I mean, i would assume that foster care => unstable home situation/significant trauma => more likely to have mental health problems. It seems like its hard to take the stat at face value without controlling for correlations