- Email: [email protected]
Kinematic analyses of the 180° standing turn: effects of age on strategies adopted by healthy young and older women
Gait & Posture 22 (2005) 119–125 www.elsevier.com/locate/gaitpost
Kinematic analyses of the 1808 standing turn: effects of age on strategies adopted by healthy young and older women Peggy Meinhart-Shibataa,*, Michelle Kramera, James A. Ashton-Millera,b,c, Carol Persadd a
Biomechanics Research Laboratory, Department of Mechanical Engineering, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA b Department of Biomedical Engineering, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA c Institute of Gerontology, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA d Department of Psychiatry, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA Accepted 2 August 2004
Abstract Standing turns are associated with an increased risk for falls and fall-related injuries in the elderly. The purpose of this study was to test the (null) hypothesis that age has no effect on the kinematics of the 1808 turn. Ten young and 10 older healthy women were asked to complete a series of 1808 turns in a standing posture after picking up a light bowl with both hands. Foot–ground reactions, insole pressures and body segment kinematics were recorded in 62 trials at 100 Hz. Turning strategies were analyzed for effects of both age and turn direction on linear and angular foot kinematics, as well as pelvic axial rotation. The older women (OW) used a preparatory stepping strategy more often (170%, p < 0.002), and employed a lower average pelvic rotation rate (21%, p < 0.011) than the younger controls. The minimum foot separation distance for OW was less in their non-preferred than in their preferred turn direction (29%, p < 0.038), thereby increasing their risk of foot– foot interference and falling when turning in their non-preferred direction. The older women were more variable in their turn execution, particularly in minimum foot separation distance (55%, p < 0.022) and the maximum rate of pelvic rotation (82%, p < 0.035). Despite the fact that these healthy older women were careful to employ a preparatory stepping strategy and slower average rotational velocities, they were also more variable in their turn execution than the young. # 2004 Elsevier B.V. All rights reserved. Keywords: Turning; Women; Aging; Falls; Kinematics
1. Introduction As the geriatric population continues to rise, the consequences of falls and fall-related injuries carry increasing socioeconomic significance. One-third of community-dwelling elderly fall each year; 5% of these experience fracture or injury requiring hospitalization [1] and expenses now exceed $20 billion annually [2]. Moreover, in 1999, falls caused 31% of all unintentional injury deaths [3]. The most common activities being performed at the time of fall-related hip fracture are walking * Corresponding author. Tel.: +1 734 936 0368; fax: +1 734 763 9332. E-mail address: [email protected] (P. Meinhart-Shibata). 0966-6362/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2004.08.005
forward and turning around. Although 40% of hip fractures occur during walking and 18% occur during turning [4], a fall while turning is 7.9 times more likely to result in hip fracture than a fall while walking straight [5]. Hence, turning carries a significant risk for hip fracture. Standing turn performance has been used to delineate elderly fallers from non-fallers. Elderly fallers take 20% longer to turn 1808 than non-fallers [6] and frail elderly fallers take six more steps to turn 3608 than non-fallers [7]. However, the reason why fallers choose to turn more slowly, and take more steps, is unknown. Possible factors include vestibular impairments [8], difficulty handling whole-body angular momentum (in the same way that linear momentum can be problematic [9,10]), or fear of foot–ground or foot–
120
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
foot interference. Inadvertent foot–foot interference could result from the increase in variability in step width and step length observed with increasing age [11]. Because of its value in predicting potential fall risk, standing turn performance has been included in routine clinical assessments of elderly by measuring the time and/or number of steps required to turn [12]. For example, 3608 turns are incorporated into the ‘Performance-Oriented Mobility Assessment’ (POMA) [13], the ‘Physical Performance’ [14], and the ‘Berg Balance’ [15] test batteries to evaluate functional balance. Furthermore, turning was a component of a mobility test that successfully predicted health outcomes in older hip fracture patients [16]. While the etiology of most falls is multifactorial [17], these activityoriented evaluations are useful for evaluating subjects who are more likely to fall while turning [18]. Despite the positive association between turning difficulty and increased risk of fall-related hip fractures, we are unaware of biomechanical analyses of standing turns. Our main goals were to characterize standing turns using kinematics and to determine whether differences in turning execution exist between healthy young and elderly women. The primary measures evaluated were self-selected stepping strategies, separation distance between the feet during the turn, and turning rate as measured by average pelvic rotational velocity about the vertical. Secondary measures included the time and number of steps to turn, foot position during double support, and variability in foot separation distance and pelvic velocity. These measures are related to potential falling mechanisms such as tripping over one’s feet, unintended aberrant foot placement, and/or unexpected angular momentum due to variability in execution. Older women were studied because they have twice the risk for fall-related hip fractures as men [19]. The hypothesis that no age differences exist in the kinematics of 1808 turns was tested with the expectation that dissimilarities would be revealed and this hypothesis would ultimately be rejected.
2. Methods Ten healthy young women (YW) aged 21.8 1.99 years and 10 healthy, community-dwelling older women (OW) aged 72.5 5.82 years participated. YW were University of Michigan students; OW were contacted through the University of Michigan Institute of Gerontology. The YW and OW heights were 160.52 6.83 and 162.01 6.70 cm, respectively. A nurse practitioner examined individuals in the OW group under supervision of a geriatrician–physician. All denied having fallen in the preceding 2 years, were not fearful of falling, and had no complaints of imbalance. Subjects gave written consent to participate in this study as approved by the governing Institutional Review Board. All subjects wore comfortably fitted athletic walking shoes (Model 800, New Balance, Inc.) and were outfitted in an overhead harness system to avoid injury in the case of a
Fig. 1. Overhead view of experimental setup showing tables (a, b) and bilaterally symmetric initial foot position. Preparatory stepping pattern and foot trajectories for a turn to the right are shown (cross-hatched pattern: 1L = first step, left foot; 2-R = second step, right foot; etc.). Left and right arrows denote darkened and illuminated cues, respectively.
