- Email: [email protected]
Quantitative assessment of the stops walking while talking test in the elderly1
838
Quantitative Assessment of the Stops Walking While Talking Test in the Elderly Esther W. de Hoon, MD, John H. Allum, DSc, PD, Mark G. Carpenter, PhD, Christian Salis, Bastiaan R. Bloem, PhD, MD, Martin Conzelmann, MD, Heike A. Bischoff, MD ABSTRACT. de Hoon EW, Allum JH, Carpenter MG, Salis C, Bloem BR, Conzelmann M, Bischoff HA. Quantitative assessment of the stops walking while talking test in the elderly. Arch Phys Med Rehabil 2003;84:838-42. Objective: To examine whether trunk sway and walking speed differ between elderly “stoppers” and “nonstoppers” during a shorter version of the stops walking while talking (SWWT) test—an observational assessment of impaired dualtask performance—and during a normal walking trial. Design: The original SWWT test was administered on the way to the test room (over a distance of 150m). Then, subjects were asked to walk 2 trials of 8m while wearing a trunk sway measuring device strapped firmly to their lower back. For the first 8-m trial, no questions were asked (control trial). During the second 8-m trial, subjects were asked an easy question (What is your age?) after walking 2m. Setting: Long-stay geriatric care unit in Switzerland. Participants: Seventeen institutionalized elderly (16 women, 1 man; mean age, 86.3y; range, 79 –93y). Subjects had to be able to walk at least 150m and to understand simple questions. Interventions: Not applicable. Main Outcome Measures: The amplitude of trunk sway angle and angular velocity in the forward-backward (pitch) and side-to-side (roll) directions and the duration of each trial were compared between the two 8-m walking trials with and without a question among subjects who did and did not come to a complete stop. Results: In the original SWWT test, 4 persons stopped walking while talking, compared with 8 persons who stopped in the short (8-m) walking trial when a question was asked. Persons who stopped during the 8-m trial when a question was asked had significantly longer walking durations (by 19s) and larger trunk roll angular displacements (by 5.5°) during trials, both with and without a question. For both stoppers and nonstoppers, duration was longer during the trial when a question was asked.
From the Department of Otorhinolaryngology, University Hospital, Basel, Switzerland (de Hoon, Allum, Carpenter); Department of Neurology, Leiden University Medical Centre, Leiden, The Netherlands (de Hoon); Department of Kinesiology, University of Waterloo, Waterloo, ON, Canada (Carpenter); Department of Orthopaedics, University of Basel, Basel, Switzerland (Salis, Bischoff); Department of Neurology, University Medical Centre St Radboud, Nijmegen, The Netherlands (Bloem); and Department of Geriatrics, Felix-Platter Spital, Basel, Switzerland (Conzelmann, Bischoff). Supported by grants from the Fundatie van de Vrijvrouwe van Renswoude te ’s-Gravenhage, Glaxo-Welcome, a NDIT/FPIT StudEx subsidy, and by the Free Academic Society of Basel. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to John H. Allum, DSc, PD, University HNO-Klinik, Petersgraben 4, CH-4031 Basel, Switzerland, e-mail: [email protected]. 0003-9993/03/8406-7207$30.00/0 doi:10.1016/S0003-9993(02)04951-1
Arch Phys Med Rehabil Vol 84, June 2003
Conclusion: A fixed and brief walking distance, coupled with a single sudden question, provided an effective method of identifying subjects who stop walking while talking. These subjects are those who have slower walking speeds and more unstable trunk control in the roll plane even under normal walking conditions. Our findings support the predictive capabilities of a brief SWWT test for the unstable and fall-prone elderly, as well as the usefulness of objective trunk sway measures to identify gait instabilities. Key Words: Accidental falls; Balance; Elderly; Rehabilitation; Walking. © 2003 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation MAJOR FOCUS OF THOSE caring for the elderly is to prevent falls, primarily by reducing the factors that conA tribute to fall risk. Predicting a tendency to fall in the elderly is, however, difficult when based on single physiologic measures. Generally, it is an accumulation of deficits, such as low muscle strength, decreased reaction times, deficits in sensory inputs contributing to balance control, and mobility impairments that lead to deficits in the functional tasks of daily living.1-8 Eventually, such impairments may lead to falls in the elderly.9-11 In view of the complex and multifactorial pathophysiology underlying falls, it is understandable that the assessment of isolated components of postural control does not predict falls well. Accumulated evidence does, however, suggest that fall risks can be estimated effectively by measuring dual-task performance, that is, by a person’s ability to execute a secondary task while standing or walking.12-14 For example, Lundin-Olsson et al15 found that the timed up and go test16 is a good predictor of fall risk if combined with