In the screening (Phase I) of the CARE study mothers with unknown mood state were screened for PPD and those with positive screens who agreed then participated in diagnostic evaluation. Willing mothers who qualified were subsequently randomly assigned for participation in the RCT (Phase II), and referred for treatment options as necessary during follow-up. Feasibility of the AHRQ framework was validated, as over 5,000 mothers were willingly screened for PPD with the EPDS. Because mothers in this study had interacted with a research nurse in the delivery hospital and had signed PCFs, they anticipated calls or mailings from the CARE study research nurses and were, therefore, receptive to being screened.
Both telephone and mailed screening approaches were effective. Consistent with other research studies (Seehusen & Clark, 2007
) there were more positive PPD screens for women who filled out the EPDS at home and returned it by mail than for those screened by a nurse via telephone. One possible explanation for this finding is that the nurse may have clarified questions, thus helping to reduce false positive responses. It is also possible that mothers who self–administered the EPDS in the privacy of their homes were more honest in their responses. To balance the pros and cons of each approach in the practice setting, we suggest that mothers complete the EPDS on their own, and that a clinician should then discuss the responses with mothers to confirm answers and follow-up as needed for positive screens.
The validity of the screening approach described here was further supported by the rate of positive EPDS screens obtained (13%), which is consistent with previously reported PPD prevalence (Gaynes et al., 2005
; O’Hara & Swain, 1996
). Although other researchers have identified specific demographic and situational risk factors for PPD, such as very young maternal age, low education and income, race/ethnicity (i.e., African American race), very low infant birth weight, prior depression history and onset of depression during pregnancy (Beeghly et al., 2003
; Mayberry et al., 2007
; Segre et al., 2006
), data from this study support only two of these risk factors. While ethnicity and race have not been consistent significant predictors of PPD in studies conducted in the United States (Yonkers et al., 2001
), in this study, significantly fewer than expected White women met criteria for a positive EPDS screen than did Hispanic, Asian American, women in the Other category, and African American women. These outcomes partially confirm some of the findings of other investigators (Beeghly at al., 2003
; Segre et al., 2006
) who found that, in the United States, race/ethnicity, specifically being African/American, increases PPD risk. Nonetheless, it is important to note our finding that Hispanic women in the screening sample had higher percentages of elevated PPD symptom scores than did White women. This finding contradicts previous research outcomes that suggested that Hispanic women have lower PPD prevalence than other mothers (Beeghly et al., 2003
; Segre et al., 2006
; Wei et al., 2007
). Certainly, continued examination of race/ethnicity, and other socio-demographic variables in relation to PPD, is appropriate given that no group can be definitively eliminated from PPD risk. Nonetheless, our findings in concert with previous research outcomes suggest that African American/Black and Hispanic mothers may face elevated PPD risk.
Our findings support those of Mayberry et al., (2007)
, which indicated that limited education may be associated with increased maternal depression risk. No such differences were found for age. The small percentage of very young mothers in the CARE study sample likely explains why age was not related to PPD symptom severity as measured by the EPDS. When young age has emerged as a predictor of PPD or young mothers have had higher rates of PPD, samples have included higher proportions of young and adolescent mothers (Mayberry et al.; Troutman & Cutrona 1990
). However, in the large community-based screening sample in this study, the mean age was 32.2 years and the mean age of women who had a diagnostic interview and who subsequently enrolled in the RCT was 31 years. It may well be that only very young age is associated with increased PPD risk due to developmental and social factors that make motherhood a particularly stressful event. Parity was not related to PPD symptom severity in this study underscoring the importance of screening and supporting all postpartum women, not just first-time mothers.
Given the findings about risk factors, clinicians and researchers alike are charged to continue to examine risk profiles, consider preventive interventions, educate childbearing women about relative risk for depression, and test PPD screening models in primary care settings for translation to practice. In addition, this study confirms the importance of identifying all mothers with EPDS scores ≥ 13 and/or who indicate that they have thoughts of self-harm, for immediate mental health referral and safety evaluation. Follow-up phone assessment by one of the study APRNs to invite the mother to have a diagnostic interview and possible enrollment in the RCT, created yet another opportunity to discuss PPD symptoms and need for mental health care.
By creating the CARE protocol to guide follow-up for every woman with probable PPD or risk of self-harm based on EPDS-10 scores, study nurses were able to contact all at-risk women to assist them in contacting their primary care providers or the psychiatric service at the delivery hospital. Systematic mental health follow-up is currently lacking in postpartum health care in the United States, and in many other countries (Gibson et al., 2009
). The mental health referral plan provided in this study was a major strength of this screening initiative. While screening alone has yet to demonstrate an improvement in mental health outcomes (Gaynes et al., 2005
), implementing follow-up by the clinician or psychiatric services has the potential to bridge this gap. However, to show effects on health outcomes, more is needed than measuring changes in PPD symptoms following psychosocial or psychopharmacologic treatment. Models are needed that demonstrate the feasibility of referral efforts for mental health evaluation and follow-up for large-scale PPD screening efforts. Our experience shows that introducing a formal PPD screening measure, such as the EPDS, begins the conversation about how a woman is feeling emotionally. Women in our study were receptive to feedback about their EPDS responses and to suggestions that they contact their PCPs, as evidenced by the willingness of the 185 women with positive screens who agreed to a diagnostic interview.
Although resistance to PPD screening by PCPs and postpartum women has been reported (Tam et al., 2002
), we can only conclude that failure to provide universal PPD screening for early case identification and mental health treatment referral is due in large measure to a sense of inadequacy on the part of providers, rather than resistance to PPD screening on the part of mothers. Our ability to screen more than 5,000 women for PPD via telephone and mail challenges primary care and other providers who care for mothers in the postpartum period and their infants to discard assumptions or fears that mothers will resist PPD screening, and to forge ahead to incorporate universal PPD screening into their practice. It is imperative that current knowledge about PPD prevalence and associated risks for infants, along with training in the use of standardized screening measures like the EPD S and mental health referral practices, be integrated into physician and nurse preparation programs by the current generation of clinical educators.
Testing screening procedures is an important public health goal. One of the strengths of this study is the size of the screening sample, which was significantly larger than samples from all studies reviewed by Gibson et al. (2009)
. In this recent comprehensive review, the largest sample reported was 876. Additionally, our sample was representative of the geographic area served by the academic medical centers and included diagnostic interview data from a subset of mothers.
We also echo the recommendation of Gaynes et al. (2005)
that the EPDS or Postpartum Screening Scale (PDSS) (Beck & Gable, 2000
) be used at present as the standard PPD screening measures due to their demonstrated substantive and psychometric properties. Widespread use of the EPDS as the most commonly employed PPD screening measure also enables users to compare results from PPD studies from the United States and around the world.
Limitations include our inability to follow up with mothers with positive PPD screens on the EPDS of ≥ 13 who declined to have a diagnostic interview. Although we did encourage these mothers to seek additional mental health evaluation, we could not force such follow-up or track outcomes. We also acknowledge that some eligible mothers may not have received study information during their postpartum stay. Thus given that PPD screening was not universal practice at the time of the study, our efforts to recruit a population of mothers after delivery could only result in a population of convenience.