New Myths about Dangerous Drinking

In May, 2014, a report was released called Alcohol’s Burden of Disease in Australia. According to the infographics helpfully released with the report, fifteen people a day die in Australia from alcohol-related causes, and a further 430 are hospitalised. The immediate media response was satisfyingly gloomy. Headlines included: “Our boozy culture is killing us” (Age), “15 Australians die each day from alcohol-related illnesses” (ABC Online), “Alcohol is killing 15 Australians a day and experts want tough new warning labels” (news.com.au) and “Australia’s daily alcohol toll: 15 deaths and 430 hospitalisations” (Conversation). Media sources were especially interested in reporting that Australia’s death rate from alcohol-related disease had apparently risen by around 60 per cent in the last decade.

But is any of this true? Are we a nation of dangerous drinkers? Are we draining our health system’s limited resources and mortgaging our country’s future to the dubious joys of alcoholic over-indulgence?

Alcohol’s Burden of Disease in Australia is a solid document of around 120 pages. It is well laid out and clearly presented. Its arguments are mostly consistent, and it is mostly honest about its limitations. However, a careful reading of the whole document—rather than relying on infographics and summaries—reveals some dubious straw men, and some holes. The report tentatively acknowledges these by saying that it’s really a broad study, it’s aiming to update older burden-of-disease reports, and it should be used as a springboard for more in-depth and concise research in specific sub-populations.

The report shows that Australians are for the most part moderate drinkers, and have been for at least the last ten years. We are putting down the spirits, and picking up the beer and cider. We mostly drink wine and beer, of which most of us consume modest amounts annually. In some cases, our drinking habits are actually protecting us from earlier death, and may be reducing our risks of some chronic conditions.

And yet none of this made it into the media reports. Instead, we had shrill calls for more restrictions on drinking. How did this happen?

To answer this question, I need to turn the report inside out, and describe it back to front. The report begins with a portrait of how Australians drink, and then moves on to show the connection between this and various chronic illnesses and causes of death. I will do the opposite and ask:

• Does the report show a real relationship between alcohol consumption and an increased risk of developing various chronic illnesses?
• If so, does the report show that Australians have the type of drinking habits that lead to increased risk? Are we on a dangerous trajectory, or not?

The report has updated Australia’s existing burden-of-disease material by, for the first time, factoring in alcohol’s relationship with the following conditions: cancer, cardiovascular diseases, digestive diseases, diabetes, infectious and parasitic diseases, injuries, and neuropsychiatric diseases. So by linking alcohol to a far broader range of causes of death, the report has naturally produced a far grimmer portrait of its effect on Australians. The alcohol-related death rate has “risen” by 60 per cent—but only in the pages of the report, thanks to the data modelling.

It’s also very hard to quantify “alcohol-related harm”. Epidemiologist Dr Belinda Lloyd summ­arises it nicely for me:

Alcohol-attributable burden for many diseases [is] still difficult to estimate because of technical, methodological and co-morbidity issues. Also for all conditions protected against by alcohol, detrimental effects co-exist (low level of drinking is protective, whereas high level of drinking is harmful). The current methodology cannot capture these detrimental effects in such conditions because they are masked at the population level.

Lloyd believes that the rate of harm is probably being under-estimated—but it could be just as easily over-estimated, because that gives certain research institutes a lot more traction in their drive for recurring funding.

Can the report’s authors support a link to increased deaths with their literature review? It’s worth checking the literature cited by the report against its own definitions of “light”, “moderate” and “heavy” drinking, and using its own model of grams of alcohol translated into standard drinks.

Here’s how it works: the report translates 10 grams of pure ethanol as one standard drink—a 30ml serve of spirits, a 100ml glass of white wine, or 375ml of mid-strength beer. So when we go to the literature cited by the report on alcohol’s relationship to cancer or cardiovascular disease, we can see that much alcohol was consumed by those studies’ subjects before a link was found to chronic disease.

Here are the results for cancer. One study found an increased risk of developing colorectal cancer in people who consume on average 50g of alcohol per day. That’s around five standard drinks a day. Another study found an increased risk in those who consume more than 30g a day (three standard drinks or more) while another found a stronger “association”—not the same as finding an increased risk—among those who drank as little as 2.5 standard drinks a week.

