Accelerometric Motion Analysis of Balance-Impaired Elderly Subjects

Betty S. Troy, Eric E. Sabelman, Deborah E. Kenney, Sandy Dunn-Gabrielli Rehabilitation R&D Center, VA Health Care System, Palo Alto CA 94304

ABSTRACT

We have previously reported studies on upper body motion analysis using a wearable accelerometric data acquisition system. Our studies have expanded to include data collection from subjects diagnosed with Parkinson's disease, post-stroke hemiplegia, hip arthroplasty, and fall-prone elderly.. This paper examines the use of accelerometry in quantitative analysis of balance-impaired elderly subjects. In particular, the activities of sit-to-stand and quiet standing are used to present apparent differences between normal and balance-impaired subjects in accelerometric data. collected.

BACKGROUND

Falling is a prevalent problem in the elderly population. The sense of balance declines with age due to combined vestibular, proprioceptive and visual losses, resulting in impaired mobility and increased risk of injurious falls. According to Rubenstein, et.al [1] approximately one third of the elderly population living at home will fall each year. About one in forty of these falls will result in hospitalization of the elderly individuals. It has also been reported that recent prior experience of a fall results in fear of falling which has been associated with a significant decrease in social and daily activity and, therefore, quality of life [2, 3].

The physical stability of "balance-impaired" elderly populations (i.e., those diagnosed with Parkinson's disease, post-stoke hemiplegia, hip arthroplasty) is further challenged as compared with normal elderly populations.

Previously, we have reported findings on the use of accelerometry as a measure of stability in the elderly and were able to present distinguishable differences between elderly and young populations [4]. In this paper, we will examine the use of accelerometry as a means of quantitative analysis of elderly balance-impaired subjects.

METHODOLOGY

We have tested subjects as they performed 65 standardized activities, including standing, reaching, bending and walking tasks and simulated activities of daily living (ADLs). The baseline healthy subject pool consisted of 18 elderly (59 to 82 years, equal number males and females) and 12 young (19 to 37 years, 8 male and 4 female). Balance-impaired subject groups included Parkinson's disease patients (20 subjects), pre- and post-surgery hip prosthesis patients (20 subjects), stroke patients (4 subjects) and fall-prone participants in a study of physical therapy to reduce fall risk (17 subjects).

A self report questionnaire, was used to document the characteristics of the subjects and to determine eligibility. Data included fall history, medical history, medication, and the Physical Self-Maintenance Scale [5], which assesses need for help with feeding, dressing, grooming, ambulation, and bathing.

Two 3-axis accelerometers were attached to the left and right corners of eyeglass frames for measuring head motion. Each sensor assembly with its cable weighs 28 grams. Another pair of 3-axis accelerometers were attached to a Velcro belt at the waist above the hip joints to detect body motion. The twelve accelerometer signals were fed to a small (5 x 8 x 15.5 cm) portable computer with battery power source, also attached to the belt. The whole system weighs slightly under 1 kg with the weight equally distributed between front and back of the belt. Technical details of the system were reported previously [6].

Each 3-axis accelerometer site yielded Z (vertical), X (antero-posterior) , and Y (lateral) signals. After calibration for individual sensor characteristics and conversion to units of gravity (1 g = 9.8 m/sec2), two calculations were performed: (1) The acceleration magnitude irrespective of direction of movement was determined by taking the vector sum of the X, Y and Z axes at a single sensor site; this value was expressed in units of gravity and normalized to eliminate the constant 1 g gravitational acceleration. (2) lateral sway, ÐZY, angles of the trunk and head relative to vertical were derived from the vector sum of the Z and, Y accelerometric signals such that ÐZY= arc tan(aZ/aY). The standard deviation (SD) of tilt angle ÐZY, which represents the root-mean-square difference from the mean over the duration of the task, was used as the measure of steadiness. Sway stability has been reported as a predictive measure of fall risk in elderly populations [7]. To eliminate artifacts at the start and end of a task and highly variable events such as false starts and stumbles, the lowest standard deviation (SD) at all four sensor sites during a contiguous 3 second interval was used. This concurrent minimum SD includes a subject's best performance even if he or she had episodes of greater unsteadiness at other times during the test, and therefore is a conservative measure for comparison between groups having varying balance stability. The 3-second interval was selected after observing that even in short (5 second) tasks, overt loss-of-balance episodes lasted no more than two seconds.

RESULTS

Currently, data from one stroke (62 year old female) and four pre-surgery hip-replacement (3 male, ages 59, 60 and 70; 1 66 year old female) subjects have been examined. Results from sit-to-stand and quiet standing activities will be discussed here as an example of analyzing a static and dynamic activity.

Figure 1. Vector magnitudes derived from accelerometric data during sit-to-stand performed by normal and impaired subjects.