a manual task, such as carrying a tray on which there is a glass of water. The stops walking when talking (SWWT) test appears to be an even better predictor of fall risk.1 The test is a functional one of impaired dual-task performance that is related to fall risk in the elderly.1 To perform the test, the examiner starts a routine conversation with the person being tested during a walk of 100 to 200m and documents whether that person completely stops walking during the conversation. This inability to walk and talk at the same time appeared to be an excellent predictor of falls in cognitively impaired elderly persons.1 Nonetheless, the test has several shortcomings. One shortcoming is that the test requires mobility skills (eg, the ability to walk more than 100m) that may be beyond the capabilities of frail institutionalized elderly persons. Another drawback is that the conversation content is not defined. A third shortcoming is that classification is limited to visual inspection of complete stops. It could be argued that the SWWT test is not objective enough to detect more subtle changes in balance or gait control during dual-task conditions17 (ie, the test is not graded enough to recognize borderline fallers). Quantitative assessment of balance control by using static or dynamic posturography can reveal subtle changes in postural control during dual-task per-
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
formance.14-16,18 However, such assessments are based only on balance control during stance and not on the gait activities of daily living; thus, they are possibly less relevant in predicting falls. We had 2 objectives in this study. First, we wanted to know whether subjects also stop walking when talking if the walking distance is shorter and is paired with a simple question. Second, we studied whether measures of trunk sway might provide objective measures with which to identify persons with impaired dual-task performance, and therefore at risk to fall. We measured dynamic balance control (in the form of trunk sway) during an abbreviated version of the SWWT test, based on an 8-m walking distance. The trunk sway measuring device we used19,20 permitted an accurate, observer-independent, assessment of balance control during a natural walking task and was well tolerated by the elderly. We also recorded the duration of the trial (as an estimate of walking velocity). METHODS Similar to Lundin-Olsson et al,1 we asked frail elderly people with weakness and an increased risk of falling9 to participate in this study. Seventeen institutionalized elderly subjects (16 women, 1 man; mean age, 86.3y; range, 79 –93y) agreed to participate. They were recruited from 2 long-stay geriatric care units within the Felix-Platter Hospital, Basel University, Basel, Switzerland. Inclusion criteria were 70 years of age or older and the ability to walk 150m with or without a walking aid. Exclusion criteria were a score below 10 on the Mini-Mental State Examination21 (MMSE), decreased corrected or uncorrected vision (sufficiently severe to interfere with gait and balance), a hearing impairment, Parkinson’s disease, spasticity, or amputation. Medical comorbidities were documented with the Charlson Comorbidity Index.22 Witnessed, oral informed consent was obtained from all participants, as required by the local ethics committee. In the study’s first phase, the original SWWT test as described by Lundin-Olsson,1 was performed. Subjects were asked to walk from their room to the examination room (distance, ⬇150m) while engaged in a routine conversation throughout the walk. Complete stops in walking were noted by the examiner. In part 2 of the study, subjects were instructed to walk a distance of 8m twice at their own speed with their normal walking aid (if needed) without stopping. One examiner walked at the same speed next to the subject. During the first 8m walk, no question was asked. At 2m into the second walk, the examiner posed a simple question (What is your age?), which all subjects answered. During each trial, trunk sway and trial duration were measured. To measure trunk sway, we used a devicea that has 2 angular-velocity transducers, which were mounted on a converted motorcycle kidney belt. The transducers were oriented on the belt so that trunk angular velocity was measured in the pitch (fore-aft) and roll (lateral) directions at the level of L1-3. Measurement samples were collected every 10ms. The transducers were connected to a personal computer by a 10-m cable. The examiner walking next to the subject carried the cable, thereby allowing the subjects to move freely over distance of the short walk. Offline, angular velocity first was integrated to yield angular displacement. From the stream of angular displacement and velocity samples, maximum excursions in the positive and negative directions were calculated for the roll and pitch measurements to yield 4 outcome variables, in addition to trial duration: peak-to-peak pitch angle and angular velocity, and peak-to-peak roll angle and angular velocity (fig 1).
