Gender differences in musculoskeletal injury rates: a function of symptom reporting?