Clinicians should measure total testosterone more than once and obtain a symptom history before making a diagnosis. Validated questionnaires are not sufficient for either diagnosis or monitoring.