Psychometric evaluation and refinement of the Prenatal Attachment Inventory

Objective: To explore the underlying structure of the Prenatal Attachment Inventory (PAI) and to assess the psychometric properties of the subscales. Background: The establishment of the mother-fetus relationship is an important developmental task for successful adaption to pregnancy, with implications for the well-being of mother and child. The PAI was developed to measure the relationship between a women and her fetus. Although originally developed as a single dimension, other researchers have suggested alternative factor structures. Methods: A self-report questionnaire, including the PAI, was administered to 775 Swedish women in late pregnancy. Psychometric assessment of the PAI was undertaken using exploratory (EFA) and confirmatory factor analysis (CFA), and Rasch analysis. Results: EFA indicated a three-factor solution, which was confirmed by CFA. The previously reported single-factor and five-factor solutions were not supported. Rasch analysis was used to form three 6-item PAI-Revised subscales (Anticipation, Interaction, Differentiation). All subscales showed good overall fit to the Rasch model and good internal consistency. The three subscales were moderately intercorrelated, sharing between 23% and 42% of their variance, suggesting they should be used separately, pending further research concerning their unique predictive power. Conclusion: These results suggest that PAI-Revised is a psychometrically sound tool, suitable for use in research and clinical settings. It can be used to guide the identification, support, and follow-up of pregnant women with low attachment. The three subscales, although related, may tap different aspects of the attachment construct, with different antecedents and consequences for the well-being of mother and child.