1 pillhead per 10,000 population who cuts 30 years off his life is the way this comes up. Statistically, it's a weak analysis method when reviewing small, correlated populations within large randomized populations. Put the article down as "well cooked".