Previous studies have supported the psychometric properties of the 22-Item Zarit Burden Interview (ZBI-22) scale among family caregivers of people with various disorders, including Parkinson's disease (PD). However, its short forms have not been psychometrically tested among PD family caregivers, and available psychometric analyses have not accounted for the ordinal nature of item-level data.
Objectives
To assess the psychometric properties of the ZBI-22 and its short forms among family caregivers of people with PD, while taking account for the ordinal nature of data.
Methods
Cross-sectional postal survey ZBI-22 data from 66 family caregiver members (59% women; mean age 69.6 years) of a local Swedish PD society branch were analyzed according to classical test theory methods based on polychoric/polyserial correlations.
Results
Missing item responses were ≤ 5%. Corrected item-total correlations were ≥ 0.42 and floor/ceiling effects were <20%, besides for the briefest (4- and 1-item) short forms (20% and 40% floor effects, respectively). Reliability was good for all scales (ordinal alpha 0.89–0.95). External construct validity was in general accordance with a priori expectations. Short forms demonstrated good criterion-related validity (rs 0.87–0.99) and discriminative ability (area under the curve, 0.91–0.98) relative to the full ZBI-22.
Conclusion
This study provides support for the reliability and validity of the ZBI-22 and its various short forms for use among PD family caregivers. In studies where caregiver burden is a central outcome, either ZBI-22 or ZBI-12 is suggested for use; other short forms can be used when caregiver burden is of less central focus or for clinical screening.