839
Statistical Analysis We reasoned that subtle balance changes would be apparent in subjects who stopped during the 8-m question trial. We therefore placed subjects into 2 groups—1 group that stopped while walking (stoppers) and 1 group that did not stop (nonstoppers). The measurement variables were analyzed after log transformation to ensure a normal distribution.19 We used a between- and within-subjects analysis of variance (ANOVA) to examine effects between stoppers and nonstoppers, or between question and no-question trials, as well as any interactions. Significant ANOVA effects were then investigated using t tests with Bonferroni adjustments for the number of variables measured. Significance was set at P less than .05. RESULTS During the 150-m walk of the original SWWT test,1 4 of 17 subjects (24%) stopped walking while talking. During the trials over the 8-m walking distance, 1 subject (6%) stopped walking with no question, whereas 8 (47%) subjects (stoppers) stopped during the question trial. The 4 subjects who stopped walking in the 150-m walk were among the 8 subjects who stopped during the 8-m question trial. No significant differences were found between the stopper and nonstopper groups in age (P⫽.66) or scores on the MMSE (P⫽.79). One person from each group (n⫽2, 14%) had reported previous falls. Figure 1 provides examples of the original trunk sway data for 2 subjects, one a stopper and the other a nonstopper. The data were recorded during the 8-m trial without a question. Apart from the obviously longer duration of the trial for the stopper, it is also apparent in the time histories and x-y plots (of pitch vs roll) that roll sway angle and angular velocities oscillations were larger for the stopper. The distinct differences in trial durations and roll sway angular deviations between individual stoppers and nonstoppers were present across the complete populations (fig 2) for both of the 8-m trials, that is, independent of whether a question was asked. Thus, walking duration was significantly longer (35.1s vs 16.2s; F1,15⫽13.48, P⬍.0023) in the stoppers compared with nonstoppers, independent of whether a question was asked. Similarly, roll angle of the trunk was significantly larger in the stoppers (13.4° vs 7.9°; F1,15⫽8.14, P⬍.0121) for the no-question 8-m walking trial, as well as for the 8-m walking trials with a question, compared with the nonstoppers (fig 2). There was also a trend for larger pitch angle of the trunk in the stoppers (12.0° vs 10.3°). However, this trend was not statistically significant. For all subjects, walking time was significantly longer (F1,15⫽5.83, P⬍.029) during the 8-m trial with a single question (mean ⫾ standard deviation, 27.03⫾18.03s), compared with the 8-m trial without a question (23.17⫾14.60s). DISCUSSION Our first finding was that a shorter version of the original SWWT test (asking a single question during an 8-m walk) provided a faster and perhaps more effective method of identifying subjects with impaired dual-task performance (classified as stoppers) with less space requirements (a walking path of 150m was not needed). In this study, 29% of subjects stopped walking when talking during the original SWWT test, a finding similar to the 21% reported by Lundin-Olsson.1 However, when subjects were required to answer a simple question during the 8-m walking distance, 47% stopped while answering the question. We suspect that the brief test had a higher diagnostic yield in identifying subjects who are prone to fall because we asked a sudden question, rather than engaging
Arch Phys Med Rehabil Vol 84, June 2003
840
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
Fig 1. Time history and x-y (roll-pitch) plots of trunk sway for a stopper subject (upper set of traces) and a nonstopper (lower set of traces) walking 8m during the no-question trial. Underneath each plot of pitch and roll time histories, samples of roll and pitch angle are plotted against one another to create an x-y plot. Similarly, the x-y plots of pitch and roll angular velocity are plotted to the right of the angle x-y plots. The envelop of the x-y plots is marked. The scales for all plots are the same for the 2 subjects. Note the differences in the durations of the time histories (14.3s vs 38.0s) and the large roll (x direction) extent of the stopper’s x-y plots (4.6° vs 15.9°/s and 29.2° vs 61.6°/s). The measurement variables used for this study were the peak-to-peak extent of the x-y plots in the roll and pitch direction and trial duration (ie, the values for time and roll amplitude listed here).
in a more predictable routine conversation. It is possible that the higher proportion of stoppers we identified was confounded by the inclusion of false-positive results (ie, subjects who stop during the brief SWWT test, but who do not fall in daily life). Elderly persons do, however, have more problems with sudden and unexpected events than with more predictable routines (eg, falls typically occur under sudden and unexpected circumstances7). Thus, our asking a question abruptly on average 6 seconds into the 8-m walking trial may mimic an event leading to a fall more effectively than a predictable conversation during a longer walk. To verify this statement, it may prove interesting to examine the dynamics of body motion related to the question event in future studies and to compare the motion with that occurring during actual falls. In addition, our question was
Arch Phys Med Rehabil Vol 84, June 2003
clearly defined in its content and timing, whereas the routine conversation in the original SWWT1 was not. For these reasons, our abbreviated version of the original SWWT test might be used more easily in a clinical setting and may help to expand the application of this test to other patient populations that are unable to walk the long distances of the original SWWT test, with or without a walking aid. The abbreviated version of the SWWT test permits easier quantification of trial duration and body movements, such as trunk sway measured in a controlled environment. In agreement with previous reports of slowed performance during dual tasking,15,19 we found that the overall trial duration was longer during the dual-task question trial compared with the noquestion trial. The increased duration for the dual-task question
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
841
pers that exceed those of younger (about 55y) acute vestibularloss subjects during gait20 appear to provide a quantitative measure of gait instability during the single- and dual-task circumstances. In contrast, during normal stance, trunk sway is generally increased with age and also with vestibular loss.19,20 Thus, increases in trunk sway deviations in people prone to fall during stance may not be related to deviations occurring during gait. CONCLUSION Our results add explanatory power to the findings of LundinOlsson et al,1 showing that persons with impaired dual-task performance have a decreased overall walking speed and poorer dynamic control of trunk roll. A prospective study with long-term follow-up of falls is needed to validate the clinical significance in terms of fall risk prediction of our short version of the SWWT test. Furthermore, future research with this short version of the SWWT test should establish thresholds of performance (both for trial duration and trunk sway) in relation to fall risks to identify subjects with an increased risk for falling at an early stage.