For breast cancer, an increased risk was found in women consuming three or more standard drinks a day, and another meta-analysis found that the risk increased among women who drank more than five standard drinks a day. Again, a “relationship”—not a firm link—was found with levels as low as one standard drink a day. The risk of developing oesophagal, mouth, nasopharynx and oropharynx cancer increased two to three times when a person consumed 50g or more of alcohol daily—five standard drinks a day.

In other words, the literature cited by the report shows that if a person is a persistent and heavy drinker, they will increase their risk of developing some cancers. The literature cited by the report does not show that alcohol consumption leads directly to an increase in the death rate from these cancers. It’s perfectly reasonable to speculate that this is happening, and to show this with data modelling, which is what the authors have done. But data modelling is just that—modelling. It doesn’t necessarily represent reality. If you change one indicator, you change the whole model.

So what about cardiovascular disease? Similar amounts of alcohol were claimed to increase the risk of developing hypertensive disease (50g to 100g daily, or five to ten standard drinks a day), ischemic heart disease (“heavy alcohol consumption”), cardiac arrhythmias (anywhere between 24g and 120g of alcohol on average per day), and stroke (60g to a whopping 150g daily). The report also noted that low to moderate drinking is actually protective against ischemic heart disease and stroke, especially in older adults—but not in those who have been binge-drinkers.

Diabetes is the same. The literature cited by the report shows that a person needs to be drinking around 48g or more daily—almost five standard drinks—to increase their risk of developing it. If they can keep it to 12g to 24g daily—just a couple of standard drinks—individuals may actually experience a protective effect.

The real surprises for me were in the literature on digestive diseases. Apparently only 1 to 2 per cent a year of heavy drinkers will develop cirrhosis of the liver; I would have thought it was more than that. Other conditions show nil risk increase at lower levels of alcohol consumption, but they rise once a person goes past the 36g daily mark—three and a half standard drinks a day.

The real moment of glory appears in the section on infectious and parasitic diseases. I have saved this till last, because it’s a beautiful example of how prominent authorities can fail to understand the human condition in all its alternations of confusion and bliss. The report tells us:

In a review meeting hosted by the Medical Research Council and the WHO in 2008, 25 international experts concluded that the causal relationship between alcohol use and risky sex was still unclear.

I will leave you to speculate about the social life of those twenty-five international experts. And yet we need only continue reading the report to find the answer to this apparently baffling question. The section on injuries, for example, shows a clear relationship between alcohol consumption and all kinds of risky behaviour, through its impact on visual functions, mental processing, vigilance and reaction time, with disastrous consequences for drivers.

The section on neuropsychiatric diseases also offers food for thought. These include epilepsy (where there may be an increased risk at four or more standard drinks per day) and an ICD-10 classification called “mental and behavioural disturbances due to use of alcohol”, which is made up of sub-categories including “alcohol intoxication”, or in rough demotic, being drunk. Between 80 and 95 per cent of all hospitalisations for neuropsychiatric conditions attributable to alcohol are caused by these “mental and behavioural disturbances”. I can see our twenty-five international experts being baffled by this causal relationship as well, although I have it on good authority that there’s a direct relationship between alcohol consumption and being drunk.

So the real relationship between alcohol and the burden of disease, according to the literature, is largely restricted to an increased risk for heavy, prolonged drinkers and binge-drinkers. But are Australians heavy, prolonged drinkers and binge-drinkers?

This is my second question, and to answer it, we need to go back to the beginning of the report. First, who are the “Australians” in question? The report has used data from the National Drug Strategy Household Survey (NDSHS) to work out how much we are drinking. This has been compared against actual alcohol sales figures, as the report’s authors argue that the NDSHS may not provide the full picture of alcohol consumption, instead under-estimating it—a national telephone-based survey is going to have trouble reaching some of our sturdiest problem drinkers.

