Reliability, Not Overdiagnosis

Letter to the Editor

Sir: The recent Journal article by Zimmerman et al. does
not show that bipolar disorder is overdiagnosed, because it mistakes
reliability and validity. All the authors showed was that,
about half the time, researchers reached different diagnoses
than clinicians who had diagnosed bipolar disorder. This merely
reflects interrater diagnostic reliability, a long-established issue
in epidemiology.