How one statistical mistake convinced the world that alcohol was good for us
Back in the 1980’s a popular subject for discussion in consumer programs, news bits and cultural commentaries was the fact that French people experienced chronic illness at a lower rate than the English did. This was especially strange considering that the French diet contained a higher proportion of foods that were known to increase the risk of cardiovascular disease and other negative health outcomes. The leading theory was that this health benefit correlated with the propensity of the French to drink a glass of wine with dinner. Wine, it was believed, had a protective effect on many bodily systems
Those with Christian leanings would sometimes quote the apostle Paul who addressed Timothy (in 1 Timothy 5:23) with these words:
“No longer drink only water, but use a little wine for the sake of your stomach and your frequent ailments” ESV.
When Paul wrote these words he was not backed up by double-blind placebo controlled studies, but spoke using the medical knowledge of the day. Physicians such as Hippocrates (5th century BC) and later Galen (2nd century AD) recommended wine for a variety of conditions. It was thought to:
• aid digestion
• stimulate the appetite
• relieve stomach discomfort
• disinfect water
• promote sleep
Alcohol's Beneficial Influences
These observations are not without merit. Alcohol is known to relieve many acute symptoms. It alleviates muscle tension, diminishes anxiety, reduces the intensity of certain tremors and eases pain and the onset of sleep. However, it does not provide these benefits without risk. We would not view Heroin, another pain and anxiety relieving substance, as harmless, but because alcohol plays such a central role in the social activities of many cultures we give it more leeway.
The acceptance we extend towards alcohol use, however, does not reflect reality when all of its risks are considered. In fact, the Nutt Drug Harm Scale (see below) identifies alcohol as the most destructive substance available, in terms of harms to the self and others.
For the vast majority of users these harms never materialize, but they are common enough that they outweigh the negative effects of all other abused substances. For alcohol, the harms to others carry the most weight. The Nutt Scale classifies these harms as follows:
Harms to the user
• Mortality
• Physical damage
• Dependence
• Mental impairment
• Loss of relationships, employment etc.
Harms to others
• Crime
• Family adversity
• Economic costs
• Community damaged
• Environmental harm
• International damage (trans-national crime, corruption, violence etc.)
• Other societal impacts
Alcohol's Harmful Influences
The harms alcohol brings to the frequent user include cardiovascular disease, cancer, liver disease and mortality. The amount of alcohol required to cause injury differs from individual to individual, but we intuitively know that we can always consume too much of a good thing. Even the overconsumption of water can be fatal through a process known as hyponatremia. In this condition the electrolytes required for neural function are diluted to the point that nerve impulses in the brain and body are interrupted. This is most commonly seen in long-distance runners who over-hydrate in an attempt to counter water loss through sweating.
Too much alcohol is obviously harmful, but are small amounts of alcohol actually protective against disease? Early evidence seemed to support this assertion. The J-Shaped graph below shows illustrates an eronious interpretation of the data that has since been corrected.
The positive health effects of wine were believed to vest in the high levels of antioxidants such as resveratrol, flavonoids and tannins contained predominantly in red wine. In addition to these factors, alcohol itself improves cardiovascular performance by reducing blood platelet stickiness (easing blood flow) and increases HDL (good) cholesterol levels. It also modestly improves insulin sensitivity in some people.
The impact of Statistical Bias
The belief that alcoholic drinks have a protective effects has an insidious influence on some drinkers because it discourages abstinence. Rather than asking should I quit, the question becomes how much can I drink before it becomes harmful. This ambiguity of messaging is an impediment to those who have trouble regulating their alcohol intake for one reason or another and for whom abstinence is the only viable solution. Thankfully the latest research has clarified the situation, as shown below.
So what is the statistical fallacy behind the J shaped graph? It turns out that it is due to the “sick quitter” effect. When measuring health outcomes in non-drinkers there are two distinct populations: those who never drank to excess and those who were forced to quit drinking due to alcohol’s harmful effects. This distinction was not made in early research, so chronic disease in the sick quitter population was included in the non-drinking cohort. Moderate drinkers appeared healthier than non-drinkers because they had never been exposed to the damages incurred through overconsumption.
Assessing the Risks of Alcohol Consumption
This discovery in association with refinements in research methods and outcomes informed today’s scientific consensus that there is no safe level of alcohol intake and that the risk of negative health outcomes increases with increased consumption. So, how much risk is there to drinking, say 15 drinks a week? This might reduce the life expectancy of 100 people by 3.5 months on average. You might be willing to exchange 3.5 months of your life for the freedom to enjoy a glass or two of wine with dinner each night, but that mean measurement has a more impactful interpretation. The 3.5 months mentioned is actually a lifespan reduction of 15 years for two individuals in the cohort, and you could be one of them.
The main takeaway from this updated perspective is this: we are all responsible for establishing the amount of risk we are comfortable with and managing our alcohol intake accordingly, but we can be confident in knowing there is no risk at all in abstaining from alcohol entirely.
0 Comments