Poster #55 - Screening for Anxiety and Depression in Mothers of Infants Who Required Neonatal Intensive Care

• In Event: 1-061 - Poster Session 02
  In Poster Session: PS 02 Section - Developmental Psychopathology

Integrative Statement

There is strong evidence indicating that exposure to adverse experiences and environments in infancy and early childhood is associated with negative effects on an individual’s life-long health and economic outcome (Center on the Developing Child, 2010; Halfon, et al., 2017). However, the quality of early infant-mother relationships may mediate the effects of exposure to early adversity on future health (Center on the Developing Child, 2010). Mothers’ mental health plays a significant role in their ability to engage in warm, development enhancing interactions with their infants (Beeghly & Tronick, 2011; Feldman et al., 2009).
A valid, easily administered, tool, which can serve as a screen for maternal mood disorders is a crucial component to universal maternal mental health screening (Gollan et al., 2017). In their review Matthey, Fisher and Rowe (2013), suggest that the ten-item, self-report Edinburgh Postnatal Depression Scale (EPDS), developed to screen for maternal depression, may identify anxiety as well (Chiu et al., 2017; Matthey, et al., 2013). Although the EPDS has been extensively validated for perinatal depression with demographically and culturally diverse women (Gibson, McKenzie-Mcharg, Shakespeare, Price, & Gray, 2009), the capacity of the EPDS to identify anxiety has not been evaluated in mothers whose children required Neonatal Intensive Care.
To probe the underlying structure of the EPDS, in our demographically diverse sample of mothers of infants and toddlers who required care in a University NICU, (N = 157; mean income $43,300 with SD 2.35, 38.2 % White, see Table 1), we conducted an exploratory factor analysis (EFA). Literature review and scree plot evaluation suggested the presence of two-three factors; therefore, three and two-factor models were tested. Maximum likelihood and generalized least squares (to address the non-normal distribution of the EPDS item 10, self-harm) estimations using an oblique (direct oblimin) rotation, with and without item 10 were compared (see Table 2). Finally, Partial Confirmatory Factor Analysis (PCFA, Giles, 2009) was used to estimate fit indices (see Table 2).
Model 1, a three-factor model with item 10 included was preferred. Model 1 explained 54.04% of the variance, with each item loading on only one factor, F1-anhedonia, F2-anxiety or F3-depression. The RMSEA, the SRMR and the Comparative Fit Index (CFI) indicated excellent fit at .04, .03, and .97 respectively (see Table 2). Models 3 and 4 (two-factor models, F1-depression and F2-anxiety), also demonstrated an acceptable fit; nevertheless, pattern matrix evaluation revealed overlapping loadings on item 9 (crying). In Models 1 and 2, item 10 was non-salient at <.30 on any factor. However, item 10 is frequently non-normally distributed and often excluded from analyses (Chiu et al., 2017; Pallant, Miller & Tennant, 2006).
To our knowledge, this study is the first to examine the factor structure of the EPDS in mothers of infants and toddlers who required NICU hospitalization. Results indicate that the EPDS is sensitive to maternal anxiety and that the factor structure of the EPDS may reflect variability in the experience of anxiety and depression in different populations of mothers (Chiu et al., 2017; Matthey et al., 2013).

Authors

• Rose L Belanger, University of Maryland, Baltimore County
  Presenting Author

• Susan Sonnenschein, UMBC
  Non-Presenting Author

• Shuyan Sun, University of Maryland Baltimore County
  Non-Presenting Author

• Brenda Hussey-Gardner, University of Maryland School of Medicine
  Non-Presenting Author