| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
E Nordin, PT, MSc, is a doctoral student, Department of Community Medicine and Rehabilitation, Physiotherapy and Geriatric Medicine, Umea University, SE-901 87 Umea, Sweden
E Rosendahl, PT, MSc, is a doctoral student, Department of Community Medicine and Rehabilitation, Physiotherapy and Geriatric Medicine, Umea University
L Lundin-Olsson, PT, PhD, is Researcher, Department of Community Medicine and Rehabilitation, Physiotherapy and Geriatric Medicine, Umea University
(ellinor.nordin{at}physiother.umu.se). Address all correspondence to Mrs Nordin
Submitted March 21, 2005;
Accepted December 7, 2005
 | Abstract |
|---|
Key Words: Cognition • Frailty • Geriatric assessment • Mobility • Reliability • Timed "Up & Go" Test • Variability
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussionConclusionReferences |
|---|
The Timed "Up & Go" Test (TUG) 1 is a simple and inexpensive method that was developed to screen basic mobility. The TUG comprises basic everyday movements: stand up from a chair, walk 3 m, turn around, walk back, and sit down again. The outcome is the time taken to perform this sequence of movements. Within research, the use of the TUG has increased over the last few years, and it is recommended by the British Geriatrics Society, the American Geriatrics Society, and Nordic geriatricians when screening for risk of falling.2,3 The TUG is reported to yield reliable and valid data for use in older people.1,4–12 However, studies that have evaluated the reliability and validity of TUG scores have included mainly medically stable subjects with good or only mildly affected cognitive functions. To our knowledge, only 2 studies13,14 have attempted to evaluate the effect of cognitive dysfunction on the performance of the TUG. Rockwood et al13 found poor test-retest reliability; however, because the time interval between the measurements was 112±72 (mean±SD) days, this issue cannot be said to have been explored correctly. Thomas and Hageman14 found excellent test-retest reliability; however, their sample of 9 subjects seems to be too small to draw conclusions regarding reliability. Therefore, it is still not known how the presence of cognitive impairment affects the reliability of TUG scores.
There is some support for the notion that cognitive impairment could have a negative influence on the reliability of geriatric assessments in older nursing home residents.15 There is a known correlation between cognitive function and gait. Slower walking speed, decreased step length, increased double-support ratio, and step-to-step variability have been found more often in people with cognitive impairment than in older adults who are healthy.16,17 This finding could be explained by a deterioration of cerebral integration and processing of sensory (visual, vestibular, and proprioceptive) information.16,17 Accordingly, cognitive impairment may result in increased variability in repeated measures of TUG assessment, as this test requires an understanding of the instructions as well as an interaction among the patient, the assessor, and the environmental setting.
There is always some variability in repeated measurements; this variability depends on the person being measured, on the assessor who does the measuring, and on the precision of the instrument.18,19 This variability can represent systematic bias (learning or fatigue effects on the test) or random error attributable to inherent biological or mechanical variations. Reliability can be considered to be the amount of measurement error that has been judged to be acceptable for the valuable use of a measurement tool,20 and the determination of reliability is a clinical decision based on evaluations that include statistical analyses. Consequently, it is important that the users of the tool be provided with the results of the evaluations so that they can make decisions appropriate for their purposes.
Reliability can be quantified as either relative or absolute. Relative reliability examines the relationship between the scores of 2 or more measurements obtained by the same method. Absolute reliability examines the variability of these measurements.19 Many studies that have evaluated methods of measurements have presented only relative reliability with an intraclass correlation coefficient (ICC),5,9,13,14 and it is assumed that a coefficient of nearly 1.00 corresponds to excellent reliability.19 However, on the basis of this type of analysis alone, it is impossible for a clinician to know what variability to expect for a particular patient and to decide whether a change in a score represents a real change in a patient’s condition. For this purpose, the results of an evaluation must be presented in the same units as the measurement or as a proportion of the measured value, that is, in terms of absolute reliability. Analyses of absolute reliability can provide a range in which a patient’s "true" value is expected to be found, as well as another range, to be used in repeated measurements, in which the value is expected to vary from one test occasion to another. These ranges represent the measurement error for a particular patient and can, depending on the chosen level of confidence, be presented as valid for about two thirds (68%) of the patients or for almost all (95%) of the patients.19,20
The quality of measures is of fundamental importance. In clinical assessments, professionals should have knowledge of the expected variability of scores obtained from the instrument in question. Accordingly, when a single TUG score is obtained, health care professionals must be aware that the patient’s true value is within a range above and below the measured value and that the expected range will be wider in a test-retest situation. This range can differ from one population to another. Thus, whether the expected range is influenced by cognitive impairment must be evaluated.
The aim of the present study was to investigate the expected variability of TUG scores in older subjects dependent in ADL and with different levels of cognitive state. The hypothesis was that cognitive impairment would increase the variability of TUG scores.
