Background: This study examined the performance of multiple anxiety scales in measuring anxiety during pregnancy, an important issue due to the possible effect of pregnancy-related symptoms on the measurement of anxiety.
Methods: Secondary data on anxiety, measured by the State-Trait Anxiety Inventory-State (STAI-S) 20-item and six-item scales, the Edinburgh Postnatal Depression Scale-Anxiety Subscale (EPDS-3A) and the Symptoms Checklist-90-Anxiety Subscale (SCL-90), were obtained from two pregnancy cohort studies. Both cohorts completed the EPDS-3A, while 3341 women completed the STAI-S and 2187 women completed the SCL-90, with 231 women participating in both cohorts. Data were analyzed using confirmatory factor analysis and Spearman correlation.
Results: The STAI-6 had adequate model fit, while the STAI-20 and the SCL-90 had inadequate model fit. Model fitness for the EPDS-3A could not be assessed due to its low number of items. The correlation between the STAI-20 and STAI-6 was excellent (r = 0.93). The correlation of EPDS-3A with other anxiety scales was low to moderate (r (STAI-20) = 0.57, r (STAI-6) = 0.53 and r (SCL-90) = 0.44). The correlation of SCL-90 with both STAI-20 and STAI-6 was low (r < 0.50).
Conclusion: Findings indicate that these scales do not measure anxiety as a single dimension and that these scales are incomparable and may conceptualize anxiety differently.
Keywords: Anxiety during pregnancy; anxiety measurement scale; confirmatory factor analysis; correlation; measurement performance; validity.