Informant Agreement in Treatment Gains for Child Anxiety

Abstract

The present study examined multiple informant agreement in reports of treatment gains in a sample of children (M age = 10.27) treated for social phobia, generalized anxiety disorder, and separation anxiety disorder. Mothers and fathers agreed on their child's improvement, and parents and children also generally agreed on the child's improvement. This agreement was typically not differentiated by child's age or gender. Mothers and fathers did not agree with teachers. Parent-child “improvement agreement” was not predicted by child's age, gender, social desirability, or maternal psychopathology. The results indicate that mothers, fathers, and children report similar magnitudes of improvement following treatment.

KEYWORDS:

• agreement
• anxiety disorders
• CBT
• childhood anxiety
• cognitive behavioral therapy
• informant consistency

Acknowledgments

This research was supported by National Institute of Mental Health Grant MH 59087 awarded to Philip C. Kendall.

Notes

Note. CBCL = Child Behavior Checklist; TRF = Teacher Report Form; CBCL/TRF-Int = Child Behavior Checklist and Teacher Report Form-Internalizing Scale; CBCL/TRF-A = Child Behavior Checklist and Teacher Report Form-Anxiety Scale; Pre = pretreatment; Post = posttreatment.

Note. GIR = Global Interference Rating obtained from parent and child during ADIS-C/P; Method (a) = improvement in GIR score computed using parent and child identified primary pretreatment diagnoses; Method (b) = improvement in GIR score computed using parent identified primary pretreatment diagnosis.

Note. CBCL = Child Behavior Checklist; Pre = pretreatment; Post = posttreatment; CBCL-I = Child Behavior Checklist-Internalizing Scale; CBCL-Anx = Child Behavior Checklist-Anxiety Scale.

^a Denotes significant paired samples t test at p < .05.

*p < .05. **p < .01.

Note. CBCL = Child Behavior Checklist; TRF = Teacher Report Form; CBCL/TRF-Int = Child Behavior Checklist and Teacher Report Form-Internalizing Scale; CBCL/TRF-A = Child Behavior Checklist and Teacher Report Form-Anxiety Scale; Younger = 7 to 10 years old at pretreatment; Older = 11 to 14 years old at pretreatment.

^a n = 77.

^b n = 40.

^c n = 62.

^d n = 55.

*p < .05. **p < .01.

Note. GIR = Global Interference Rating obtained from parent and child during ADIS-C/P; Pre = pretreatment; Post = posttreatment.

^a Denotes significant paired samples t test at p < .05.

Note. Method (a) = improvement in GIR score computed using parent and child identified primary pretreatment diagnoses; Age = child's age at initial assessment; Gender = child's gender; RCMAS Lie = child's pretreatment Revised Children's Manifest Anxiety Scale-Lie Scale raw score; Mother BDI = mother's pretreatment total BDI-II score; Mother Anx = mother has a primary anxiety disorder diagnosis; Mother Dep = mother has a primary depressive disorder diagnosis; Method (b) = improvement in GIR score computed using parent identified primary pretreatment diagnosis.

^a n = 117.

^b n = 77.

When presented, κ coefficients will be interpreted per the criteria described in Mannuzza et al. (1989). κ values >0.74 represent “excellent” agreement, κ values between 0.60 and 0.74 represent “good” agreement, κ values between 0.40 and 0.59 represent “fair” agreement, and κ values <0.40 represent “poor” agreement (Mannuzza et al., 1989).

Previous research has evaluated parent-child agreement longitudinally (see Safford, Kendall, Flannery-Schroeder, Webb, & Sommer, 2005). These authors evaluated agreement between separate parent and child diagnostic interviews at two time points: prior to treatment and upon completion of treatment. We similarly examined longitudinal agreement in the present sample and our results were comparable to previously published reports (e.g., generally poor to moderate agreement).

When presented, r will be interpreted per the criteria described by Cohen (1988, 1992) where r = .10 represents a small effect, r = .30 represents a moderate effect, and r = .50 represents a large effect.

Youths ages 7–10 and ages 11–14 were classified into younger and older groups, respectively, based on developmental psychology and in order to remain consistent with the analytic strategy utilized by the parent RCT from which these data were collected (see Kendall et al., 2008).