Postural control in the elderly: An analysis of test-retest and interrater reliability of the COP-COM variable☆☆☆★

Article Outline

• Abstract
• Methods
  • Subjects
  • Data collection procedure
  • Test-retest
  • Interrater study
  • Instrumentation
  • Data analysis
• Results
  • Test-retest
  • Interrater
• Discussion
• Conclusion
• Acknowledgements
• Suppliers
• Appendix: Summary of test-retest coefficients of different variables of postural sway in quiet standing (feet side by side)
• References
• Copyright

Abstract 

Corriveau H, Hébert R, Prince F, Raîche M. Postural control in the elderly: an analysis of test-retest and interrater reliability of the COP-COM variable. Arch Phys Med Rehabil 2001;82:80-5. Objectives: To estimate the test-retest and interrater reliability of the center of pressure-center of mass (COP-COM) variable of postural control in the elderly. Design: The biomechanic variable COP-COM, which represents the distance between the COP and the COM, was determined from 2 AMTI force platforms and 3 OPTOTRAK® position sensors. Measurements were taken in quiet position, double leg stance, and eyes open and eyes closed conditions. Setting: Laboratory environment. Participants: Forty-five healthy subjects, 8 patients with diabetes neuropathy, and 7 stroke survivors, all of whom were at least 60 years old. Interventions: Subjects were evaluated on 2 separate occasions within 7 days by the same evaluator to determine test-retest reliability. Interrater reliability was determined the same day. Main Outcome Measures: The biomechanic variable COP-COM, which represents the distance between the COP and the COM in terms of root mean square. The mean of 4 trials of the COP-COM variable for each condition was used for statistical analysis. Intraclass correlation coefficients (ICCs) were used. Results: The COP-COM variable has good reliability for both the test-retest and interrater studies, but its reliability varies according to the direction of the COP-COM. For the test-retest and interrater studies, the ICC ranged from .89 to .93 in the anteroposterior direction and from .74 to .79 in the mediolateral direction. Conclusion: The equivalence of the test-retest and interrater coefficients obtained suggests that the measurement error of the COP-COM variable is mainly linked to the biologic variability of this measure over a short period of time. Using the mean of 4 trials stabilizes the COP-COM variable enough to be potentially used to evaluate clinical change. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation

Keywords:  Accidental falls, Aged, Cerebrovascular accidents, Diabetic neuropathics, Posture, Rehabilitation

FALLS AND THEIR CONSEQUENCES are a major health problem in elderly people. Of people over the age of 65 years who live in the community, 30% to 35% fall at least once a year, and this proportion increases to 50% by the age of 80 years.¹, ² Furthermore, 40% to 50% of fallers experience 2 or more falls.² Many risk factors have been identified as a potential precursor to falls.², ³ They are classified as intrinsic (those related to the individual) and extrinsic (those related to the environment). Among the intrinsic factors, poor balance is probably the most frequent.², ³, ⁴, ⁵ Because of its importance, a systematic evaluation of balance is needed to identify specific problems.⁶ For this purpose, clinical and laboratory evaluations have been proposed.⁷

In a laboratory setting, posturography is generally used to quantify the amount of postural sway. Measures of postural sway are based on the outcome signals from a force platform or are determined by directly evaluating head, limb, and trunk movements with a video-based system or accelerometry.⁷, ⁸ The measures most commonly used are various parameters of the center of pressure (COP) and the proportion of vertical force on each leg. The COP is a weighted average of all the pressures over the surface area in contact with the ground. The net COP is the integrated control variable of the center of mass (COM).⁹ The displacement of the COP can be used as a measure of stabilizing postural reactions in quiet standing, as well as in expected and unexpected perturbation.

Computerized posturography for the assessment of postural control is becoming more popular in clinical settings; thus, establishing the validity and reliability of the biomechanic variables is critical. Because the reliability coefficient places an upper limit on an instrument's validity, reliability might be studied first.¹⁰ Moreover, clinicians need reliable measures to identify individuals at risk for falling and to assess the effectiveness of balance-intervention programs. An instrument's reliability is related to its precision and how much measurement error it may have. Resolution and accuracy determine the precision of a laboratory instrument according to the supplier specifications. The resolution of the OPTOTRAK® position sensor^a is 0.01mm, and its accuracy in root mean square (RMS) is 0.1mm. The resolution of the AMTI force platform^b is .08μV/(V · N). The accuracy of the COP measured in the present study is 0.2mm. Such precision is satisfactory, but good precision does not necessarily mean that the measure is reliable.

