Tuesday 17 June 2014

Who believes in the Total Consumption Model?

Punishing the Majority, the paper John Duffy and I wrote for the IEA, has been well received, but I'd like to respond to one criticism made by James Nicholls when I was on the radio with him over the weekend. The criticism really takes two forms: firstly, that nobody in the alcohol research community believes in the Total Consumption Model any more (and so we are attacking a straw man), and secondly that government policy is not aimed at reducing per capita alcohol consumption.

On the first point, we note in the report that 'Nothing in this paper will seem new or controversial to those who have followed this academic debate in the last forty years.' This is true and it would be great if everybody accepted that reducing per capita alcohol consumption is not the key to reducing heavy and harmful drinking. Unfortunately, there are plenty of influential groups that still endorse the theory, either explicitly or in so many words.

This 2010 NICE document, for example, which was commissioned by the government to advise the government, offers a texbook explanation of the Total Consumption Model and cites its biggest British proponent, Geoffrey Rose.

Population-wide interventions

The PDG believes that most of the recommendations will have a greater impact on those who drink irresponsibly. However, taken together, they are very likely to improve the health of the population as a whole. As indicated by the Rose hypothesis, a small reduction in risk among a large number of people may prevent many more cases, rather than treating a small number at higher risk. A whole-population approach explicitly focuses on changing everyone's exposure to risk (Rose 2008). In this instance, the number of people who drink a heavy or excessive amount in a given population is related to how much the whole population drinks on average. Thus, reducing the average drinking level, via population interventions, is likely to reduce the number of people with severe problems due to alcohol.

And here is Alcohol Focus Scotland, the state-funded temperance group...

Harmful drinking has become so normal and acceptable that the problems it causes to other people are often overlooked. That’s why we need alcohol policies for the whole population. If we all drink less, then harms will come down across the board. Drinking less is in all our interests.

And here is a communication from the European Commission (which advised the Scottish government that minimum pricing would be illegal):

It is worth noting that total alcohol consumption is a key health indicator in the EU (ECHI 46) as a proxy for the level of alcohol related harm in a Member State.

For this reason, the EU has a specific target of reducing per capita alcohol consumption from 10.2 litres to 9 litres per year by 2020. Note that it sets goals for consumption, rather than harm.

Moreover, although we have been told that minimum pricing has nothing to do with the Total Consumption Model, one of its main proponents, Tim Stockwell said (in an report promoting minimum pricing) that it is central to it...

The [Sheffield] model is based on two fundamental elements that are well established in the much larger literature on the relationship between alcohol consumption and alcohol-related harms:  

(i) When the price of alcohol increases consumption by most drinkers goes down, including, critically, consumption by hazardous and harmful drinkers;

(ii) When population alcohol consumption declines rates of alcohol-related harms also.

All subsequent debate about the Sheffield Model has centred on the degree of certainty regarding the size of these effects... High quality reviews confirm that when total consumption of alcohol in the population declines, consumption among heavier drinkers is reduced and, further; rates of alcohol-related mortality also decline [32]. The Sheffield Model applied these general principles specifically to the UK and provided numerical estimates of the benefits.

It should be noted that reference 32 is Alcohol: No Ordinary Commodity, a book that revived the Total Consumption Model for the 21st century, complete with the usual policy recommendations with regards to price, advertising and availability. Its summary asserts that "As the per capita consumption in a population increases the consumption of the heaviest drinkers also rises, as does the prevalence of heavy drinkers and the rate of alcohol-related harm".

The Total Consumption Model is therefore not dead, it just smells funny

But is it part of government policy? At the EU level, yes, but the EU has limited authority in this area. In Britain, perhaps not so much (although we are not just talking about Britain in the report). The government is naturally influenced by the influencers, such as NICE, but famously ignores them when they feel like it (eg. with minimum pricing). It is probably fair to say that politicians in the UK are not ideologically wedded to the Total Consumption Model, indeed they probably don't know what it is. But let's look at what they do in practice.

The 'cornerstone policies' of the Total Consumption Model involve raising taxes, restricting advertising and limiting availability. Britain certainly has high alcohol taxes and its advertising regime is more restrictive than most. It can be therefore seen to be following the 'whole population' approach even if it is not ideologically committed to it. I accept that availability has been the exception in recent years. The 2003 Licensing Act liberalised the sale of alcohol and would not have been brought in by a government that was primarily interested in reducing overall consumption.

Nevertheless, overall consumption has fallen since the Licensing Act came in. I don't think this was necessarily the intention, but it shows that the dogma of 'affordability, advertising and availability' doesn't stand up in the real world. Still more tellingly, the decline in consumption (and the rise in prices) has not coincided with a commensurate decline in alcohol-related harm. Thus, the fundamental belief that less per capita consumption = less harm has not been borne out in the UK over the last decade. Why? In large part, because it has been the 'wrong' people who have been cutting down on their drinking.

