Best Practices: Community-Based Postpartum Depression Screening: Results From the CARE Study

The evidence-based AHRQ screening framework recommends formal postpartum screening, diagnostic evaluation for those with positive screens, and random group assignment to the treatment or control group with follow-up evaluations ( 4 ). Applying this framework, research nurses visited mothers recovering on the postpartum units of two Harvard teaching hospitals from August 2004 through September 2007. The study was approved by the institutional review boards of Partners Human Research Committee and Boston College. All mothers who received a diagnostic interview and who enrolled in the study provided written informed consent.

Research nurses visited mothers who had been previously identified by staff as meeting all eligibility criteria: English-speaking; medically stable and sufficiently recovered from labor and birth; and uncomplicated delivery of full-term, singleton infants cared for in the newborn nursery. Interested mothers were approached by research nurses and asked to complete a form that gave researchers permission to contact them for depression screening at four weeks postpartum. Postpartum "blues" resolve by two weeks, and postpartum depression is defined as beginning by four weeks ( 5 ).

Research nurses telephoned all mothers who signed a permission form and who did not return an opt-out card. The ten-item Edinburgh Postnatal Depression Scale (EPDS) ( 6 ) and the brief version of the Mothers Information Tool (MIT) were administered by telephone. The MIT was developed to elicit demographic information and responses to a series of open-ended questions. If the nurse failed to contact a mother by the end of the fifth week, the EPDS and MIT were mailed. The authors of the EPDS recommended use of the lower score of 9 or 10 to reduce failed detection to less than 10%. Therefore, if an EPDS score was ≥10 on either the phone or mail-in screen, she was informed about the research study and invited to undergo a diagnostic interview with possible enrollment in the clinical trial. Cox and colleagues ( 6 ) and others ( 7 ) suggested that a score of ≥13 indicates referral to mental health care. Data from the study reported here support the recommendations of Gaynes and colleagues ( 4 ) and others ( 7, 8, 9 ) that the EPDS merits continued use as a standard postpartum depression screening measure.

If a mother's EPDS score was ≥13 or her response to EPDS item 10 (thoughts of self-harm) was anything other than zero, the nurse encouraged her to contact her clinician immediately if she was currently receiving mental health treatment or to contact her primary care provider or the Department of Psychiatry at the delivery hospital for follow-up evaluation. An advanced practice psychiatric nurse also contacted her to provide a brief evaluation and to assist her to seek follow-up. If nurses were unable to contact a mother with an EPDS score of ≥13 by phone, a letter was sent with CARE study contact information that suggested she contact her doctor, midwife, or another clinician.

All mothers with an EPDS score of ≥10 were invited to participate in the Structured Clinical Interview for DSM-IV administered by the advanced practice nurse in their homes. When conducted by a qualified clinician, this interview is the "gold standard" for confirming a depression diagnosis ( 8 ). Mothers with confirmed diagnoses of minor or major depression were invited to enroll in the clinical trial. Enrolled mothers whose scores indicated risk of self-harm (≥13) were offered the same assistance described above.

Over two-and-a-half years, 5,169 mothers aged 14 to 49 were screened by telephone or mail. Of this group, 674 (13%) mothers scored ≥10 on the EPDS. The prevalence rate for the mothers who self-administered and mailed in the EPDS (N=260, 19%) was higher than for those who were screened with the EPDS by a nurse on the phone (N=425, 11%).

Most mothers screened were Caucasian (N=3,382, 66%), followed by Hispanic (N=620, 12%), black (N=449, 9%), Asian or Pacific Islander (N=401, 8%), and "other" (N=286, 5%). Thirty-one mothers did not indicate their race-ethnicity. The mean±SD age was 32.2±5.4, and the mean years of education were 16.3±2.9. A total of 2,478 (48%) were primiparas. A smaller proportion of white mothers (N=381, 11%) screened positive on the EPDS compared with Hispanic mothers (N=113, 18%), Asian-American mothers (N=69, 17%), mothers in the "other" racial-ethnic group (N=46, 16%), and black mothers (N=65, 15%) (p<.001). Compared with mothers who had at least a high school education, a greater proportion of those who were not high school graduates screened positive (p<.001).

Of the 674 mothers who were offered a diagnostic interview and subsequent enrollment in the nine-month study, 267 (40%) declined for a variety of reasons. A total of 185 mothers with positive EPDS screens received diagnostic interviews. Of this group, 111 (60%) were given a diagnosis of major depression and 24 (13%) were given a diagnosis of minor depression. Forty-one of the mothers who had a diagnostic interview (22%) did not receive a diagnosis of depression and were excluded from the study.

The first of the four open-ended questions asked mothers whether clinicians had asked about their emotional state. Of the 4,419 mothers screened who had seen a clinician by four- to six-weeks postpartum, 2,806 (63%) reported that they had been asked by the clinician about how they were feeling emotionally. In contrast, 1,613 (37%) reported that they were not asked about their emotional state by a clinician. No significant differences in age, education, or depression status (yes or no) were found between those who were and were not asked. Multiparous mothers were asked less often than primiparous mothers (p<.001). Asian, Caucasian, and Hispanic mothers were asked less often than black mothers and mothers in the "other" racial-ethnic category (p=.012). A total of 178 mothers with an EPDS score of ≥10 were not asked by a clinician about their emotional health.

Clinicians who were most often identified were pediatricians, obstetricians, nurse midwives, and nurse practitioners. When mothers were asked to elaborate, most reported that the clinician provided a specific checklist or asked specific questions about how they were feeling. Others reported being asked only general or informal questions about their emotional health.

The third question asked how the mother responded to the clinician's question about her emotional state. Most reported that their response was "fine," "good," "ok," "stable," or "had no problem." Other responses were categorized as "has social or health care support," feels sad or overwhelmed," "feels stress from other sources," "is seeking treatment for postpartum depression," or "feels stress related to breastfeeding or parenting." Seventy-three mothers with an EPDS score of ≥10 denied feeling sad when asked about their emotional state by the clinician.

The fourth question asked how the clinician followed up. Most mothers reported that the clinician offered some type of unspecified intervention, a psychiatric consultation or referral, advice, or emotional support. A handful stated that they were offered support but did not act on it. However, 91 mothers with elevated EPDS scores or who told the clinician that they felt "sad" were not offered any intervention.