Did the study take into account the prevalence of each substance? Alcohol would be the most prevalent by a wide margin. If it's normalized for prevalence however, then the results are mighty interesting.

Yes, it was normalized. It was a very comprehensive, well thought out study. Of course the researchers could still get the results they wanted by choosing which metrics to include in the comparisons, but the choices attempted to balance individual cost (physical, mental) with social (family suffering, divorce) and societal (burden on society from underperforming people, shared health costs, etc.)

The takeway was that alcohol was one of the worst, along with crack cocaine and heroin; ecstacy was really not bad, and weed was a bit worse but still nowhere near as overall costly/damaging as alcohol.