Alright, so I went back through the Rule 48 paper on concussions. Thoughts are basically unfiltered, plus I’m a born pessimist so you have to account for my personal negative bias towards almost anything in general. I haven’t edited this for readability, is written conversationally, and my ADHD is shining bright today so read at your own peril.
First things I look at: authors, funding, and the ‘call room view.’
Donaldson: Couldn’t find anything. Research assistant, stats consultant, or PhD candidate likely.
Asbridge: Associate Professor at Dalhousie University in Nova Scotia. Sociology background. Does a lot of work in substance abuse epidemiology. Social science guy basically. Also makes me happy to pronounce his name Assbridge in my mind cause I’m still 13 at heart.
Cusimano: MD, Brain surgeon, researcher, rocket scientist etc…. Pretty legit guy.
-Side Note: Asbridge and Cusimano also have a paper coming out in JAMA in a few days on "Prevalence and correlates of traumatic brain injuries among adolescents." Looks interesting. Could also point to Rule 48 paper being a "throw away" publication.
Canadian Instiute of Health Research Team Grant in Applied Injury Research. Ties to Sport Canada. Shouldn’t be a source of bias.
Call Room View:
So I do what every resident does, read the abstract and conclusions to see if its worth reading or if I can use the info quickly. When I dig into a paper like this I see if the abstract and what the data suggests match up. Nothing too major in this one. Their flawed conclusions are consistent from paper to abstract.
Onto the paper…
Intro: yada yada expected stuff. Hypothesis’ of 1) rule 48, if enforced, would reduce concussion incidence and 2) NHL would have higher concussion incidence than OHL. I can see why they would hypothesize #2 but disagree with their reasoning behind it of stricter head shot rules in the OHL. Younger brains are more vulnerable brains and concussion prone players have been filtered out to a greater extent in the NHL than OHL.
Retrospective cross sectional anaylsis. Basically they look at 10 weeks of games in each of the 3 seasons 09-10-11 for both the NHL and OHL. Rule 48 was introduced in 10 and broadened in 11. So a year before, then blindside implementation, then further coverage of shots to the head, and looking at the difference in rates. Average week for then had 40 NHL games and 25 OHL games so around 1200 NHL games during the period. Got penalty info from ohl.com and nhl.com.
Concussions officially reported by team were only definite concussions recorded. They used a proxy of internet data and reporter info for possible concussions and included non-nasal facial fractures as possible concussions. Obviously that’s not ideal but probably the best you can hope for at this point. What irked me about this is they backed the use of third party injury info with an NBA study, not one of the several studies done in hockey on this issue. Injuries are more clear cut in the NBA and the incidence of concussion is too low to have this support your use of collecting this type of injury info. Not disagreeing with using the media, just their supporting evidence.
Including facial fractures just pissed me off. They included them because "such injuries are caused by the application of substantial force to the head and are associated with a number of intracranial injuries" and cited a large trauma study which stated THE COMPLETE FUCKING OPPOSITE FOR CONCUSIONS and even included the sentence "No associate was found for concussion only" in the damn abstract. Come on man.
They did look at secondary contact of the head or body with the boards or ice which was nice.
They also looked at penalty data.
Used standard stuff.
First off, fuck table 1. You publish a paper, get publicity all over the NHL community from it, and you put this piece of shit and hard to organize table in your paper? I understand wanting to copy and paste the table straight from STATA, but your peers have to read that. Lazy, and bias’ me further to towards the "throw away study" angle. Anyways..
Concussion rates increased significantly from 09 to 10 but were stable from ’10 to ‘11 seasons in the NHL (confidence interval (CI) 0.42-0.86). OHL rate was stable the whole time. Concussion incidence was significantly higher in the NHL at 5.23/100 games vs 5.05/100 OHL (CI 1.01-1.50). This difference can be explained by the ’10 season uptick.
Penalty rates were lower in NHL vs OHL 9.4/game vs 13.9/game (CI 0.66-0.69) and the proportion of fights and ‘aggressive’ penalties were higher in the OHL. This was true in the relative risk ratio (RRR) for total over 3 years and for each individual year. They did not included the rates of penalties in each individual year, which should have been included. Was the uptick in concussion rates due to less penalties being called in the ’10 season? More penalties being called? No clue, cause they didn’t include the data. When someone doesn’t include basic info like you wonder what they are hiding and if they have an agenda by doing so.
Then they looked at 123 ‘randomly selected concussions or suspected concussions’ to try and find the cause. How did they decide 123 was the right number to look at? What’s the split of concussions vs. suspected concussions? Was each year evenly represented in both number and concussion vs. suspected concussions? Who knows.
Concussions that resulted from penalties, fine, or suspension accounted for 28.4% of concussions and 36.8% of suspected concussions. Since we don’t know the breakdown of these incidents, if we use the max and say all were concussions or all were possible concussions, a max of 35 concussions and 45 possible concussions were assessed penalties. Even at these max numbers the sample size is small. The most common penalty assessed was fighting at 32.3%, check to the head at 11.8%, and other aggressive penalties 44.1% and non-aggressive penalties 11.8%. Table below nicely shows how small the sample size is.
They also did a regression for risk factors such as position, age, height etc.. Prior concussion history was not looked at. Defensemen, # of fights per 10 games, and games on roster were all positive risk factors while age, minutes per game, and height and weight were negative risk factors.
So rule 48 didn’t decrease overall concussion rate because blindsiding is not a common cause of concussion. But it has reduced blindside concussions.
Authors talked about being surprised at the OHL concussion rate being so high. Touched on it before but age and less filtering of concussion prone players are likely causes. Authors concluded the increased rate of fighting and aggressive penalty calls in OHL is cause of high concussion rate. May have some part, but I think age is more important factor. They might shit a brick if they realize using their methods, suspected concussions would be greatly underestimated due to smaller degree of media attention and injury speculation when a 4th liner from Guelph gets his bell rung as compared to an NHL player.
Authors also conclude lack of enforcement of rule 48 as a cause because it is called 10-fold less often per game than the similar OHL rule. Not sure how you can conclude that. Also they don’t mention or even consider that, from their own data, rule 48 HAS decreased number of concussions from blindside hits. It seemingly has done what it was meant to do.
Then they say that a ban on fighting or increased penalties should be considered. Just….No. How can you honestly conclude from your own data that banning fighting will meaningfully reduce concussion rate? 11 incidences of concussions or possible concussions during 3 years? Smaller rate than getting hit by a puck. Never mind the fact that we don’t know how many of those 11 were concussions or just possible concussions. If a player suffers the dreaded “upper body injury” during a fight, the false hit rate for possible concussion will be higher than any other penalty. Shoulder could have been popped out, hand could be too bruised to close on a stick…doesn’t matter it could get defined as a possible concussion by their methods. Plus, facial fractures also are thrown into their data, which further dilutes it.