fall. Because a directional preference for spontaneous turns has been demonstrated [20], we first asked subjects to turn in a circle. We designated the turn direction the subject chose as the ‘preferred direction’ (PD), the contralateral direction was designated as the ‘non-preferred direction’ (NPD). Subjects performed 62 trials, each initiated by an illuminated arrow pointing to the right or left. Subjects used both hands to pick up a light-weight bowl from a waisthigh table, executed a 1808 turn in the indicated direction, and placed the bowl on a table 2 m behind their starting position (Fig. 1). Subjects carried the bowl because 1808 turns are often executed in confined spaces, like a kitchen, to move objects from one place to another. Subjects were instructed to perform trials at a comfortable speed as if in their kitchen preparing a meal. Forty-eight trials consisted of turns to the right or left. Fourteen trials were turns with a sudden direction change upon an audio tone. The results of these latter trials, presented intermittently throughout the experiment, are not presented here. Trials were organized in a pre-determined randomized order unknown to subjects. The time to complete a trial was measured digitally as the time between the visual cue appearance and interference of a light beam as the bowl was placed on the second table. Actual turn initiation was identified by recording when the normal force from one of two forceplates (Model OR6, AMTI, Watertown, MA), situated beneath each foot, changed by 5% of the output extant during quiet standing. The number of steps for each trial was established by insole pressure sensors (thickness 0.18 mm; F-Scan, Tekscan, Inc., South Boston, MA). The beginning and ending of each step was determined using a force threshold of 15% body weight. Any foot–foot interference was identified by electrical contact between strips of thin adhesive aluminum, covering
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
the medial aspect of each shoe causing a voltage change to be recorded. Body segment kinematics were measured at 100 Hz via an Optotrak (Northern Digital Corporation, Waterloo, Ontario) dual 3020 optoelectronic camera system. Groups of three coplanar but non-collinear infrared-emitting diodes (IREDS) formed six rigid triads and established local coordinate systems. Triads were fixed to both medial and lateral aspects of the upper last of each shoe and the most anterior prominence of each iliac crest. With subjects seated, the locations of the triads affixed to the shoes, and points separated by 1 cm along the margin of each shoe sole, were recorded by Optotrak hand digitization establishing the relative position between the sole and corresponding shoe markers. Using this information and transformation methods similar to Startzell and Cavanagh [21], exact shoe sole placement during double support and trajectory during single support was reconstructed for all frames (Fig. 1). Kinematic data also permitted the quantification of pelvic rotational behavior. Pelvic position was represented by the origin and direction of a vector perpendicular to and bisecting the line joining the origins of the two local coordinate systems at each anterior iliac crest. This averaged coordinate does not provide a specific measure such as center of mass, but can identify the change in linear and rotational position of the pelvic region. Unfortunately, one YW and OW held the bowl in a manner that frequently obstructed the markers located on the iliac crests making it impossible to obtain sufficient pelvic data. Therefore, pelvis kinematic results (only) were limited to nine YW and nine OW. Kinetic and kinematic data were filtered using a fourthorder, low-pass, Butterworth filter with break point at 8 Hz. The analysis of the 1808 turn established a general turning strategy for each group. One quantifiable measure of turning behavior was the minimum (Euclidean) foot separation distance at each sample interval. Comparison of all frames revealed the overall minimum foot separation distance and how close the feet came to contacting one another during each trial. The variability in minimum foot separation distance was later analyzed to estimate accuracy of foot trajectory and placement. While variability may not kinematically describe a standing turn, any group differences yield insight into how age affects turning execution. Another foot kinematic measure was the ‘included foot angle’, or the angle between two vectors representing the direction of each foot from heel to toe during double support. The maximum included foot angle reflected the maximum included angle observed over all trials for each subject. Rotations were calculated with respect to the global yaxis (Fig. 1). The absolute value of rotations was used to combine turns in both directions. To measure the time to complete only the turning portion of the task, the time between movement initiation and pelvic rotational changes of 458, 908, 1358, and 1808 was recorded. Rotation data were numerically differentiated to obtain the angular velocity for each trial. Maximum pelvic velocity, a measure reflecting
121
the maximum angular momentum generated, was analyzed along with average pelvic rotational velocity. By averaging over all trials, the outcome measure (minimum foot separation distance, average and maximum rate of rotation, number of steps, and total time) mean was determined for each of the 10 subjects (n = 9 for rotational measures). The kurtosis and skewness of these samples were calculated to test for normality. Age effects and withinsubject differences in PD and NPD turns were investigated. Mann–Whitney and Wilcoxon Ranks non-parametric tests were utilized for analyses involving non-normal distributions. Unless otherwise stated, all other analyses consisted of two-sample or paired t-tests as appropriate. Identical methods were used to investigate effects on outcome measure variability after calculating the standard deviation of each parameter for each subject and grouping according to age. To further examine minimum foot separation distance and interactions among age, turn direction, and practice, individual trials were used in a repeated-measures analysis of variance (rm-Anova). Approximately 48 trials for each subject were used, resulting in 947 samples for this analysis. A value of p < 0.05 was considered significant for all results.
3. Results Age effects were evident in the preparatory stepping strategy used by OW, whereby a small first step was commonly taken in the direction of the turn by the contralateral foot (Fig. 1). OW used this strategy in 65% of the trials, whereas YW used it in only 24% of trials. For turns in the NPD this age difference was accentuated (Table 1). No effect of turn direction on stepping strategy was evident in either group. Not one instance of foot–foot contact was recorded in these healthy women. Overall, YW and OW revealed similar minimum foot separation distances (Table 1). However, when PD and NPD turns were analyzed within each age group, there was no direction effect in the YW, but OW brought their feet closer together during NPD turns (Table 1). The rm-ANOVA showed that turn direction, not age, was the only factor affecting this parameter. The borderline interaction between age and turn direction reflected that the OW had a greater tendency to bring their feet closer together during NPD turns than did the YW (Table 2). Across all trials, age affected the average pelvic rotational velocity: the YW and OW turned at an average velocity of 85.98/s and 67.98/s, respectively (Table 1). The turn direction did not influence the average pelvic rotational velocity in either group. Finally, there was no age or direction effect on maximum pelvic rotational velocity. Secondary measures analyzed included the number of steps and time required to complete the task, maximum included foot angle during double support, and variability in kinematic measures. Overall, the OW and YW showed no
122
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
Table 1 Mean (S.D.) primary and secondary outcome measures are shown for young and elderly women (YW, OW, respectively) in the preferred direction (PD), nonpreferred direction (NPD) by group, and for the trials combined for direction (ALL) YW Primary measures Proportion of trials with preparatory strategy
PD NPD ALL
0.24 (0.29) 0.24 (0.35) 0.24 (0.24)
OW 0.56 (0.37)a 0.72 (0.38)b 0.65 (0.25)c
Average minimum foot separation distance (mm)
PD NPD ALL
34.1 (5.1) 31.9 (13.9) 33.0 (13.4)
49.2 (17.6)d 34.9 (13.1) 42.1 (12.5)
Average pelvic rotational velocity (8/s)
PD NPD ALL
87.2 (13.2) 84.4 (10.9) 85.9 (11.6)
66.4 (16.1)e 70.2 (15.4)f 67.9 (14.9)g
Maximum pelvic rotational velocity (8/s)
PD NPD ALL
167.1 (21.8) 164.0 (21.6) 165.5 (19.7)
Secondary measures Number of steps
149.1 (26.7) 142.8 (47.2) 148.0 (26.0)
PD NPD ALL
4.01 (0.61) 3.85 (0.78) 3.93 (0.67)
4.53 (1.02) 4.37 (1.02) 4.45 (0.89)
Total time (s)
PD NPD ALL
3.99 (0.63) 3.95 (0.58) 3.97 (0.60)
4.46 (1.25) 4.37 (1.18) 4.40 (1.19)
Maximum included foot angle (8)
PD NPD ALL
150.7 (17.4) 149.2 (16.6) 155.4 (17.7)
125.1 (14.8)h 136.6 (19.1) 138.0 (16.9)i
Variability of minimum foot separation distance (mm)
PD NPD ALL
11.8 (5.0) 10.9 (4.4) 12.7 (4.6)
16.0 (5.3) 17.4 (7.6)j 19.7 (7.6)k
Variability of maximum pelvic rotational velocity (o/s)
PD NPD ALL
19.6 (3.7) 19.8 (10.1) 21.8 (5.7)
39.7 (30.4) 29.8 (15.3) 39.7 (22.6)l
Quantities in italics represent samples with non-normal distributions. a p = 0.041 for age. b p = 0.024 for age. c p = 0.002 for age. d p = 0.038 for turn direction. e p = 0.009 for age. f p = 0.037 for age. g p = 0.011 for age. h p = 002 for age. i p = 0.037 for age. j p = 0.030 for age. k p = 0.022 for age. l p = 0.035 for age.
