Published studies were eligible if they were randomised controlled trials (RCTs) or prospective cohort studies that included residents with dementia in the study population of nursing home residents where psychoactive medication use was studied and falls and gait parameters were measured. Studies that excluded residents with advanced dementia from participation in the study were excluded.
In the included studies, the proportion of nursing home residents with cognitive impairment or dementia ranged from 22% to 82% (where reported). In two studies it was unclear which patients were from nursing homes and which were in intermediate care facilities. Most residents were female. Mean age was over 80 years. Most studies ascertained falls from medical records or nursing home charts and from incidence reports. Falls were categorised as recurrent, injurious, occurring in day time, occurring at night time or falls not otherwise specified. None of the studies provided data on the association of gait and psychoactive drug use. The Minimum Data Set was used by most studies to ascertain psychoactive drug use. Psychoactive drugs included hypnotics, sedatives, anti-anxiety drugs, benzodiazepines, anti-psychotics and antidepressants.
Two reviewers independently performed selection of studies in a two-stage process (reviewing of titles and abstracts and screening of full text papers). Differences were resolved by discussion.