Postpartum depression in the NICU: an examination of the factor structure of the Postpartum Depression Screening Scale

• Published in: Advances in neonatal care Vol. 14; no. 6; p. 424
• Main Authors: Blucker, Ryan T, Gillaspy, Jr, J Arthur, Jackson, Dennis, Hetherington, Carla, Kyler, Kathy, Cherry, Amy, McCaffree, Mary Anne, Gillaspy, Stephen R
• Format: Journal Article
• Language: English
• Published: United States 01.12.2014
• Subjects: Adult; Depression, Postpartum - diagnosis; Depression, Postpartum - psychology; Factor Analysis, Statistical; Female; Humans; Intensive Care Units, Neonatal; Intensive Care, Neonatal; Mothers - psychology; Mothers - statistics & numerical data; Prospective Studies; Psychiatric Status Rating Scales - standards; Psychometrics; Reproducibility of Results; Retrospective Studies; United States; United States
• ISSN: 1536-0911
• DOI: 10.1097/ANC.0000000000000135

The purpose of this research was to examine the construct validity of scores from the Postpartum Depression Screening Scale administered to mothers of infants in a neonatal intensive care unit. Two samples (n = 385 and n = 110) of mothers with infants in a neonatal intensive care unit in the south-central region of the United States completed the Postpartum Depression Screening Scale 2 weeks postpartum. Both samples were similar in race and education level but differed according to marital status. Study 1 was retrospective and descriptive, whereas study 2 was prospective and descriptive. In study 1, confirmatory factor analysis was used to evaluate the originally proposed 7-factor structure and a 1-factor model. Because of overly high correlations between 3 of the factors, a revised 5-factor model was also tested. In study 2, confirmatory factor analysis was again used to test the construct validity or goodness of fit of the 7-factor and 5-factor models. Goodness-of-fit indices and factor pattern coefficients. In study 1, high correlations between 3 of the factors in the 7-factor model did not converge. Thus, a 5-factor model was also tested. This model had reasonable fit: χ= 1339.70 (550); P < 0.01; comparative fit index = 0.85; root mean square error of approximation = 0.06; and 90% confidence interval for root mean square error of approximation = 0.058 to 0.067. In study 2, confirmatory factor analysis was again used to test the construct validity or goodness of fit of the 7-factor and 5-factor models. Results for the 7-factor model indicated unacceptable fit: χ (539) = 959.10; P < 0.01; comparative fit index = 0.76; and root mean square error of approximation = 0.09. The 5-factor model was also poor: χ (550) = 992.95; P < 0.01; comparative fit index = 0.75; and root mean square error of approximation = 0.09. Although the construct validity of the 7-factor model of the PDSS was not supported for this sample of mothers, additional factor analytic work was used to develop and provide initial validation of a 5-factor model. Future research should continue to explore the unique experiences of mothers in the NICU who experience postpartum depression.