difference in the number of steps or time required to complete a trial (Table 1). No effect of turn direction on either measure was revealed for either group. The YW displayed larger maximum included foot angles than the OW (Table 1). The maximum included foot angle was also unaffected by turn direction. The time history of pelvic rotation from turn initiation to completion of 1808 rotation for each subject’s first PD turn was plotted by age group (Fig. 2). The YW completed this rotation within 1.5–3.0 s with two or three steps. OW were more likely to take three or more steps, following their
tendency to use smaller included foot angles than the YW. All but one OW curve ended beyond the average time for YW, and for OW the times for the rotations were more variable, with a range of 1.5–6 s. We note two interesting facts. First, the two longest-duration turns were performed by the two oldest subjects. Their respective ages were 83 and 81 years; the next oldest subject was 73 years. Second, the shortest duration trial for the OW is indicative of one OW behaving generally like the YW. Similar results (not shown) were obtained for initial NPD turns. The time required to turn was examined for all trials, and age-associated time
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
123
Table 2 Repeated measures ANOVA results for minimum foot separation distance data F
p-value
Main effect Age Turn direction Trial
2.231 6.146 1.066
0.153 0.023 0.385
Factor interaction Age turn direction Age trial Turn direction trial Age turn direction trial
3.704 1.016 0.423 0.857
0.070 0.441 0.985 0.637
Note: ‘Trial’ factor represents practice effect.
differences increased as the turn neared completion, although not significantly (Fig. 3). The kinematic variability was the final measure examined. The OW were more variable in their minimum foot separation distance than were the YW (Table 1, Fig. 4). Turn direction did not affect the variability of minimum foot separation distance for either group. There was no effect of age or turn direction on the variability of average rotational velocity. However, the OW were more variable than the YW in their maximum pelvic rotational velocity (Table 1).
4. Discussion We have investigated the effect of age on turning behavior by comparing the kinematics of two groups of healthy women. Age-related differences were revealed in stepping strategy, minimum foot separation distance, and
Fig. 2. Pelvic rotation time history with respect to the global y-axis (Fig. 1) for the first PD trial for each subject. Symbols signify completion of successive steps according to legend at right. Arrows represent the mean time for each group.
Fig. 3. Average time required for subjects to complete quartiles of total pelvic rotational change for all turns. Bars denote corresponding S.D.
average pelvic rotational velocity, indicating a tendency by the OW to be more cautious when turning. Despite their carefulness, the OW exhibited more variability in minimum foot separation distance and maximum pelvic rotational velocity. Strong evidence for caution in the OW is their increased use of the preparatory stepping strategy. This strategy reduces the angle of lower extremity external rotation required for subsequent steps, as confirmed by smaller included foot angles exhibited by the OW (Table 1). The preparatory step also permits a larger anterior/posterior dimension of their bipedal base of support than would result if one step with the ipsilateral foot were used to achieve the same absolute angle of rotation. The OW also used smaller included foot angles in order to achieve their turn. Without measuring maximum ranges of hip motion, we cannot determine whether the reduced angles
Fig. 4. Minimum foot separation distance frequency histograms. Data grouped according to age (YW, OW) and turn direction (PD, NPD).
124
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
can be attributed to age-differences in the motor control of turning, or a decline in hip range of motion with age. However, hip rotational range of motion in both men and women continues to decrease even after the age of 70 years [22]. Therefore, age-related limitations in hip external rotation could be an explanation for the smaller included foot angles measured in these OW. An additional age effect on turn execution was the trend toward the OW maintaining increased minimum foot separation distances. This parameter reflected age effects in subjects’ abilities to handle turns in opposing directions. For example, the feet of the OW came significantly closer in the NPD turns, thereby increasing their risk for foot–foot interference and for a trip or stumble during a NPD turn. More evidence for OW conservatism is their use of reduced average pelvic rotational velocities. This likely resulted in less whole-body angular momentum to arrest at the end of the turn, assuming that the torso rotates en-bloc. These decreased comfortable turning speeds parallel the decrease in comfortable walking speed with age in OW, for example, as shown by Chen et al. [23]. Moreover, they are consistent with the reticence of the elderly to generate high velocities and momentum in challenging tasks, such as reaching [10]. Older adults use different strategies to develop momentum [24], and may generate smaller momenta because they are either unable to develop lower extremity joint torques as rapidly [25] or because they have more difficulty arresting movements than younger adults [26]. To establish the validity of our results, we compared the time required for OW to turn 1808 (2.8 s from turn initiation) with reported values. The present OW were faster than published elderly non-faller times of 4.8 and 8.2 s for 1808 and 3608, respectively [6,7]. Differences may be explained by disparities in subject group composition and turn time measurement. For example, Imms and Edholm measured the time to turn 1808 as the time taken to pass through a beam, turn around, and pass back through the beam. Their measurement therefore includes some linear motion. Also, their average age for non-fallers was 77.4 years compared to our average of 72.5 years [6]. Lipsitz et al. [7] subject group possibly took longer because they were over a decade older and more frail than our subjects, including those with cognitive impairment, assistive device use, or some degree of dependency. In an earlier study of turning behavior, three groups were examined: young subjects with no difficulty turning (YNDT), elderly subjects with no difficulty turning (ENDT), and elderly subjects with difficulty turning (EDT) as identified by a self-reported questionnaire (Thigpen et al. [12]). For just the turning portion of the Timed ‘Up and Go’ Test [28], the ENDT took more time and steps than the YNDT. While some overlap was present, the turning characteristics and strategies of the ENDT group were more variable than in the YNDT group [12]. Our results corroborate these findings. Our results indicate that even physically active and welladapted OW showed differences in the kinematics of 1808
turn execution. We expect that frail elderly would exhibit even more caution in their turning strategies. But the question remains why turning is associated with so many hip fractures. This study cannot answer that question. But, despite being more cautious, the minimum foot separation distance and maximum pelvic rotational velocity of the OW reflected greater average variability than in the YW. Increased step variability is a hallmark of fallers [27], and it may underlie difficulty with turns. The present study has a number of limitations. It is unknown whether the present age differences extend to turns through angles other than 1808. One can also not know whether these results may be extrapolated to men. But, smaller hip ranges of motion in males [22] may alter their turning strategy even more with age. Turns in this study were carried out on industrial carpeting. It is unknown if turning strategies would differ on hard surfaces or on deeper pile carpet. Finally, the small sample sizes used limited the statistical power available. Although we determined that YW and OW turned at different speeds, the difficulty of their respective turns was comparable because subjects completed the task at their comfortable speed. Given that Fitts demonstrated that movement accuracy generally decreases with increasing movement speed, [29] more kinematic variability might have been found had we asked the subjects to turn in a hurry. This may be worth examining in the future, since hurrying is known to be associated with a higher risk for falls [30].