Fig 2. Differences in gait measures for stoppers and nonstoppers. The height of the columns represents the mean values of durations and the peak-to-peak amplitude of trunk roll and pitch angular deviations measured during each of the 8-m walking trials, one with a question and one without. NOTE. Error bars indicate the standard error of the mean. *Significantly different means of the stoppers versus nonstoppers.
trial can be partially explained by the fact that some subjects completely stopped for a few seconds. However, it is worth noting that stoppers also walked slower than nonstoppers during the no-question trial. More interestingly, differences were observed in trunk roll sway amplitude between subjects classified as stoppers when compared with nonstoppers. Thus, stoppers had a significantly larger trunk roll angular displacement during both the no-question and the question walking trials, indicating more lateral instability and a tendency to fall sideways. Stoppers and nonstoppers could not be separated by other variables, including cognitive ability, that we measured. The validity of our findings is substantiated by our identification of similar changes in subjects with balance deficits compared with their age-matched controls. These subjects also had longer gait trial times and larger trunk roll angle deviations.20 Furthermore, trunk sway during gait trials is generally reduced with age.19 Thus, the measures of trunk sway (peakto-peak trunk roll angle during gait) we observed in the stop-
References 1. Lundin-Olsson L, Nyberg L, Gustafson Y. “Stops walking when talking” as a predictor of falls in elderly people [letter]. Lancet 1997;349:617. 2. Tinetti ME, Baker DI, Garrett PA, Gottschalk M, Koch ML, Horwitz RI. Yale FICSIT: risk factor abatement strategy for fall prevention. J Am Geriatr Soc 1993;41:315-20. 3. Tinetti ME, Baker DI, McAvay G, et al. A multifactorial intervention to reduce the risk of falling among the elderly people living in the community. N Engl J Med 1994;331:821-7. 4. Tinetti ME, McAvay G, Claus E. Does multiple risk factor reduction explain the reduction in fall rate in the Yale FICSIT Trial? Frailty and Injuries: Cooperative Studies of Intervention Techniques. Am J Epidemiol 1996;144:389-99. 5. Fiatarone MA, O’Neill EF, Doyle N, et al. The Boston FICSIT Study: the effects of resistance training and nutritional supplementation on physical frailty in the oldest old. J Am Geriatr Soc 1993;41:333-7. 6. Province MA, Hadley EC, Hornbrook MC, et al. The effects of exercise on falls in elderly patients. A preplanned meta-analysis of the FICSIT Trials. Frailty and Injuries: Cooperative Studies of Intervention Techniques. JAMA 1995;273:1341-7. 7. Connel BR, Wolf SL. Environmental and behavioral circumstances associated with falls at home among healthy elderly individuals. Atlanta FICSIT Group. Arch Phys Med Rehabil 1997;78: 179-86. 8. Shumway-Cook A, Gruber W, Baldwin M, Liao S. The effect of multidimensional exercises on balance, mobility, and fall risk in community-dwelling older adults. Phys Ther 1997;77:46-57. 9. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med 1988;319: 1701-7. 10. Tinetti ME, Powell L. Fear of falling and low self-efficacy: a case of dependence in elderly persons. J Gerontol 1993;48 Spec No: 35-8. 11. Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of admission to a nursing home. N Engl J Med 1997;337:1279-84. 12. Lajoie Y, Teasdale N, Bard C, Fleury M. Attentional demands for static and dynamic equilibrium. Exp Brain Res 1993;97:139-44. 13. Maki BE, McIlroy WE. Influence of arousal and attention on the control of postural sway. J Vest Res 1996;6:53-5. 14. Andersson G, Yardley L, Luxon L. A dual-task study of interference between mental activity and control of balance. Am J Otol 1998;19:632-7. 15. Lundin-Olsson L, Nyberg L, Gustafson Y. Attention, frailty, and falls: the effect of a manual task on basic mobility. J Am Geriatr Soc 1998;46:758-61.
Arch Phys Med Rehabil Vol 84, June 2003
842
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
16. Podsiadlo D, Richardson S. The timed “up & go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991;39:142-8. 17. Bloem BR, Grimbergen YA, Cramer M, Valkenburg VV. “Stops walking when talking” does not predict falls in Parkinson’s disease [letter]. Ann Neurol 2000;48:268. 18. Thapa PB, Gideon P, Fought RL, Kormicki M, Ray WA. Comparison of clinical and biomechanical measures of balance and mobility in elderly nursing home residents. J Am Geriatr Soc 1994;42:493-500. 19. Gill J, Allum JH, Carpenter MG, et al. Trunk sway measurements of postural stability during clinical balance tests: effects of age. J Gerontol A Biol Sci Med Sci 2001,56:M438-47.
Arch Phys Med Rehabil Vol 84, June 2003
20. Allum JH, Held-Ziolkowska M, Adkin AL, Carpenter MG, Honegger F. Trunk sway measures of postural stability during clinical balance tests: effects of a unilateral vestibular deficit. Posture Gait 2001;14:227-37. 21. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-98. 22. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83. Supplier a. SwayStar™; Balance Int Innovations, CH-3807 Iseltwald, Switzerland.
