A sharing by Alan Ho Chok Chan
Statistics is the science and art of presenting data, or sets of figures compiled directly ( by surveys) or indirectly ( sourcing from other people’s figures: known sometimes as Meta-analysis.)
One thing I learned from medical school days was : figures, datas and graphs can be manipulated to present it in good light or bad, depending on your view of the subject, your agenda, your vested interests.
Take a middling number, 50%. It can be looked upon as half empty, or half full, and the readers swayed in the direction you wanted them to go.
If you want to be famous overnight, collect lots and lots of statistics, dress it up in lots of rocket scientists’ jargon and methodology, draw the conclusions you wanted, and send it off to the Lancet.
Which was what the group from Institute of Health Metrics and Evaluation from the University of Washington did, under the auspices of Bill & Melinda Gates Foundation, a tome of 21-page masterly survey from 195 countries spanning the period from 1990 to 2016. Just the collaborators alone numbered in the hundreds.
The conclusion : The safest level of drinking is NONE.
Whenever such bold conclusions are made, whether justifiable, it guarantees to provoke an uproar and a tsunami of counter-arguments.
Critique from the sideline
Firstly this is a meta-analysis, not a double-blind controlled, prospective study with two cohorts : drinkers vs non-drinkers with a long term follow up.
When one gathers data from others’ study, many factors are beyond one’s control :
The quantity of alcohol consumed is the first problem. In this study, the authors used the amount of alcohol sales in a country as a gauge of average consumption. Therein lies the first flaw. The total quantity consumed in one year is highly variable ( how much is actually consumed, how much is kept or cellared, how much is wasted ?) Then there is always the unknown figure of how much is consumed by tourists and other transient populations. There is further the unknown figures of illicit consumption, contrabands, and private home brewings. Multiply this figure by 26 ( study spanned from 1990 -2016 ) and imagine the discrepancy/ error between estimated and actual consumption!
Then the question of subdividing the population into non-drinkers and light, moderate and heavy drinkers. Drinkers vary in their habits: not everyone abide by 1, 2, or 3 drinks a day strictly. It depends on the occasion and the company. ( and whether there’s a friendly, jovial wine waiter named Alan !) There is the further complication known as the Sick Quitters’ syndrome, by which I take it to mean : If you are sick, you quit, then you resume drinking once you recover, all of which makes the estimation of alcohol consumption an inexact science, especially when data is extrapolated from 195 countries from Russia to Papua New Guinea ! Different continents, different people and culture, different socio-economic groups.
How to standardise one drink at 10g pure ethanol boggles the mind.
Secondly the authors kept harping on the figure (one year’s figure,2016 ) of 2.8 million deaths, world-wide, as attributable to alcohol use. And also the very dubious statement that alcohol consumption was the leading risk factor for Premature Death and Disability. They quoted alcohol’s role in 27 health risks ranging from shrinking brains, heart, liver, pancreatic damage, sexual dysfunction and infertility, diabetes, thinning bones, lung infections, behavioural changes, hallucinations, slurred speech, muscle cramps and changes in coordination, birth defects, diarrhoea, stomach distress, fatigue and numbness, and of course, the final diagnostic wastebasket: Cancer.
A lot of accusations, but where are the unbiased figures ?
The only attributable leading causes of death in the study are :
In the 15-49 year age group : Tuberculosis, road accidents, and self-inflicted injuries ( suicide)
And in the above 50 age group : Cancer ( unspecified )
Can anyone among the readers make head or tail out of the conclusions? I cannot.
Perhaps a fairer way to look at the effects of alcohol on health would be this :
In non-drinkers, out of a population of 100,000, 914 people will develop one of the 27 diseases in question in any one year.
In those who drink ONE drink a day, 918 people will be affected. An increase of 4 out of 100,000 people
And In those who drink 2 drinks daily : 977 ( 63 per 100,000 more compared to the teetotallers )
You can draw your own conclusions from these.
My views on drinking.
I accept there are risks associated with ethanol.
But all pleasurable activities also carry risks : driving, running, swimming, trekking, even eating.
The logical answer is not abstinence. It is moderation.
And when it comes to moderation nothing tops the dictum of : A bottle of wine when opened, begs to be shared !
Cheers !
Dr Alan Ho Chok Chan is a Paediatrician in private Family Practice. He also spends time golfing, swimming, playing tennis, wine tasting, playing guitar and singing. He is also a bibliophile and voracious reader.