Acknowledgements The support of National Institute on Aging Grants AG 10542 and AG 08808, the National Science Foundation, and the Rackham School of Graduate Studies, University of Michigan is acknowledged. The authors are grateful for the assistance of Diane Scarpace for subject screening under the supervision of Neil B. Alexander, and Janet Kemp, Martin Stenzel, Adriana Figueroa, and Alicia Gehle for technical support.
References [1] Tinetti M, Speechley M, Gintar SF. Risk factors for falls among elderly persons living in the community. N Engl J Med 1988;319:1701–6. [2] Englander F, Hodson TJ, Terregrossa RA. Economic dimensions of slip and fall injuries. J Forensic Sci 1996;41(5):733–6. [3] Stevens J. Falls among older adults: public health impact and prevention strategies. Generations 2002;26(4):7–13. [4] Nevitt M, Cummings SR. Type of fall and the risk of hip and wrist fractures: the study of osteoporotic fractures. J Am Geriatr Soc 1993;41(11):1226–34. [5] Cumming R, Klineberg RJ. Fall frequency and characteristics and the risk of hip fractures. J Am Geriatr Soc 1994;42(7):774–8. [6] Imms F, Edholm OG. Studies of gait and mobility in the elderly. Age Ageing 1981;10:147–56.
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125 [7] Lipsitz L, Johsson PV, Kelley MM, Koestner JS. Causes and correlates of recurrent falls in ambulatory frail elderly. J Gerontol 1991;46(4):M114–22. [8] Sloane P, Baloh RW, Honrubia V. The vestibular system in the elderly: clinical implications. Am J Otolaryngol 1989;10(6):422–9. [9] Pai Y, Naughton BJ, Chang RW, Rogers MW. Control of body centre of mass momentum during sit-to-stand among young and elderly adults. Gait Posture 1994;2:109–16. [10] Kozak K, Ashton-Miller JA, Alexander NB. The effect of age and movement speed on maximum forward reach from an elevated surface: a study in healthy women. Clin Biomech 2003;18:190–6. [11] Grabiner P, Biswas ST, Grabiner MD. Age-related changes in spatial and temporal gait variables. Arch Phys Med Rehabil 2001;82: 31–5. [12] Thigpen M, Light KE, Creel GL, Flynn SM. Turning difficulty characteristics of adults aged 65 years or older. Phy Ther 2000;80(12):1174–87. [13] Tinetti M. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc 1986;34:119–26. [14] Reuben D, Siu AL. An objective measure of physical function of elderly outpatients. The physical performance test. J Am Geriatr Soc 1990;38(10):1105–12. [15] Berg K, Wood-Dauphinee SL, Williams JT, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Can 1989;41:304–11. [16] Fox K, Hawkes WG, Hebel JR, Felsenthal G, Clark M, Zimmerman SI, Kenzora JE, Magaziner J. Mobility after hip fracture predicts health outcomes. J Am Geriatr Soc 1998;46:169–73. [17] Tinetti M. Preventing falls in elderly persons. N Engl J Med 2003;348(1):42–9. [18] Topper A, Maki BE, Holliday PJ. Are activity-based assessments of balance and gait in the elderly predictive of risk of falling and/or type of fall? J Am Geriatr Soc 1993;41(5):479–87.
125
[19] DeGoede K, Ashton-Miller JA, Schultz AB. Fall-related upper body injuries in the older adult: a review of the biomechanical issues. J Biomech 2003;36:1043–53. [20] Yazgan M, Leckman JF, Wexler BE. A direct observational measure of whole body turning bias. Cortex 1996;32:173–6. [21] Startzell J, Cavanagh PR. A three-dimensional approach to the calculation of foot clearance during locomotion. Hum Movement Sci 1999;18:603–11. [22] James B, Parker AW. Active and passive mobility of lower limb joints in elderly men and women. Am J Phys Med Rehabil 1989;68(4):162– 7. [23] Chen H, Ashton-Miller JA, Alexander NB, Schultz AB. Effects of age and available response time on ability to step over an obstacle. J Gerontol 1994;49(5):M227–33. [24] Polcyn A, Lipsitz LA, Kerrigan DC, Collins JJ. Age-related changes in the initiation of gait: degradation of central mechanisms for momentum generation. Arch Phys Med Rehabil 1998;79(12):1582–9. [25] Thelen DG, Schultz AB, Alexander NB, Ashton-Miller JA. Effects of age on rapid ankle torque development. J Gerontol 1996;51A(5):M226–32. [26] Vrtunski P, Patterson MB, Hill GO. Factor analysis of choice reaction time in young and elderly subjects. Percept Mot Skills 1984;59:659– 76. [27] Mbourou G, Lajoie Y, Teasdale N. Step length variability at gait initiation in elderly fallers and non-fallers, and young adults. Gerontology 2003;49:21–6. [28] Mathias S, Nayak USL, Isaacs B. Balance in elderly patients: the ‘getup and go test’. Arch Phys Med Rehabil 1986;67:387–9. [29] Schmidt R, Lee TL. Motor Control and Learning, A Behavioral Emphasis. Champaign, IL: Human Kinetics; 1999. [30] Berg W, Alessio HM, Mills EM, Tong C. Circumstances and consequences of falls in independent community-dwelling older adults. Age Ageing 1997;26(4):261–8.
