american football could be the next generation's smoking in hospitals

Ta-Nehisi Coates, longtime football fan--but more importantly, football dad--shares a pretty jarring story:

Six plays into the game, another Brave was removed after a hard hit. An official with the Tantasqua team said the eyes of one of the boys were rolling back in his head.
But the game, an obvious mismatch between teams from neighboring towns in central Massachusetts, went on, with Southbridge building a 28-0 lead in the first quarter. The game went on without the officials intervening. It went on despite the fact that the Braves, with three of their players already knocked out of the game, no longer had the required number of players to participate. Even with what are known as "mercy rules"—regulations designed to limit a dominant team's ability to run up scores—the touchdowns kept coming, and so did the concussions. When the game ended, the final score was 52-0, and five preadolescent boys had head injuries, the last hurt on the final play of the game.

That's a Pop Warner game, based on the description in the article, in the "Peewee" division (9-12 years old, 75-120 pounds!)

A lot has been made over the ill health effects of repeated head traumas suffered by NFL players. But it is looking more like the problems of concussions are not at all limited to the extremes of size and speed.

I've mentioned a couple of times that I saw a presentation a few months back that gave me pause about my own football fandom. I'm still not at liberty to discuss the particulars of the study. However, I do think I can mention a few things, and they really do bear mentioning sooner rather than later.

1) It's a longitudinal study, meaning that they are following a cohort over time (which is why the data aren't published yet, and probably won't be for a few more years.) This is a really powerful study design, because you have baseline data for every person in the study--much better than the standard epidemiological type study in which you look at groups of people after the fact (e.g., "people who played football for X years" and "people who played soccer for X years") and make inferences about relative risk based on how many people in each group have a particular health outcome. If 5% of your football players have (say) migraines* and only 2.5% of your soccer players do, then you can say playing football is associated with doubling your risk of having migraines. But it doesn't tell you anything about an individual's risk of migraine, or what specifically leads to that outcome.

2) Early data on the outcomes they are looking at are astonishingly robust, and they  correlate almost perfectly among the subjects with how many hits they take over the course of their playing careers. 

3) The cohort are high school players.


*This study does not look at migraine, at least to my knowledge.