One point that hasn't been addressed so far is how misleading this study is. It's looking at death rates related to prescription painkiller deaths doing a time-series analysis 1999-2010 and then correlating results to states that made medical marijuana legal over that time period. It doesn't say how they eliminated any other number of causes during that period. This is merely showing correlation not causation.
As an example people who have more sex make the most money. While we would all like that to be true there's no science proving causation http://blogs.scientificamerican.com/roots-of-unity/sex-makes...
Most of the examples given in the comments here are anecdotal. Not that I disagree with the overall tone, but this is not the study to hold up and say "This is why we should legalize"