Identifying Fallers: The Importance of Age-specific Base-of-Support Models to Quantify Balance

4268
LH Sloot1; M Millard2
1. Translational and Clinical Research Institute, National Institute for Health and Care Research (NIHR) Newcastle Biomedical Research Centre (BRC), Newcastle University; 2. Institute for Sport and Movement Science, University of Stuttgart
Scientific Research
SP - Falls
View DOI

Abstract

Introduction

Biomechanical balance models can help identify fall-prone individuals. The often-used extrapolated center-of-mass model describes how well the body is controlled relative to the feet’s functional base-of-support (fBOS). Often, a single foot marker is used to define the fBOS, reducing balance accuracy. Therefore, we created age-specific fBOS models and evaluated changes in the fBOS size with age, frailty, fall history and fear.

Methods

We assessed 38 young (26±5 yrs), 14 middle-aged (50±7 yrs), and 34 older participants (76±6 yrs; 20 non-frail (Clinical Frailty Scale=1&2); 10 managing well (CFS=3); 4 mildly frail (CFS=4); Fear of falling: Falls Efficacy Scale; n=6 reporting fall <6 months). Ground forces and foot markers were recorded during standing, slow circular maximum leaning. The fBOS is the convex hull enclosing the centre-of-pressure points (45% body weight, no foot movement), normalised to foot dimensions.

Results

The median fBOS was only 22[9-36]% of the foot outline for young and middle-aged participants, and decreased to 13[6-28]% in older participants (-42%, pranksum=0.001), corresponding to a 6.5cm difference in margin-of-stability when using the fBOS edge compared with heel marker for an average 30cm long foot. The median fBOS size was 38% larger in non-frail older participants (16[7-28]% versus 11[6-21]%, pRankSum=0.02) and 61% larger in those without a fall history (13[8-28]% versus 8[6-21]%; pRankSum=0.03), but not related to fear of falling (τ=-0.14, pKendall=0.30).

Conclusions

The results underline the importance of an accurate fBOS model to identify fall-prone persons, with a decline in the fBOS related to falls. The relation to age and frailty indicates that different sub-models are needed. To address study limitations, we are studying whether these static fBOS models work during dynamic conditions in a larger group of fallers. The age-specific fBOS models are publicly available to retrospectively fit to marker data to increase the accuracy of dynamic balance analysis.

Comments

Dear cath_38580,

I am indeed continuing this research. I am applying these group-specific foot support area (base of support) models to a combined fait dataset of over 200 people across the lifespan to understand how mechanical dynamic balance declines, when it starts, and how it different between sexes and frailty levels. 

My interest is to understand where this age-related decline in base of support comes from, is it muscle strength / control, haptic or other sensory feedback, response time, and can we train this and see the support area improve again. Another interest is to understand the support area during movements, and quantifying balance decline during other daily movements. 

Thank you for your interest! 

Liz

Dear Esma,

I think the calculation of margin of stability like metrics should indeed use a more precise estimation of the actual (of functional as I call my approach) base of support. We know a marker placed at the exterior of the foot is not representative of the BOS which is currently mostly done. I would like to point out that there are alternative models to the xCOM model with different model (violations) that might be more appropriate to different daily movements (as we increasingly realize that steady gait might not be the most balance challenging movement to study). Thus, picking the most appropriate body and foot models can considerably increase the accuracy of margin-of-stability measures, as my result show. 

We do provide the code to attach (also retrospectively) the fBOS models for the different age groups to the most common marker models (and we can make it work for other models). 

It would be interesting to see what dynamic balance does when we more consistently and accurately define the BOS - this is what i'm currently working on. 

Are you working with the MoS / xCOM?

This is a poster presented at a BGS conference. It constitutes research undertaken by the authors  and queries should be addressed to them. The BGS do not endorse or recommend any of the content. BGS do not make any warranties or guarantees in relation to the content.  If you wish to comment on the poster you will need to create a website account, sign in and submit for the authors to respond.