Bias in co-occurring research

An interesting new study on drinking and depression:

"Not all studies have found a significant relationship between drinking and depression," said Kathryn Graham, senior scientist at the Centre for Addiction and Mental Health, "and some have found a relationship for one gender but not the other. In our study, we included two quite different types of measures of depression. We also used four clearly different types of alcohol consumption measures that examined both drinking pattern as well as overall consumption." Graham is the corresponding author for the study.

I hate to sound jaded, but when I read this it felt a little like a search for conclusive proof of a strong relationship between the two and self-medication as the probable explanation. I figured I was just too jaded and read on...

Results indicate that measurement and gender are key issues in interpreting findings on the relationship between alcohol and depression. Specifically, depression is primarily related to drinking larger quantities per occasion, is unrelated to drinking frequency, and these effect are stronger for women than for men.

"Depression is most strongly related to a pattern of binge drinking," said Graham. "A pattern of frequent but low quantity drinking is not associated with depression. In fact, those who usually drink less than two drinks per occasion and never drink as much as five drinks are less depressed -- for both measures of depression -- than former drinkers. This relationship with drinking pattern is greater for women than for men."

Second, the overall relationship between depression and alcohol consumption is stronger for women than for men, but only when depression is measured as meeting a clinical diagnosis of major depression. Conversely, there is no gender difference when depression is measured as recent depressed feelings, which is commonly done in research on this topic.

The first conclusion is interesting--it makes a lot of sense that depression would be more strongly associated with heavy drinking episodes rather than the frequency of drinking. The second finding also is not surprising, women self-report depressive symptoms at higher levels then men, and the study is based on a phone survey.

Finally comes the self-medication hypothesis:

"This pattern of associations is more consistent with women using alcohol to counteract depression -- by high-quantity drinking and intoxication -- than with chronic alcohol consumption tending to make women depressed," said Wilsnack. "However, a vicious circle could possibly begin with drinking in response to depression....

The bottom line, said Wilsnack, is that "clinical depression may encourage some women to drink large amounts of alcohol in hopes of numbing depressed feelings, with risks of alcohol abuse and dependence. Therefore, clinicians treating women for depression really need to be concerned about women's use of alcohol, because of the risks that women may try to medicate their moods with alcohol."

This in spite of the fact that the source article itself says "these cross-sectional data do not provide information about temporal ordering or causation".