G-STRIDE inertial system to explore gait factors that are associated with frailty in older adults with falls

doi:10.21203/rs.3.rs-3563181/v1

Download PDF

Research Article

G-STRIDE inertial system to explore gait factors that are associated with frailty in older adults with falls

https://doi.org/10.21203/rs.3.rs-3563181/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose

Falls and frailty share an intrinsic relationship and World Health Organization recommends health initiatives focused on the early detection of both to increase individual resilience. The objective of the study is to compare clinical characteristics, functional performance, and gait parameters estimated with the G-STRIDE inertial sensor, between different frailty groups in older adults with and without falls and secondly to identify characteristics of prefrail and frail subjects that are associated with falls.

Method

We conducted an observational, multicenter case-control study in older adults with and without falls. Participants were included in the outpatient clinic in two general public hospitals and three public nursing homes from September 2021 to March 2022. We gathered clinical variables and several gait parameters using G-STRIDE. Frailty status segmentation was based on the Short Physical Performance Battery (SPPB) criteria.

Results

There were 163 participants (72% women and mean age of 82.6 ± 6.2 years), 53% fallers and 70 were fit, 52 were prefrail, and 41 were frail. Frailty and pre-frailty people had more falls and G-STRIDE parameters (mean values and coefficient of variability) show significant differences between frailty groups. Step speed during the swing phase and mean walking speed measured by G-STRIDE identify those prefrail and frail participants with falls.

Conclusion

G-STRIDE inertial sensor show significant differences between frailty groups and identify gait characteristics of prefrail and frail adults with higher risk of falls suggesting its potential implementation into routine assessment for detecting early frailty.

frailty

gait analysis

inertial sensors

falls

Aim

To assess the relationship between the kinematic data obtained with a wearable sensor (G-STRIDE) and frailty status in older adults with and without falls and to identify gait parameters in frail and pre-frail people with higher risk of falls.

Findings

G-STRIDE is a low-cost device that evaluates more than fifteen gait variables during walking test. It detects differences between frailty groups in the same way as the usual functional tests and describe gait characteristics of frail and pre-frail subjects that identify those in the group of fallers. Additionally, is easy to use and timeless, no adverse events related, no subjectivity interpretation, and no time limitation as it can be done out of the clinic.

Message

G-STRIDE wearable sensor shows good characteristics to be implemented in clinical practice for early frailty assessment

Population ageing is a reality of recent decades, especially in developed countries. People over 65 years will reach 25.2% of the Spanish population in 2033¹. Biological ageing and sociodemographic changes determine an increase in chronic diseases, functional disability and health resources. In this context, the concept of frailty has gained interest as screening and intervention on it can lead to healthy aging.

Frailty is a clinical condition characterized by reduced capacities in multiple physiological systems, determining a state of increased vulnerability and susceptibility to adverse health-related outcomes including disability, falls, hospitalisation, institutional care, or even death².

Around 10% of people aged over 65 years have frailty, rising to between a quarter and a half of those aged over 85³, and it is more frequent in women and nursing home residents⁴.

There are two major approaches to defining frailty. For the first one, frailty can be considered the result of deficit accumulation⁵. The second one, proposed by Linda Fried defines Frailty Phenotype according to five criteria² being the locomotor system crucial as three criteria depend on it (gait, physical activity, and muscle strength).

Falls are one of the main “Geriatric Syndromes”, which have a complex multifactorial etiology and are intrinsically linked to frailty^6–9. Patients who suffer falls with consequences, such as fractures or those with recurrent falls are especially vulnerable and, therefore, present a greater risk of new falls¹⁰.

Given this close relationship, World Health Organization (WHO) recommends the implementation of public health initiatives focused on the early detection of frailty and falls to increase individual resilience by improving their health status¹¹.

Multidimensional assessment of frail patients or those with falls have to take special attention to physical performance, gait analysis, and balance and it is usually done using tests such as gait speed¹², Time up and go test (TUG)¹³ or Physical Performance Battery (SPPB)¹⁴. In addition to being useful tools in frailty or falls assessment, they have also proven to be useful for frailty screening^15–18. But all these tests have some limitations; data is potentially subject to bias due to the subjectivity of the explorer, patient collaboration attitude, and characteristics of the physical space where the assessments are conducted and also multidimensional frailty assessment requires a considerable amount of time, which is limited within daily clinical practice.

The use of electronic devices, wearable sensors, and other technologies to measure physical activity or evaluate gait functions is relatively new but with a growing interest in the past ten years^16–20. Although some of the parameters obtained by these technologies seem to provide relevant information for frailty and falls evaluation, as previous results obtained in G-STRIDE pilot study suggested²¹, and they give additional advantages as they are cheap, easy-to use and accurate, it is also true that some aspects need further research to determine the most relevant clinical features, cut-off points, or how to interpret some results²². So, they are not currently used in clinical practice for falls or frailty assessment and are mainly reserved for research purposes.

In particular, little is known about the association between gait parameters, other than gait speed and frailty status, being a major area of interest as spatiotemporal parameters of gait may enhance the accuracy of frailty diagnosis²³.

In this context, the objectives of this study are:

Compare clinical characteristics, functional performance, and gait parameters (estimated with the G-STRIDE inertial sensor) between different frailty groups in older adults with and without falls.

Identify what characteristics of prefrail and frail subjects are associated with falls.

This is an observational, multicenter case-control study in older adults with and without falls. Participants were included in the outpatient clinic in two general public hospitals and three public nursing homes from September 2021 to March 2022.

The study protocol was approved by the Ethics Review Committee of University Hospital La Paz (Registration Number: PI-4486) and informed consent was signed by all participants.

The inclusion criteria for patients with falls are a modification of the criteria proposed by the American Geriatrics Society (AGS) and the British Geriatrics Society (BGS)²⁴. We define the “Fallers Group” as those adults over 70 years who had one of the next circumstances: -One fall with consequences in the last year (requiring medical attention); -Two or more falls in the same period. -Gait and balance disorder. -Fear of falling or post-fall syndrome.

