Validation of the Zarit Burden Interview in caregivers of non-demented elderly

Introduction Even if the Zarit Burden Interview (ZBI) is widely used and has extensively been studied in different populations, it has only been formally validated in informal caregivers of elderly with dementia. However, literature shows that the burden of these populations is quite different, due to the behavior characteristics of the care receiver. Aim of the study To perform a complete validation of the scale in informal caregivers of home-dwelling elderly without dementia. The 12-item ZBI was chosen, due to its alleged user-friendliness. Methods Acceptability was evaluated by the percent of missing values of the questionnaires. Content validity was assessed by both Content Validity Index (CVI) and Cohen’s kappa multirater index. Construct validity was studied by the means of the known groups technique. Reliability was assessed by the test-retest reproducibility, by calculating intraclass-correlation coefficient (ICC). Results Content validity indexes for the 12 questions scored from 0.64 to 0.95. As for Cohen’s kappa multirater index, questions rated from 0.53 (question 3) and 0.95. Acceptability, construct validity and reliability data collection are still in process, but will probably be completed before the September EDCNS Conference. Discussion incl. Conclusion Initial findings shows different validity of ZBI for in informal caregivers of elderly without dementia than those for elderly with dementia. Practical relevance This study stresses the importance to validate scales for specific study populations, even it was initially validated but in a different population. For instance, the first results of the content validity show that some of the ZBI questions are not relevant for informal caregivers of elderly without dementia. Research implications Formal validation has to be performed for scales used in a different population. A scale in order to evaluate burden for informal caregivers of home-dwelling elderly without dementia must be validated and, if necessary developed