absence.3 With binge drinking on the rise, these figures are likely to increase.4 In the United States, the total cost of alcohol use has been estimated at $12-30 billion per year, although other figures criticised for being inflated range as high as $148 billion per year.1,5 The rates of medically certified sickness absence seem to be higher in never drinkers, former drinkers, and current heavy drinkers than in current light drinkers.6 In the workplace, a person with a hangover may experience impaired memory and visual-spatial skills and may be at risk.7,8 Other symptoms in varying combinations may include lightheadedness, nausea, and concentration difficulties. The symptoms seem to be due to a combination of ethanol's main metabolic product acetaldehyde, congeners including methanol, endocrine and immune system disturbances, dehydration, and sleep disturbance.1,9 Substantially increased risks of all cause mortality can occur even in people drinking less than recommended maximums, particularly among younger adults.10,11 An increased risk of strokes has been observed in young adults, particularly on occasions when alcohol intake is higher than average.12 In addition, during celebrations at Christmas, for instance, alcohol consumption increases and may lead to a rise in fatal alcohol poisonings by 0.4% for every 1% increase in the sales of spirits.13

These considerations emphasise the need for safe and effective preventive and therapeutic measures. A plethora of “hangover cures” is on offer.14-16 Searching the internet on Google.com (search term: hangover cure, accessed 20 Jan 2005) retrieved in excess of 325 000 hits. The box gives a flavour of what is on offer. The aim of this systematic review was to assess the clinical evidence on the effectiveness of any medical intervention for preventing or treating alcohol hangover.

That's a lot of billions. A lot of incentive to create some kind of solution, yet it seems practically no one is doing this. The authors are just about as negative as I am:

The paucity of randomised controlled trials is in stark contrast to the plethora of “hangover cures” marketed on the internet. This confirms the unreliability of the internet in healthcare matters.33 Our findings show that no compelling evidence exists to suggest that any conventional or complementary intervention is effective for preventing or treating alcohol hangover. Encouraging findings for their main outcome measures exist for γ linolenic acid from B officinalis, a yeast based combination preparation, and tolfenamic acid. However, only single randomised controlled trials for each of the tested interventions were available, most were of small sample size, and all used unvalidated symptom scores. Independent replications of these studies are therefore necessary. The lack of a sensitive standard outcome measure to assess the physiological and subjective effects of alcohol hangover may be one of the reasons for the small body of evidence. The development and initial validation of the hangover symptoms scale will hopefully encourage further systematic research and will aid the integration of trial data.34

The second author of the study Joris C.Verster published a follow-up paper up in 2021:

• Verster, J. C., van Rossum, C. J., & Scholey, A. (2021). Unknown safety and efficacy of alcohol hangover treatments puts consumers at risk. Addictive Behaviors, 122, 107029.

It is important that hangover products are both safe and effective. The aims of the current study were to evaluate (a) the ingredients of currently marketed hangover treatments, (b) whether companies make disease modification claims for these products, and (c) the extent and quality of any independent scientific evidence on their efficacy and safety. Of eighty-two hangover products identified, the most common ingredients were vitamin B, vitamin C, milk thistle extract (silymarin), dihydromyricetin (DHM), and N-acetyl L-cysteine (NAC), often in combination. Fifty-one products (62.2% of the 82 evaluated products) contained one or more vitamins of which the dose exceeded the corresponding daily recommended intake level. For 9 (28.1%) of 32 products that reported the dose of Vitamin B3 and 2 (8.0%) of 25 products that reported the dose of Vitamin B9 the corresponding tolerable upper intake level was exceeded. Further, in many other cases the dose of other ingredients was not reported (e.g., dosages of DHM and NAC were not reported by 59% and 73% of the products containing these ingredients), and corresponding tolerable upper limits are unknown. A review of scientific literature revealed no peer-reviewed human data demonstrating either safety or efficacy of any of the 82 evaluated hangover products. Further, the product name and/or package/insert included explicit disease modification claims in 64.6% of the products. Finally, 45.1% of the products contain NAC as an ingredient. As NAC is registered as a drug by the US Food and Drug Administration (FDA), it is prohibited as an ingredient in dietary supplements or foods. We conclude that, in the interest of consumers, independent research supporting the safety and efficacy of hangover treatments should be a minimum requirement for hangover treatment claims irrespective whether the products are registered as medicinal drugs or dietary supplements.