The participants without falls were volunteers over 70 years that gave informed consent.

Exclusion criteria for the study were terminal illness with a life expectancy of fewer than six months.

Clinical assessment was carried out in a single visit and following data were registered: sociodemographic characteristics, physical activity, weight, height, Body Mass Index (BMI), cognitive status by Deterioration Scale from Reisberg (GDS)²⁵, walking speed (WS)¹² Time up and go test (TUG)¹³ and fear of falling syndrome using the Short Falls Efficacy Scale - International (Short FES-1)²⁶.

Frailty Assessment by Short Physical Performance Battery Scale

Short Physical Performance Battery Scale (SPPB) assesses physical performance through three domains: balance, strength, and gait measurements¹⁴.

This tool has demonstrated its ability to predict adverse health events in older populations, and it is considered one of the best performance tests to identify frail subjects^15–18.

Recently it has been recommended by the Spanish Government as the screening tool for frailty in older adults²⁷ and also in Madrid Community for Health Promotion and Frailty Prevention Strategy²⁸.

We consider fit subjects when the participant obtains 10 or more points, prefrail when obtains between 7 and 9, and frail with 6 or fewer points.

Gait analysis. The G-STRIDE system.

The G-STRIDE device is based on the device presented in a previous paper²¹. It is comprised of an inertial sensor (IMU) and processing electronics that allows obtaining kinematic gait-related variables when the device is attached to the foot. Raw inertial data is stored in an SD memory card for subsequent offline processing. This Raw IMU data (accelerations and angular rates) is post-processed for the estimation of spatiotemporal and cinematic parameters, which describe the gait pattern of each subject. The architecture of the G-STRIDE system for gait analysis can be separated into two development blocks: Hardware for IMU data registration and Software for gait analysis.

The device is lightweight with dimensions 57x27x32 mm, and during the tests with the device attached to the shoe with an elastic band, no subject had any complications or problems derived from the use of the device (Fig. 1).

After the walking test of each participant, the different gait parameters for each gait cycle (step) are stored in a database and finally are post-processed (remove outliers, or steps not in a straight line). This database containing all the IMU recordings, estimated parameters in Excel format, and the Python code to process the IMU signals are available in a Zenodo repository²⁹. The variables estimated by the G-STRIDE using the IMU on the foot of each participant are described in Table 1. Parameters are computed as the mean on a step-by-step basics, but we also estimate the standard deviation (STD) and CV (Coefficient of Variability), which is a measure of relative variability and is the ratio of the standard deviation to the mean value.

Table 1

Sensor-based gait estimated variables.
Parameter	Description
Total distance(m)	The total distance travelled during long the free walk is measured in meters.
Total time(s)	The total time taken in the long walk measured in seconds.
Total steps	The total number of steps in the free walk.
Gait Cycle Time (GCT)	The mean Gait Cycle Time (GCT) measured in seconds
Velocity(m/s)	The mean walking speed computed over the total detected steps measured in meters per second.
Cadence(steps/min)	The mean cadence measured number of steps per minute.
Swing time(% GCT)	Swing time (from toe-off to heel strike) as percentage of GCT.
Stance-Loading time(% GCT)	From heel-strike to start of foot-flat time as percentage of GCT.
Stance-FootFlat time(% GCT)	Foot-flat (from start to end) time as percentage of GCT.
Stance-Pushing time(% GCT)	From end of foot-flat to toe-off time as percentage of GCT.
Heel strike angle(deg)	The maximum pitch angle at heel strike measured in degrees.
Toe-off angle(deg)	The maximum pitch angle at toe-off is measured in degrees.
Stride Length-SL(m)	The Stride length (distance from one stance position to the next stance of the same foot) is measured in meters.
StepSpeed(m/s)	The forward speed of foot only during the swing phase is measured in meters per second.
2D Path(m)	The path length of the foot in the horizontal plane during a step (always equal or larger than SL).
3D Path(m)	The path length of the foot in 3D space during a step (always equal to or larger than SL and 2D path)
Clearance(m)	The clearance or maximum height of the foot with respect to ground during the swing phase.

Statistical Analysis

Statistical analysis was carried out with SPSS v.27 (Copyright© 2013 IBM SPSS Corp.) and the R language for statistical computing (R Foundation for Statistical Computing, Vienna, Austria)³⁰ together with the Rstudio Integrated Development Environment for R (Rstudio, Boston, MA)³¹.

The normality of the sample was checked using the Kolgomorov-Smirnov test. Continuous variables were analysed using the mean and standard deviation, and the frequency and percentages of categorical variables. Subsequently, parametric tests were used to measure the differences between independent samples (T-Student); the Chi-Squared test was used to analyse the differences in percentages.

To study which characteristics of prefrail and frail patients are associated with falls we propose to perform logistic regression.

There were 163 participants with a mean age of 82.6 ± 6.2 years and 118 (72%) were women. According to SPPB criteria, there were 70 fit, 52 prefrail, and 41 frail.

Of the total sample, 86 subjects had falls (53%). In the robust group, 23 subjects had falls (33%), 31 prefrail subjects had falls (60%), and 32 frail subjects had falls (78%). Table 2 shows the basal characteristics of the sample.

