I do, because it normalizes a bit for how much time they spend on the floor.
One player might average 10 points a game and another 5 points per game... but if the first player gets 22 minutes to the second player's 11 minutes, they really score at about the same rate while they're on the floor.
In a more real example, Issa Thiam averages 0.4 steals per game, while Eugene averages 0.5... but when you take into account their relative time on the court, Eugene has nearly twice as many steals per minute played as Issa does. Just looking at overall "per game" numbers, you don't see how much better Eugene was at disrupting passes/handlers.
Or compare Gettys' 1.2 assists per game with Eugene's 1.0.... the per game numbers make it look like Gettys' was slightly better at setting up his teammates for points. Adjusting for minutes played, though, Eugene had about 50% more assists per minute played than Gettys did.
When you have guys who only play 10-14 minutes per game, the "per game" numbers don't really tell the whole story. The "per 40" numbers don't either, but they give a little different context and perspective.
When looking at the "per 40" numbers, though, you have to ignore players who see limited action because their numbers get skewed pretty dramatically. For instance, Khalil Batie has 5.7 assists per 40 min... but he only played 14 min all season, with 2 assists.