The effect of activity and environment on dynamic fall risk

• Applicant: Professor Dr. Jochen Klenk
• Subject Area: Biogerontology and Geriatric Medicine
• Term: from 2018 to 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 401741517

Falls are the major cause of injury and disability in older people, leading to serious health and social consequences including fractures and institutionalization. However, knowledge about falls and triggering fall risk factors is limited due to insufficient data about real-world falls. Currently, studies are mainly based on imprecise self-reports from fallers, as direct observation of falls is difficult. Furthermore, fall risk factors are dynamic and may change within short time intervals. For this reason, accuracy of present fall risk models is weak and efficiency of fall prevention measures is limited. Body-worn sensors including accelerometers and gyroscopes enable the continuous biomechanical observation of falls. During the European FARSEEING project, we built the world largest database of more than 200 sensor-recorded real-world falls. We also conventionalized a dynamic fall risk model considering the causal conditions at the fall event, including pre-fall activity and environmental factors.The aim of this project is to operationalise the dynamic fall risk model using data from body-worn sensors. We hypothesise, that the model will facilitate the instantaneous estimation of the individual fall risk of a person in different situations. During this project, we will focus on three common ecologically valid real-world fall scenarios based on data from the FARSEEING database (falling in a bus while the bus is stopping, falling due to a trip over an obstacle and falling during walking backwards while opening a door). In a first step, the fall-related pre-fall activities (standing, walking forward, walking backwards) and important modifying environmental factors (e.g. opening the door) will be identified and characterised. Algorithms will be developed for the automatic detection of these factors. In a second step, an experimental environment will be designed to simulate the selected fall scenarios under varying conditions to investigate the fall risk related to each considered risk factor in an experimental study. The main outcome defining a fall will be the registration of >30% of the participant’s body weight on a load cell, which is connected with a safety harness, protecting the participant from falling. Based on the results of this experiment and the developed algorithms, we will establish a mathematical model combining all risk factors to instantaneously estimate the individual fall risk in the considered scenarios. This model will then be validated in a final experiment.The project will prove the concept of a dynamic fall risk model and provide the foundation to expand it to further scenarios, overcoming the limitations of present fall risk models. The continuous and valid monitoring of fall risk and the identification of individual risk factors will timely guide effective personalized fall prevention interventions to reduce burden of disease, institutionalisation and health care costs.

• DFG Programme: Research Grants
• Co-Investigator: Professor Dr. Clemens Becker