Table 2

Sample characteristics by frailty groups
Characteristic	Overall, N = 163†	fit, N = 70†	prefrail, N = 52†	frail, N = 41†	p-value‡
Age	82.6 (6.2) / [70.0\|98.0]	80.3 (5.4) / [70.0\|96.0]	82.7 (5.9) / [73.0\|94.0]	86.4 (6.1) / [70.0\|98.0]	< 0.001
Sex					0.5
Female	118 / 163 (72%)	51 / 70 (73%)	35 / 52 (67%)	32 / 41 (78%)
Weight	64.3 (13.1) / [33.1\|105.0]	67.1 (12.2) / [41.6\|105.0]	63.1 (14.1) / [33.1\|90.0]	61.0 (12.6) / [37.6\|93.0]	0.049
Height (m)	1.57 (0.10) / [1.32\|1.84]	1.59 (0.11) / [1.37\|1.84]	1.57 (0.10) / [1.32\|1.81]	1.53 (0.08) / [1.37\|1.72]	0.009
IMC	26.2 (5.0) / [16.5\|42.5]	26.6 (4.5) / [19.5\|42.5]	25.5 (5.4) / [16.9\|36.4]	26.2 (5.4) / [16.5\|41.2]	0.4
GDS	2.0 (1.6) / [1.0\|7.0]	1.4 (0.9) / [1.0\|5.0]	2.4 (1.8) / [1.0\|6.0]	2.7 (1.9) / [1.0\|7.0]	< 0.001
Living_Site					< 0.001
Nursing home	53 / 163 (33%)	12 / 70 (17%)	18 / 52 (35%)	23 / 41 (56%)
Terrain					0.006
Flat Terrain	110 / 163 (67%)	42 / 70 (60%)	32 / 52 (62%)	36 / 41 (88%)
Mix Terrain	53 / 163 (33%)	28 / 70 (40%)	20 / 52 (38%)	5 / 41 (12%)
Fallers	86 / 163 (53%)	23 / 70 (33%)	31 / 52 (60%)	32 / 41 (78%)
Test_Site					< 0.001
Hospital	64 / 163 (39%)	25 / 70 (36%)	21 / 52 (40%)	18 / 41 (44%)
Nursing home	52 / 163 (32%)	13 / 70 (19%)	18 / 52 (35%)	21 / 41 (51%)
Home	47 / 163 (29%)	32 / 70 (46%)	13 / 52 (25%)	2 / 41 (4.9%)
†Mean (SD) / [Minimum\|Maximum]; n / N (%)
‡Kruskal-Wallis rank sum test; Pearson's Chi-squared test

The participants reported that the device was comfortable and did not disturb during walking, although we did not investigate comfort in depth.

Functional and G-STRIDE variables according to frailty status

Table 3 shows the results of the conventional tests used in the assessment of the subjects (WS in 4-meter walking test, TUG, and FES-1). All show significant differences between the three frailty groups. We also found significant differences in all mean values of gait parameter measurements detected by G-STRIDE device and Fig. 2 shows differences between three groups according to the gait speed (4-meter walking test) and step speed and velocity measured by G-STRIDE.

Table 3

Descriptive functional and G-STRIDE IMU variables according to frailty status
Characteristic	Overall, N = 163†	fit, N = 70†	prefrail, N = 52†	frail, N = 41†	p-value‡
Speed_4m_walk	0.8 (0.3) / [0.1\|1.7]	1.1 (0.2) / [0.7\|1.7]	0.8 (0.2) / [0.4\|1.5]	0.5 (0.2) / [0.1\|1.0]	< 0.001
TUG	15.0 (6.7) / [5.3\|36.0]	10.8 (3.3) / [5.3\|23.0]	15.8 (5.5) / [7.4\|29.2]	21.0 (7.4) / [12.1\|36.0]	< 0.001
FES1	10.4 (4.5) / [7.0\|32.0]	8.9 (3.0) / [7.0\|21.0]	10.6 (3.8) / [7.0\|21.0]	12.9 (6.0) / [7.0\|32.0]	< 0.001
†Mean (SD) / [Minimum\|Maximum]; n / N (%)
‡Kruskal-Wallis rank sum test; Pearson's Chi-squared test
Characteristic	Overall, N = 163†	robust, N = 70†	prefrail, N = 52†	frail, N = 41†	p-value‡
Total distance (m)	1,006.5 (610.4) / [73.0\|2,684.8]	1,322.0 (616.9) / [366.5\|2,684.8]	988.3 (494.4) / [298.4\|2,571.7]	490.9 (305.8) / [73.0\|1,193.4]	< 0.001
Total time (s)	1,281.6 (424.2) / [338.6\|2,184.4]	1,415.5 (377.8) / [739.4\|2,184.4]	1,322.7 (403.8) / [564.8\|2,020.5]	1,000.8 (400.5) / [338.6\|1,672.8]	< 0.001
Total steps	1,082.6 (413.6) / [213.0\|2,185.0]	1,252.8 (375.0) / [634.0\|2,185.0]	1,097.6 (360.7) / [485.0\|1,907.0]	773.1 (367.9) / [213.0\|1,707.0]	< 0.001
Gait Cycle Time - GCT (s)	1.2 (0.2) / [0.9\|1.7]	1.2 (0.2) / [0.9\|1.7]	1.2 (0.2) / [1.0\|1.7]	1.4 (0.2) / [1.0\|1.7]	< 0.001
stance - Pushing time (% GCT)	17.8 (2.7) / [8.1\|26.8]	18.6 (2.9) / [12.6\|26.8]	18.0 (2.1) / [12.6\|21.5]	16.1 (2.6) / [8.1\|22.9]	< 0.001
Swing (% GCT)	28.3 (3.5) / [16.1\|34.8]	30.2 (2.2) / [23.3\|34.8]	28.4 (2.8) / [19.2\|33.5]	24.9 (3.7) / [16.1\|31.0]	< 0.001
Stance - Loading time (% GCT)	11.0 (2.3) / [4.1\|16.6]	12.2 (2.0) / [6.3\|16.6]	10.9 (1.7) / [7.6\|15.0]	9.0 (2.1) / [4.1\|13.1]	< 0.001
Stance - FootFlat time (% GCT)	43.0 (7.6) / [26.7\|69.8]	38.8 (5.6) / [26.7\|51.1]	42.7 (5.2) / [32.3\|59.1]	50.6 (7.6) / [34.8\|69.8]	< 0.001
Toe off angle (deg)	-52.3 (13.5) / [-84.2\|-19.2]	-59.7 (12.4) / [-84.2\|-34.1]	-51.5 (11.0) / [-80.9\|-26.3]	-40.9 (9.7) / [-71.7\|-19.2]	< 0.001
Heel strike angle (deg)	14.8 (7.0) / [0.0\|38.0]	18.9 (6.3) / [6.1\|38.0]	14.1 (5.5) / [3.9\|30.8]	8.8 (4.8) / [0.0\|19.1]	< 0.001
Cadence (steps/min)	49.2 (7.5) / [24.0\|66.4]	52.1 (6.2) / [35.3\|66.4]	49.1 (6.1) / [26.4\|61.4]	44.3 (8.7) / [24.0\|61.1]	< 0.001
StepSpeed (m/s)	0.8 (0.3) / [0.1\|1.6]	1.0 (0.3) / [0.4\|1.6]	0.8 (0.2) / [0.3\|1.4]	0.5 (0.2) / [0.1\|1.0]	< 0.001
StrideLength - SL (m)	0.9 (0.3) / [0.3\|1.5]	1.1 (0.2) / [0.5\|1.5]	0.9 (0.2) / [0.5\|1.4]	0.7 (0.2) / [0.3\|1.1]	< 0.001
3D Path (m)	1.0 (0.3) / [0.3\|1.7]	1.2 (0.2) / [0.6\|1.7]	1.0 (0.2) / [0.5\|1.5]	0.7 (0.2) / [0.3\|1.1]	< 0.001
2D Path (m)	0.9 (0.3) / [0.3\|1.6]	1.1 (0.2) / [0.6\|1.6]	0.9 (0.2) / [0.5\|1.4]	0.7 (0.2) / [0.3\|1.1]	< 0.001
Clearance (m)	0.16 (0.12) / [0.02\|0.55]	0.22 (0.13) / [0.06\|0.52]	0.13 (0.11) / [0.04\|0.55]	0.11 (0.07) / [0.02\|0.27]	< 0.001
Velocity (m/s)	0.8 (0.3) / [0.1\|1.5]	0.9 (0.3) / [0.4\|1.5]	0.7 (0.2) / [0.3\|1.3]	0.5 (0.2) / [0.1\|0.9]	< 0.001
†Mean (SD) / [Minimum\|Maximum]
‡Kruskal-Wallis rank sum test