So in other words, capitalism is trying to solve the problem. Consumers are buying anti-hangover products, but they don't know if they work -- we don't know if they work. They might make things worse. The issue, I think, is that capitalism's ability to improve well-being depends on the costumers' ability to tell whether a product is good or not. In the case of say cars that breakdown or not, that have good fuel economy or not, the costumer is able to figure out these facts, and consequently avoid buying bad cars. In the case of medicine with only weak effects, it is not possible to tell from personal experience or anecdotes from others whether a product works or not. However, as the consumer has no proven effective option, they might as well gamble on some of these potentially efficacious products provided they are relatively cheap and relatively harmless. Thus, capitalism leads to a plethora of cheap products of unknown efficacy because that's the best it can do in this situation.

It would appear that what the consumer needs is some kind of consumer research unit that investigates matters of interest to the consumer. There are in fact many such institutes, but they are of doubtful reliability due to corruption issues. I mean: companies in one way or another pay money to these organizations to get them to provide positive sales materials for their products, and the research institutes will tend to deliver these. It is hard to see how the market could fund a such institute without corruption. The research findings it would produce could be easily spread and thus could not easily be sold, so where should it get money from? It would seem to be a good case for a market failure intervention. Some kind of government unit that is tasked with investigating the most pressing consumer research questions. I mean, we do have such research units in some universities. The problem is that they are run by people like Brian Wansink. Often when we try to correct for a market failure, we get another failure downstream. In this case, money was given to people like Wansink, but Wansink didn't have good incentives to publish rigorous research on actual consumer questions, such as reducing hangovers. He had a lot of incentives to publish as much research on possible on random implausible stuff that provided good press for his university. Good press and lots of studies means more research grant money.

OK, so maybe what we need is just a single good research team somewhere that will produce proper science when given the funds. Well, it would appear that Joris C. Verster is involved with that team. They have a paper out:

• C Verster, J., Stephens, R., Penning, R., Rohsenow, D., McGeary, J., Levy, D., ... & Young, M. (2010). The alcohol hangover research group consensus statement on best practice in alcohol hangover research. Current drug abuse reviews, 3(2), 116-126.

Alcohol-induced hangover, defined by a series of symptoms, is the most commonly reported consequence of excessive alcohol consumption. Alcohol hangovers contribute to workplace absenteeism, impaired job performance, reduced productivity, poor academic achievement, and may compromise potentially dangerous daily activities such as driving a car or operating heavy machinery. These socioeconomic consequences and health risks of alcohol hangover are much higher when compared to various common diseases and other health risk factors. Nevertheless, unlike alcohol intoxication the hangover has received very little scientific attention and studies have often yielded inconclusive results. Systematic research is important to increase our knowledge on alcohol hangover and its consequences. This consensus paper of the Alcohol Hangover Research Group discusses methodological issues that should be taken into account when performing future alcohol hangover research. Future research should aim to (1) further determine the pathology of alcohol hangover, (2) examine the role of genetics, (3) determine the economic costs of alcohol hangover, (4) examine sex and age differences, (5) develop common research tools and methodologies to study hangover effects, (6) focus on factor that aggravate hangover severity (e.g., congeners), and (7) develop effective hangover remedies.

There is a 10 year follow-up too reporting on their progress. Really, from reading their recent work, it involves combating useless science. Take their letter to the editor about a recent randomized controlled trial. They write:

As reported by the authors, paired samples t-test analysis of seven male participants revealed a significant reduction in stress (P = 0.039) following the 600 mg dosage. There were no other significant effects of the 600 mg dosage. According to related-samples Wilcoxon sign rank tests, the 1200 mg dosage resulted in significant improvements in nausea (P = 0.013) headache (P = 0.010) and hangover severity (P = 0.043), while improvements in anxiety approached statistical significance (P = 0.052). However, using two-sided tests, the 600 mg dosage no longer significantly improves stress (P = 0.078) and the 1200 mg dose no longer significantly improves hangover severity (P = 0.086). The significant effects of the 1200 mg dosage on nausea (P = 0.026) and headache (P = 0.020) remain. After applying a Bonferroni’s correction for multiple comparisons of individual symptoms with P < 0.010 as cut off for statistical significance (five comparisons; i.e. treatment vs placebo for hangover, nausea, stress, headache, and anxiety), no significant differences concerning L-cysteine and hangover symptom severity remain.

