Cannabis intoxication and fatal road crashes in France: population based case-control study

There is one thing about the findings of the study by Laumon et al.
that seems to be at odds with the overall claim. The study states that
2.9% of French drivers are found to have significant cannabis levels,
whilst cannabis is a causal factor in 2.5% of fatal accidents. This
suggests that cannabis use causes no greater proportion of accidents than
would be expected if it did not increase risk. This discrepancy may be due
to unknown or unmeasured confounding variables. It is a fundamental
limitation of case-control methodology that it cannot exclude confounding
variables that we either are not aware of or cannot measure. Even the
dose-response effect does not prove causality if the confounding variable
is correlated with both effect and putative cause. Is it possible, for
example, that increased cannabis levels are correlated with personality
factors that are independently associated with greater accident risk?
This study certainly moves us closer to demonstrating the risk caused by
cannabis, but the case is not yet proven.

Competing interests:
None declared