In addition, we observed significant differences between frailty groups in coefficient of variability (CV) of different gait parameters: Gait Cycle Time (p < 0.001), Swing(p < 0.001), Stance foot flat(p < 0.001), Heel strike angle(p < 0.001), Cadence (p = 0.040), Step speed(p < 0.001), Stride length(p = 0.007), 2D Path (p = 0.011), and Clearance(p < 0.001).

Characteristics of Frailty and Pre-frailty people related to falls event

A logistic regression analysis was carried out to determine which characteristics of frail and pre-frail subjects identify those in the group of fallers.

In the frail group, the variables that G-STRIDE detects to falls are Swing (CutOff = 29.1, Z= -2.22, p < 0.02), Step Speed (CutOff = 0.79, Z= -2.21, p < 0.02), velocity (CutOff = 0.73, Z= -2.18, p < 0.02) y StepSpeed CV (CutOff = 0.17, Z= -2.06, p < 0.03) (Table 4).

Table 4

Characteristics of frailty and prefrailty people related to falls events: Logistic regression analysis.
	FRAIL					PREFRAIL
variable	CutOff	intcpt	coef	z	p	CutOff	intep	coef	z	p
Total distance (m)	1102.003	2.514	-0.002	-1.79	0.073	1093.013	3.871	-0.004	-3.19	0.001*
Total time (s)	2305.233	2.319	-0.001	-1.02	0.308	1509.646	4.201	-0.003	-2.87	0.004*
Total steps	1786.694	2.345	-0.001	-1.24	0.216	1230.798	5.019	-0.004	-3.25	0.001*
Gait Cycle Time - GCT (s)	0.910	-2.756	3.027	1.58	0.114	1.140	-5.443	4.775	1.96	0.050
Stance - Pushing time (% GCT)	-27.281	0.799	0.029	0.20	0.844	19.477	5.489	-0.282	-1.86	0.062
Swing (% GCT)	29.107	12.574	-0.432	-2.22	0.026*	30.494	6.014	-0.197	-1.72	0.085
Stance - Loading time (% GCT)	52.400	1.533	-0.029	-0.16	0.870	13.696	1.912	-0.140	-0.82	0.412
Stance - Foot Flat time (% GCT)	35.370	-3.223	0.091	1.49	0.136	39.127	-4.665	0.119	1.90	0.058
Toe off angle (deg)	-88.634	2.389	0.027	0.69	0.489	-57.103	4.578	0.080	2.51	0.012*
Heel strike angle (deg)	37.400	1.676	-0.045	-0.56	0.572	17.300	2.331	-0.135	-2.18	0.029*
Cadence (steps/min)	61.459	5.022	-0.082	-1.62	0.105	53.096	5.854	-0.110	-1.88	0.059
Step Speed (m/s)	0.796	4.572	-5.745	-2.21	0.027*	0.897	4.445	-4.954	-2.91	0.003*
Stride Length - SL (m)	0.976	4.328	-4.434	-1.89	0.058	1.022	4.166	-4.076	-2.55	0.010*
3D Path (m)	1.039	4.450	-4.285	-1.87	0.061	1.089	4.248	-3.900	-2.66	0.007*
2D Path (m)	0.995	4.381	-4.404	-1.89	0.058	1.039	4.191	-4.033	-2.55	0.010*
Clearance (m)	-0.276	0.921	3.339	0.55	0.582	0.912	-4.093	-4.093	-1.43	0.152
Velocity (m/s)	0.737	4.550	-6.172	-2.18	0.029*	4.393	-5.272	-5.272	-2.81	0.004*
* indicates statistical significance at the α = 0.001 level.

In the pre-frail group, the G-STRIDE detects that the most important values during the gait cycle would be total distance, total time, total number of steps, step speed, stride length, the toe-off phase, heel strike, and velocity (Table 4). When analysing the height compensated data, the variables that are detected are total steps (CutOff = 1183.5, Z= -2.78, p < 0.005), Gait cycle time (CutOff = 1.20, Z = 2.17, p < 0.02), Stance pushing time (CutOff = 18.44, Z= -1.98, p < 0.04), Cadence (CutOff = 50.3, Z= -2.10, p < 0.03), Step speed (CutOff = 0.86, Z= -1.97, p < 0.04), velocity (CutOff = 0.79, Z= -2.04, p < 0.04), Gait cycle time CV (CutOff = 0.07, Z = 2.17, p < 0.03), Swing CV (CutOff = 0.07, Z = 2.05, p < 0.04).