So allowing for this, how do Australians measure up? Most of the Australian men who responded to the NDSHS survey (just over 70 per cent) estimated their consumption of alcohol at between 0g and 40g of pure alcohol daily, while most women (around 66 per cent) are in the 0g to 20g bracket. So for men, that’s anywhere from none to around four drinks a day, which is a pretty broad spectrum, while women are anywhere between no drinks and two standard drinks a day.

Nationally on average, we are estimated to drink around ten litres of pure alcohol per person per year—litres, not grams, which makes conversions to standard drinks a bit more challenging. These are the figures from the NDSHS, but corrected by factoring in alcohol sales figures from each state. A litre of pure ethanol weighs 789 grams, so it still works out at around two standard drinks a day.

Males may drink as much as fourteen litres a year (around three standard drinks a day), and females as little as 6.5 litres a year (around one and a half standard drinks a day). These levels have remained reasonably stable for the last ten years, with “no significant increases in PCA” (per capita consumption of alcohol). Australia is ranked as a 2 by the World Health Organisation, which puts it on a par with New Zealand and Canada, neither of which are proverbial for their raucous and problematic drinking culture.

Nationally, we are drinking less beer and more wine, and half of all men and a third of all women have only ever had one binge (four or more drinks in one session). Just over 9 per cent of all men and 13 per cent of all women are total abstainers, and a further 7 per cent and 9 per cent respectively describe themselves as “former drinkers”. We sit mostly in the category of enjoying one to three standard drinks per day (this particular table is confusing, as it scores alcohol by grams for normal consumption, and then by standard drinks for bingeing).

But it’s when we break it down at state and territory level that some appalling anomalies immediately appear. In the Northern Territory, both men and women drink far more than anywhere in the rest of the country, and especially towards the heavier end of the scale—more than four to six drinks daily. The Northern Territory has the highest rate of binge-drinkers in the country (21 per cent of males, and 7 per cent of females, compared to the national average of 13 per cent of males and 3 per cent of females). It has the highest overall rate of alcohol-attributable deaths and hospitalisations in the country—the national average is 3.9 per cent of deaths and 1.8 per cent of hospitalisations, but in the Northern Territory, the rate is 11.8 per cent of deaths and 2.7 per cent of hospitalisations.

The Northern Territory suffers the most lost years of life, working days and disability burden from alcohol abuse for both men and women, far above the national average. For almost every identified disease and condition in the report, the Northern Territory outstrips the rest of the country in deaths and hospitalisations. Deaths from injuries related to alcohol are more than three times the national average for both men and women, and hospitalisations for the same problem are double the national average.

Chart after chart in the report shows an alarming spike for the Northern Territory—sometimes Western Australia and Queensland give it a run for its money, but for the most part the Northern Territory emerges head and shoulders above the rest of the nation as a place where alcohol is doing damage, proportionately speaking (and even allowing for its relatively small population distorting these figures). Yet the report only briefly mentions this outstanding anomaly from time to time, and devotes a bare paragraph to it in its final findings.

The report was produced by three organisations: Turning Point (affiliated with Monash University), the Foundation for Alcohol Research and Education (FARE) and the Victorian Health Promotion Foundation (VicHealth). It’s questionable whether any of these organisations is really impartial on this issue. Without impugning any individual, I’d suggest that for all three organisations it’s in their interest to find the direst results possible.

Turning Point’s mission is to promote “the health and wellbeing of individuals and communities living with, and affected by alcohol and other drug-related harms”. It aims to do this in part by “applying research to promote change and contribute to policy making”. Fair enough. If they’re serious about this, then their own research shows that they need to move immediately from Victoria to the Northern Territory.

FARE describes itself as an independent, not-for-profit organisation “working to stop the harm caused by alcohol”. They have been doing this for over ten years by “supporting world-leading research, raising public awareness, and advocating for changes to alcohol policy”. Their mission and staff will also find a natural home—and be exercised like a racehorse—once they decide to relocate to Darwin.

VicHealth—well, what is there to say, except that they live and breathe to deliver “a healthier Victoria” by “providing transformational expertise and insights to government”. If by “healthier”, they mean “non-alcoholic”, they have succeeded admirably. According to the report, Victorians have the lowest rates of alcohol-related problems in the country. (I can see the number-plates now: Victoria—The Wowser State.) I’d suggest again that maybe their work is done and it’s time for them to relocate and change their name to NTHealth.

