Background and aims: The purpose of our study was to examine whether severity of dementia, behavioral and psychological symptoms and depression can predict falls among nursing home residents, such as demographic variables, activities of daily living, and use of psychotropic drugs, when potential confounders are controlled for. Methods: 1147 nursing home residents were examined in this one-year follow- up study. All residents were examined with the Physical Self-Maintenance scale (Activities of Daily Living— ADL), Clinical Dementia Rating Scale (CDR), Neuropsychiatric Inventory (NPI) and Cornell Scale for Depression in Dementia. Demographic data, gender, education, physical health and use of medication were collected from medical records. Results: 40% of participants had at least one fall during the one-year follow-up period. Bivariate survival analysis revealed that low level of education, severe dementia, severe behavioral and psychological symptoms, severe depression, greater functional impairment, age, worsening in physical health, and use of sedatives, significantly predict one or more falls. Multivariate Cox regression analyses showed that age, higher scores on NPI and CDR, use of sedatives and dependency in ADL were all, independently of each other, predictors of an increased risk of falling. Conclusions: Having a high NPI score was identified as a significant and independent predictor of falls. Since falling is a common event which causes considerable morbidity and mortality in older people, these findings are important for healthcare and for the individuals concerned. To prevent falling in nursing homes, special attention must be paid to residents with severe dementia, to behavioral symptoms and use of sedatives.
