Diagnostic validity of the STRATIFY and Downton instruments for evaluating the risk of falls by hospitalised acute-care patients: a multicentre longitudinal study.

Abstract

Falls are major adverse events in hospitals. The appropriateness of using risk assessment instruments for falls in hospitals has recently been questioned, although the research performed in this respect presents some methodological shortcomings. The purpose of the present study is to evaluate the accuracy of the Downton and STRATIFY instruments to determine the risk of falls and to predict their incidence in acute care hospitals in the public health system in Andalusia (Spain). A longitudinal, multicentre prospective study was made of a cohort of patients recruited between May 2014 and March 2016. The risk of falls was assessed using each of the above instruments during the first 24 h after hospital admittance, with later re-evaluations every 72 h until discharge. Descriptive statistics were obtained, bivariate and multivariate analysis were performed. The diagnostic validity of the process was assessed by calculations of sensitivity, specificity, positive and negative predictive values and ratios of positive and negative likelihood. ROC curve analysis was performed for both instruments. For this study, 1247 patients were recruited, of whom 977 completed all the follow-up assessments. Twenty-three of these patients (2.35%) suffered 24 falls. ROC curve analysis showed that the optimal cut-off point for each assessment instrument was below that described by the authors: AUC STRATIFY = 0.69 (95% CI: 0.57-0.8); AUC Downton = 0.6 (95% CI: 0.48-0.72). With a cut-off point of 1, the sensitivity of STRATIFY was 47.6% and its specificity, 85%. With a cut-off point of 2, Downton presented a sensitivity of 66.7% and a specificity of 55.3%. The Downton and STRATIFY falls risk assessment instruments presented little utility as means of detecting the risk of falls among a sample of adult patients admitted to acute care hospitals. Fall prevention in hospitals should be based on the study of individual risk factors.