The results reveal that conventional and G-STRIDE functional screening show significant differences between the three frailty groups. Falls are more frequent in pre-frail and frail subjects, and it is possible to identify gait parameters using G-STRIDE device in pre-frail and frail subjects at higher risk of falls.

We found that frail and pre-frail subjects have significantly more falls. Previous studies have described similar findings^32–34. The relationship between frailty and falls is close and there are numerous studies demonstrating the ability to identify or detect frailty and risk of falls with functional or clinical tools. The study by Bartosch et al.³⁵ finds that frailty as measured by the Frailty Index is a predictor of falls and recommends its use in patients with falls. The study by Middleton et al.³⁶ analysed the association of mortality, falls, and fractures with frailty using eFRAGICAP based on the Frailty Index, finding that frail subjects had a higher risk of adverse events. Seong-Hi Park's review of falls assessment tools finds TUG and Tinetti Test as the most appropriate in the community setting with a sensitivity > 0.7 but in general, they recommend the use of several tools simultaneously to improve falls prediction³⁷. In our study, we found differences between pre-frail, frail and fit subjects in functional tests (gait speed and TUG ) but also in other clinical aspects such as cognitive impairment measured by the GDS, suggesting the importance of the cognitive sphere in both frailty and falls or fear of falling measured by the FES-1 which also underlines the importance of the self-perceived risk of falling in frail subjects as other authors have also published^38,39. However, clinical functional assessment with both scales and physical tests has limitations as it is based on the patient's perception or the assessor's interpretation.

The results of this study indicate that the G-STRIDE device detects differences between frailty groups in the same way as the usual functional tests, as has been proven in previous publications²¹. Studies have been published using sensor devices, although under different testing conditions (different locations of the sensor, duration of the tests, number and type of variables to be determined), which makes it still difficult to protocolise their use²². G-STRIDE collects a large number of gait parameters while the subject walks freely and we find that frail and pre-frail subjects walk smaller distances, have a smaller number of steps and the time they walk is shorter. Furthermore, all the parameters evaluated by the device show significant differences between the three frailty groups, as well as the "coefficient of variability" (CV) that allows this differentiation between the three groups. Other authors^40,41 also attach great importance to variability, asymmetry, and irregularity in gait performance as significant and early indicator of frailty and falls. A recent review²³ concluded that the parameters that best differentiate between frailty subgroups are gait speed, in particular during habitual walking, cadence, step width variability, step length during habitual walking, and double support time during fast walking. Furthermore, it concluded that variability and dual tests are predictive of the development of frailty.

G-STRIDE device shows two gait parameters that identify pre-frail and frail subjects in the group of falls risk; StepSpeed and Velocity are related to the gait speed component, which we know is an indicator of frailty and adverse events, considered the sixth vital sign by many authors. Mohler et al.⁴² found that among frail and pre-frail older adults, only balance and physical activity were predictive of future falls, but gait parameters were not, but the study had limitations as a small number of frail participants and exclusion criteria for neurological disorders.

Other studies have found a relationship between gait parameters detected by technology and frailty status. In the study by Montero-Odasso et al.⁴⁰, step width variability during normal gait was significantly increased in the prefrail group compared to the fit one and during fast walking, most of the variability parameters showed significant differences between groups.

Schwenk et al.⁴⁴ found that gait speed, hip sway, and steps/day were the most sensitive parameters for prefrailty identification, and stride length and double support were the most sensitive gait parameters for discriminating between frailty levels^43, but falls were not studied regarding frailty status.

In this context, G-STRIDE device shows good characteristics to be implemented in clinical practice for frailty assessment as it is easy to use, does not require any specialized personal or knowledge, have no adverse events related, high accuracy, no subjectivity interpretation, and no time limitation as it can be done out of the clinic.

Our study has some limitations. Firstly, the sample size is not excessively large; however, the homogeneity of the sample and the type of statistical analysis ensure a reliable prediction of the results. In addition, there are more women and a larger group of fit subjects, which may be related to the fact that most of them come from community settings where the prevalence of frailty is lower than in nursing homes. All this could have partly conditioned the results, but on the other hand, the study has the strength of analysing a large number of variables in subjects from two very different environments and with a very high average age, which is of great value from the point of view of research in the elderly population.

Finally, these results suggest new lines of research to confirm current results, to identify cut-off points to distinguish fit from prefrail subjects and those with higher risk of falls.

In conclusion, the results of this study confirm that falls are more frequent in frail or pre-frail subjects, there are clinical, functional and gait differences between them and G-STRIDE is an easy and accurate tool that may identify those pre-frail and frail subjects at higher risk of falling.

Acknowledgments

We thank all the participants of the study for their collaboration, time and dedication.

Disclosure statement

The authors declare that they have no conflicts of interest concerning this article.

