Alcohol Deaths and Minimum Unit Pricing

News

Alcohol Deaths and Minimum Unit Pricing

By Colin Angus, University of Sheffield (Twitter: @VictimOfMaths)

Yesterday the National Records for Scotland published data on alcohol-specific deaths in Scotland in 2019. These figures are of interest to anyone looking to prevent alcohol deaths. One of the most significant policies which has been put in place at a national level in Scotland to help reduce deaths and improve health is Minimum Unit Price. We asked Colin Angus, from the University of Sheffield team which developed the model for the 50p per unit MUP in place in Scotland since 2018, to look at what the alcohol-specific death stats tell us about MUP.

Yesterday, National Records of Scotland have published new data which tells us, for the first time, how many alcohol-specific deaths there were in Scotland in 2019. This is an important number for several reason, not least because levels of mortality attributable to alcohol in Scotland are among the highest in the world – the age-standardised rate of alcohol-specific deaths in 2018 was almost double the rate in England, for example. This number has fallen by around a third from its peak in the early 2000s, but this fall stopped in 2012 and alcohol-specific deaths have been rising gradually since then.

Annual deaths from alcohol specific causes and the annual percentage change

The second reason that these numbers are important and have attracted media attention is that they represent the first full year of data which is available since Scotland introduced its landmark Minimum Unit Pricing (MUP) legislation in May 2018. Modelling work undertaken by me and my colleagues at the University of Sheffield back in 2016 estimated that a 50p MUP would save 121 deaths each year. Yesterday’s publication shows us that there were 116 fewer alcohol-specific deaths in Scotland in 2019 compared to 2018, and 100 fewer than in 2017, before MUP had been introduced.

So, does that mean that my colleagues and I were right? Does this data prove that MUP is working? Should you all be asking me for the lottery numbers now? In this blog I’m going to go through the new statistics, what they do and don’t tell us, and what this might mean for the success or failure of MUP.

Alcohol-specific vs alcohol related conditions

The first thing to clarify is what exactly today’s numbers represent. Alcohol is a risk factor for a huge range of health conditions, from liver disease and alcohol poisoning to cardiovascular disease and numerous forms of cancer. For some of these conditions, alcohol is the only cause – you cannot get alcoholic liver disease if you don’t drink alcohol – while for others, excessive alcohol consumption serves to increase a risk which is already there – any of us could develop oral cancer, but the chances of that happening are greater for heavier drinkers.

Today’s figures reflect the number of alcohol-specific deaths in Scotland in 2019. Alcohol-specific deaths are those from conditions which are solely caused by alcohol, with around two-thirds of these coming from alcoholic liver disease. What that means is that we can be sure that all of the deaths reported today were because of alcohol.

But it also means that this is a significant underestimate of the total number of deaths in Scotland which were caused by alcohol, since it does not include deaths from conditions where alcohol is one of several risk factors. The most recent evidence from the Global Burden of Disease study estimates that there were a total of 3,177 deaths in Scotland caused by alcohol in 2019.

What this means is that while today’s figures are shocking in their own right, they only reflect around 1/3 of the true mortality burden of alcohol. This means that alcohol-specific deaths are a very useful indicator, but they are a long way from giving us the full picture about the harms of alcohol. All of our modelling work on MUP and our estimates on the number of lives it would be expected to save were based on this wider definition of mortality capturing all deaths caused by alcohol, not purely those categorised as alcohol-specific. So our estimate that MUP would save 121 lives each year was based on a considerably larger estimate of the number of deaths caused by alcohol before MUP was introduced than the 1,000 or so alcohol-specific deaths identified in NRS’s data.

Reductions in some harms take time. Often years.

A second key point is that the clinical evidence tells us that while reducing your alcohol consumption is likely to be beneficial for your health, for some health conditions it can take several years for those benefits to be felt as a reduction in risk. In the most extreme case, a heavy drinker who cuts down their alcohol intake in 2019 won’t expect to see a reduction in their cancer risk for a decade, and the full benefit won’t be felt until 2039. This means that, while we would expect there to be short-term health benefits from reduced alcohol consumption (be that as a result of MUP or anything else), the full benefits will not be realised for some years. When we incorporated this clinical evidence into our modelling work, we found that we’d expect less than half of the ‘full effect’ of MUP on mortality to be seen in the first year after the policy, and two-thirds of the effect to be seen after 3 years.

