Background: Psychiatric symptoms and behavioral disturbances are highly prevalent in the residents of nursing homes. The Neuropsychiatric Inventory (NPI) is a commonly used scale for the assessment of such symptoms in diverse settings. We have conducted a study of the reliability and the validity of the Norwegian version of the NPI nursing home version (NPI-NH).
Methods: The reliability study comprised 41 patients. We established inter-rater reliability between raters with various levels of health education using kappa statistics. Fifty patients were included in the validity study. The patients were examined by a physician, who also rated the patient's behavior using "behavioral pathology in Alzheimer's disease" (BEHAVE-AD). Subsequently, a research nurse performed a standardized interview using the NPI and the Cornell scale. Concurrent validity of the NPI and the BEHAVE-AD was analyzed.
Results: Internal consistency, as measured by Cronbach's alpha was above 0.8. Inter-rater reliability was, except for one item, between 0.85 and 1.0 across assessors with different levels of health education. All correlations between the NPI and the BEHAVE-AD were significant, ranging from 0.38 to 0.72. The weakest correlations were between items assessing affective and anxiety symptoms.
Conclusion: The Norwegian version of the NPI-NH is a reliable and valid instrument for assessing psychiatric symptoms and behavioral disturbances in the residents of nursing homes. The investigation of depressive symptoms merits particular attention.