Wine is bad for you again but the conflicting alcohol advice is enough to make anyone need a drink

The public should stop fooling themselves that a daily glass of wine is “somehow healthy”, according to Dr Sarah Hartz, author of the latest piece of research chronicling our on-again, off-again relationship with drinking.

Simultaneously flying in the face of dozens of studies showing its benefits and corroborating mounds of data on its risks, it shows that drinking on four or more days a week increases risk of an early death by 20 per cent.

Each time a seemingly contradictory piece of research is trotted out the cliche is to cry “why can’t scientists make up their minds?” and anyone who built a diet around 2014’s superlative “a bottle of wine a day is not bad for you and abstaining is worse than drinking” can feel particularly aggrieved.

But the latest study, published in the journal Alcoholism, does actually address some of drinking’s split personality.

People who drank moderately (one or two drinks) on three days a week or fewer, did show some signs of benefits to their cardiovascular health and this is the area most positive reports focus on.

However, the dose makes the poison, as the saying goes, and this study shows that those benefits – relating to alcohol increasing HDL cholesterol – are lost among more frequent drinkers.

What’s more, the study confirms that alcohol only increases the risk of other serious diseases like cancer, as well as other potential causes of death like accidents and violence.

These issues apply to health claims about everything from chocolate to live music.

The most fundamental is about risk.

The 20 per cent increase in early death is an average taken across 400,000 people included in the study and there will be huge variation between individuals based on their lifestyle or even genetics.

It is also an example of “relative risk”, essentially how much you overall chance (absolute risk) of dying has changed.

As the study points out, a healthy twenty-something who exercises every day and only eats boiled chicken and kale has a lower risk of dying at any given point than someone in their seventies – even if the twenty-something is a moderate drinker.

These studies are also blighted by coincidences that can only be partly controlled for with statistics.

People who are heavier drinkers are also likely to be smokers, have worse diets and exercise less. Conversely, people who abstain from drinking entirely may do so because of other, unrelated, health conditions which mean they are more likely to die early or become unwell.

The gold standard of scientific evidence, and the proof required for every new drug, is to take two groups and randomly assign half to receive the treatment, half to receive a placebo.

But this is impractical in many cases. Putting aside the costs, in a hypothetical study of alcohol where drinks were all served unmarked containers, the participants would figure out they were in the placebo group by about 8pm on Friday night.

There are many other problems, and the media’s own culpability is a subject for another article.

But as a conclusion, the general consensus is that serious harms don’t kick in until you’re regularly drinking more than 14 units a week (the UK guideline amount), though drinking less might not make you healthier.

However, health risks are just one line in the calculations that rule our lives, and when most people knock off work on Friday, the words of Dr Sarah Hartz rarely hold as much clout as the one-word text saying “pub?”.