Inter- and intrarater consistency in diagnosis of clinic-referred children
The current study examined consistency among raters' descriptions of children, looking for consensus among parent, child, and clinician. In addition, demographic variables (sex, age, ethnicity, and family structure) were examined for differences in interrater consistency. Parent ratings were measured via the Child Behavior Checklist (Achenbach & Edelbrock, 1983) and the Conners Parent Rating Scale (Conners, 1989). The former yielded T-scores on scales of Internalizing and Externalizing behavior problems, while the latter provided T-scores on narrow-band scales measuring conduct and learning problems, impulsive and neurologically based hyperactivity, anxiety, and somatic complaints. Children were asked to complete the Nowicki-Strickland Locus of Control Inventory and the Coopersmith Self-Esteem Inventory. The clinician's rating was measured by the DSM-III-R Axis I diagnosis obtained via clinical interview.^ The subjects were 61 boys and girls, ages 5-16, referred for treatment to the psychiatric outpatient clinic at a metropolitan municipal hospital during 1989-91. The data were analyzed to determine level of inter- and intrarater consistency in diagnosing children as well as whether significant differences were found when the group is divided by sex, age, ethnicity, and a family situation. Correlation coefficients were obtained for level of agreement between parent and child, parent and psychiatrist, and psychiatrist and child, in addition to intrarater agreement on the different measures. Group differences were measured by t-tests, Anovas, Chi-squares, and Scheffe post hoc comparisons.^ Inter- and intrarater agreement followed the anticipated pattern. Correlations were highest when the same rater evaluated similar traits. Evidence of bidirectional or general pathology was also found. However, agreement between raters was slight. Thus, each rater was determined to be individually reliable but to present a unique perspective as to the child's emotional state.^ Preexisting demographic differences were few, as were differences in interrater agreement based on the demographics. However, several specific differences were noted between boys and girls, and younger versus older children. Findings suggest that, to improve diagnostic accuracy and to engage children and adolescents in their treatment, psychiatric clinics should use structured, quantifiable, information from parents and children together with clinical data in formulating diagnoses and treatment plans. ^
Bracha J Sacks,
"Inter- and intrarater consistency in diagnosis of clinic-referred children"
(January 1, 1992).
ETD Collection for Pace University.