Wednesday, September 30, 2015

Alcohol, heart attack and unwarranted correlations

Let's assume that we want to design a medical research on cancer: we set up an experimental and a control group, made up of 50 individuals each, and we check their health for a certain period of time, let's say, 25 years. We order the members of the experimental group to say “Merry Christmas!” as soon as they get up from their bed every day in the morning; members of the control group should say nothing. After 25 years we compare the results. Let's imagine that 2% (1 out of 50) of people belonging to the first group died of cancer, whereas the amount of people belonging to the control group who died for the same reason is 10% (5 out of 50).
Conclusions: saying “Merry Christmas” every morning reduces the risk of cancer disease.
What a nonsense, you may say. Indeed it is. Unfortunately it's not unusual to find similar mistakes in social and medical researches. Actually the real mistake rests often in the the interpretation that journalists, especially those with limited knowledge of methodology of social and medical research, give to such studies. The recent news about beer, women and heart attack is a clear example of this approach. A Swedish study, based on data collected over a 32-year period on 1462 participants, found that women who drank moderate amounts of beer were at a reduced risk of getting heart attacks. Online news magazine jumped immediately to the following conclusions:

Why women should drink beer: Two pints a week slashes the risk of heart attack by a third
(DailyMail)

Women who drink 2 pints of beer a week cut heart attack risk by a third
(Mirror UK)


As a matter of fact, things are not as straightforward as these headlines claim. The Swedish researchers simply made a correlation of two things (namely: drinking beer and suffering from heart attack), but such correlation can be arbitrary (just as believing that saying “Merry Christmas” every morning has the power to avoid cancer). It is possible that, among the women who drank and didn't have heart problems, a considerable amount of them went to the pub on foot. In this case, what might have prevented them from suffering from heart disease was not the drinking, but the physical activity. We can also assume that those who drink have a more social active life, which is an index of healthy lifestyle. It's not my intention to put down altogether the Swedish research. I'm just saying that we cannot come to convincing conclusions based on a relatively small study like this one (featuring less than 1500 participants in a limited geographical region). The fact that enthusiastic online journalists trumpeted the study results (probably without even reading the whole research) and the benefits of beer, shows three interesting things: 
1) alcohol lobbyists are powerful and always on the lookout for new groups of consumers. Now it's the turn of women (by the way, I am sure a new sensational study on the benefits of alcohol on pregnant women will follow soon...)
2) I detect here a certain disapproval towards people who promote healthy lifestyle and try to highlight the danger of drinking alcohol. For instance, I read the news following a link on FB. The link was posted by some FB female friends of mine who celebrated the good (?) news as a “slap in the face" to those who claim that drinking alcohol is always unhealthy. Such misleading news hit considerably the ego of those suffering from “cognitive dissonance” (a situation involving conflicting attitudes, beliefs or behaviors: drinking is dangerous vs I like drinking).
3) Beer, wine and alcoholic beverages are very popular in our society, therefore many people don't accept the fact that their consume brings poisoning effects on our body and mind, even if it contributes to specific economic sectors (but causes even larger losses on health budgets).

No comments:

Post a Comment