Research report
Validation of the Edinburgh Postnatal Depression Scale for men, and comparison of item endorsement with their partners

https://doi.org/10.1016/S0165-0327(00)00236-6Get rights and content

Abstract

Background: The Edinburgh Postnatal Depression Scale (EPDS) has been validated and used extensively in screening for depression in new mothers, both in English speaking and non-English speaking communities. While some studies have reported the use of the EPDS with fathers, none have validated it for this group, and thus the appropriate cut-off score for screening for depression or anxiety caseness for this population is not known. Method: Couples were recruited antenatally and interviewed at six weeks postpartum. EPDS scores and distress caseness (depression or anxiety disorders) for 208 fathers and 230 mothers were determined using the Diagnostic Interview Schedule. Results: Analyses of the EPDS for fathers using distress caseness (depression or anxiety disorders) as the criterion shows that a cut-off of 5/6 has optimum receiver operating characteristics. Furthermore acceptable reliability (split-half and internal consistency) and validity (concurrent) coefficients were obtained. For mothers the optimum cut-off screening value to detect distress caseness was 7/8. Item analysis revealed that fathers endorsed seven of the ten items at lower rates to mothers, with the most significant being that referring to crying. Conclusions: The EPDS is a reliable and valid measure of mood in fathers. Screening for depression or anxiety disorders in fathers requires a two point lower cut-off than screening for depression or anxiety in mothers, and we recommend this cut-off to be 5/6.

Introduction

The Edinburgh Postnatal Depression Scale (EPDS; Cox et al., 1987) has been used extensively in studies investigating mood disturbance in the perinatal period for women. It consists of 10 self-report items, eight addressing depressive symptoms (e.g. sadness, self-blame) and two inquiring about anxiety symptoms (e.g., feeling worried or anxious and feeling scared or panicky). It is used as a screening instrument for antenatal and postnatal depression, and has been validated for English-speaking women in Britain (Cox et al., 1987, Murray and Cox, 1990) and Australia (Boyce et al., 1993) using diagnostic criteria for major or minor depression (DSM111R) as the criterion measure. Using a cut-off of 12/13, sensitivity in detecting major or minor depression in new mothers has been reported as being between 80%–100% (Cox et al., 1987, Harris et al., 1989, Boyce et al., 1993), though one study (Murray and Carothers, 1990) obtained sensitivity of only 67.7% using this cut-off with their sample of 646 women. It is possible that the different receiver operating characteristics were a reflection of administering the EPDS by mail, rather than by person, and hence the diagnostic interview was administered several days after the women had completed the EPDS. A cut-off of 14/15 was found to be optimal for antenatal screening for major depression (Murray and Cox, 1990).

The impetus behind creating the EPDS was the unsuitable nature during the perinatal period of some items found in commonly used self-report measures of depression, such as the Beck Depression Inventory (Beck et al., 1961) or Centre for Epidemiological Studies–Depression scale (CES–D; Radloff, 1977). Asking new mothers if they have had disturbed sleep, and scoring a positive response as indicating the presence of a depressive symptom, is incorrect without inquiring further, as clearly all new mothers have disturbed sleep. The same rationale applies to questions about fatigue, changes in libido and weight, and concentration difficulties. Because of this, and results comparing scores on the BDI and caseness classification on a structured diagnostic interview, Hopkins et al. (1989) concluded that “the BDI may not be an appropriate instrument for diagnosing depression in a postpartum sample” (p. 251). Thus the items in the EPDS have been worded so that positive responding by women indicates the presence of mood symptoms not ordinarily associated with new parenthood – such as difficulty sleeping due to feeling unhappy, or feeling anxious or worried for no good reason. There have been several reports of the usefulness of the EPDS, both in its acceptability to the women being screened for PND, and by the increase in the detection of PND when Health Visitors or Early Childhood Nurses use it routinely in their practice (e.g., Holden, 1996, Schaper et al., 1994, Tully et al., 1998).

This argument for using the EPDS for new mothers, rather than using a different self-report measure of mood that has poor items for new mothers, applies equally to new fathers. Items which refer to a loss of interest in sex, fatigue, weight changes, or disturbed sleep or concentration difficulties (i.e., those in the BDI and CES–D) are examples of those which might also be experienced by new fathers as a normal part of early parenthood, and not necessarily be an indicator of mood disturbance. Not only would such items lack face validity for new fathers, in the same way that such items have lacked credibility with women (Pop et al., 1992), but the resulting scores will erroneously inflate their level of negative affect. While there have not yet been any reports indicating concern by either clinicians or fathers with such scales, this may in part be a reflection of the comparative lack of work on the assessment of mood in new fathers. Given this, we believe there is a need for an instrument, such as the EPDS, to detect distress in fathers that does not classify normal experiences of new fatherhood as being indicative of affect symptomatology, and which is likely to have greater face validity for the participants than some other measures of affect.

