An automated, electronic assessment tool can accurately classify older adult postural stability

Abstract

Current methods of balance assessment in the clinical environment are often subjective, time-consuming and lack clinical relevance for non-ambulatory older adults. The objective of this study was to develop a novel method of balance assessment that utilizes data collected using the Microsoft Kinect 2 to create a Berg Balance Scale score, which is completely determined by statistical methods rather than by human evaluators. 74 older adults, both healthy and balance impaired, were recruited for this trial. All participants completed the Berg Balance Scale (BBS) which was scored independently by trained physical therapists. Participants then completed the items of the “Modified Berg Balance Scale” in front of the Microsoft Kinect camera. Kinematic data collected during this measurement was used to train a feed-forward neural network that was used to assign a Berg Balance Scale score. The neural network model estimated the clinician-assigned BBS score to within a median of 0.93 points for the participants in our sample population (range: 0.02–5.69). Using low-cost depth sensing camera technology and a clinical protocol that takes less than 5 min to complete in both ambulatory and non-ambulatory older adults, the method outlined in this manuscript can accurately predict a participant’s BBS score and thereby identify whether they are deemed a high fall risk or not. If implemented correctly, this could enable fall prevention services to be deployed in a timely fashion using low-cost, accessible technology, resulting in improved safety of older adults.

Introduction

In 2012, older adults in the US were treated for an estimated 1.76 million falls in the emergency room (Burns et al., 2016). Falls are the leading cause of fatal injury, can increase hospital stays and readmissions, and are an important factor affecting morbidity, mortality, and general independence among older adults (Perell et al., 2001). The cost of falls in older adults in the US is high. In 2015, the direct medical costs of fatal falls for people aged 65 years and over totalled $637.5 million (USD), and $31.3 billion (USD) was spent on injuries following non-fatal falls (Burns et al., 2016). Given the risk and cost of falls and fall-related injuries, the assessment of balance and identification of fall risk should be a critical part of routine clinical care for older adults (Xu et al., 2018). Indeed, risk assessment and stratification initiatives may be able to decrease the incidence of falls in older adults, but often require clinical visits that are costly and logistically difficult for both clinicians and patients.

The Berg Balance Scale (BBS) is a commonly used measure of balance within the clinical setting (Berg et al., 1995). However, the BBS has demonstrated floor and ceiling effects in the past (Blum and Korner-Bitensky, 2008), and has limited application in non-ambulatory individuals given most components of the test require the individual to be standing. We previously proposed the modified BBS (MBBS) as an assessment of seated balance for non-ambulatory people, which may enable balance to be safely assessed without the need for supervision (Dehbandi et al., 2017). The BBS is also a time-consuming assessment (15–20 min), whereas the MBBS can be completed in 5–6 min, which may be particularly advantageous given clinical consults are time limited.

Computerised posturography is a more sensitive measure but typically requires expensive equipment (e.g. force platforms) and trained staff to administer the test and interpret the data collected. Furthermore, the set-up, calibration and data collection are often time-consuming, which might limit its clinical applicability. There are several low-cost, automated, electronic devices that are widely available and have the capacity to measure gait and balance, but until recently, their capacity to sensitively and reliably quantify human motor behaviour has not been established. A comprehensive automated kinematic analysis might offer an innovative solution to the current risk landscape that is the measurement and interpretation of balance in the clinical setting.

The Microsoft Kinect 2 (MK2, Microsoft Corporation, Redmond, WA, USA) is a combined high-definition (HD) video camera and active infrared (IR) camera with a depth sensor designed for 3-dimensional (3D) body tracking. It accurately records the movements of the head, trunk and limbs in 3D space by tracking the displacement of 25 inferred anatomical landmarks; or “centroids”. This technology has widespread application in assessing balance and gait control. It has been shown to accurately assess static (Clark et al., 2015) and dynamic (Eltoukhy et al., 2017) balance, and gait (Mentiplay et al., 2015), including multiple components of the timed-up-and-go (Vernon et al., 2015).

When collected under appropriate conditions, data from the MK2 can be paired with machine learning algorithms to sensitively identify clinically relevant information. Our previous work demonstrated that data collected using the MK2 can be used to discriminate between three distinct states of postural stability in 12 healthy adults performing a modified version of the MBBS (Dehbandi et al., 2017). Taken together, previous work with the MK2 suggests that it could offer a novel approach to identifying individuals with impaired balance. In addition, its low-cost, accessibility and customizability make it, and similar technologies, of relevance to the telemedicine and rehabilitation communities.

The aim of this study was to evaluate the capability of the MK2 to predict the BBS score of older adults with varying levels of postural stability from a few simple, seated movements performed in front of the camera. The research makes progress towards a brief, accurate measure of balance that could be applied both within and away from the clinical setting. This could alleviate time-pressures placed on clinicians/therapists and enhance the robustness of the clinical assessment of balance, and thereby reduce the risk of falling.