A new meta-analysis of 57 psychological studies found that essentially identical patient interviews can lead to different diagnoses for the same exact patient.