Item Analysis of the Behavior Rating Inventory of Executive Function-Adults Version Using the Rasch Model in Two Forms for People With Schizophrenia in Taiwan.

BACKGROUND

The Behavior Rating Inventory of Executive Function-Adults version (BRIEF-A) consists of two forms, self-report and informant-report. Each form encompasses nine scales to assess executive functions. The study aimed to examine the unidimensionality of each scale and item-level agreement between the two forms in people with schizophrenia.

METHODS

A total of 197 people with schizophrenia and 194 caregivers were recruited from a single psychiatric center. Among the people with schizophrenia, 44.2% were male, with a mean age of 44.0 years. As for the caregivers who completed the informant-report form, 27.8% were male, with a mean age of 60.4 years. The unidimensionality of each BRIEF-A scale was assessed using Rasch analysis. Weighted kappa and percentage of agreement were used to examine the item-level agreement between the two forms.

RESULTS

The infit and outfit mean squares of all items on each scale of the self-report form were 0.65-1.48. Except for the Plan/Organize scale, the items of the other eight scales in the informant-report form met the standards for infit and outfit mean squares (0.69-1.44). Principal component analysis of the standardized residuals of the nine scales were 1.4-1.8 and 1.5-2.1 for the self-report and informant-report forms, respectively. Weighted kappa and percentage of agreement between the two forms were 0.02-0.33 and 36.8%-63.4% respectively.

CONCLUSION

The BRIEF-A has acceptable unidimensionality for people with schizophrenia. The slight to fair item-level agreement between the two forms necessitates cautious explanation of results and incorporation of multiple informants' perspectives to ensure a more thorough and accurate assessment of executive functions.