Falls are a major health problem in nursing home residents with dementia. In nursing homes one-third of all falls results in an injury. In order to take tailor-made preventive measures in time, the fall risk profile of each individual nursing home resident should be periodically evaluated. A systematic evaluation of fall risk should include an assessment of major contributing components, including an assessment of balance and gait, and the use of psychotropic drugs. The purpose of this study was a) to evaluate the feasibility and predictive validity of different assessment methods for balance and gait impairments in a population of nursing home residents with dementia with a specific view to predicting falls in the short term, i.e., three months, and b) to quantify the additive contribution of psychotropic drugs to fall risk in nursing home residents with dementia. Part one of this thesis describes how ambulatory nursing home residents with moderate to severe dementia participated in a prospective cohort study. This study showed that gait velocity is a feasible and valid predictor of a fall within three months. In part two of this thesis the additive contribution of psychotropic drugs to fall risk in nursing home residents with dementia is quantified. Daily drug use and daily falls were recorded over a two-year period. Fall risk increased already at low dosages of psychotropics; it increased further with dose increments, and with combinations of these drugs. There also was a significant dose-response relationship between the use of selective serotonin reuptake inhibitors (SSRIs) and the risk of an injurious fall. The risk increased further with the use of a SSRI in combination with a hypnotic or sedative. New treatment protocols might be needed that take into account the dose-response relationship between psychotropics and falls, and between SSRIs and injurious falls.