Tuesday, May 30, 2006

How To Read Medical Studies

This is something journalists and other popularizers often have trouble doing. Consider this left-column summary of a blogpost in the Broadsheet:

Back away from the chardonnay!
A drink a day reduces men's heart disease risk, but not women's. What a buzz kill.

This refers to a recent study on the effects of alcohol on heart disease risk. But the study findings are actually not what the above summary implies:

A team of researchers in Denmark studied over 50,000 men and women aged 50 - 65. Details on alcohol intake and drinking frequency over the preceding years were collected and the participants were then monitored for an average of five years.

During the study period, coronary heart events were recorded and the results were adjusted to take into account known risk factors, such as smoking, diet and physical activity.

Over the course of the study, women consumed an average of 5.5 alcoholic drinks per week, while men consumed 11.3. Meanwhile 1,283 men and 749 women developed coronary heart disease.

The study found that women who drank alcohol on at least one day a week had a lower risk of coronary heart disease, compared to women who drank alcohol less than once a week.

However the risks were similar for drinking on one day a week (36% reduced risk) or on seven days a week (35% reduced risk). This suggests that the amount of alcohol consumed is more important than drinking frequency among women.

In contrast, risks for men were lowest among the most frequent drinkers. Men who drank one day a week had a 7% reduced risk, while men who drank every day had a 41% reduced risk. This suggests that it does not matter how much men drink, as long as they drink every day.

Note that women who drank on seven days a week had a 35% reduction in the risk. So a drink a day indeed does appear to reduce women's risk of heart disease. It just doesn't seem to reduce it any more than drinking once a week.

The reference to "amount of alcohol consumed" suggests that the women who drank only once a week did so rather liberally. They may even have been drunk. So an alternative way to interpret this study is to say that it's a healthy thing for women in this age group to get really soused every seven days. Not a buzz kill, after all.

Just kidding...