Problem Drinking Studies

Scientists who claim that a glass of wine a day will help you live longer may be all wet. An analysis of a mountain of data from previous studies on the subject, presented in this month's issue of the journal Addiction, suggests that there's no clear benefit from moderate drinking.

After reviewing many studies touting a link between moderate drinking and longer life, epidemiologist Kaye Fillmore of the University of California, San Francisco, says she found "all kinds of potentially serious problems." One of the most egregious, she says, was that researchers had often lumped two kinds of nondrinkers together--those who never drank ("abstainers") and those who had quit ("former drinkers")--that could have much different life expectancies. And several studies did not appear to take into consideration major risk factors that influence life expectancy, such as socioeconomic status and overall health.

Fillmore and colleagues pooled raw data from 10 such studies and analyzed them more rigorously. Unlike the previous efforts, her team divided the nondrinkers into abstainers and former drinkers, analyzing each group separately. In the end, they found no significant difference in life expectancy between abstainers and moderate drinkers. The researchers also found a hidden factor that could help explain the earlier results: Abstainers and former drinkers were both more likely to be in poorer health and of a lower socioeconomic status than light drinkers.

But Fillmore's study has some problems of its own, asserts one epidemiologist studying the alcohol-mortality link. Jian-Min Yuan of the University of Southern California in Los Angeles notes that Fillmore's sample group consisted mostly of people under 25--a group more prone to alcohol-related accidents. The adults, the most important group for assessing alcohol's long-term health consequences, numbered only about 6000 people--not enough for such a complex analysis, he asserts. "Based on these numbers, it's very difficult to analyze using as many variables as they did," he says. But Yuan agrees that his colleagues need to be more careful in devising studies with proper controls.

Posted in Math