Cannabis cohort studies

Every few months an article will crop up telling us that cannabis is ruining people’s education, making us schizophrenic or generally corrupting our precious youth. But how is it that people ever draw conclusions on such matters when experiments are completely untenable? Cohorts. This is a word that strikes fear into the funding bodies that fuel scientific research. They require decades of time, huge numbers of people and large sums of money. Cohort studies are longitudinal in nature; they measure people’s progression over long periods of time. Therefore, they allow researchers to try and tease apart the often rather attached chicken and egg variables. But are cohort studies all they are cracked up to be? They suffer the problem that correlation doesn’t imply causation, yet this doesn’t seem to stop national newspapers from using them to disseminate over simplified messages. The Dunedin cohort has interviewed and tested over 1,000 people in New Zealand since the early ‘70s. This cohort has produced data for several major scientific findings, such as the link between various genes and mental disorders. From this same cohort, researchers from King’s College London reported this year that “regular” adolescent cannabis use leads to a decline of around 6 IQ points by adulthood. The press unsurprisingly latched onto this result and reported it without touching on the possible shortcomings of cohort studies. It is true that people diagnosed with cannabis dependence at three interview points during the longitudinal study showed an average drop of 6 IQ points from their pre-cannabis using IQ. However, only about 10% of cannabis users become dependent, so it is foolish to claim that cannabis users in general show a large decrease in IQ by adulthood. Admittedly the study did a good job at taking account of other, potentially confounding factors, such as other drug use and schizophrenia. However, it did not consider the impact of depression, which is related to cannabis use and may have detrimental effects on educational achievement and IQ. This highlights a general issue with these correlational, cohort studies: you can never completely isolate one variable’s impact on the outcome. Similarly, an Australian cohort from the ‘90s has recently been used to investigate the effect adolescent cannabis use has on the development of depression and anxiety. Daily use in teenage girls increased the likelihood of these mood disorders by a factor of four, a seemingly unnerving relationship. However, this finding must again be interpreted with a pinch of salt because although other drug use was controlled for, there are a multitude of other factors that can play in to both the development of cannabis use and mood disorders. This factor of four probably isn’t due completely to cannabis use. Cohort studies are certainly a whole lot better than simple one-off correlation studies which cannot probe the temporal relationship between two variables. Yet cohort studies must still be interpreted cautiously. It is reasonable to claim that very heavy cannabis use in adolescence does contribute to cognitive decline and this kind of research is critical in helping to develop evidence based drug policy. However, the media must stop portraying these hard-to-interpret cohort studies as simple, undeniable fact.