Instituto Nacional de Estadística. Proyecciones de Población 2018 [Internet]. 2018; Available from: https://www.ine.es/prensa/pp_2018_2068.pdf.
Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Group., Cardiovascular Health Study Collaborative Research. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci [Internet]. 2001;56(3):M146-156. doi: 10.1093/gerona/56.3.m146.
Rubenstein LZ. Falls in older people: epidemiology, risk factors and strategies for prevention. Age Ageing [Internet]. 2006, 35 Suppl 2 (SUPPL.2) doi: 10.1093/ageing/afl084.
Bergen G, Stevens MR, Burns ER. Falls and Fall Injuries Among Adults Aged ≥65 Years -United States, 2014. MMWR Morb Mortal Wkly Rep [Internet]. 2016; 65(37):993-8. doi: 10.15585/mmwr.mm6537a2.
Rockwood K, Mitnitski A. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci [Internet]. 2007;62(7):722-7. doi: 10.1093/gerona/62.7.722.
Lu Z, Er Y, Zhan Y, Deng X, Jin Y, Ye P, Duan L.. Association of Frailty Status with Risk of Fall among Middle-Aged and Older Adults in China: A Nationally Representative Cohort Study. J Nutr Health Aging. [Internet] 2021;25(8):985-92. doi: 10.1007/s12603-021-1655-x.
Chittrakul J, Siviroj P, Sungkarat S, Sapbamrer R. Physical Frailty and Fall Risk in Community-Dwelling Older Adults: A Cross-Sectional Study. J Aging Res [Internet]. 2020; 2020:3964973. doi: 10.1155/2020/3964973.
GKojima G. Frailty as a Predictor of Future Falls Among Community-Dwelling Older People: A Systematic Review and Meta-Analysis. J Am Med Dir Assoc [Internet]. 2015;16(12):1027-33. doi: 10.1016/j.jamda.2015.06.018.
Cheng MH, Chang SF. Frailty as a Risk Factor for Falls Among Community Dwelling People: Evidence From a Meta-Analysis. J Nurs Scholarsh [Internet]. 2017;49(5):529-536. doi: 10.1111/jnu.12322.
Gallo da Silva TT, Melo Filho J, Biesek S, Vojciechowski AS, Borba VZC, Gomes ARS. Accuracy of Tools to Differentiate Single From Recurrent Fallers Pre-Frail Older Women. Front Public Health [Internet]. 2022; 10:716851. doi: 10.3389/fpubh.2022.716851.
WHO. 2021. Step safely: strategies for preventing and managing falls across the life-course. Geneva: World Health Organization; 2021. Licence: CC BY-NC-SA 3.0 IGO. Avalaible from: https://www.who.int/publications/i/item/978924002191-4.
Studenski S, Perera S, Patel K, Rosano C, Faulkner K, Inzitari M, et al. Gait speed and survival in older adults. JAMA [Internet]. 2011;305(1):50-8. doi: 10.1001/jama.2010.1923.
Podsiadlo D, Richardson S. The Timed “Up & Go”: A Test of Basic Functional Mobility for Frail Elderly Persons. J Am Geriatr Soc [Internet]. 1991; 39: 142-8. https://doi.org/10.1111/j.1532-5415.1991.tb01616.x.
Guralnik JM, Simonsick EM, Ferrucci L, Glynn RJ, Berkman LF, Blazer DG, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol [Internet]. 1994;49(2):M85-94. doi: 10.1093/geronj/49.2.m85.
Abizanda P, Romero L, Sánchez-Jurado PM, Atienzar-Núñez P, Esquinas-Requena JL, García-Nogueras I. Association between Functional Assessment Instruments and Frailty in Older Adults: The FRADEA Study. J Frailty Aging [Internet]. 2012;1(4):162-8. doi: 10.14283/jfa.2012.25.
Marzetti E, Cesari M, Calvani R, Msihid J, Tosato M, Rodriguez-Mañas L, et al.; SPRINTT Consortium. The "Sarcopenia and Physical fRailty IN older people: multi-componenT Treatment strategies" (SPRINTT) randomized controlled trial: Case finding, screening and characteristics of eligible participants. Exp Gerontol. 2018 Nov;113:48-57. doi: 10.1016/j.exger.2018.09.017.
Jung HW, Baek JY, Jang IY, Guralnik JM, Rockwood K, Lee E, et al. Short Physical Performance Battery as a Crosswalk Between Frailty Phenotype and Deficit Accumulation Frailty Index. J Gerontol A Biol Sci Med Sci. 2021 Nov 15;76(12):2249-2255. doi: 10.1093/gerona/glab087.
Ramírez-Vélez R, López Sáez de Asteasu M, Morley JE, Cano-Gutierrez CA, Izquierdo M. Performance of the Short Physical Performance Battery in Identifying the Frailty Phenotype and Predicting Geriatric Syndromes in Community-Dwelling Elderly. J Nutr Health Aging. 2021;25(2):209-217. doi: 10.1007/s12603-020-1484-3.
Chen M, Wang H, Yu L, Yeung EHK, Luo J, Tsui KL, et al. A Systematic Review of Wearable Sensor-Based Technologies for Fall Risk Assessment in Older Adults. Sensors (Basel) [Internet]. 2022;22(18):6752. doi: 10.3390/s22186752.
Brognara L, Mazzotti A, Di Martino A, Faldini C, Cauli O. Wearable Sensor for Assessing Gait and Postural Alterations in Patients with Diabetes: A Scoping Review. Medicina (Kaunas). 2021 Oct 22;57(11):1145. doi: 10.3390/medicina57111145. PMID: 34833363; PMCID: PMC8621058.
García-Villamil G, Neira-Álvarez M, Huertas-Hoyas E, Jiménez A.R., Rodríguez-Sánchez C. A Pilot Study to Validate a Wearable Inertial Sensor for Gait Assessment in Older Adults with Falls. Sensors (Basel) [Internet]. 2021;21(13):4334. doi: 10.3390/s21134334.
Ruiz-Ruiz, L.; Jimenez, A.R.; Garcia-Villamil, G.; Seco, F. Detecting Fall Risk and Frailty in Elders with Inertial Motion Sensors: A Survey of Significant Gait Parameters. Sensors (Basel) [Internet]. 2021; 21, 6918. https://doi.org/10.3390/s21206918).
Schwenk M, Howe C, Saleh A, Mohler J, Grewal G, Armstrong D, Najafi B. Frailty and technology: a systematic review of gait analysis in those with frailty. Gerontology [Internet]. 2014;60(1):79-89. doi: 10.1159/000354211. Epub 2013 Aug 15.
