Agreement and predictive power of fall risk assessment methods in the hospitalized elderly population: a prospective observational study

Abstract

Introduction: According to the World Health Organization (WHO), Brazil will reach the sixth position in the world in number of elderly people by 2025. The increase in longevity has brought about demographic changes not only in Brazil, but worldwide.
The aging process promotes physiological changes and functional losses, bringing more vulnerability. In 2007, WHO conceptualized the fall as "as an event that results in a person inadvertently resting on the ground or some other lower level". Decreased
postural reflexes, disturbances in proprioception, vision, motor and gait are examples of some factors that contribute to the occurrence of falls, which together increase the risk of falls in the elderly. In order to avoid accidents, identification and accurate
assessment using tested, valid and reliable instruments is essential. Objective: This study aims to investigate the reliability and validity of fall risk screening methods in hospitalized elderly. Methods: Longitudinal, observational, prospective study. Participants were assessed using five clinical instruments: Functional Independence Measure (FIM), Morse Fall Scale (MFS), St. Thomas's Risk Assessment Tool in Falling Elderly Inpatients (STRATIFY) and Johns Hopkins Fall Risk Assessment Tool (JH-FRAT) and use of polypharmacy (POLIF). The prospective occurrence of falls was registered in the medical record. The performance of the methods was evaluated for their accuracy, sensitivity, specificity, positive and negative predictive values, together with the 95% confidence interval. Results: Partial analysis of data collected between Feb/21 and Oct/21 with 75 participants (median [P25-P75]) 67 (64-71) years, 39 (52%) men. In prospective follow-up, 5 (7%) participants reported at least one fall during their hospital stay. The STRATIFY method (accuracy = 92% [83⎯97]) was more accurate for tracking a faller, followed by POLYF (accuracy = 80% [69⎯88]), MFS (accuracy = 76% [65-84]), JH-FRAT (accuracy = 75% [64⎯83]) and FMI (accuracy = 75% [64⎯83]). STRATIFY also showed the highest positive (40% [12⎯77]), and negative (96% [88⎯99]) predictive value. Conclusion: The findings on accuracy and related probabilities suggest that
most of the investigated methods are valid for tracking the risk of falling in hospitalized elderly, with emphasis on STRATIFY. The continuity of the study will allow the analysis of agreement between methods for screening low/high risk of falls in this population.