Measurement error includes 3 types of variability: intrasession retest, intersession retest, and interrater. Intrasession reliability is the immediate retest reliability, which is related to the random variability of the measurement per se. Test-retest reliability includes the stability of the phenomena over a longer period of time and the variations related to the procedures. Interrater reliability includes the procedure and adds the variations related to the standardization of the procedure, ie, the extent to which the procedure is applied similarly by different observers. To maintain the integrity of metrological definitions, we use the term interrater, even though in the context of posturography, the term intertechnicians would be more relevant.

To date, few metrologic studies have been conducted to estimate the reliability of various biomechanic measures of balance in quiet standing, particularly for elderly persons with and without disabilities. The reliability of different variables measuring postural control in quiet standing in a laboratory setting is shown in the appendix. Depending on the criterion reliability coefficient recommended, measurements had poor reliability, particularly in regard to test-retest and interrater reliability, and they also have considerable variability. Recently, computerized measurements and feedback systems used to assess and train static and dynamic balance performance have become more widely available.¹¹, ¹², ¹³

The clinical value of some parameters is questionable. For example, the better reliability of vertical reaction force does not really provide a rationale for choosing force reaction measures to evaluate postural control in quiet standing. Weight bearing on 1 leg rather than the other does not necessarily mean that the subjects have problems controlling their balance. The main disadvantage of the force analysis platform providing the COP measure is that it measures the secondary consequences of swaying movements, not the movements themselves.¹⁴ Increasing such COP parameters as path length, area, displacement, or velocity does not necessarily link to postural instability. These variables may be indicative of underlying neural or sensorimotor dysfunction, but COP movements may successfully stabilize the COM. Subjects with high COP velocity values may be quite stable in the sense that the COP does not approach the limits of the base of support, but may require frequent postural corrections to achieve this degree of stability.¹⁵ An estimate of the COM variable is somewhat difficult to obtain and has been crudely estimated, even if it seems to be related to postural imbalance.⁹

To overcome these limitations, it has been suggested that the combined interpretation of COP and COM displacements provides better insight into the assessment of balance than COP and COM taken separately.⁹, ¹⁶, ¹⁷ Thus, a new biomechanic variable (COP-COM) has been proposed; this measure represents the scalar distance at a given time between the COP and COM. The movement of the COP therefore varies with the movement of the COM to keep the COM over the base of support. Thus, the COP must be continuously moving back and forth around the COM to keep the COM over the base of support.⁹, ¹⁸ The major argument for using the COP-COM amplitude is that the difference between the COP and COM is proportional to the horizontal acceleration of the COM during quiet standing.⁹ Furthermore, using the COP-COM amplitude provides an estimate of the efficacy of postural control.⁹, ¹⁹

In previous intrasession variability study,²⁰ researchers established that a mean of 4 trials of the COP-COM variable is necessary to produce acceptable anteroposterior (AP) and mediolateral (M/L) measurement results (intraclass correlation coefficients [ICC] > .94 in AP direction, ICC > .90 in M/L direction). Two sources of error are included in the COP-COM variable: one associated with the COP and the other with the COM. The marker placements used to calculate the COM vary each time because of clothing and the error of each rater in estimating the anatomic points where the markers have to be placed. This manipulation of the markers adds variations related to the procedures and influences test-retest and interrater reliability.

Because the COP-COM variable seems a promising measure, more information is needed about the measurement error before it can be used with confidence. The objective of the present study was to estimate the test-retest reliability and interrater reliability of the COP-COM variable with healthy elderly persons and elderly persons with impairments.