I am concerned that the “Findings” section of the report is not entirely accurate. For example, it launches out bravely and states, “Alcohol has detrimental health effects in a number of disease categories”, but avoids any qualifying statement about the heavy and prolonged drinking needed to produce even a slight increase in risk, not in actual disease. There is a table in the appendices which purports to show this relationship by using a column “summarising” each article’s findings, but these summaries are too blunt. Given that their chosen literature showed that it took heavy and prolonged drinking to increase the risk of developing various diseases, simply using the word “detrimental” as a summary is not accurate.

What else is missing? The report tells us that fifteen Australians die each day from alcohol-related conditions. This may not be entirely substantiated, but even if it’s true, the report doesn’t tell us who they are. There is no data on age groups anywhere in the report, even though their data was collected from sources that used age groups.

Those fifteen people a day could all be over the age of eighty, dying in bed of a collection of health problems common to old age, but surrounded by their progeny and having lived a long and happy and productive life, enjoying their one to three standard drinks a day. Or it could be fifteen teenagers, snuffing out short and unhappy lives in road accidents and suicides. We don’t know who and where the fifteen people a day are, so it’s a meaningless figure. But it’s very helpful if you want to sell a scare story about the terrible impact of alcohol on Australian society.

The other outstanding omission throughout the report is the word Aboriginal. I can accept that a telephone-based survey may not have collected data on a person’s ethnicity, but this information is readily available from hospital and death statistics, both of which have been used to compile the report. We know from countless other data sources, including the Australian Bureau of Statistics, that Aboriginal people die at greatly increased rates of precisely the conditions highlighted by the report, and at far younger ages than the rest of the population. We also know that Aboriginal people make up 30 per cent of the Northern Territory’s population, rather than the 4 per cent which is the average in most other states. The report’s authors feel free to speculate about other relationships not substantiated by their data, so why not this one?

It’s the sale of this story that is the most concerning. How did a problem so obviously specific to the Northern Territory turn into a national problem for all of us, including Victoria, the Wowser State? The media flurry around the report’s release—with its handy infographics—brought with it the now-inevitable calls for stricter government control of alcohol use.

For example, Dr Stephen Parnis, emergency physician and vice-president of the AMA, said:

The Australian medical profession is profoundly concerned that the harms of alcohol are getting worse—domestic violence, depression, deaths from trauma, obesity—the list goes on. It is time to deal with this epidemic, and denial will only allow it to get worse. The evidence of harm is absolutely conclusive.

And yet none of these issues was actually covered by the report. I don’t doubt that Dr Parnis sees a lot of the ill-effects of drunkenness in his emergency department, but that’s not what the report was describing.

Rob Moodie, Professor of Public Health at the University of Melbourne, said the report showed the nation’s alcohol problem was worse than previously thought:

I would like to see sensible restrictions on alcohol advertising, promotions and sponsorship … It’s not about trying to stop everyone enjoying alcohol, instead it’s about reducing the impact harmful forms of drinking have on everyone else.

But if the report is correct, we’re pretty much staying within reasonable limits already, especially in Victoria (where Professor Moodie lives), which is one of the driest states in the country. Professor Steve Allsop, Director of the National Drug Research Institute at Curtin University, agreed with Dr Moodie: “What stands out most in the report is that the burden of alcohol is not just felt by the individual alcohol consumer. There is significant burden on the whole of Australia.” But the report actually shows that there’s a huge problem in the Northern Territory, which may or may not be affecting Aboriginal people disproportionately.

It’s very clear from the report—even with its tendency to jump between drinks, grams of ethanol and litres of pure alcohol—that there is a problem in Australia with alcohol consumption, but it’s not a national one. It’s a shame that this report has been used to drive the usual agenda of more government control of the sale and consumption of alcohol. I can only hope that the report instead ends up being used for its original purpose—to stimulate further and more detailed research into population sub-groups. These sub-groups continue to suffer while privileged academics ask for “sensible” restrictions on people who don’t actually need them.