Sit-To-Stand

According to Schenkman, et.al. [8], there are typically four stages evident in performance of a "normal" sit-to-stand. They are as follows: (1.) flexion momentum = initiation of movement to just before buttocks are lifted from the seat; (2.) momentum transfer = buttocks lift from the seat to maximum ankle dorsiflexion; (3.) extension = just after maximal ankle dorsiflexion to when the hips first cease to extend; (4.) stabilization = just after hip-extension to when all motion associated with stabilization from rising is complete.

In Figure 1, the plot of the vector magnitude at the head of three subjects performing a sit-to-stand are superimposed. The normal elderly male's trace depicts all four stages, as expected. Stage 1 begins at the beginning of motion to the inflection point of the second positive peak as the head accelerates forward. Momentum transfer ends at the bottom of the negative peak as the subject's center of mass travels anteriorly and upward. Extension ends and stabilization begins after the rise of the negative peak.

The female stroke subject's trace is similar to the normal elderly male's trace in form. Each of the significant peaks are present, except with less amplitude and more frequent fluctuations throughout the activity. The settling period also extends further than that for the normal subject. This subject has left hemiparesis. She has also been diagnosed with diabetes, high blood pressure, and arthritis. She indicated that she has required some assistance in performing activities of daily living, but is relatively independent and ambulatory with a cane. This subject was able to perform the sit-to-stand shown in Figure 1 without assistance.

In the sit-to-stand trace of the female hip replacement subject, the normal significant peaks are not apparent. And, even more erratic fluctuations throughout the trace are present. This subject required a right hip replacement due to degenerative arthritis. She indicated feeling a great deal of pain during testing.

Figure 2. Sway angle at the head vs. sway angle at the hip of normal elderly and hip replacement patients during quiet standing.

Quiet Standing

The best three second standard deviations of the sway angles (ÐZY) at the head were plotted against standard deviations at the body for three male hip arthroplasty subjects tested during quiet standing before surgery (Fig. 2). We would expect to find subjects with greater sway at the head and at the body than normal elders (the shaded box), represents normal baseline elderly sway ranges (head=0.110 degrees; body=0.119 degrees). All three subjects remained outside of the normal elderly sway range.

DISCUSSION

The goal of this paper was to examine the use of accelerometry for the analysis of balance-impaired elderly subjects. We were able to identify abnormal sit-to-stand body motions (i.e., increased frequency of fluctuations and increased settling time) which may be indicative of decreased steadiness. We were also able to identify excessive sway during quiet standing outside of "normal" limits based on our normal elderly baseline population. In general, the existence of balance impairments were distinguishable in accelerometric data of both sit-to-stand and quiet standing activities. We plan to continue data collection and analysis on these impaired populations in an effort to study the subtle balance effects of different disorders and to quantify the changes in balance as a patient's condition improves or worsens.

REFERENCES

1. Rubenstein, L.Z., Josephson, K.R., "Causes and Prevention of Falls in Elderly People", Falls, Balance and Gait Disorders in the Elderly, B Vellas, M Toupet, L Rubenstein, JL Albarede, Y Christen, eds. Elsevier, Paris: 1992, pp. 21-38.

2. Franzoni, S., Ronzzini, R. Boffelli, S., Frisoni, G. B., Trabucchi, M., "Fear of Falling in Nursing Home Patients", Gerontology (Switzerland) 40(1), 1994, pp. 38-44.

3. Arfken, C.L., Lach, H.W., Birge, S.J., Miller, J.P., "The Prevalence and Correlates of Fear of Falling in Elderly Persons Living in the Community", American Journal of Public Health, 84(4), Apr 1994, pp. 565-570.

4. Sabelman, E.E., Hoy, M.G., Winograd, C.H., "Upper Body Motion and Balance Stability in the Elderly", Proceedings of the RESNA 14th Annual Conference, Kansas City, MO 1991, paper 1.1, pp.1-3.

5. Adapted by Brody, Elaine & Lawton, M. Powell from Langley-Porter Physical Self-Maintenance Scale, Philadelphia Geriatric Center, 5301 Old York Road, Philadelphia, PA 19141.

6. Sabelman, E.E., Gadd, J.J., Kenney, E.E., Merritt, P.M., Winograd, C.H., "Balance Diagnosis Using a Wearable Upper Body Motion Analysis Computer", Proceedings of the RESNA International 1992 Conference, paper 5.5, pp.81-83.

7. Maki, B.E., Holliday, P.J., Topper, A.K., "A Prospective Study of Postural Balance and Risk of Falling in an Ambulatory and Independent Elderly Population", Journal of Gerontology, 49(2), Mar 1994, pp.M72-84.

8. Schenkman, M.L., Berger, R., Riley, P.O., Mann, R.W., Hodge, W.A., "Total Body Dynamics During Rising to Standing from Sitting", Physical Therapy 70, 1990, pp.638-651.

ACKNOWLEDGEMENTS

Supported by VA Rehabilitation R&D Merit Review project E601-2RA.

VA Rehabilitation R&D Center, 3801 Miranda Ave #153, Palo Alto CA 94304 Accelerometric Motion Analysis of Balance-Impaired Elderly Subjects