ABSTRACT
AIM: The purpose of this article is to present the research protocol of a systematic review about fall risk assessment tools in acute hospitalized patients. BACKGROUND: Various risk assessment tools for falls have been developed, but with uncertainties derived from validation in heterogeneous environments and variations in their sensitivity, specificity and predictive validity. DESIGN: Systematic review and meta-analysis. METHODS: Two independent reviewers will extract data in a blinded process. Quality of studies will be assessed using various standardized instruments. A meta-analysis will be performed if applicable. For all studies, sensitivity, specificity, positive and negative predictive values, together with the positive and negative likelihood ratios and Youden index will be calculated. The diagnostic odds ratio of the studies and the hierarchical summary Relative Operating Characteristic curve and bivariate model will be applied. Calculations will be made from random effects models. Forest-plot diagrams for sensitivity and specificity and likelihood ratios, in addition to Cochrane's Q test and the I(2) statistic, will be calculated. Funding of the review was confirmed in December 2010. DISCUSSION: The results of this review will help to clarify some uncertainties provoked by earlier research findings and enable informed choice of a validated, reproducible instrument for assessing the risk of falls by hospital patients, so that preventive action may be taken to minimize this risk.
