The fraction of teens with BMI < x is not a measure of how many teens have eating disorders. To pick the simplest possible confounding factor - teens dying of c

The fraction of teens with BMI < x is not a measure of how many teens have eating disorders. To pick the simplest possible confounding factor - teens dying of cancer would be false positives.