This is probably the most important stat about drugs, however, it falls victim to a pet peeve of mine. For cigarettes and alcohol and drugs of abuse (the ones

This is probably the most important stat about drugs, however, it falls victim to a pet peeve of mine. For cigarettes and alcohol and drugs of abuse (the ones without accepted medical applications - street heroin, cocaine, etc.), the analysis is pretty useful. However, with prescription drugs, simply counting the number of deaths is misleading. ADR's are tragic, and we should be doing all we can to learn more to prevent them. However, a correct analysis, at least when attempting to optimize the way it seems he is trying to, is to total the number of deaths caused by drugs, and then subtract the number of deaths prevented by these drugs. All optimizations should be addressed through that lens. I think the author would agree with this, I just didn't see any mention of the fact that while drug deaths are tragic, there are also many deaths prevented by current application of drugs, and ignoring the opportunity cost of changing drug application (potentially reducing the number of lives saved) is potentially dangerous.