Abstract
The Edinburgh Postnatal Depression Scale (EPDS) is increasingly used in public health and social service programs serving
postpartum women of racially, ethnically, and socioeconomically diverse backgrounds at risk for depression. However, we know
little about its factor structure across groups of women with implications for measuring symptom levels in research. This
study evaluated the underlying structure of the EPDS using a confirmatory factor analyses model comparison approach of five
factor models from the literature in a purposive community sample of 169 postpartum African American women of low socioeconomic
status. Participants were identified through an exhaustive review of local health department program files dated August 2006
to August 2010 in a Midwestern state of USA. Tuohy and McVey’s (Br J Clin Psychol 47:153–169, 2008) three-factor model (depression, anxiety, and anhedonia) demonstrated the best fit to the data with a nonsignificant Satora–Bentler
scaled chi-square value (21.70, df = 24, p = 0.60) and the lowest root mean square error of approximation (0.00) and standardized root mean square residual (0.05) values.
The results call for further study of the factor structure of the EPDS in other racial and ethnic groups and cautious use
of the EPDS among perinatal women of racially, ethnically, and socioeconomically diverse backgrounds until its factorial invariance
is better understood.
postpartum women of racially, ethnically, and socioeconomically diverse backgrounds at risk for depression. However, we know
little about its factor structure across groups of women with implications for measuring symptom levels in research. This
study evaluated the underlying structure of the EPDS using a confirmatory factor analyses model comparison approach of five
factor models from the literature in a purposive community sample of 169 postpartum African American women of low socioeconomic
status. Participants were identified through an exhaustive review of local health department program files dated August 2006
to August 2010 in a Midwestern state of USA. Tuohy and McVey’s (Br J Clin Psychol 47:153–169, 2008) three-factor model (depression, anxiety, and anhedonia) demonstrated the best fit to the data with a nonsignificant Satora–Bentler
scaled chi-square value (21.70, df = 24, p = 0.60) and the lowest root mean square error of approximation (0.00) and standardized root mean square residual (0.05) values.
The results call for further study of the factor structure of the EPDS in other racial and ethnic groups and cautious use
of the EPDS among perinatal women of racially, ethnically, and socioeconomically diverse backgrounds until its factorial invariance
is better understood.
- Content Type Journal Article
- Category Original Article
- Pages 1-10
- DOI 10.1007/s00737-012-0260-8
- Authors
- Patricia A. Lee King, Department of Social Work, Helen Bader School of Social Welfare, University of Wisconsin-Milwaukee, PO Box 786, Milwaukee, WI 53201, USA
- Journal Archives of Women’s Mental Health
- Online ISSN 1435-1102
- Print ISSN 1434-1816