Injuries in **any** sport are a complex set of issues.  Most people think of
injury as an event, but, in reality is merely part of a process with many
elements.  Some of the elements in this process include type of sport (e.g.
contact, collision, combative sports, etc.), field conditions, previous injury
(and the degree of rehabilitation), musculoskeletal maturity, training
techniques, fitness, etc.  In many sports, including hockey and soccer, a
significant number of injuries are the direct result of rules violations (in
soccer, in fact, the person committing a foul is **more** likely to be
injured!) and may be prevented by stricter rule enforcement.
 
A recent article in the British Journal of Sports Medicine reported on injury
rates on an elite Swedish hockey team.  85% of the injuries found in 4 seasons
were due to trauma (as opposed to overuse or "misuse").  30% of the injuries
involved the head and face.  The authors concluded that stricter rule
enforcement might reduce the number of injuries.
 
The main problems with these kind of studies relate to operational
definitions.  Depending on how "injury" is defined, it is hard to compare
diferent studies.  In this paper, the main flaw is that they included injuries
that did not occur during games or during practice!  I have a hard time
accepting an injury as being caused by a sport when it is unrelated to the
sport. Nonetheless, the injury rate her was 2.6 per 1000 practice hours per
player and 74.1 per 1000 game hours per player.  In most other studies, injury
rates are higher during practice than in games (yes, this is true for football
and other collision sports).  Think about it.  Most time is spent in practice,
so the exposure probability is higher.  So what is happening in hockey?  You
can think of many explanations: practice time is spent more on drills with
less contact; rules are inconsistently enforced (from my reading of the
HOCKEY-L, I guess this is far fetched); intensity differences between game day
and practice sessions.
 
I don't have any articles comparing injury rate and severity for hockey and
other sports (at least none that compare similar levels of skill, development,
etc.).  Most everything is anecdotal and subject to lots of bias.  We have a
tracking system at MSU, but it is painful to extract information for different
teams.  One of my long range goals is to sort out the injury and exposure
rates for our teams, but it ain't gonna happen soon.  Sorry, not enough hours
in the day!