Quantitative Assessment of the Stops Walking While Talking Test in the Elderly Esther W. de Hoon, MD, John H. Allum, DSc, PD, Mark G. Carpenter, PhD, Christian Salis, Bastiaan R. Bloem, PhD, MD, Martin Conzelmann, MD, Heike A. Bischoff, MD ABSTRACT. de Hoon EW, Allum JH, Carpenter MG, Salis C, Bloem BR, Conzelmann M, Bischoff HA. Quantitative assessment of the stops walking while talking test in the elderly. Arch Phys Med Rehabil 2003;84:838-42. Objective: To examine whether trunk sway and walking speed differ between elderly “stoppers” and “nonstoppers” during a shorter version of the stops walking while talking (SWWT) test—an observational assessment of impaired dualtask performance—and during a normal walking trial. Design: The original SWWT test was administered on the way to the test room (over a distance of 150m). Then, subjects were asked to walk 2 trials of 8m while wearing a trunk sway measuring device strapped firmly to their lower back. For the first 8-m trial, no questions were asked (control trial). During the second 8-m trial, subjects were asked an easy question (What is your age?) after walking 2m. Setting: Long-stay geriatric care unit in Switzerland. Participants: Seventeen institutionalized elderly (16 women, 1 man; mean age, 86.3y; range, 79 –93y). Subjects had to be able to walk at least 150m and to understand simple questions. Interventions: Not applicable. Main Outcome Measures: The amplitude of trunk sway angle and angular velocity in the forward-backward (pitch) and side-to-side (roll) directions and the duration of each trial were compared between the two 8-m walking trials with and without a question among subjects who did and did not come to a complete stop. Results: In the original SWWT test, 4 persons stopped walking while talking, compared with 8 persons who stopped in the short (8-m) walking trial when a question was asked. Persons who stopped during the 8-m trial when a question was asked had significantly longer walking durations (by 19s) and larger trunk roll angular displacements (by 5.5°) during trials, both with and without a question. For both stoppers and nonstoppers, duration was longer during the trial when a question was asked.
From the Department of Otorhinolaryngology, University Hospital, Basel, Switzerland (de Hoon, Allum, Carpenter); Department of Neurology, Leiden University Medical Centre, Leiden, The Netherlands (de Hoon); Department of Kinesiology, University of Waterloo, Waterloo, ON, Canada (Carpenter); Department of Orthopaedics, University of Basel, Basel, Switzerland (Salis, Bischoff); Department of Neurology, University Medical Centre St Radboud, Nijmegen, The Netherlands (Bloem); and Department of Geriatrics, Felix-Platter Spital, Basel, Switzerland (Conzelmann, Bischoff). Supported by grants from the Fundatie van de Vrijvrouwe van Renswoude te ’s-Gravenhage, Glaxo-Welcome, a NDIT/FPIT StudEx subsidy, and by the Free Academic Society of Basel. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to John H. Allum, DSc, PD, University HNO-Klinik, Petersgraben 4, CH-4031 Basel, Switzerland, e-mail: [email protected]. 0003-9993/03/8406-7207$30.00/0 doi:10.1016/S0003-9993(02)04951-1
Arch Phys Med Rehabil Vol 84, June 2003
Conclusion: A fixed and brief walking distance, coupled with a single sudden question, provided an effective method of identifying subjects who stop walking while talking. These subjects are those who have slower walking speeds and more unstable trunk control in the roll plane even under normal walking conditions. Our findings support the predictive capabilities of a brief SWWT test for the unstable and fall-prone elderly, as well as the usefulness of objective trunk sway measures to identify gait instabilities. Key Words: Accidental falls; Balance; Elderly; Rehabilitation; Walking. © 2003 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation MAJOR FOCUS OF THOSE caring for the elderly is to prevent falls, primarily by reducing the factors that conA tribute to fall risk. Predicting a tendency to fall in the elderly is, however, difficult when based on single physiologic measures. Generally, it is an accumulation of deficits, such as low muscle strength, decreased reaction times, deficits in sensory inputs contributing to balance control, and mobility impairments that lead to deficits in the functional tasks of daily living.1-8 Eventually, such impairments may lead to falls in the elderly.9-11 In view of the complex and multifactorial pathophysiology underlying falls, it is understandable that the assessment of isolated components of postural control does not predict falls well. Accumulated evidence does, however, suggest that fall risks can be estimated effectively by measuring dual-task performance, that is, by a person’s ability to execute a secondary task while standing or walking.12-14 For example, Lundin-Olsson et al15 found that the timed up and go test16 is a good predictor of fall risk if combined with a manual task, such as carrying a tray on which there