Identifying postpartum depression: are 3 questions as good as 10?

BACKGROUND Postpartum depression is the most common medical problem that new mothers face. Anxiety is a more prominent feature of postpartum depression than of depression that occurs at other times in life. Routine, universal screening significantly improves detection in primary health care settings. Thus, an ultrabrief scale that could be incorporated into a general health survey or interview would be useful. OBJECTIVE We tested the hypothesis that, during the first 6 postpartum months, the 3-item anxiety subscale of the Edinburgh Postpartum Depression Scale is a better ultrabrief depression screener than 2 Edinburgh Postpartum Depression Scale questions that are almost identical to the widely used Patient Health Questionnaire. METHODS A cohort of 199 14- to 26-year-old participants in an adolescent-oriented maternity program completed the Edinburgh Postpartum Depression Scale at well-child visits during the first 6 postpartum months. Three subscales of the Edinburgh Postpartum Depression Scale were examined as ultrabrief alternatives: the anxiety subscale (3 items; Edinburgh Postpartum Depression Scale-3), the depressive symptoms subscale (7 items; Edinburgh Postpartum Depression Scale-7), and 2 questions that resemble the Patient Health Questionnaire (Edinburgh Postpartum Depression Scale-2). The reliability, stability, and construct validity of the Edinburgh Postpartum Depression Scale and 3 subscales were compared. Criterion validity was assessed by comparison with a score of >/=10 on the full, 10-item Edinburgh Postpartum Depression Scale. RESULTS A total of 41 mothers (20.6%) met study criteria for referral for evaluation of depression (Edinburgh Postpartum Depression Scale-10 score >/= 10). The Edinburgh Postpartum Depression Scale-3 exhibited the best screening performance characteristics, with sensitivity at 95% and negative predictive value at 98%. It identified 16% more mothers as depressed than the Edinburgh Postpartum Depression Scale did. The performance of the Edinburgh Postpartum Depression Scale-2 was markedly inferior, with sensitivity at 48% to 80%. Moreover, the Edinburgh Postpartum Depression Scale-2 was unreliable for mothers who had not been depressed in the past. CONCLUSION The brevity, reliability, and operating characteristics of the Edinburgh Postpartum Depression Scale-3 make it an attractive postpartum depression screening tool for primary health care settings in which the goal is to detect depression, not to assess its severity. Validation by diagnostic psychiatric interview is needed.