Back to Article Outline

Methods 

Subjects 

Healthy elderly subjects were recruited from a database of people who showed an interest in studies carried out at the research center.²¹ They were then contacted by phone to verify their eligibility and to make an appointment if they were interested in participating. The eligibility criteria were to be at or over the age of 60 years, living independently in the community, and having no neurologic or musculoskeletal impairments (eg, history of stroke, transient ischemia attacks, Parkinson's disease, lower extremity joint replacement). These criteria were verified with a standardized phone questionnaire and updated before the subject was evaluated in the laboratory. Subjects were excluded if they reported visual or somatosensory impairments or reported at least 1 fall in the past year. Additionally, none of the healthy elderly subjects was currently taking medication known to affect balance adversely.

Seven subjects had hemiplegia or paresis after a stroke. At the time of recruitment, they had completed rehabilitative training and were therefore considered stable in their neurologic recovery. They had to be able to maintain independent unsupported stance for 2 minutes and to report no cardiac, respiratory, or medical condition that would interfere with the testing protocol. Eight peripheral diabetic neuropathy (PDN) subjects were also evaluated. Using a polyneuropathy severity index developed by Valk et al²² 3 PDN subjects' polyneuropathy was classified as mild, 4 as moderate, and 1 as severe. All subjects were living in the community and could walk independently. All subjects gave informed consent to participate.

Data collection procedure 

Each subject was evaluated in the laboratory. An experienced physical therapist and an experienced kinesiologist collected the data. Before the study, all measurement procedures were developed and standardized on the basis of feasibility and reproducibility. A training period ensured conformity between the raters on uniformity of the position of the markers on the body and the position of the feet on the platform.

Test-retest 

At the first evaluation, the subjects' characteristics were collected. For the test-retest study, each subject was evaluated on 2 occasions (time 1, time 2) in a laboratory environment by the same evaluator, at the same time of day, under the same conditions, and with an interval between 3 and 7 days according to the availability of the subject. The procedure was the same as in a previous study¹⁸, ²⁰ that determined the reliability of 1 measure of the COP-COM; it is summarized here. Subjects stood quietly on 2 adjacent force platforms. Measurements were made in a double leg stance with feet at pelvis width. Some degree of hip external rotation (max 14°) was permitted to minimize discomfort and constraints on the subject's preferred position.²³ All subjects wore flat-soled shoes usually used for walking or sports activities. To ensure that this position remained constant, tracings were taken of foot placement, and subjects were required to remain within these tracings during all the trials. They were instructed to look straight ahead with their head erect and to maintain balance. Their arms were placed in a comfortable position hanging at their sides. Four successive trials lasting 120 seconds with eyes open (EO) with a rest period of approximately 5 minutes between trials were done. Then after a rest of 10 minutes, all subjects performed the 2-minute standing position 4 times with eyes closed (EC). The same protocol with the same condition order was used the second day of the test-retest evaluation.

Interrater study 

In the present interrater reliability study, both evaluations were done on the same day as the retest session with a reasonable rest time of 30 minutes between the 2 evaluations. By this procedure, we intended to evaluate the influence of marker placement on the reliability of the COP-COM variable. After the rest period, a second evaluator replaced all the markers at the end of the standardized procedure. The EO and EC conditions were repeated in the same order as in the test-retest. Again, all subjects performed the 2-minute standing position 4 times with EO and EC.

Instrumentation 

Anthroprometric tables from Dempster et al²⁴ were used to estimate the total COM. Full details of the anthropometric data and marker placement protocol were previously published¹⁸ and are summarized here. The model consists in 4 segments of the trunk, 1 pelvis, 2 thighs, 2 legs and feet, 2 upper arms, 2 forearms (including hands), and 1 head. Twenty-one infrared light-emitting diodes were attached bilaterally to anatomic landmarks to define a 14-segment model. Three OPTOTRAK^a sensors recorded marker displacement, and ground reaction forces and moments were recorded by 2 AMTI^b force platforms. Both devices were interfaced to an A/D converter with a sampling rate of 20Hz. The natural frequency of the COP-COM in quiet standing is 1.04Hz in the AP direction and 1.14Hz in the M/L direction.⁹ Because the process signal must be sampled at a frequency at least twice (sampling theorem) the highest frequency observed, 20Hz is adequate.²⁵, ²⁶ Data were processed on a Pentium computer using MATLAB® software, version 5.1.^c