is a glass of water. The stops walking when talking (SWWT) test appears to be an even better predictor of fall risk.1 The test is a functional one of impaired dual-task performance that is related to fall risk in the elderly.1 To perform the test, the examiner starts a routine conversation with the person being tested during a walk of 100 to 200m and documents whether that person completely stops walking during the conversation. This inability to walk and talk at the same time appeared to be an excellent predictor of falls in cognitively impaired elderly persons.1 Nonetheless, the test has several shortcomings. One shortcoming is that the test requires mobility skills (eg, the ability to walk more than 100m) that may be beyond the capabilities of frail institutionalized elderly persons. Another drawback is that the conversation content is not defined. A third shortcoming is that classification is limited to visual inspection of complete stops. It could be argued that the SWWT test is not objective enough to detect more subtle changes in balance or gait control during dual-task conditions17 (ie, the test is not graded enough to recognize borderline fallers). Quantitative assessment of balance control by using static or dynamic posturography can reveal subtle changes in postural control during dual-task per-
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
formance.14-16,18 However, such assessments are based only on balance control during stance and not on the gait activities of daily living; thus, they are possibly less relevant in predicting falls. We had 2 objectives in this study. First, we wanted to know whether subjects also stop walking when talking if the walking distance is shorter and is paired with a simple question. Second, we studied whether measures of trunk sway might provide objective measures with which to identify persons with impaired dual-task performance, and therefore at risk to fall. We measured dynamic balance control (in the form of trunk sway) during an abbreviated version of the SWWT test, based on an 8-m walking distance. The trunk sway measuring device we used19,20 permitted an accurate, observer-independent, assessment of balance control during a natural walking task and was well tolerated by the elderly. We also recorded the duration of the trial (as an estimate of walking velocity). METHODS Similar to Lundin-Olsson et al,1 we asked frail elderly people with weakness and an increased risk of falling9 to participate in this study. Seventeen institutionalized elderly subjects (16 women, 1 man; mean age, 86.3y; range, 79 –93y) agreed to participate. They were recruited from 2 long-stay geriatric care units within the Felix-Platter Hospital, Basel University, Basel, Switzerland. Inclusion criteria were 70 years of age or older and the ability to walk 150m with or without a walking aid. Exclusion criteria were a score below 10 on the Mini-Mental State Examination21 (MMSE), decreased corrected or uncorrected vision (sufficiently severe to interfere with gait and balance), a hearing impairment, Parkinson’s disease, spasticity, or amputation. Medical comorbidities were documented with the Charlson Comorbidity Index.22 Witnessed, oral informed consent was obtained from all participants, as required by the local ethics committee. In the study’s first phase, the original SWWT test as described by Lundin-Olsson,1 was performed. Subjects were asked to walk from their room to the examination room (distance, ⬇150m) while engaged in a routine conversation throughout the walk. Complete stops in walking were noted by the examiner. In part 2 of the study, subjects were instructed to walk a distance of 8m twice at their own speed with their normal walking aid (if needed) without stopping. One examiner walked at the same speed next to the subject. During the first 8m walk, no question was asked. At 2m into the second walk, the examiner posed a simple question (What is your age?), which all subjects answered. During each trial, trunk sway and trial duration were measured. To measure trunk sway, we used a devicea that has 2 angular-velocity transducers, which were mounted on a converted motorcycle kidney belt. The transducers were oriented on the belt so that trunk angular velocity was measured in the pitch (fore-aft) and roll (lateral) directions at the level of L1-3. Measurement samples were collected every 10ms. The transducers were connected to a personal computer by a 10-m cable. The examiner walking next to the subject carried the cable, thereby allowing the subjects to move freely over distance of the short walk. Offline, angular velocity first was integrated to yield angular displacement. From the stream of angular displacement and velocity samples, maximum excursions in the positive and negative directions were calculated for the roll and pitch measurements to yield 4 outcome variables, in addition to trial duration: peak-to-peak pitch angle and angular velocity, and peak-to-peak roll angle and angular velocity (fig 1).