Panel on Prevention of Falls in Older Persons, American Geriatrics Society and British Geriatrics Society. Summary of the Updated American Geriatrics Society/British Geriatrics Society clinical practice guideline for prevention of falls in older persons. J Am Geriatr Soc [Internet]. 2011; 59(1):148-57. doi: 10.1111/j.1532-5415.2010.03234.x.
Auer S, Reisberg B. The GDS/FAST Staging System. International Psychogeriatrics. 1997 dec;9(S1):167–171.Available from: https://www.cambridge.org/core/product/identifier/S1041610297004869/type/journal_article.
Kempen GI, Yardley L, van Haastregt JC, Zijlstra GA, Beyer N, Hauer K, et al. The Short FES-I: a shortened version of the falls efficacy scale-international to assess fear of falling. Age Ageing [Internet]. 2008;37(1):45-50. doi: 10.1093/ageing/afm157.
[Ministry of Health]. [Update of the consensus document on prevention of frailty in older people] [Internet] Madrid; 2022. Available at: chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.sanidad.gob.es/areas/promocionPrevencion/envejecimientoSaludable/fragilidadCaidas/docs/ActualizacionDoc_FragilidadyCaidas_personamayor.pdf
[Plan for Comprehensive Care for Frailty and Promotion of Healthy Longevity in Older People in the Community of Madrid 2022-2025] [Internet]. Madrid: september/2022. ISBN: 978-84-451-4007-9. Available at: https://www.comunidad.madrid/transparencia/informacion-institucional/planes-programas/plan-atencion-integral-fragilidad-y-promocion-longevidad
Garcia-Villamil GSTRIDE: A database of frailty and functional assessments with inertial gait data from elderly fallers and non-fallers populations. 2022. [Online]. Available: https://doi.org/10.5281/zenodo.6883292.
R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria; 2022. Available from: https://www.R-project.org/.
RStudio Team. RStudio: Integrated Development Environment for R. Boston, MA; 2022. Available from:http://www.rstudio.com/.
Zhang Q, Zhao X, Liu H, Ding H. Frailty as a predictor of future falls and disability: a four-year follow-up study of Chinese older adults. BMC Geriatr [Internet]. 2020;20(1):388. doi: 10.1186/s12877-020-01798-z.
Duarte GP, Santos JLF, Lebrão ML, Duarte YAO. Relationship of falls among the elderly and frailty components. Rev Bras Epidemiol [Internet]. 2019; 21Suppl 02 (Suppl 02):e180017. doi: 10.1590/1980-549720180017.supl.2.
Del Brutto OH, Mera RM, Peinado CD, Zambrano M, Sedler MJ. Frailty and Risk of Falls in Community-Dwelling Older Adults Living in a Rural Setting. The Atahualpa Project. J Frailty Aging [Internet]. 2020;9(3):150-154. doi: 10.14283/jfa.2019.36.
Bartosch PS, Kristensson J, McGuigan FE, Akesson KE. Frailty and prediction of recurrent falls over 10 years in a community cohort of 75-year-old women. Aging Clin Exp Res [Internet]. 2020;32(11):2241-2250. doi: 10.1007/s40520-019-01467-1.
Middleton R, Poveda JL, Orfila Pernas F, Martinez Laguna D, Diez Perez A, Nogués X, et al. Mortality, Falls, and Fracture Risk Are Positively Associated With Frailty: A SIDIAP Cohort Study of 890 000 Patients. J Gerontol A Biol Sci Med Sci [Internet]. 2022;77(1):148-154. doi: 10.1093/gerona/glab102.
Park SH. Tools for assessing fall risk in the elderly: a systematic review and meta-analysis. Aging Clin Exp Res [Internet]. 2018;30(1):1-16. doi: 10.1007/s40520-017-0749-0.
Esbrí-Víctor M, Huedo-Rodenas I, López-Utiel M, Navarro-López JL, Martínez-Reig M, Serra-Rexach JA, et al. Frailty and Fear of Falling: The FISTAC Study. J Frailty Aging [Internet]. 2017;6(3):136-140. doi: 10.14283/jfa.2017.19.
Gazibara T, Kurtagic I, Kisic-Tepavcevic D, Nurkovic S, Kovacevic N, Gazibara T, et al. Falls, risk factors and fear of falling among persons older than 65 years of age. Psychogeriatrics [Internet]. 2017;17(4):215-223. doi: 10.1111/psyg.12217.
Montero-Odasso M, Muir SW, Hall M, Doherty TJ, Kloseck M, Beauchet O. Gait variability is associated with frailty in community-dwelling older adults. J Gerontol A Biol Sci Med Sci [Internet]. 2011; 66(5):568-76. doi: 10.1093/gerona/glr007.
Pradeep Kumar D, Toosizadeh N, Mohler J, Ehsani H, Mannier C, Laksari K. Sensor-based characterization of daily walking: a new paradigm in pre-frailty/frailty assessment. BMC Geriatr [Internet]. 2020;20(1):164. doi: 10.1186/s12877-020-01572-1.
Mohler MJ, Wendel CS, Taylor-Piliae RE, Toosizadeh N, Najafi B. Motor Performance and Physical Activity as Predictors of Prospective Falls in Community-Dwelling Older Adults by Frailty Level: Application of Wearable Technology. Gerontology [Internet]. 2016;62(6):654-664. doi: 10.1159/000445889.
Schwenk M, Mohler J, Wendel C, D'Huyvetter K, Fain M, Taylor-Piliae R, et al. Wearable sensor-based in-home assessment of gait, balance, and physical activity for discrimination of frailty status: baseline results of the Arizona frailty cohort study. Gerontology [Internet]. 2015;61(3):258-67. doi: 10.1159/000369095.
Schwenk M, Mohler J, Wendel C, D'Huyvetter K, Fain M, Taylor-Piliae R, et al. Wearable sensor-based in-home assessment of gait, balance, and physical activity for discrimination of frailty status: baseline results of the Arizona frailty cohort study. Gerontology [Internet]. 2015;61(3):258-67. doi: 10.1159/000369095.

