Drinking in lockdown - what's the story now?

This month saw the release of alcohol sales figures for the UK during lockdown. They told a dramatic story. In the 17 weeks to 11 July, the number of alcohol units sold fell by a whopping 35%.

The total volume of alcohol sold during lockdown (the 17 weeks to 11 July 2020) fell to 1.3bn litres, down from 2bn the previous year, data from Nielsen Scantrack and the CGA found, despite value sales through the major retailers rising £1.9bn over the same period.

Sales of booze at the supermarket during the four month period hit £7.7bn, it said, however with the on-trade remaining shut, the overall volume of alcohol bought in the UK was far lower than last year, despite the increase in value.

Nielsen’s senior client business partner Gemma Cooper pointed out that around 46% of booze sold in the UK was sold through the on-trade and although spend had undoubtedly shifted to the off-trade, it was not enough to make up for the shortfall.

Note that Nielsen tracks the off-trade and CGA tracks the on-trade, so both bases are covered in these figures, in contrast to the slightly hysterical reports during the lockdown about alcohol sales 'soaring' by 31% (based solely on the off-trade).

You wouldn't guess that the UK had just seen its biggest peacetime decline in alcohol consumption from reading the papers today.

This comes from the Royal College of Psychiatrists who want more funding for alcohol and drug treatment services. They claim that 'Public Health England’s latest data on the indirect effects of Covid-19 found that over 8.4 million people are now drinking at higher risk, up from just 4.8 million in February'.

This statistic has been repeated by the BBC and others, but it doesn't appear in the report published by the RCP today. It looks like something added to spice up the press release. If so, it certainly had the desired effect.

But is it true? The source seems to be this webpage from Public Health England. PHE's data show that the proportion of adults drinking above the guidelines has risen slightly, from 22.7% to 23.5%.

If we exclude the people drinking between 14 and 21 units - a safe level, in practice - the number has risen from 14.5% to 16.4%. There has also been a rise in the heaviest drinkers (50 units or more) from 7.5% to 9.2%, but also in the number of nondrinkers (34.8% to 39.6%).

None of these figures come close to the doubling claimed by the RCP. For that, we need to look at PHE's alternative drinking data which is based on the Alcohol Use Disorders Identification Test (AUDIT), a survey which looks at behaviour and psychology as well as consumption. A score of more than seven classes you as a 'hazardous drinking'.

Here, we can clearly see a near-doubling since February. Numbers tend to be low in February and it would have been better to compare April to June with the same period in 2019, but there is nevertheless an obvious spike in the second quarter of 2020.

If this is the RCP's source - and it makes sense that they would use a psychiatric survey rather than alcohol consumption figures alone - their estimate is about right.

What does this tell us? Firstly, it shows that lockdown was a boring and/or stressful time for many people and that we shouldn't repeat it. 

Secondly, it shows - once again - that the single distribution theory of alcohol use is wrong. According to this theory, also known as the Ledermann hypothesis, there is a distribution of alcohol consumption which moves left or right as a whole. If per capita consumption (which is what the mean average drinker consumes, by definition) increases, the whole curve moves to the right and the number of heavy drinkers at the tail increases and the heaviest drinkers consume even more alcohol. If it moves to the left, everybody drinks less.

Graph taken from here

This theory gives support to the Whole Population Approach which focuses on driving down per capita alcohol consumption by whatever means necessary. The Whole Population Approach was most explicitly outlined by Geoffrey Rose and it underpins the 'public health' lobby's agenda not only with regards to alcohol but also to sugar, salt, blood pressure and many other risk factors. It officially underpins the Scottish government's approach to alcohol.

Put simply, there are two key assumptions:

1. The number of heavy drinkers, and therefore the amount of alcohol-related harm, is directly linked to per capita alcohol consumption. When consumption goes up, harm goes up. When consumption goes down, harm goes down.

2. Per capita alcohol consumption is driven by price, availability and advertising. If prices go up, availability goes down and advertising is restricted, consumption will go down and, as per the assumption above, heavy drinking and alcohol-related harm will decline.

You only have to look at the UK in the last twenty years to see that this theory is wrong. Alcohol consumption has gone down significantly but alcohol-related deaths have stayed at about the same level. (The temperance lobby would also say that alcohol-related hospital admissions have risen sharply, but those figures are worthless.)

The pandemic has given us an exceptional natural experiment. Alcohol advertising, like most advertising, declined during lockdown. Availability declined enormously as ten of thousands of pubs, clubs and restaurants were shut down for three and a half months. Prices stayed much the same (although one could argue that the average price fell as a result of the temporary abolition of the more expensive on-trade).

What was the effect? Per capita consumption did indeed decline (by 35%), but the bell curve did not move as one organism. Some people stopped drinking completely, other people drank less and some people drank more. If the AUDIT survey is any guide, an unprecedented decline in per capita consumption was accompanied by an unprecedented rise in 'hazardous drinking'.

We shall have to wait to see if this led to an increase in actual harm (if it doesn't then perhaps we should reevaluate the definition of hazardous drinking), but things are not looking good for supporters of the Whole Population Approach.

And why would they? The amount of alcohol individuals consume is driven by personal choice and circumstances, not the retail environment. It was ever thus.