Lucerne Fall Risk Prediction Score for Stroke Rehabilitation

Overview

In Switzerland, every year around 16'000 people suffer a stroke. Stroke represents the third most common cause of death in Switzerland and leads to impairments (e.g., motoric, cognitive, sensory) resulting in disability. People with disabilities after stroke should have access to specialised interprofessional rehabilitation settings. During inpatient rehabilitation, 15-36% of the patients experience one or more falls. It is well known that stroke is an important risk factor for falls. On average stroke patients fall 1.77 times more than the age- and gender-matched controls over 13 months. Falling events during inpatient stroke rehabilitation result in an extension of rehabilitation stay of about eleven days. Wong et al. (2016) suspect that a reduction in the activity level due to falls, fear of falling again as well as changes in discharge conditions could be the reason for this extended length of stay. Walsh et al. (2018) demonstrate that patients who fall once within the first year after stroke cause € 8'600 and recurrent fallers € 12'700 higher healthcare costs. Fall risk factors after stroke are well investigated. Campbell \& Matthews (2010) have collected multiple factors for falls in inpatient stroke rehabilitation from 1990 to 2009 in an integrative review. A newer systematic review points out physical function, hemi-attention, and stability as the most important factors for falls in inpatient stroke rehabilitation. However, none of the included studies showed a validated prediction model with acceptable performance. Hence, further investigations regarding the impact of various valid and reliable fall risk assessments at admission in inpatient rehabilitation are needed. The neurorehabilitation team of LUKS systematically assesses the patient's functions and activity to design patient-specific, evidence-based rehabilitation. Therefore, a population-specific fall risk model based on standardized assessments performed in the clinical routine would help to identify patients with a high risk of falling during rehabilitation without the need of implementing an existing model with a low performance. Aim of the study The main aim of this study is to establish a multivariable prediction model for falls during inpatient rehabilitation in acute and subacute stroke patients admitted to the Clinic for Neurology and Neurorehabilitation of the Kantonsspital Luzern (LUKS) in Lucerne, Switzerland. The secondary aim is to explore the value of the mini-BESTest as a fall predictor in a subgroup consisting of patients who are ambulatory at admission to the Clinic for Neurology and Neurorehabilitation.