Download PDF

Version 1

posted

You are reading this latest preprint version

Characteristic	Overall, N = 163†	fit, N = 70†	prefrail, N = 52†	frail, N = 41†	p-value‡
Speed_4m_walk	0.8 (0.3) / [0.1\|1.7]	1.1 (0.2) / [0.7\|1.7]	0.8 (0.2) / [0.4\|1.5]	0.5 (0.2) / [0.1\|1.0]	< 0.001
TUG	15.0 (6.7) / [5.3\|36.0]	10.8 (3.3) / [5.3\|23.0]	15.8 (5.5) / [7.4\|29.2]	21.0 (7.4) / [12.1\|36.0]	< 0.001
FES1	10.4 (4.5) / [7.0\|32.0]	8.9 (3.0) / [7.0\|21.0]	10.6 (3.8) / [7.0\|21.0]	12.9 (6.0) / [7.0\|32.0]	< 0.001
†Mean (SD) / [Minimum\|Maximum]; n / N (%)
‡Kruskal-Wallis rank sum test; Pearson's Chi-squared test
Characteristic	Overall, N = 163†	robust, N = 70†	prefrail, N = 52†	frail, N = 41†	p-value‡
Total distance (m)	1,006.5 (610.4) / [73.0\|2,684.8]	1,322.0 (616.9) / [366.5\|2,684.8]	988.3 (494.4) / [298.4\|2,571.7]	490.9 (305.8) / [73.0\|1,193.4]	< 0.001
Total time (s)	1,281.6 (424.2) / [338.6\|2,184.4]	1,415.5 (377.8) / [739.4\|2,184.4]	1,322.7 (403.8) / [564.8\|2,020.5]	1,000.8 (400.5) / [338.6\|1,672.8]	< 0.001
Total steps	1,082.6 (413.6) / [213.0\|2,185.0]	1,252.8 (375.0) / [634.0\|2,185.0]	1,097.6 (360.7) / [485.0\|1,907.0]	773.1 (367.9) / [213.0\|1,707.0]	< 0.001
Gait Cycle Time - GCT (s)	1.2 (0.2) / [0.9\|1.7]	1.2 (0.2) / [0.9\|1.7]	1.2 (0.2) / [1.0\|1.7]	1.4 (0.2) / [1.0\|1.7]	< 0.001
stance - Pushing time (% GCT)	17.8 (2.7) / [8.1\|26.8]	18.6 (2.9) / [12.6\|26.8]	18.0 (2.1) / [12.6\|21.5]	16.1 (2.6) / [8.1\|22.9]	< 0.001
Swing (% GCT)	28.3 (3.5) / [16.1\|34.8]	30.2 (2.2) / [23.3\|34.8]	28.4 (2.8) / [19.2\|33.5]	24.9 (3.7) / [16.1\|31.0]	< 0.001
Stance - Loading time (% GCT)	11.0 (2.3) / [4.1\|16.6]	12.2 (2.0) / [6.3\|16.6]	10.9 (1.7) / [7.6\|15.0]	9.0 (2.1) / [4.1\|13.1]	< 0.001
Stance - FootFlat time (% GCT)	43.0 (7.6) / [26.7\|69.8]	38.8 (5.6) / [26.7\|51.1]	42.7 (5.2) / [32.3\|59.1]	50.6 (7.6) / [34.8\|69.8]	< 0.001
Toe off angle (deg)	-52.3 (13.5) / [-84.2\|-19.2]	-59.7 (12.4) / [-84.2\|-34.1]	-51.5 (11.0) / [-80.9\|-26.3]	-40.9 (9.7) / [-71.7\|-19.2]	< 0.001
Heel strike angle (deg)	14.8 (7.0) / [0.0\|38.0]	18.9 (6.3) / [6.1\|38.0]	14.1 (5.5) / [3.9\|30.8]	8.8 (4.8) / [0.0\|19.1]	< 0.001
Cadence (steps/min)	49.2 (7.5) / [24.0\|66.4]	52.1 (6.2) / [35.3\|66.4]	49.1 (6.1) / [26.4\|61.4]	44.3 (8.7) / [24.0\|61.1]	< 0.001
StepSpeed (m/s)	0.8 (0.3) / [0.1\|1.6]	1.0 (0.3) / [0.4\|1.6]	0.8 (0.2) / [0.3\|1.4]	0.5 (0.2) / [0.1\|1.0]	< 0.001
StrideLength - SL (m)	0.9 (0.3) / [0.3\|1.5]	1.1 (0.2) / [0.5\|1.5]	0.9 (0.2) / [0.5\|1.4]	0.7 (0.2) / [0.3\|1.1]	< 0.001
3D Path (m)	1.0 (0.3) / [0.3\|1.7]	1.2 (0.2) / [0.6\|1.7]	1.0 (0.2) / [0.5\|1.5]	0.7 (0.2) / [0.3\|1.1]	< 0.001
2D Path (m)	0.9 (0.3) / [0.3\|1.6]	1.1 (0.2) / [0.6\|1.6]	0.9 (0.2) / [0.5\|1.4]	0.7 (0.2) / [0.3\|1.1]	< 0.001
Clearance (m)	0.16 (0.12) / [0.02\|0.55]	0.22 (0.13) / [0.06\|0.52]	0.13 (0.11) / [0.04\|0.55]	0.11 (0.07) / [0.02\|0.27]	< 0.001
Velocity (m/s)	0.8 (0.3) / [0.1\|1.5]	0.9 (0.3) / [0.4\|1.5]	0.7 (0.2) / [0.3\|1.3]	0.5 (0.2) / [0.1\|0.9]	< 0.001
†Mean (SD) / [Minimum\|Maximum]
‡Kruskal-Wallis rank sum test

G-STRIDE inertial system to explore gait factors that are associated with frailty in older adults with falls

Status:

Version 1

Abstract

Purpose

Method

Results

Conclusion

Figures

Key summary points

INTRODUCTION

MATERIALS AND METHODS

Frailty Assessment by Short Physical Performance Battery Scale

Statistical Analysis

RESULTS

Functional and G-STRIDE variables according to frailty status

Characteristics of Frailty and Pre-frailty people related to falls event

DISCUSSION

Declarations

References

Status:

Version 1