Predicting the risk of falling – efficacy of a risk assessment tool compared to nurses' judgement: a cluster-randomised controlled trial [ISRCTN37794278]

Cluster-randomised controlled trial with nursing home clusters randomised either to the use of a standard fall risk assessment tool alongside nurses' clinical judgement or to nurses' clinical judgement alone. Standard care of all clusters will be optimised by structured education on best evidence strategies to prevent falls and fall related injuries. 54 nursing home clusters including 1,080 residents will be recruited. Residents must be ≥ 70 years, not bedridden, and living in the nursing home for more than three months. The primary endpoint is the number of participants with at least one fall at 12 months. Secondary outcome measures are the number of falls, clinical consequences including side effects of the two risk assessment strategies. Other measures are fall related injuries, hospital admissions and consultations with a physician, and costs.Fall prevention in the elderly has been recognised as an important area of research and intervention [1]. Numerous studies have presented combinations of fall related risk factors and a number of risk assessment tools have been developed [2,3]. However, few of these tools are based on rigorous research. The minority has been adequately tested for accuracy [4]. Reproducibility and transportability have been rarely investigated [5]. An own recently conducted systematic review including 27 publications on 25 non-laboratory fall risk assessment tools found that only 13 instruments have been validated in different settings [6]. Treatment paradox has been discussed by only four publications although it seems to be an important source of bias in accuracy studies which use the number of fallers as reference standard. However, there is no other measure to use as the gold standard for determining the validity of a fall risk assessment tool. Treatment paradox is difficult to overcome as it would be unethical to discourage fall prevention measures in the clinical setting in order to test risk assessment tools. None of the publications