In conclusion, large parts of the 'public health' community remain wedded to the Total Consumption Model (or some variation thereof) and whilst the government may not be consciously wedded to the theory, it endorses most of the policies that are associated with it. The model, however, does not stand up in theory or in practice.


Vova said...

Nobody would try to pretend that per-capita consumption cannot be an important and useful indicator, often related to alcohol-related harm (which is what makes it useful)but to make an indicator a target is really pretty stupid. An interesting and sensible quote from Judith Innes 25 years ago -

(1)-the measure must be pertinent to questions of concern
(2)-the concepts underlying the measures must be clear and agreed
(3)-the measure must relate to the concept which it is assumed to, and do so in a well understood way
(4)-the methods to produce the measure must provide reliable results, measuring what they purport to without hidden or unexpected bias
(5)-the measure must be understandable and understood in its concept and limitations
(6)-major parties to discussion on opposite sides must accept the measure
(7)- it should relate to more complex analytical models'

From that point of view per-capita consumption is a good indicator, but unfortunately the neo-prohibitionists pretend the relationship in (3) is causal from proxy to problems, which it isn't, gloss over the limitations in (5), don't give a **** about (6) and pretend that they have satisfactory tested models for (7) which they of course don't.

James Nicholls said...

This post is misleading. On LBC, I did indeed point out that the IEA report was based on the incorrect claims that the 'cornerstone' of UK alcohol policy was whole population models of harm. I did not, however, say ‘nobody in the research community believes in the Total Consumption Model’. We didn’t, in fact, discuss the research community’s views on the Total Consumption Model.

Chris appears to have confused our LBC discussion with a subsequent exchange on Twitter where it was pointed out (not by me) that that Total Consumption Model, as characterised in the IEA report, was subject to ongoing critical discussion among alcohol researchers. That is not news, nor does it undermine the argument that there is a known correlation between population level consumption and alcohol harms.

Unfortunately, then, this post adds a second straw man (a non-existent statement about the research community) to an existing one (the supposed UK Government’s use of whole population models as the ‘cornerstone’ of their alcohol policy).

I would prefer it, Chris, if you corrected your post.

Vova said...

>>Total Consumption Model, as characterised in the IEA report, was subject to ongoing critical discussion among alcohol researchers.

What is the point of 'critical discussion' among people who don't understand the details,and have already decided what to think?

And will James please admit that control of pcc *should not* be the cornerstone of alcohol [policy? No, I didn't think so.

James Nicholls said...

John, you appear to know how I will respond to your question before you have actually posted it! (Also, according to your IEA report, you know in advance how the team carrying out the harm paradox research at LJMU will assess their data. Could you give me your prediction for the World Cup - I'll put a tenner on it...)

The point of 'critical discussion', as I understand it, is to work out the strengths and weaknesses of any given argument in order to refine, redesign or abandon it as appropriate. That seems to have been a feature of the work around the TCM over the years, as your IEA report demonstrates. There may be some people who accept versions of TCM uncritically, but I can't speak for them. There are also plenty of researchers who focus more strongly on subgroup analysis. It's not black and white, is the point.

It's a bit odd that the IEA report begins by asserting whole population approaches *are* the cornerstone of UK policy; then Chris says that, in some respects they *can* perhaps be seen as a factor, and now you are asking whether they *should* be the cornerstone. Which is it, I wonder?

The IEA report asserts (again, wrongly I believe) that population approaches exclude targeted measures such as brief interventions and specialist treatment. If that were the case then no, they shouldn't be the cornerstone. But that simply isn't the case. In Scotland - where whole population approaches are key to policy - they have introduced the only national IBA programme in the world ... *precisely because* that is part of any population-based framework.

Anyway, my question about the IEA report on LBC was primarily that it appeared not to have understood the political framing of alcohol policy in the UK over the last couple of decades. The 2004 AHRSE explicitly rejected population approaches, so did the 2007 National Strategy, and even MUP was included in the 2012 Strategy as a measure targeted at binge drinkers. Our levels of alcohol taxation are motivated by revenue generation; our advertising restrictions are motivated by child protection and related issues; and the 2003 Licensing Act can't in any conceivable sense be construed as expressing a population approach.

Therefore, if alcohol harm hasn't reduced, it *cannot* be because the UK Government adopted population approaches - leaving open the possibility, however, that it may be because they have rejected them.

I think that's a fair criticism.

Vova said...

I was right about James anyway...

>>will James please admit that control of pcc *should not* be the cornerstone of alcohol policy? No, I didn't think so.

And indeed he didn't.

But I'm not really in the prediction racket. Perhaps he should ask the Sheffield Alcohol Research Group to run their World Cup Model for him.