Data analysis 

The RMS amplitudes were calculated for the COP-COM variable in both the AP and M/L directions. The RMS is related to the mean of the distance between COP and COM. We used the mean of 4 trials of the COP-COM variable for each condition measurement for the statistical analysis to determine the reliability coefficients. The mean and standard deviation (SD) describe the characteristics of the sample. Test-retest and interrater reliability of COP-COM were estimated using ICCs that compare within-subject variability with between-subject variability.²⁷ The ICC_2,1 described by Shrout and Fleiss²⁸ was chosen because it considers random effects over time for the retest study. The ICC_2,1 was also used for the interrater study because it considers random effects over time and the effect of the evaluators. For each ICC, the 95% confidence interval (CI) was calculated to take sampling variation into account. Paired t tests on the difference in the scores obtained at time 1 and time 2 and for rater A and B (interrater) were used to verify the absence of systematic bias between the measurements. The test of egality of 2 correlations based on the Fisher transformation was done to verify if the ICCs of the healthy and impaired subjects were significantly different.²⁹

Back to Article Outline

Results 

The study sample consisted of 45 healthy elderly persons and 15 elderly persons with impaired postural control from PDN or stroke. Careful examination of data collection and processing procedures showed a technical problem in the record of 1 impaired subject at the first evaluation. Consequently, test-retest reliability was estimated on 59 subjects. Interrater reliability was estimated using all the subjects because there were no technical problems. The subjects' characteristics are in table 1.

Table 1: Subject Characteristics

Values are means (SD) or number.

No significant difference existed in the ICC between healthy and impaired elderly in both test-retest and interrater reliability, as we can see by examining the CIs (which overlap). This lack of significant difference was also confirmed by the Fisher test.

Test-retest 

Table 2 shows RMS scores for each condition (EO, EC) and both directions (AP, M/L), mean differences in scores, t test results, and ICCs.

Table 2: Test-Retest Reliability Analysis of the COP-COM Variable

The ICCs and their 95% CIs showed excellent temporal stability in the AP direction (.89, .90) and moderate stability in the M/L direction (.74, .72). The difference between the ICCs in the 2 directions (AP vs M/L) was statistically significantly different in both EO (p = .01) and EC (p = .003) conditions. In the AP direction in the EO condition, paired t tests on the mean difference between the scores on the 2 measurements showed the presence of a small but statistically significant systematic bias (p < .008). In the M/L direction, there was no systematic bias.

Interrater 

For the interrater study, the ICCs were also not statistically significantly different between the healthy (n = 45) and the impaired subjects (n = 15). Table 3 shows that the ICCs were excellent in the AP direction and good in the M/L direction.

Table 3: Interrater Reliability Analysis of the COP-COM Variable

For the whole population evaluated, the difference between the ICCs in the 2 directions (AP vs M/L) was statistically significantly different in both EO (p = .01) and EC (p = .002) conditions. No systematic bias was observed between rater A and rater B.

There was no statistically significant difference between the test-retest and interrater ICCs in the AP direction (EO, p = .58; EC, p = .33) and M/L direction (EO, p = .61; EC, p = .38). Furthermore, vision did not affect the reliability of the COP-COM variable in either the test-retest (AP, p = .80; M/L, p = .41) or interrater (AP, p = .48; M/L, p = .94) study.

Back to Article Outline

Discussion 

The results obtained indicated good reliability of the COP-COM variable for both the test-retest and interrater studies, but varied significantly according to the direction of the COP-COM (AP vs M/L). It is accepted in the literature that postural measures are split in the AP and M/L directions because of the differences between balance control in the sagittal and frontal planes, both from a biomechanic¹⁸ and a neural control perspective.³⁰ In the present study, the directional difference in the reliability coefficient may be because the subjects stood with their feet side by side at pelvic width. In the context of the inverted pendulum model, stability is generated by ankle torque application in the AP direction. As the precision of the sensory information decreases or the capacity of the motor control system deteriorates, different people move differently in this direction to control their balance. The resulting variability in the AP direction increases the ICC. However, with pelvic width distance between the feet in the frontal plane, there is an intrinsic stability, reducing the COM excursion using different systems rigidity.¹⁸ As a result of both these situations, the ICCs are higher in the AP direction.