Several studies have reported the use of the EPDS with men. Ballard et al. (1994) used an early version of the EPDS which contained 13 items, and reported that a cut-off of 12/13 produced a sensitivity of 85.7%, and specificity of 75% in their sample of 178 fathers at six weeks postpartum, using the Psychiatric Assessment Schedule (PAS; Dean et al., 1983) as the criterion measure. It is unclear whether they used just depression as caseness, or included those with other diagnoses, such as anxiety disorders.

Areias et al. (1996) translated the 10-item EPDS into Portuguese, and reported that in their sample of 42 fathers this instrument was unable to discriminate between major and minor depression (whereas it could for the mothers), using the Schedule for Affective Disorders (Spitzer et al., 1978) to determine depression caseness. They also found that it was not particularly sensitive. While these investigators did not specify optimum cut-off scores, they did report the various statistics for the different EPDS scores. At what appears to be the optimum threshold for their sample of women (9/10: sensitivity of 71%, specificity of 89%, positive predictive value of 79%), the comparative statistics for the men were 40%, 81% and 36%.

Lane et al. (1997) used the 10-item EPDS on a sample of 181 men. They did not use a criterion measure of caseness, but just used the cut-off of 12/13 for both women and men, without discussing why this figure was chosen for the fathers. In a similar manner Thorpe (1993) also reported rates of high scorers (>12) on the EPDS for fathers at 19–22 months postpartum, but qualified this by saying that no suitable cut-off could be calculated for the men due to no cases of depression being detected on the diagnostic interview.

Dragonas et al. (1992) report the use of a modified EPDS for Greek fathers without describing the modifications, nor discussing cut-off scores.

Section snippets

Varying EPDS cut-off scores

The cut-off score to screen for depression on the EPDS has been shown to vary for women from different cultures (Arabic: Barnett et al., 1999 (9/10); Ghubash and Abou-Saleh, 1997 (11/12); Vietnamese: Barnett et al., 1999 (9/10); Swedish: Wickberg and Hwang, 1996 (11/12); French: Guedeney and Fermanian, 1998 (11/12); Chinese: Lee et al., 1998 (6/7 or 9/10); Japanese: Yoshida et al., 1997 (4/5); Italian: Carpiniello et al., 1997 (9/10)). These differing thresholds are likely to be at least partly

Participants

Couples (N=251) were recruited from the evening Preparation for Parenthood classes held in a public hospital in South West Sydney, as part of a larger study examining the effectiveness of an intervention for postnatal distress. These couples had paid to attend.

Measures

The EPDS (Cox et al., 1987); the CES–D (Radloff, 1977); the Diagnostic Interview Schedule (DIS: Robins et al., 1981).

Procedure

Mothers and fathers who had been recruited antenatally were sent various questionnaires, including the EPDS, at 6–7 weeks

Participants

The sample sizes for the data analyses vary from 200 to 218 for the men, 230 to 238 for the women, and 212 to 218 for couples. These numbers vary depending upon whether the analyses inspect complete self-report data (EPDS, CES–D), caseness data (depression and anxiety modules), or a combination of either of these.

Drop-outs prior to the six-week interview consisted of 34 men and 13 women. These were due to either inability to contact the participants, or failure by the participant to be home on

Discussion

The results show that the EPDS is both reliable and valid for fathers. It discriminates between distressed and non-distressed fathers, using caseness of either just depression or depression and anxiety. Fathers whose partner was currently distressed also scored higher on the EPDS than fathers whose partner was not distressed. This is in line with other research that has found that distressed mothers often have distressed partners (e.g. Lovestone and Kumar, 1993, Ballard et al., 1994, Morgan et

Acknowledgements

This project was funded by a grant from the Commonwealth Department of Health and Family Services, Australia.

References (42)

  • A.T. Beck et al.

    An inventory for measuring depression

    Arch. Gen. Psychiatry

    (1961)
  • D. Blazer et al.

    Psychiatric disorders: a rural/urban comparison

    Arch. Gen. Psychiatry

    (1985)
  • P. Boyce et al.

    Edinburgh Postnatal Depression Scale: validation for an Australian sample

    Aust. NZ J. Psychiatry

    (1993)
  • B. Carpiniello et al.

    Validation of the Edinburgh Postnatal Depression Scale in Italy

    J. Psychosom. Obstet. Gynaecol.

    (1997)
  • J. Cox et al.

    Detection of postnatal depression: development of the 10 item Edinburgh Postnatal Depression Scale

    Br. J. Psychiatry

    (1987)
  • C. Dean et al.

    Comparison of research diagnostic systems in an Edinburgh community sample

    Br. J. Psychiatry

    (1983)
  • T. Dragonas et al.

    Transition to fatherhood: a cross-cultural comparison

    J. Psychosom. Obstet. and Gynaecol.

    (1992)
  • R. Ghubash et al.

    Postpartum psychiatric illness in Arab culture: prevalence and psychosocial correlates

    Br. J. Psychiatry

    (1997)
  • B. Harris et al.

    The use of rating scales to identify post-natal depression

    Br. J. Psychiatry

    (1989)
  • J. Holden

    Using the Edinburgh Postnatal Depression Scale in clinical practice

  • J. Holden

    The role of health visitors in postnatal depression

    Int. Review Psychiatry

    (1996)
  • Cited by (0)

    View full text