Importantly, the study "assumes" that participants performed a "maximum number of hours per week" rather than measuring the actual hours per week. Because they "assume" the amount of time spent exercising, their injuries per 1000 hours metric is simply wrong.

Specifically, this study conflicts with several other studies that found significantly higher rates of injury (and which this study acknowledges), including 2 studies which found a 20% rate of injury and 50% rate of injury for Crossfit athletes, respectively, using controlled observational samples rather than the self-reported, self-selected sample of this study.

Anecdotally, I don't know any current Crossfitters without overuse injuries--including those who go to the "safe" gyms. I also know a larger number of former Crossfitters, all of whom stopped because of injury.