The COP-COM variable is more reliable than the other variables used in other studies, even though they used the average of several trials for the data analysis (see appendix). Furthermore, only the COM acceleration variable shows better reliability in the AP than the M/L direction.³¹ It seems logical that the results of these reliability coefficients were in the same direction, because the COP-COM variable is highly correlated with the acceleration of the COM.⁹ Nevertheless, the reliability of the COP-COM variable is better than just the COM acceleration. The COP-COM signal is considered to be like an error signal that the balance control system senses.⁹ However, the choice of the COP-COM outcome variable used in the present study may be an issue in the partial failure to get a good ICC in the M/L direction. When considering direction-specific instability, it may be important to allow for differences between the sagittal and frontal planes. Nevertheless, the frontal plane seems important to discriminate between fallers and nonfallers. Maki³² investigated several laboratory measures and suggested that COP velocity in quiet stance in the M/L direction is the most predictive measure of falls in the elderly. So, even with the noise in the measure in the M/L direction, this variable would distinguish between fallers and nonfallers.

Statistically significant systematic biases between the 2 measures were found only in the AP direction with EO and EC conditions in the test-retest study. This bias is not relevant for 2 reasons. First, that this bias is negative implies that the second measure is more unstable. However, the interrater reliability study does not contain a bias but it does contain a test-retest procedure that did not confirm this instability. Therefore, it cannot be fatigue that increases the amplitude of displacement (COP-COM). Second, the difference between T₁ and T₂ is .005cm, which is less than the minimal metrically detectable change (MMDC) that can be found with the COP-COM variable. This MMDC is based on the measurement error associated with an outcome instrument and is defined as the 95% CI of the standard error of measurement (±1.96 S_e).³³ The MMDC for the AP direction with EO was previously calculated as ±.01cm.²⁰ Consequently, the bias obtained is below the measurement error (.01cm) and is not metrically significant. Furthermore, using a paired t test with a sample size of 59 and a power of 80%, it is possible to detect a small, significant difference of 35% of the SD.³⁴ In the present test-retest study, the SD of the mean difference is .014cm; consequently, this test is very powerful and it is possible to detect a significant difference of .005cm.

The results of the present study showed unexpectedly that the interrater reliability coefficients were equivalent to the test-retest reliability. We studied both test-retest and interrater reliabilities because both measures are clinically important. Test-retest reliability is crucial in evaluating the effects of an intervention on postural control, and interrater reliability gives the stability of the measure when 2 different raters evaluate the same patient. Theoretically, in addition to the variability attributable to the measure and to the subject (found in test-retest studies), the difference in raters is expected to influence the result.²⁷ Our results showed that the procedure is easily reproducible. The reliability found for the COP-COM variable was not influenced by inaccuracies in the experimental setup or recording system. Furthermore, an inaccuracy in marker placement over anatomic landmarks or clothes, depending on the location of the markers, slightly influences the reliability of the COP-COM variable. Moreover, in the test-retest, the second session was conducted a maximum of 1 week after the first evaluation. Using this procedure, we assumed that the postural control of our population was stable over that period of time. The stroke subjects were in the chronic stage and their rehabilitation was finished. Therefore, they were supposed to be clinically stable.¹³ PDN is a chronic condition and is not supposed to change from day to day. The equivalence of the test-retest and the interrater coefficients obtained suggests that the measurement error of the COP-COM variable may be linked to the biologic variability of this measure over a short period of time.

Back to Article Outline

Conclusion 

Postural control is a dynamic phenomenon that changes from day to day. Using the mean of 4 trials of the COP-COM variable stabilizes the COP-COM variable enough to get a reliable measure over a short period of time. An inverted pendulum model involving simultaneous COM and COP records is required to understand how the trajectory and acceleration of the COM in both AP and M/L directions is controlled to maintain postural balance. Using the COP-COM variable provides the opportunity to measure reliably postural stability in quiet standing.