839
Statistical Analysis We reasoned that subtle balance changes would be apparent in subjects who stopped during the 8-m question trial. We therefore placed subjects into 2 groups—1 group that stopped while walking (stoppers) and 1 group that did not stop (nonstoppers). The measurement variables were analyzed after log transformation to ensure a normal distribution.19 We used a between- and within-subjects analysis of variance (ANOVA) to examine effects between stoppers and nonstoppers, or between question and no-question trials, as well as any interactions. Significant ANOVA effects were then investigated using t tests with Bonferroni adjustments for the number of variables measured. Significance was set at P less than .05. RESULTS During the 150-m walk of the original SWWT test,1 4 of 17 subjects (24%) stopped walking while talking. During the trials over the 8-m walking distance, 1 subject (6%) stopped walking with no question, whereas 8 (47%) subjects (stoppers) stopped during the question trial. The 4 subjects who stopped walking in the 150-m walk were among the 8 subjects who stopped during the 8-m question trial. No significant differences were found between the stopper and nonstopper groups in age (P⫽.66) or scores on the MMSE (P⫽.79). One person from each group (n⫽2, 14%) had reported previous falls. Figure 1 provides examples of the original trunk sway data for 2 subjects, one a stopper and the other a nonstopper. The data were recorded during the 8-m trial without a question. Apart from the obviously longer duration of the trial for the stopper, it is also apparent in the time histories and x-y plots (of pitch vs roll) that roll sway angle and angular velocities oscillations were larger for the stopper. The distinct differences in trial durations and roll sway angular deviations between individual stoppers and nonstoppers were present across the complete populations (fig 2) for both of the 8-m trials, that is, independent of whether a question was asked. Thus, walking duration was significantly longer (35.1s vs 16.2s; F1,15⫽13.48, P⬍.0023) in the stoppers compared with nonstoppers, independent of whether a question was asked. Similarly, roll angle of the trunk was significantly larger in the stoppers (13.4° vs 7.9°; F1,15⫽8.14, P⬍.0121) for the no-question 8-m walking trial, as well as for the 8-m walking trials with a question, compared with the nonstoppers (fig 2). There was also a trend for larger pitch angle of the trunk in the stoppers (12.0° vs 10.3°). However, this trend was not statistically significant. For all subjects, walking time was significantly longer (F1,15⫽5.83, P⬍.029) during the 8-m trial with a single question (mean ⫾ standard deviation, 27.03⫾18.03s), compared with the 8-m trial without a question (23.17⫾14.60s). DISCUSSION Our first finding was that a shorter version of the original SWWT test (asking a single question during an 8-m walk) provided a faster and perhaps more effective method of identifying subjects with impaired dual-task performance (classified as stoppers) with less space requirements (a walking path of 150m was not needed). In this study, 29% of subjects stopped walking when talking during the original SWWT test, a finding similar to the 21% reported by Lundin-Olsson.1 However, when subjects were required to answer a simple question during the 8-m walking distance, 47% stopped while answering the question. We suspect that the brief test had a higher diagnostic yield in identifying subjects who are prone to fall because we asked a sudden question, rather than engaging
Arch Phys Med Rehabil Vol 84, June 2003
840
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
Fig 1. Time history and x-y (roll-pitch) plots of trunk sway for a stopper subject (upper set of traces) and a nonstopper (lower set of traces) walking 8m during the no-question trial. Underneath each plot of pitch and roll time histories, samples of roll and pitch angle are plotted against one another to create an x-y plot. Similarly, the x-y plots of pitch and roll angular velocity are plotted to the right of the angle x-y plots. The envelop of the x-y plots is marked. The scales for all plots are the same for the 2 subjects. Note the differences in the durations of the time histories (14.3s vs 38.0s) and the large roll (x direction) extent of the stopper’s x-y plots (4.6° vs 15.9°/s and 29.2° vs 61.6°/s). The measurement variables used for this study were the peak-to-peak extent of the x-y plots in the roll and pitch direction and trial duration (ie, the values for time and roll amplitude listed here).
in a more predictable routine conversation. It is possible that the higher proportion of stoppers we identified was confounded by the inclusion of false-positive results (ie, subjects who stop during the brief SWWT test, but who do not fall in daily life). Elderly persons do, however, have more problems with sudden and unexpected events than with more predictable routines (eg, falls typically occur under sudden and unexpected circumstances7). Thus, our asking a question abruptly on average 6 seconds into the 8-m walking trial may mimic an event leading to a fall more effectively than a predictable conversation during a longer walk. To verify this statement, it may prove interesting to examine the dynamics of body motion related to the question event in future studies and to compare the motion with that occurring during actual falls. In addition, our question was
Arch Phys Med Rehabil Vol 84, June 2003
clearly defined in its content and timing, whereas the routine conversation in the original SWWT1 was not. For these reasons, our abbreviated version of the original SWWT test might be used more easily in a clinical setting and may help to expand the application of this test to other patient populations that are unable to walk the long distances of the original SWWT test, with or without a walking aid. The abbreviated version of the SWWT test permits easier quantification of trial duration and body movements, such as trunk sway measured in a controlled environment. In agreement with previous reports of slowed performance during dual tasking,15,19 we found that the overall trial duration was longer during the dual-task question trial compared with the noquestion trial. The increased duration for the dual-task question
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
841
pers that exceed those of younger (about 55y) acute vestibularloss subjects during gait20 appear to provide a quantitative measure of gait instability during the single- and dual-task circumstances. In contrast, during normal stance, trunk sway is generally increased with age and also with vestibular loss.19,20 Thus, increases in trunk sway deviations in people prone to fall during stance may not be related to deviations occurring during gait. CONCLUSION Our results add explanatory power to the findings of LundinOlsson et al,1 showing that persons with impaired dual-task performance have a decreased overall walking speed and poorer dynamic control of trunk roll. A prospective study with long-term follow-up of falls is needed to validate the clinical significance in terms of fall risk prediction of our short version of the SWWT test. Furthermore, future research with this short version of the SWWT test should establish thresholds of performance (both for trial duration and trunk sway) in relation to fall risks to identify subjects with an increased risk for falling at an early stage.
