Reliance on Diagnostic Codes
The study relies on diagnostic codes for CUD and schizophrenia, which may not fully capture the complexity of these conditions. Individuals with subthreshold symptoms or those who have not sought treatment would be missed, potentially leading to an underestimation of the true association.
While the study controls for several factors, it does not account for all potential confounders, such as tobacco use (which has been linked to psychosis), frequency and amount of cannabis use, age of first use, or specific THC content of cannabis products. These unmeasured confounders could influence the observed association.
While the study utilizes national registers and a large sample size, it is observational and thus cannot establish causality. It remains possible that the association between CUD and schizophrenia is influenced by unmeasured factors or that the relationship is bidirectional (i.e. schizophrenia increases risk of cannabis use).
Potential Over-adjustment
Although unlikely, there is a possibility of over-adjustment in the model by including other psychiatric disorders as covariates. Some of these other disorders could potentially be mediators (intermediate steps) in the causal pathway between CUD and schizophrenia, thus somewhat attenuating the true direct effect of CUD.