It's hopeless tier research. I suggest that researchers do international, multi-lab collaborations, where they use planned analyses, and they ask industry to sponsor some of the research. The analysis code will be written beforehand, and the data published with the research. Given the impossibility of cheating under this setup, funding from industry cannot bias the results, and should be aggressively pursued. Any company that really thinks their product works will be happy to sponsor this research as the ensuing positive PR would surely generate a lot of sales ("Groundbreaking academic study finds that X is the best treatment for hangovers, 100 scientists agree" headlines will follow). If companies don't think their products really work, they should stay away. I suspect most will stay away. Capitalist companies don't care so much if products work, they care if they can be sold.

The main good note I can think of ending on, is that this hangover research team has started their own clinical trials, two of which are published in the special issue that has the 10 year follow-up. The first study is n=13, p=.042, so probably useless. The second is n=20, p>.05, so probably also useless. These studies are too small to be informative. I suggest that a large naturalistic randomized trial could relatively easily be done. Get a bunch of different treatments into capsule form. Mix them up, some of them being placebo. Put each of them in numbered individual pockets in a capsule wrapper thingy. Distribute these to subjects along with their IDs. Have people decide on their own which capsule number(s) to try. It doesn't matter which people choose first because their numbers are randomly assigned. Have them take some surveys about their symptoms. Compare the effects of each treatment on the outcome. It doesn't matter if sometimes multiple of them are taken together, as their combinations will also be random. In fact, this allows for tests of interactions with a sufficient number of subjects. This study would be essentially a large-scale randomized trial in naturalistic settings, and should be easily scaleable to 100s of people. I bet one could get students to participate in this kind of research. After all, they are already interested in drinking, they are already on campuses, they already want the treatment so their homework isn't so dreadful the day after. Simply travel around the country giving talks at each university about this ongoing experiments, and recruit students at the end of the talk. What prevents this?

PS. I know there are readers out there who want to say "Aha, but wouldn't hangover treatments be net negative for society if they just cause people to drink more, so they get more liver cancer, cirrhosis etc., even if they no longer get so many hangovers? Maybe the market is playing 3d chess and is secretly very effective.". That same 2005 review already thought of this objection and their reply is:

Ethical concerns may relate to research in this area. It is conceivable that positive trials might lead to considerable media interest and industry marketing, which ultimately might lead to an increase in alcohol consumption. However, little evidence exists to show that alleviation of hangover symptoms results in increased alcohol consumption.1 Conversely, no conclusive evidence shows that hangover effectively deters alcohol consumption.

PPS. Some will say: "But Emil, Wikipedia does list some reasonable looking evidence on congeners and hangovers. Why did you ignore this?". Have you looked at the most impressive looking research? It's a study by... Verster (truly a one man army in this field), which used an uncontrolled survey approach of students, and was only published as an abstract from a conference. This kind of evidence is not that convincing, we need something better. What about having a hangover research party at university? Entrance is free BUT on the condition that one enrolls in the study. Each subject will be assigned an ID. When they go to the bar, they will be given a random mixed drink that the bartender prepares without them knowing. OK, they can perhaps get to choose which mixer they want in. The bartender notes the type of alcohol in a database of some sort for that student's ID. Next customer. The day after, each subject fills out a questionnaire about their hangover symptoms. Then it's a matter of a regression analysis since we know what each student was given, and this was randomized. Of course, some students will have been drinking something else that night, and this gives noise in the results, but it doesn't give bias because the drinks were still randomized inside the party. Thus, the regression analysis should provide us with a ranking of what is good and what is bad. Sure, you can come up with some plausible issues, say, students pouring out alcohols they didn't like, sharing with each other. More noise in the data, but probably not enough to invalidate the design.