Sometimes you just have to find your own evidence.

Tuesday's Wall Street Journal Health Journal column earns a show-me-the-evidence rating of 4.5 soup cans (out of a possible 5).

In The Case for Alcoholics Anonymous: It Works Even if the Science Is Lacking [login required], columnist Kevin Helliker shows how difficult it is to apply formal, evidence-based methods to programs such as substance-abuse interventions. He is responding to the Cochrane Library's March release of a systematic review titled Alcoholics Anonymous and other 12-step programmes for alcohol dependence. Their study concluded that this therapeutic approach is no more or less effective than the alternatives.

No, that's not what it means. Findings like this inevitably lead some people to believe that a particular treatment was "proven" ineffective. Helliker writes: "Could this mean that A.A., the world-wide fellowship of recovering alcoholics, doesn't work? What it actually means is that the pursuit of evidence-based medicine sometimes produces conclusions and headlines that are misleading." The column does a good job of explaining that AA will probably never go through the gold-standard, double-blind randomized clinical trial -- particularly since "Such a trial could raise ethical questions if, for instance, a newcomer were steered to an alternative treatment - including possibly a control group receiving no treatment at all." Helliker also explains that associative studies often provide useful evidence, saying that "A multitude of studies show that A.A. attendance is associated with reduced drinking and higher social functioning. ...Unlike randomized clinical trials, associative studies don't prove cause, but are often treated as powerful evidence."

About the soup-can rating. The column provides an appropriate level of detail for a business newspaper, and includes a table of quick facts about AA attendance. Our reasons for not assigning a full 5 soup cans are 1) the lack of links to additional information sources, and 2) the absence of an online forum for discussing how to collect and weigh evidence about these programs. The author's email address is provided, however.