Major alcohol study’s claim that drinking is good for older people is disputed

Colin Angus | August 2022 | 8 minutes

The research published in The Lancet, as part of the Global Burden of Disease project, came to some conclusions that contradict previous studies on the effects of alcohol. It claimed that safe levels of drinking can vary depending on a person’s age.


Colin Angus, a research fellow from The School of Health and Related Research at the University of Sheffield, sheds some light on how the research was conducted and what it really means.

This study aimed to estimate the level of alcohol consumption at which health risks are lowest. For the majority of health conditions, this level is 0 (i.e., complete abstinence) as the risks of harm increase with any amount of drinking. However, for a small number of conditions, most notably cardiovascular disease, the evidence incorporated into this new study suggests that low levels of drinking may reduce your risks of harm, however, the existence of these so-called ‘protective effects’ is hugely disputed.

In order to estimate the overall level of risk associated with drinking at a certain level, it is necessary to add up the risks from each of the different health conditions associated with alcohol, taking account of the prevalence of these different conditions in the population. It is also important to consider the age structure of the population, as well as how drinking patterns vary between individuals.

This process would produce a curve showing the relative risk of experiencing any ill-health at different levels of drinking compared to a non-drinker. This new study repeats this process for 204 countries around the world. So far so good.

However, where things start to unravel is when the authors decide to repeat this analysis separately by age. This is because older age groups suffer more harm from cardiovascular conditions, for which low levels of drinking may be protective. While younger age groups tend to suffer a much greater proportion of their health harms through injuries, for which there are no protective effects. As a result, estimates come out at lower levels for younger people. Indeed, the results suggest that the level of drinking at which alcohol harms are minimised for 20–24-year-olds in Western Europe is 0. It is this finding on which the suggestion of lower drinking guidelines for younger people is based.

The study has another fundamental flaw: it removes any nuance between the different types of risk. These curves represent the relative risks of harm for drinkers compared to non-drinkers within the same age group. They tell you absolutely nothing about the difference in risks between age groups or the absolute risks that people are facing as a consequence of their drinking. To illustrate the problem, when we conducted similar analyses in 2016 we found that although drinking increased the risk for young people, the risk for older people was already so much higher that this increase was relatively small in comparison to the increase for older people. Simply put, a small percentage of a big number can be larger than a big percentage of a small number.

On top of this, there are several major methodological problems, including the fact that the authors do not account for drinking patterns. This is key because people drinking more alcohol but less frequently face greater immediate risks (e.g., from injuries) due to higher levels of intoxication than people drinking less alcohol more regularly. Beyond that, evidence suggests that even infrequent heavy drinking can remove any protective effects of alcohol consumption against cardiovascular risks.

Finally, in the press release accompanying the article, the study’s senior author is quoted as saying “our message is simple: young people should not drink, but older people may benefit from drinking small amounts”. This statement is problematic for at least two reasons.

Firstly, it fails to acknowledge the relative magnitude of risk involved for younger people. We face risks all the time in our everyday lives and the UK drinking guidelines were set in relation to a risk threshold defined to represent an ‘acceptable’ level of risk. Therefore, young people choosing to drink within these guidelines are not at a substantially increased risk of suffering harm as a result of their drinking compared to non-drinkers.

Secondly, it seems to imply that older people who do not drink should be actively encouraged to consume small amounts of alcohol for the associated health benefits. We do not know with any degree of certainty whether light drinking is beneficial or not and it would be irresponsible to make clinical decisions on this basis. Further, as 65–74-year-olds have become the heaviest drinking age group, according to the latest data from the Health Survey for England, we should be particularly careful about promoting public health messages that suggest drinking is beneficial in this age group.

Ultimately, the results of this study are valuable and interesting, but the conclusions that the authors have drawn from these results are extremely problematic and do not align with the wider scientific evidence.