Kinematic analyses of the 1808 standing turn: effects of age on strategies adopted by healthy young and older women Peggy Meinhart-Shibataa,*, Michelle Kramera, James A. Ashton-Millera,b,c, Carol Persadd a
Biomechanics Research Laboratory, Department of Mechanical Engineering, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA b Department of Biomedical Engineering, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA c Institute of Gerontology, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA d Department of Psychiatry, GGB 3208, University of Michigan, Ann Arbor, MI 48109-2125, USA Accepted 2 August 2004
Abstract Standing turns are associated with an increased risk for falls and fall-related injuries in the elderly. The purpose of this study was to test the (null) hypothesis that age has no effect on the kinematics of the 1808 turn. Ten young and 10 older healthy women were asked to complete a series of 1808 turns in a standing posture after picking up a light bowl with both hands. Foot–ground reactions, insole pressures and body segment kinematics were recorded in 62 trials at 100 Hz. Turning strategies were analyzed for effects of both age and turn direction on linear and angular foot kinematics, as well as pelvic axial rotation. The older women (OW) used a preparatory stepping strategy more often (170%, p < 0.002), and employed a lower average pelvic rotation rate (21%, p < 0.011) than the younger controls. The minimum foot separation distance for OW was less in their non-preferred than in their preferred turn direction (29%, p < 0.038), thereby increasing their risk of foot– foot interference and falling when turning in their non-preferred direction. The older women were more variable in their turn execution, particularly in minimum foot separation distance (55%, p < 0.022) and the maximum rate of pelvic rotation (82%, p < 0.035). Despite the fact that these healthy older women were careful to employ a preparatory stepping strategy and slower average rotational velocities, they were also more variable in their turn execution than the young. # 2004 Elsevier B.V. All rights reserved. Keywords: Turning; Women; Aging; Falls; Kinematics
1. Introduction As the geriatric population continues to rise, the consequences of falls and fall-related injuries carry increasing socioeconomic significance. One-third of community-dwelling elderly fall each year; 5% of these experience fracture or injury requiring hospitalization [1] and expenses now exceed $20 billion annually [2]. Moreover, in 1999, falls caused 31% of all unintentional injury deaths [3]. The most common activities being performed at the time of fall-related hip fracture are walking * Corresponding author. Tel.: +1 734 936 0368; fax: +1 734 763 9332. E-mail address: [email protected] (P. Meinhart-Shibata). 0966-6362/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2004.08.005
forward and turning around. Although 40% of hip fractures occur during walking and 18% occur during turning [4], a fall while turning is 7.9 times more likely to result in hip fracture than a fall while walking straight [5]. Hence, turning carries a significant risk for hip fracture. Standing turn performance has been used to delineate elderly fallers from non-fallers. Elderly fallers take 20% longer to turn 1808 than non-fallers [6] and frail elderly fallers take six more steps to turn 3608 than non-fallers [7]. However, the reason why fallers choose to turn more slowly, and take more steps, is unknown. Possible factors include vestibular impairments [8], difficulty handling whole-body angular momentum (in the same way that linear momentum can be problematic [9,10]), or fear of foot–ground or foot–
120
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
foot interference. Inadvertent foot–foot interference could result from the increase in variability in step width and step length observed with increasing age [11]. Because of its value in predicting potential fall risk, standing turn performance has been included in routine clinical assessments of elderly by measuring the time and/or number of steps required to turn [12]. For example, 3608 turns are incorporated into the ‘Performance-Oriented Mobility Assessment’ (POMA) [13], the ‘Physical Performance’ [14], and the ‘Berg Balance’ [15] test batteries to evaluate functional balance. Furthermore, turning was a component of a mobility test that successfully predicted health outcomes in older hip fracture patients [16]. While the etiology of most falls is multifactorial [17], these activityoriented evaluations are useful for evaluating subjects who are more likely to fall while turning [18]. Despite the positive association between turning difficulty and increased risk of fall-related hip fractures, we are unaware of biomechanical analyses of standing turns. Our main goals were to characterize standing turns using kinematics and to determine whether differences in turning execution exist between healthy young and elderly women. The primary measures evaluated were self-selected stepping strategies, separation distance between the feet during the turn, and turning rate as measured by average pelvic rotational velocity about the vertical. Secondary measures included the time and number of steps to turn, foot position during double support, and variability in foot separation distance and pelvic velocity. These measures are related to potential falling mechanisms such as tripping over one’s feet, unintended aberrant foot placement, and/or unexpected angular momentum due to variability in execution. Older women were studied because they have twice the risk for fall-related hip fractures as men [19]. The hypothesis that no age differences exist in the kinematics of 1808 turns was tested with the expectation that dissimilarities would be revealed and this hypothesis would ultimately be rejected.
2. Methods Ten healthy young women (YW) aged 21.8 1.99 years and 10 healthy, community-dwelling older women (OW) aged 72.5 5.82 years participated. YW were University of Michigan students; OW were contacted through the University of Michigan Institute of Gerontology. The YW and OW heights were 160.52 6.83 and 162.01 6.70 cm, respectively. A nurse practitioner examined individuals in the OW group under supervision of a geriatrician–physician. All denied having fallen in the preceding 2 years, were not fearful of falling, and had no complaints of imbalance. Subjects gave written consent to participate in this study as approved by the governing Institutional Review Board. All subjects wore comfortably fitted athletic walking shoes (Model 800, New Balance, Inc.) and were outfitted in an overhead harness system to avoid injury in the case of a
Fig. 1. Overhead view of experimental setup showing tables (a, b) and bilaterally symmetric initial foot position. Preparatory stepping pattern and foot trajectories for a turn to the right are shown (cross-hatched pattern: 1L = first step, left foot; 2-R = second step, right foot; etc.). Left and right arrows denote darkened and illuminated cues, respectively.