Back to Article Outline

Acknowledgements 

We thank Lise Trottier for the statistical analysis.

Back to Article Outline

Suppliers 

a. Northern Digital Inc, 103 Randall Dr, Waterloo, Ont N2L 1C5, Canada.

b. Advanced Mechanical Technology Inc, 176 Waltham St, Watertown, MA 02172.

c. Math Works Inc, 3 Apple Hill Dr, Natick, MA 01760.

Back to Article Outline

Appendix: Summary of test-retest coefficients of different variables of postural sway in quiet standing (feet side by side) 

Abbreviations: COG, center of gravity.

Back to Article Outline

References 

1. Vellas BJ, Wayne SJ, Garry PJ, Baumgartner RN. A two-year longitudinal study of falls in 482 community-dwelling elderly adults. J Gerontol A Biol Sci Med Sci. 1998;53:M264–M274
   • View In Article
2. O'Loughlin J, Robitaille Y, Boivin JF, Suissa S. Incidence of and risk factor for injurious falls among the community-dwelling elderly. Am J Epidemiol. 1993;137:342–354
   • View In Article
   • MEDLINE
3. Myers AH, Young Y, Langlois JA. Prevention of falls in the elderly. Bone. 1996;18(Suppl):87S–101S
   • View In Article
   • MEDLINE
4. Shumway-Cook A, Baldwin M, Polissar NL, Gruber W. Predicting the probability of falls in community-dwelling older adults. Phys Ther. 1997;77:812–819
   • View In Article
   • MEDLINE
5. Tideiksaar R. Preventing falls: how to identify risk factors, reduce complications. Geriatrics. 1996;51:43–46 49-50, 53, quiz 54-5
   • View In Article
   • MEDLINE
6. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med. 1988;319:1701–1707
   • View In Article
   • MEDLINE
   • CrossRef
7. Horak FB. Clinical assessment of balance disorders. Gait Posture. 1997;6:76–84
   • View In Article
   • CrossRef
8. Prieto TE, Myklebust JB, Myklebust BM. Characterization and modelling postural steadiness in the elderly: a review. IEEE Trans Rehabil Eng. 1993;1:26–34
   • View In Article
   • CrossRef
9. Winter DA. A.B.C. of balance during standing and walking. Waterloo (Ont): Waterloo Biomechanics; 1995;
   • View In Article
10. Pedhazur E, Schmelkin LP. Measurement, design, and analysis. Broadway (NJ): Lawrence Erlbaum; 1991;
    • View In Article
11. Brouwer B, Culham EG, Liston RA, Grant T. Normal variability of postural measures: complications for the reliability of relative balance performance outcomes. Scand J Rehabil Med. 1998;30:131–137
    • View In Article
    • MEDLINE
    • CrossRef
12. Ford-Smith CD, Wyman JF, Elswich RK, Fernandez T, Newton RA. Test-retest reliability of the Sensory Organisation Test in noninstitutionalized older adults. Arch Phys Med Rehabil. 1995;76:77–81
    • View In Article
    • Abstract
    • Full-Text PDF (2126 KB)
    • CrossRef
13. Liston RAL, Brouwer BJ. Reliability and validity of measures obtained from stroke patients using Balance Master. Arch Phys Med Rehabil. 1996;77:425–430
    • View In Article
    • Abstract
    • Full-Text PDF (696 KB)
    • CrossRef
14. Stevens DL, Tomlinson GE. Measurement of human postural sway. Proc R Soc Med. 1971;64:653–654
    • View In Article
    • MEDLINE
15. Maki BE, Holliday PJ, Fernie GR. Ageing and postural control: a comparison of spontaneous- and induced-sway balance tests. J Am Geriatr Soc. 1990;38:1–9
    • View In Article
    • MEDLINE
16. Murray MP, Seireg A, Scholz RC. Centre of gravity, centre of pressure and supportive forces during human activities. J Appl Physiol. 1967;23:831–838
    • View In Article
17. Spaepen AJ, Peeraer L, Willems EJ. Center of gravity and center of pressure in stabilometric studies. Agressologie. 1979;20:117–118 b
    • View In Article
