If it caused 10% of individuals to have symptoms then they’d detect a p value like those, the other 90% could be pleased as pudding for all the p value would care. I doubt it’s that high, my comment is just that the quoted p value and association does not actually provide any support for the statement

And the first of your sources does say that 7-19% of officers get ptsd but also that 25% kill or seriously injured someone, which is a larger number than get ptsd. so that suggests 6-18% officers kill or seriously injure someone and don’t get ptsd. Assuming the ones that get ptsd are the same as the ones doing bodily harm, unrealistic but conservative, that means half of officers who do that could easily be just dandy.