fall. Because a directional preference for spontaneous turns has been demonstrated [20], we first asked subjects to turn in a circle. We designated the turn direction the subject chose as the ‘preferred direction’ (PD), the contralateral direction was designated as the ‘non-preferred direction’ (NPD). Subjects performed 62 trials, each initiated by an illuminated arrow pointing to the right or left. Subjects used both hands to pick up a light-weight bowl from a waisthigh table, executed a 1808 turn in the indicated direction, and placed the bowl on a table 2 m behind their starting position (Fig. 1). Subjects carried the bowl because 1808 turns are often executed in confined spaces, like a kitchen, to move objects from one place to another. Subjects were instructed to perform trials at a comfortable speed as if in their kitchen preparing a meal. Forty-eight trials consisted of turns to the right or left. Fourteen trials were turns with a sudden direction change upon an audio tone. The results of these latter trials, presented intermittently throughout the experiment, are not presented here. Trials were organized in a pre-determined randomized order unknown to subjects. The time to complete a trial was measured digitally as the time between the visual cue appearance and interference of a light beam as the bowl was placed on the second table. Actual turn initiation was identified by recording when the normal force from one of two forceplates (Model OR6, AMTI, Watertown, MA), situated beneath each foot, changed by 5% of the output extant during quiet standing. The number of steps for each trial was established by insole pressure sensors (thickness 0.18 mm; F-Scan, Tekscan, Inc., South Boston, MA). The beginning and ending of each step was determined using a force threshold of 15% body weight. Any foot–foot interference was identified by electrical contact between strips of thin adhesive aluminum, covering
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
the medial aspect of each shoe causing a voltage change to be recorded. Body segment kinematics were measured at 100 Hz via an Optotrak (Northern Digital Corporation, Waterloo, Ontario) dual 3020 optoelectronic camera system. Groups of three coplanar but non-collinear infrared-emitting diodes (IREDS) formed six rigid triads and established local coordinate systems. Triads were fixed to both medial and lateral aspects of the upper last of each shoe and the most anterior prominence of each iliac crest. With subjects seated, the locations of the triads affixed to the shoes, and points separated by 1 cm along the margin of each shoe sole, were recorded by Optotrak hand digitization establishing the relative position between the sole and corresponding shoe markers. Using this information and transformation methods similar to Startzell and Cavanagh [21], exact shoe sole placement during double support and trajectory during single support was reconstructed for all frames (Fig. 1). Kinematic data also permitted the quantification of pelvic rotational behavior. Pelvic position was represented by the origin and direction of a vector perpendicular to and bisecting the line joining the origins of the two local coordinate systems at each anterior iliac crest. This averaged coordinate does not provide a specific measure such as center of mass, but can identify the change in linear and rotational position of the pelvic region. Unfortunately, one YW and OW held the bowl in a manner that frequently obstructed the markers located on the iliac crests making it impossible to obtain sufficient pelvic data. Therefore, pelvis kinematic results (only) were limited to nine YW and nine OW. Kinetic and kinematic data were filtered using a fourthorder, low-pass, Butterworth filter with break point at 8 Hz. The analysis of the 1808 turn established a general turning strategy for each group. One quantifiable measure of turning behavior was the minimum (Euclidean) foot separation distance at each sample interval. Comparison of all frames revealed the overall minimum foot separation distance and how close the feet came to contacting one another during each trial. The variability in minimum foot separation distance was later analyzed to estimate accuracy of foot trajectory and placement. While variability may not kinematically describe a standing turn, any group differences yield insight into how age affects turning execution. Another foot kinematic measure was the ‘included foot angle’, or the angle between two vectors representing the direction of each foot from heel to toe during double support. The maximum included foot angle reflected the maximum included angle observed over all trials for each subject. Rotations were calculated with respect to the global yaxis (Fig. 1). The absolute value of rotations was used to combine turns in both directions. To measure the time to complete only the turning portion of the task, the time between movement initiation and pelvic rotational changes of 458, 908, 1358, and 1808 was recorded. Rotation data were numerically differentiated to obtain the angular velocity for each trial. Maximum pelvic velocity, a measure reflecting
121
the maximum angular momentum generated, was analyzed along with average pelvic rotational velocity. By averaging over all trials, the outcome measure (minimum foot separation distance, average and maximum rate of rotation, number of steps, and total time) mean was determined for each of the 10 subjects (n = 9 for rotational measures). The kurtosis and skewness of these samples were calculated to test for normality. Age effects and withinsubject differences in PD and NPD turns were investigated. Mann–Whitney and Wilcoxon Ranks non-parametric tests were utilized for analyses involving non-normal distributions. Unless otherwise stated, all other analyses consisted of two-sample or paired t-tests as appropriate. Identical methods were used to investigate effects on outcome measure variability after calculating the standard deviation of each parameter for each subject and grouping according to age. To further examine minimum foot separation distance and interactions among age, turn direction, and practice, individual trials were used in a repeated-measures analysis of variance (rm-Anova). Approximately 48 trials for each subject were used, resulting in 947 samples for this analysis. A value of p < 0.05 was considered significant for all results.
3. Results Age effects were evident in the preparatory stepping strategy used by OW, whereby a small first step was commonly taken in the direction of the turn by the contralateral foot (Fig. 1). OW used this strategy in 65% of the trials, whereas YW used it in only 24% of trials. For turns in the NPD this age difference was accentuated (Table 1). No effect of turn direction on stepping strategy was evident in either group. Not one instance of foot–foot contact was recorded in these healthy women. Overall, YW and OW revealed similar minimum foot separation distances (Table 1). However, when PD and NPD turns were analyzed within each age group, there was no direction effect in the YW, but OW brought their feet closer together during NPD turns (Table 1). The rm-ANOVA showed that turn direction, not age, was the only factor affecting this parameter. The borderline interaction between age and turn direction reflected that the OW had a greater tendency to bring their feet closer together during NPD turns than did the YW (Table 2). Across all trials, age affected the average pelvic rotational velocity: the YW and OW turned at an average velocity of 85.98/s and 67.98/s, respectively (Table 1). The turn direction did not influence the average pelvic rotational velocity in either group. Finally, there was no age or direction effect on maximum pelvic rotational velocity. Secondary measures analyzed included the number of steps and time required to complete the task, maximum included foot angle during double support, and variability in kinematic measures. Overall, the OW and YW showed no
122
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
Table 1 Mean (S.D.) primary and secondary outcome measures are shown for young and elderly women (YW, OW, respectively) in the preferred direction (PD), nonpreferred direction (NPD) by group, and for the trials combined for direction (ALL) YW Primary measures Proportion of trials with preparatory strategy
PD NPD ALL
0.24 (0.29) 0.24 (0.35) 0.24 (0.24)
OW 0.56 (0.37)a 0.72 (0.38)b 0.65 (0.25)c
Average minimum foot separation distance (mm)
PD NPD ALL
34.1 (5.1) 31.9 (13.9) 33.0 (13.4)
49.2 (17.6)d 34.9 (13.1) 42.1 (12.5)
Average pelvic rotational velocity (8/s)
PD NPD ALL
87.2 (13.2) 84.4 (10.9) 85.9 (11.6)
66.4 (16.1)e 70.2 (15.4)f 67.9 (14.9)g
Maximum pelvic rotational velocity (8/s)
PD NPD ALL
167.1 (21.8) 164.0 (21.6) 165.5 (19.7)
Secondary measures Number of steps
149.1 (26.7) 142.8 (47.2) 148.0 (26.0)
PD NPD ALL
4.01 (0.61) 3.85 (0.78) 3.93 (0.67)
4.53 (1.02) 4.37 (1.02) 4.45 (0.89)
Total time (s)
PD NPD ALL
3.99 (0.63) 3.95 (0.58) 3.97 (0.60)
4.46 (1.25) 4.37 (1.18) 4.40 (1.19)
Maximum included foot angle (8)
PD NPD ALL
150.7 (17.4) 149.2 (16.6) 155.4 (17.7)
125.1 (14.8)h 136.6 (19.1) 138.0 (16.9)i
Variability of minimum foot separation distance (mm)
PD NPD ALL
11.8 (5.0) 10.9 (4.4) 12.7 (4.6)
16.0 (5.3) 17.4 (7.6)j 19.7 (7.6)k
Variability of maximum pelvic rotational velocity (o/s)
PD NPD ALL
19.6 (3.7) 19.8 (10.1) 21.8 (5.7)
39.7 (30.4) 29.8 (15.3) 39.7 (22.6)l
Quantities in italics represent samples with non-normal distributions. a p = 0.041 for age. b p = 0.024 for age. c p = 0.002 for age. d p = 0.038 for turn direction. e p = 0.009 for age. f p = 0.037 for age. g p = 0.011 for age. h p = 002 for age. i p = 0.037 for age. j p = 0.030 for age. k p = 0.022 for age. l p = 0.035 for age.