Modelling isn’t the same as predicting.

The final important thing to consider when comparing today’s alcohol-specific deaths to the estimated impacts of MUP from our modelling work, is that we did not attempt to forecast what would happen. Our modelling work was an estimate of the impact of MUP all else being equal. What this means is that we estimated the impact of MUP if nothing else changed in the world.

Of course, this is a very strong assumption. Nothing ever stays the same for long. Young people’s alcohol consumption has fallen hugely in recent years, while the drinking of people in their 50s and 60s has risen sharply. These trends are likely to be an important component of the rising levels of alcohol-related harm, but are not considered in our modelling. The implication of this is that we need to consider the impact of these trends when assessing the impact of Minimum Unit Pricing. NHS Health Scotland’ recent study which found that MUP had led to a fall in drinking in Scotland did this by comparing changes in alcohol purchasing in Scotland since MUP was introduced to changes over the same period in England. This research established that alcohol sales in Scotland fell by 3.6% in the year after MUP was introduced, but that sales were actually 4-5% lower than we would expect them to have been without MUP.

What do the 2019 numbers tell us?

The first thing to say is that seeing a fall in alcohol-specific deaths is unequivocally a good thing, and a reduction of 10.2% is extremely promising. In fact, once you adjust for changes in the age of the population using a process called ‘age standardisation’, the fall is more like 10.9%. Our modelled estimates suggested an overall fall of 7.4% in all alcohol-related deaths, after 20 years, so these numbers suggest our modelling work may well be conservative in its effect estimates.

But before we get too carried away, let’s have a look at the details of the new data.

The largest contributor to alcohol-specific mortality in the UK is alcoholic liver disease. Encouragingly, deaths from this have fallen in both 2018 and 2019, while deaths from mental and behavioural disorders related to alcohol dependence are still at similar levels to 2017.

This is a similar pattern to what we saw in the recent alcohol-related hospital statistics. It is too early to tell whether this reflects some negative side-effect of MUP on dependent drinkers, random ‘noise’, or something else entirely. It’s certainly something we’ll want to keep an eye on in the next few years.

Alcohol specific deaths in Scotland by cause

Age and mortality

If we look at how trends in alcohol-specific deaths vary with age, we can see that deaths have fallen quite sharply across ages 45-74, which are the age groups that suffer the largest harm burden from alcohol. This is what our modelling suggested – that the targeted nature of MUP means that you’d expect to see the biggest improvements in health among those groups suffering the greatest harm.

Age patterns in alcohol specific deaths in Scotland over the last 15 years

Different areas, different trends

Finally, we can look at how trends in alcohol-specific deaths vary between local authorities in Scotland. Some authorities are much bigger than others, so it makes sense to convert the numbers to rates per 100,000 population to allow us to compare them. Overall this is a much more mixed picture.

Some areas have seen a sharp downturn in alcohol-specific deaths in 2019, while others have seen an increase. We have to be careful not to over-interpret these numbers, as there is inevitably going to be more ‘noise’ when looking at smaller populations, but this suggests that we still need to be cautious about claiming these new figure as proof that MUP ‘works’.

Local variations and trends in alcohol-specific deaths

Cautious optimism for now

Overall, I think it’s fair to say that these numbers are good news, and provide cautious optimism that the reductions in alcohol consumption we’ve seen in Scotland are beginning to be reflected in improvements in health. Indeed, if anything these numbers suggest that our modelled estimates of the impact of MUP might be on the conservative side.

However, just as we need to be careful not to dismiss evidence which goes against our expectations, so we need to be cautious about overinterpreting numbers which meet or exceed what we would have hoped.

These figures are a really encouraging sign, but they represent only a part of the overall picture, and are, by their nature, subject to some random fluctuation from year to year. The true proof that MUP has been effective will be a sustained fall in these numbers, and for that, we’re going to have to wait a little while yet.

25 November 2020.