The diagnostic utility of separation anxiety disorder symptoms: an item response theory analysis

J Abnorm Child Psychol. 2014;42(3):417-28. doi: 10.1007/s10802-013-9788-y.

Abstract

At present, it is not clear whether the current definition of separation anxiety disorder (SAD) is the optimal classification of developmentally inappropriate, severe, and interfering separation anxiety in youth. Much remains to be learned about the relative contributions of individual SAD symptoms for informing diagnosis. Two-parameter logistic Item Response Theory analyses were conducted on the eight core SAD symptoms in an outpatient anxiety sample of treatment-seeking children (N = 359, 59.3 % female, M Age = 11.2) and their parents to determine the diagnostic utility of each of these symptoms. Analyses considered values of item threshold, which characterize the SAD severity level at which each symptom has a 50 % chance of being endorsed, and item discrimination, which characterize how well each symptom distinguishes individuals with higher and lower levels of SAD. Distress related to separation and fear of being alone without major attachment figures showed the strongest discrimination properties and the lowest thresholds for being endorsed. In contrast, worry about harm befalling attachment figures showed the poorest discrimination properties, and nightmares about separation showed the highest threshold for being endorsed. Distress related to separation demonstrated crossing differential item functioning associated with age-at lower separation anxiety levels excessive fear at separation was more likely to be endorsed for children ≥9 years, whereas at higher levels this symptom was more likely to be endorsed by children <9 years. Implications are discussed for optimizing the taxonomy of SAD in youth.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Adolescent
  • Age Factors
  • Anxiety, Separation / diagnosis*
  • Anxiety, Separation / psychology
  • Female
  • Humans
  • Interview, Psychological / methods*
  • Logistic Models
  • Male
  • Models, Psychological
  • Reproducibility of Results
  • Sex Factors