difference in the number of steps or time required to complete a trial (Table 1). No effect of turn direction on either measure was revealed for either group. The YW displayed larger maximum included foot angles than the OW (Table 1). The maximum included foot angle was also unaffected by turn direction. The time history of pelvic rotation from turn initiation to completion of 1808 rotation for each subject’s first PD turn was plotted by age group (Fig. 2). The YW completed this rotation within 1.5–3.0 s with two or three steps. OW were more likely to take three or more steps, following their
tendency to use smaller included foot angles than the YW. All but one OW curve ended beyond the average time for YW, and for OW the times for the rotations were more variable, with a range of 1.5–6 s. We note two interesting facts. First, the two longest-duration turns were performed by the two oldest subjects. Their respective ages were 83 and 81 years; the next oldest subject was 73 years. Second, the shortest duration trial for the OW is indicative of one OW behaving generally like the YW. Similar results (not shown) were obtained for initial NPD turns. The time required to turn was examined for all trials, and age-associated time
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
123
Table 2 Repeated measures ANOVA results for minimum foot separation distance data F
p-value
Main effect Age Turn direction Trial
2.231 6.146 1.066
0.153 0.023 0.385
Factor interaction Age turn direction Age trial Turn direction trial Age turn direction trial
3.704 1.016 0.423 0.857
0.070 0.441 0.985 0.637
Note: ‘Trial’ factor represents practice effect.
differences increased as the turn neared completion, although not significantly (Fig. 3). The kinematic variability was the final measure examined. The OW were more variable in their minimum foot separation distance than were the YW (Table 1, Fig. 4). Turn direction did not affect the variability of minimum foot separation distance for either group. There was no effect of age or turn direction on the variability of average rotational velocity. However, the OW were more variable than the YW in their maximum pelvic rotational velocity (Table 1).
4. Discussion We have investigated the effect of age on turning behavior by comparing the kinematics of two groups of healthy women. Age-related differences were revealed in stepping strategy, minimum foot separation distance, and
Fig. 2. Pelvic rotation time history with respect to the global y-axis (Fig. 1) for the first PD trial for each subject. Symbols signify completion of successive steps according to legend at right. Arrows represent the mean time for each group.
Fig. 3. Average time required for subjects to complete quartiles of total pelvic rotational change for all turns. Bars denote corresponding S.D.
average pelvic rotational velocity, indicating a tendency by the OW to be more cautious when turning. Despite their carefulness, the OW exhibited more variability in minimum foot separation distance and maximum pelvic rotational velocity. Strong evidence for caution in the OW is their increased use of the preparatory stepping strategy. This strategy reduces the angle of lower extremity external rotation required for subsequent steps, as confirmed by smaller included foot angles exhibited by the OW (Table 1). The preparatory step also permits a larger anterior/posterior dimension of their bipedal base of support than would result if one step with the ipsilateral foot were used to achieve the same absolute angle of rotation. The OW also used smaller included foot angles in order to achieve their turn. Without measuring maximum ranges of hip motion, we cannot determine whether the reduced angles
Fig. 4. Minimum foot separation distance frequency histograms. Data grouped according to age (YW, OW) and turn direction (PD, NPD).
124
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125
can be attributed to age-differences in the motor control of turning, or a decline in hip range of motion with age. However, hip rotational range of motion in both men and women continues to decrease even after the age of 70 years [22]. Therefore, age-related limitations in hip external rotation could be an explanation for the smaller included foot angles measured in these OW. An additional age effect on turn execution was the trend toward the OW maintaining increased minimum foot separation distances. This parameter reflected age effects in subjects’ abilities to handle turns in opposing directions. For example, the feet of the OW came significantly closer in the NPD turns, thereby increasing their risk for foot–foot interference and for a trip or stumble during a NPD turn. More evidence for OW conservatism is their use of reduced average pelvic rotational velocities. This likely resulted in less whole-body angular momentum to arrest at the end of the turn, assuming that the torso rotates en-bloc. These decreased comfortable turning speeds parallel the decrease in comfortable walking speed with age in OW, for example, as shown by Chen et al. [23]. Moreover, they are consistent with the reticence of the elderly to generate high velocities and momentum in challenging tasks, such as reaching [10]. Older adults use different strategies to develop momentum [24], and may generate smaller momenta because they are either unable to develop lower extremity joint torques as rapidly [25] or because they have more difficulty arresting movements than younger adults [26]. To establish the validity of our results, we compared the time required for OW to turn 1808 (2.8 s from turn initiation) with reported values. The present OW were faster than published elderly non-faller times of 4.8 and 8.2 s for 1808 and 3608, respectively [6,7]. Differences may be explained by disparities in subject group composition and turn time measurement. For example, Imms and Edholm measured the time to turn 1808 as the time taken to pass through a beam, turn around, and pass back through the beam. Their measurement therefore includes some linear motion. Also, their average age for non-fallers was 77.4 years compared to our average of 72.5 years [6]. Lipsitz et al. [7] subject group possibly took longer because they were over a decade older and more frail than our subjects, including those with cognitive impairment, assistive device use, or some degree of dependency. In an earlier study of turning behavior, three groups were examined: young subjects with no difficulty turning (YNDT), elderly subjects with no difficulty turning (ENDT), and elderly subjects with difficulty turning (EDT) as identified by a self-reported questionnaire (Thigpen et al. [12]). For just the turning portion of the Timed ‘Up and Go’ Test [28], the ENDT took more time and steps than the YNDT. While some overlap was present, the turning characteristics and strategies of the ENDT group were more variable than in the YNDT group [12]. Our results corroborate these findings. Our results indicate that even physically active and welladapted OW showed differences in the kinematics of 1808
turn execution. We expect that frail elderly would exhibit even more caution in their turning strategies. But the question remains why turning is associated with so many hip fractures. This study cannot answer that question. But, despite being more cautious, the minimum foot separation distance and maximum pelvic rotational velocity of the OW reflected greater average variability than in the YW. Increased step variability is a hallmark of fallers [27], and it may underlie difficulty with turns. The present study has a number of limitations. It is unknown whether the present age differences extend to turns through angles other than 1808. One can also not know whether these results may be extrapolated to men. But, smaller hip ranges of motion in males [22] may alter their turning strategy even more with age. Turns in this study were carried out on industrial carpeting. It is unknown if turning strategies would differ on hard surfaces or on deeper pile carpet. Finally, the small sample sizes used limited the statistical power available. Although we determined that YW and OW turned at different speeds, the difficulty of their respective turns was comparable because subjects completed the task at their comfortable speed. Given that Fitts demonstrated that movement accuracy generally decreases with increasing movement speed, [29] more kinematic variability might have been found had we asked the subjects to turn in a hurry. This may be worth examining in the future, since hurrying is known to be associated with a higher risk for falls [30].
