This paper aims to assess the relationship between national drug policies and drug use prevalence rates in seven countries between 1996 and 2016. Our study departs from the extant literature by using a novel set of drug policy indexes encompassing six different policy dimensions: consumption, possession, traffic, harm reduction, treatment and prevention. We find that, in the case of cannabis, policy changes in the direction of a less criminally-oriented approach towards consumption and possession contribute to a decrease in prevalence rates; by contrast, a less criminally-oriented approach towards the traffic of cannabis is associated with increases in prevalence rates. The results for cocaine are somewhat different: whilst drug policy changes in the direction of a less criminally-oriented approach towards consumption decreases prevalence rates, the opposite is true for possession. Therefore, our results for cocaine are in stark contrast to those obtained for cannabis and we find that similar (in nature) drug policy changes have an impact on prevalence rates which differs across drugs. Finally, in what concerns ecstasy, we did not find evidence of a relationship between drug policy and prevalence rates.