Posted by Larry Hoover on December 30, 2002, at 9:22:18
In reply to Beardedlady, 2 marijuana/schizo links for you, posted by comftnumb on December 26, 2002, at 11:49:57
> http://www.healthscout.com/static/news/510418.html
>
> http://www.xpresssites.com/buffalo/buffalo/xpspecialsections/healthandwellness/story_318702.asp
>
> Summary:
>
> Smoking pot increases the risk of the psychiatric disorder by about 30 percent (1st article, Swedish study). There is genetic evidence that marijuana use can result in schizophrenia or a significantly increased risk of schizophrenia (2nd article, Japanese Study). Cannabis used before age 18 raises the risk of schizophrenia six-fold (2nd article, Swedish Study).I've got to add a critical commentary to this information. First, let's consider the first article linked above. It states that:
""It's not as good an explanation than the possibility that cannabis itself causes schizophrenia," says Dr. Stanley Zammit, a psychiatrist at the University of Wales College of Medicine in Cardiff and lead author of the new paper.
Zammit and his colleagues compared schizophrenia rates and marijuana use among more than 50,000 Swedish conscripts who participated in a 1969-1970 survey. Of those, 362, or about 0.7 percent, had been hospitalized with the disorder by 1996.
Of the 11 percent of the entire group that admitted ever trying marijuana, 73, or 1.4 percent, went on to be hospitalized for schizophrenia. The odds of having the disease grew as pot use climbed, reaching nearly a sevenfold increase in the men who used it 50 times or more but tried no other illegal drugs, the study says.
The effect was strongest among soldiers who developed schizophrenia within five years of entering the military. It held after the researchers accounted for use of alcohol and other drugs."
The author's conclusion of a causative link between marijuana use and schizophrenia is scientifically inexcusable. They have inserted their own beliefs into the equation, a practice that I find to be reprehensible as it leads to all kinds of misrepresentations of the data. In essence, they have made a political statement based on incomplete evidence.
The existence of a correlation between two variables can have four possible explanations. In symbolic terms: the existence of A leads to B; the existence of B leads to A; both A and B are caused by a separate and unmeasured variable; the relationship is a chance artifact of the sampling process.
Even the selection of the wording of the correlation demonstrates the bias of the reviewers of the data. In this case, marijuana is pre-selected as the independent variable, and schizophrenia is selected as the dependent one, implying that marijuana use leads to schizophrenia.
Now, because of the sample size, it is pretty unlikely that the correlation is due to chance, but there is no evidence that any of the other three possible relationships should be given precedence over the other two. Even the paragraph that suggests that, "The effect (of marijuana) was strongest among soldiers who developed schizophrenia within five years of entering the military." can be seen to strongly support the converse relationship, that those already developing schizophrenia would conceivably be more likely to self-medicate with cannabis.
It is a scientific axiom that correlation is not causation, and the authors of this study have committed a major error in judgment to have used that term. In an epidemiological study (one where no manipulation of variables occurs, i.e. subjects were not *randomly* assigned to marijuana use or non-use and then observed), that error is particularly egregious. There is an inherent confound in that those with a predisposition to schizophrenia may self-select into the group of marijuana users, fundamentally distorting the relationship between the two variables.
I'm going to deal with the second article, and other research, in another post.
Lar
poster:Larry Hoover
thread:133136
URL: http://www.dr-bob.org/babble/20021230/msgs/133784.html