The report by Anderson and colleagues (2014) highlighted several age and sex-specific associations between medical marijuana policies and suicide rates and we also sought to examine whether these would remain significant in our more comprehensive results. Thus, Table 3 reports the results of our analyses stratified by sex and age group. These stratified analyses utilized the fully adjusted Model 3 and the refined Model 4. For comparison, we also show the results of the partially adjusted Model 1, which was most similar to that employed in the report from Anderson and colleagues (2014). In the fully adjusted Model 3, the associations between medical marijuana policy and suicide risk among the twelve age-by-sex groups were non-significant with only one exception: men over 60 (OR=1.04; 95% CI: 1.005, 1.105; p=.04). This association would suggest that medical marijuana policy increases risk for suicide; however, it would not meet nominal significance criteria after adjustment for multiple testing and therefore will not be considered further. In the refined model (Model 4), the ORs were again very similar to the previous model, with slightly larger confidence intervals (Table 2).