18. Winter DA, Patla AE, Prince F, Ishac M, Gielo-Perckak K. Stiffness control of balance in quiet standing. J Neurophysiol. 1998;80:1211–1221
    • View In Article
    • MEDLINE
19. Benvenuti F, Mecaaci R, Gineprari I, Bandinelli S, Benvenuti E, Ferrucci L, et al.  Kinematic characteristics of standing disequilibrium: reliability and validity of a posturographic protocol. Arch Phys Med Rehabil. 1999;80:278–287
    • View In Article
    • Abstract
    • Full-Text PDF (1255 KB)
    • CrossRef
20. Corriveau H, Hébert R, Prince F, Raîche M. Intrasession reliability of the “Center of Pressure minus Center of Mass” variable of postural control in the healthy elderly. Arch Phys Med Rehabil. 2000;81:45–48
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (63 KB)
    • CrossRef
21. Desrosiers J, Hébert R, Bravo G, Dutil E. Hand sensitivity of healthy older people. J Am Geriatr Soc. 1996;44:974–978
    • View In Article
    • MEDLINE
22. Valk GD, Nauto JJP, Striners RLM, Bertelsman FW. Clinical examination versus neurophysiological examination in the diagnosis of diabetic polyneuropathy. Diabetic Med. 1992;9:716–721
    • View In Article
23. McIlroy WE, Maki BE. Preferred placement of the feet during quiet stance: development of standardized foot placement for balance testing. Clin Biomech. 1997;12:66–70
    • View In Article
24. Dempster WT, Gabel WC, Felts WJL. The anthropometry of manual work space for seated subjects. Am J Phys Anthropol. 1959;17:289–317
    • View In Article
    • MEDLINE
    • CrossRef
25. Nigg BM, Herzog W. Biomechanics of the musculo-skeletal system. New York: Wiley; 1994;
    • View In Article
26. Winter DA. Biomechanics and motor control of human movement. Chichester: Wiley; 1990;
    • View In Article
27. Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. 2nd ed. Oxford: Oxford Med Pub; 1995;
    • View In Article
28. Shrout PE, Fleiss JL. Intraclass correlation: uses in assessing rater reliability. Psychol Bull. 1979;86:420–428
    • View In Article
    • CrossRef
29. Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. Boston: PWS-Kent; 1988;
    • View In Article
30. Geurts CH, Nienhuis B, Mulder TW. Intrasubject variability of selected force-platform parameters in the quantification of postural control. Arch Phys Med Rehabil. 1993;74:1144–1150
    • View In Article
    • MEDLINE
31. Moe-Nilssen R. Test-retest reliability of trunk accelerometry during standing and walking. Arch Phys Med Rehabil. 1998;79:1377–1385
    • View In Article
    • Abstract
    • Full-Text PDF (800 KB)
    • CrossRef
32. Maki BE. Biomechanical approach to quantifying anticipatory postural adjustments in the elderly. Med Biol Eng Comput. 1993;21:355–362
    • View In Article
33. Hébert R, Spiegelhalter DJ, Brayne C. Setting the minimal metrically detectable change on disability rating scales. Arch Phys Med Rehabil. 1997;78:1305–1308
    • View In Article
    • Abstract
    • Full-Text PDF (665 KB)
    • CrossRef
34. Machin D, Campbell M. Statistical tables for the design of clinical trials. Oxford: Blackwell Scientific; 1987;
    • View In Article
35. Goldie PA, Bach TM, Evans OM. Force platform measures for evaluating postural control: reliability and validity. Arch Phys Med Rehabil. 1989;70:510–517
    • View In Article
    • MEDLINE
36. Haas BM, Whitmarsh TE. Inter- and intra-tester reliability of the Balance Performance Monitor in a non-patient population. Physiother Res Int. 1998;3:58–68
    • View In Article
    • MEDLINE
    • CrossRef
37. Kamen G, Patten C, Duke Du C, Sison S. An accelerometry-based system for the assessment of balance and postural sway. Gerontology. 1998;44:40–45
    • View In Article
    • MEDLINE
    • CrossRef