Fig 2. Differences in gait measures for stoppers and nonstoppers. The height of the columns represents the mean values of durations and the peak-to-peak amplitude of trunk roll and pitch angular deviations measured during each of the 8-m walking trials, one with a question and one without. NOTE. Error bars indicate the standard error of the mean. *Significantly different means of the stoppers versus nonstoppers.
trial can be partially explained by the fact that some subjects completely stopped for a few seconds. However, it is worth noting that stoppers also walked slower than nonstoppers during the no-question trial. More interestingly, differences were observed in trunk roll sway amplitude between subjects classified as stoppers when compared with nonstoppers. Thus, stoppers had a significantly larger trunk roll angular displacement during both the no-question and the question walking trials, indicating more lateral instability and a tendency to fall sideways. Stoppers and nonstoppers could not be separated by other variables, including cognitive ability, that we measured. The validity of our findings is substantiated by our identification of similar changes in subjects with balance deficits compared with their age-matched controls. These subjects also had longer gait trial times and larger trunk roll angle deviations.20 Furthermore, trunk sway during gait trials is generally reduced with age.19 Thus, the measures of trunk sway (peakto-peak trunk roll angle during gait) we observed in the stop-
References 1. Lundin-Olsson L, Nyberg L, Gustafson Y. “Stops walking when talking” as a predictor of falls in elderly people [letter]. Lancet 1997;349:617. 2. Tinetti ME, Baker DI, Garrett PA, Gottschalk M, Koch ML, Horwitz RI. Yale FICSIT: risk factor abatement strategy for fall prevention. J Am Geriatr Soc 1993;41:315-20. 3. Tinetti ME, Baker DI, McAvay G, et al. A multifactorial intervention to reduce the risk of falling among the elderly people living in the community. N Engl J Med 1994;331:821-7. 4. Tinetti ME, McAvay G, Claus E. Does multiple risk factor reduction explain the reduction in fall rate in the Yale FICSIT Trial? Frailty and Injuries: Cooperative Studies of Intervention Techniques. Am J Epidemiol 1996;144:389-99. 5. Fiatarone MA, O’Neill EF, Doyle N, et al. The Boston FICSIT Study: the effects of resistance training and nutritional supplementation on physical frailty in the oldest old. J Am Geriatr Soc 1993;41:333-7. 6. Province MA, Hadley EC, Hornbrook MC, et al. The effects of exercise on falls in elderly patients. A preplanned meta-analysis of the FICSIT Trials. Frailty and Injuries: Cooperative Studies of Intervention Techniques. JAMA 1995;273:1341-7. 7. Connel BR, Wolf SL. Environmental and behavioral circumstances associated with falls at home among healthy elderly individuals. Atlanta FICSIT Group. Arch Phys Med Rehabil 1997;78: 179-86. 8. Shumway-Cook A, Gruber W, Baldwin M, Liao S. The effect of multidimensional exercises on balance, mobility, and fall risk in community-dwelling older adults. Phys Ther 1997;77:46-57. 9. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med 1988;319: 1701-7. 10. Tinetti ME, Powell L. Fear of falling and low self-efficacy: a case of dependence in elderly persons. J Gerontol 1993;48 Spec No: 35-8. 11. Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of admission to a nursing home. N Engl J Med 1997;337:1279-84. 12. Lajoie Y, Teasdale N, Bard C, Fleury M. Attentional demands for static and dynamic equilibrium. Exp Brain Res 1993;97:139-44. 13. Maki BE, McIlroy WE. Influence of arousal and attention on the control of postural sway. J Vest Res 1996;6:53-5. 14. Andersson G, Yardley L, Luxon L. A dual-task study of interference between mental activity and control of balance. Am J Otol 1998;19:632-7. 15. Lundin-Olsson L, Nyberg L, Gustafson Y. Attention, frailty, and falls: the effect of a manual task on basic mobility. J Am Geriatr Soc 1998;46:758-61.
Arch Phys Med Rehabil Vol 84, June 2003
842
ASSESSMENT OF THE STOPS WALKING WHILE TALKING TEST, de Hoon
16. Podsiadlo D, Richardson S. The timed “up & go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991;39:142-8. 17. Bloem BR, Grimbergen YA, Cramer M, Valkenburg VV. “Stops walking when talking” does not predict falls in Parkinson’s disease [letter]. Ann Neurol 2000;48:268. 18. Thapa PB, Gideon P, Fought RL, Kormicki M, Ray WA. Comparison of clinical and biomechanical measures of balance and mobility in elderly nursing home residents. J Am Geriatr Soc 1994;42:493-500. 19. Gill J, Allum JH, Carpenter MG, et al. Trunk sway measurements of postural stability during clinical balance tests: effects of age. J Gerontol A Biol Sci Med Sci 2001,56:M438-47.
Arch Phys Med Rehabil Vol 84, June 2003
20. Allum JH, Held-Ziolkowska M, Adkin AL, Carpenter MG, Honegger F. Trunk sway measures of postural stability during clinical balance tests: effects of a unilateral vestibular deficit. Posture Gait 2001;14:227-37. 21. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-98. 22. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83. Supplier a. SwayStar™; Balance Int Innovations, CH-3807 Iseltwald, Switzerland.