Maybe there's such a thing as "Reefer Madness" after all.

We always thought that it was just a propaganda movie from the 1930s. But an analysis in this week's Lancet of previously published studies found that marijuana smokers were 40% more likely than non-smokers to develop psychotic symptoms, even when they weren't high. And the risk increased with the frequency that users toked up.

A commentary published along with the article estimates that, if the findings are correct, roughly 800 cases of schizophrenia in the U.K. could be prevented each year, if no one smoked marijuana.

The analysis reviews dozens of studies that looked at the rate of mental health problems among marijuana smokers. (The study, along with the editorial and a commentary, are here, on the Web site of the British newspaper the Guardian.)

As the researchers themselves note, a major problem with the data is the possibility that people who are more likely to develop psychotic symptoms in the first place are also more likely to smoke marijuana. The studies parsed in the analysis went to different lengths to address this problem. Several excluded patients who showed psychotic symptoms at the outset.

The results of the analysis were enough to convince the authors that the risk is real, however. And, perhaps more significantly, the findings also persuaded the editors of the Lancet, which is among the world's most influential medical journals, to reverse course on the effects of marijuana. In an commentary accompanying the study, the editors write:

      In 1995, we began a Lancet editorial with the since much-quoted words: "The smoking of cannabis, even long term, is not harmful to health." Research published since 1995, including [the] systematic review in this issue, leads us now to conclude that cannabis use could increase the risk of psychotic illness. … governments would do well to invest in sustained and effective education campaigns on the risks to health of taking cannabis.