Acknowledgements The support of National Institute on Aging Grants AG 10542 and AG 08808, the National Science Foundation, and the Rackham School of Graduate Studies, University of Michigan is acknowledged. The authors are grateful for the assistance of Diane Scarpace for subject screening under the supervision of Neil B. Alexander, and Janet Kemp, Martin Stenzel, Adriana Figueroa, and Alicia Gehle for technical support.
References [1] Tinetti M, Speechley M, Gintar SF. Risk factors for falls among elderly persons living in the community. N Engl J Med 1988;319:1701–6. [2] Englander F, Hodson TJ, Terregrossa RA. Economic dimensions of slip and fall injuries. J Forensic Sci 1996;41(5):733–6. [3] Stevens J. Falls among older adults: public health impact and prevention strategies. Generations 2002;26(4):7–13. [4] Nevitt M, Cummings SR. Type of fall and the risk of hip and wrist fractures: the study of osteoporotic fractures. J Am Geriatr Soc 1993;41(11):1226–34. [5] Cumming R, Klineberg RJ. Fall frequency and characteristics and the risk of hip fractures. J Am Geriatr Soc 1994;42(7):774–8. [6] Imms F, Edholm OG. Studies of gait and mobility in the elderly. Age Ageing 1981;10:147–56.
P. Meinhart-Shibata et al. / Gait & Posture 22 (2005) 119–125 [7] Lipsitz L, Johsson PV, Kelley MM, Koestner JS. Causes and correlates of recurrent falls in ambulatory frail elderly. J Gerontol 1991;46(4):M114–22. [8] Sloane P, Baloh RW, Honrubia V. The vestibular system in the elderly: clinical implications. Am J Otolaryngol 1989;10(6):422–9. [9] Pai Y, Naughton BJ, Chang RW, Rogers MW. Control of body centre of mass momentum during sit-to-stand among young and elderly adults. Gait Posture 1994;2:109–16. [10] Kozak K, Ashton-Miller JA, Alexander NB. The effect of age and movement speed on maximum forward reach from an elevated surface: a study in healthy women. Clin Biomech 2003;18:190–6. [11] Grabiner P, Biswas ST, Grabiner MD. Age-related changes in spatial and temporal gait variables. Arch Phys Med Rehabil 2001;82: 31–5. [12] Thigpen M, Light KE, Creel GL, Flynn SM. Turning difficulty characteristics of adults aged 65 years or older. Phy Ther 2000;80(12):1174–87. [13] Tinetti M. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc 1986;34:119–26. [14] Reuben D, Siu AL. An objective measure of physical function of elderly outpatients. The physical performance test. J Am Geriatr Soc 1990;38(10):1105–12. [15] Berg K, Wood-Dauphinee SL, Williams JT, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Can 1989;41:304–11. [16] Fox K, Hawkes WG, Hebel JR, Felsenthal G, Clark M, Zimmerman SI, Kenzora JE, Magaziner J. Mobility after hip fracture predicts health outcomes. J Am Geriatr Soc 1998;46:169–73. [17] Tinetti M. Preventing falls in elderly persons. N Engl J Med 2003;348(1):42–9. [18] Topper A, Maki BE, Holliday PJ. Are activity-based assessments of balance and gait in the elderly predictive of risk of falling and/or type of fall? J Am Geriatr Soc 1993;41(5):479–87.
125
[19] DeGoede K, Ashton-Miller JA, Schultz AB. Fall-related upper body injuries in the older adult: a review of the biomechanical issues. J Biomech 2003;36:1043–53. [20] Yazgan M, Leckman JF, Wexler BE. A direct observational measure of whole body turning bias. Cortex 1996;32:173–6. [21] Startzell J, Cavanagh PR. A three-dimensional approach to the calculation of foot clearance during locomotion. Hum Movement Sci 1999;18:603–11. [22] James B, Parker AW. Active and passive mobility of lower limb joints in elderly men and women. Am J Phys Med Rehabil 1989;68(4):162– 7. [23] Chen H, Ashton-Miller JA, Alexander NB, Schultz AB. Effects of age and available response time on ability to step over an obstacle. J Gerontol 1994;49(5):M227–33. [24] Polcyn A, Lipsitz LA, Kerrigan DC, Collins JJ. Age-related changes in the initiation of gait: degradation of central mechanisms for momentum generation. Arch Phys Med Rehabil 1998;79(12):1582–9. [25] Thelen DG, Schultz AB, Alexander NB, Ashton-Miller JA. Effects of age on rapid ankle torque development. J Gerontol 1996;51A(5):M226–32. [26] Vrtunski P, Patterson MB, Hill GO. Factor analysis of choice reaction time in young and elderly subjects. Percept Mot Skills 1984;59:659– 76. [27] Mbourou G, Lajoie Y, Teasdale N. Step length variability at gait initiation in elderly fallers and non-fallers, and young adults. Gerontology 2003;49:21–6. [28] Mathias S, Nayak USL, Isaacs B. Balance in elderly patients: the ‘getup and go test’. Arch Phys Med Rehabil 1986;67:387–9. [29] Schmidt R, Lee TL. Motor Control and Learning, A Behavioral Emphasis. Champaign, IL: Human Kinetics; 1999. [30] Berg W, Alessio HM, Mills EM, Tong C. Circumstances and consequences of falls in independent community-dwelling older adults. Age Ageing 1997;26(4):261–8.