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
10; dependence in 1 or more personal ADL, according to the Katz index22; and eligibility and ability to perform the TUG without personal physical assistance on 3 occasions within 7 days. The characteristics of the 78 participants are shown in Table 1. Sixty-two percent were women, and the mean age was 84.8 years. The mean MMSE score was 18.7 points. As indicated by the suggested classifications,23 19 participants had no cognitive impairment, 23 had mild impairment, and 36 had severe impairment. According to the Barthel Index, 72 of them were dependent when taking a shower or bath, 38 of them were totally or partly dependent when getting dressed, and 7 required assistance when eating. In addition, there were wide ranges of balance abilities and of walking speeds. The great majority of participants were independent walkers indoors, and more than half normally used a walking aid.
|
The MMSE21 is a screening tool that includes 11 questions in 6 sections, each representing a different cognitive domain or function (orientation, registration, attention and calculation, recall, language, and copying). The maximum score is 30. A score of 23 points or less has been considered to be evidence of cognitive impairment, scores between 18 and 23 points indicate mild impairment, and scores of 17 or less indicate severe impairment.23,28 We chose a cutoff for MMSE scores of 10 for inclusion in the study. Our reason for using this criterion was based on our clinical experience that people with very low MMSE scores have difficulties in following instructions. This reasoning also was in accordance with the findings of Mozley et al,29 who reported this level to be important for the ability to communicate and interact in a meaningful way.
In order to present a more complete clinical description and understanding of the studied sample, the functional capacity of the subjects was assessed by use of the Berg Balance Scale, analysis of gait speed, and the Barthel Index of ADL, and their medical state was assessed by recording of diagnosis and prescribed drugs. The Berg Balance Scale is a measure of balance; it contains 14 different items, including sitting, standing with different bases of support or with eyes closed, turning, and picking an object up from the floor.24 The maximum score is 56, indicating good balance abilities for an older person. Gait speed (in meters per second) was measured by use of a stopwatch and a 2.4-m walk at a speed comfortable for the participants.25,26 Participants were instructed to start from standing, using their ordinary walking aid, and to walk to a goal approximately 4 to 5 m away. The distance of 2.4 m was marked discretely with tape on the wall. In addition, the physical therapists interviewed the staff at the facilities with regard to the residents’ ADL according to the Barthel Index. This instrument measures the degree of independence in 3 categories of everyday living: self-care (washing and showering, dressing, eating, and toilet visits), continence of bowel and bladder, and mobility (transfer to and from a chair, toilet, and a tub or shower; walking indoors; and climbing stairs). The maximum score of 20 indicates independence in personal ADL.27 A registered nurse at each facility completed a questionnaire regarding diagnosis and prescribed drugs. This questionnaire was reviewed by a specialist in geriatric medicine before completion of the diagnosis.
The residents participated in 3 test sessions of TUG on different days, all within 1 week: TUG A (day 1), TUG B (day 2), and TUG C (day 3). To avoid any change that might normally occur in a frail subject over a day, for example, as a result of medication or food intake, each session of TUG assessment was carried out at approximately the same time of day as the preceding session for that subject. The study physical therapists assessed TUG A and TUG B (intrarater reliability), and the local authority physical therapists assessed TUG C (interrater reliability). With this procedure, multiple paired observations were made.
In accordance with the directions given by Podsiadlo and Richardson,1 the TUG was performed twice in each test session, that is, 1 trial and 1 timed performance, with only a brief seated rest in between. The physical therapist first demonstrated how to perform the TUG and then instructed the participants to stand up from a chair, walk 3 m at a comfortable speed, cross a line on the floor, turn around, walk back, and sit down again. An armchair of standard height, approximately 45 cm, was used. The starting position was sitting with the back against the chair back, with arms resting on the arm support of the chair, and with the walking aid near at hand. Participants wore their usual indoor footwear and chose whether to use their walking aid when performing the TUG. No personal physical assistance was allowed. The physical therapist either was standing next to the chair or accompanied the participant if required for reasons of safety, such as in the case of gait difficulties. In the case of a slight hesitation from the participant on how to proceed with the TUG, the physical therapist gave additional verbal cuing, such as "turn around" or "sit down." The use of verbal cuing during the test was recorded in the test protocol as present or not present. The TUG was timed with an ordinary stopwatch. The start and stop of timing were fixed firmly. Timing started when the participant’s back was no longer in contact with the back of the chair. Timing stopped when their buttocks touched the seat of the chair again. The result, in seconds to 1 decimal place, was recorded.
The results of different test sessions were recorded on separate documents. Physical therapists were unaware of the previous results. The information transferred was limited to date and time of the previous test session as well as what kind of walking aid, if any, was used in TUG A. The same walking aid was used for all test sessions. The residents were given written and oral information about the study. The participants or their proxies, when appropriate because of cognitive impairment, gave their consent for participation.
Data Analysis
The differences in times for TUGs between test sessions were calculated by subtracting the TUG B score from the TUG A score (intrarater reliability) and by subtracting the TUG C score from the TUG A score (interrater reliability). These calculations would result in a negative time difference if the second performance (TUG B or TUG C) were slower than performance on TUG A and a positive difference if it were faster. The distribution of TUGs and the difference in times between TUGs were skewed; therefore, the data were transformed logarithmically before any further analyses.
Linear regression analysis was used to evaluate the effect of cognition on the difference in times between TUGs. The dependent variables were the difference between TUG A and TUG B for intrarater analyses and the difference between TUG A and TUG C for interrater analyses. The independent variables were the MMSE score and the MMSE score dichotomized (cutoffs between 17 and 18 and between 23 and 24). Analyses were done with and without adjustments for TUG A. The rationale behind the adjustment was that the time needed to perform the TUG was considered to be a possible confounder for the difference between 2 TUG performances.
Logistic regression analysis was used to evaluate the effect of cognitive function in the presence of verbal cuing given during the performance of the TUG. The dependent variable was verbal cuing (yes or no), and the independent variable was the MMSE score. Analyses were done with and without adjustments for the TUG score because the time needed to perform the TUG was considered to be a possible confounder for verbal cuing.
The reliability of TUG scores was analyzed by use of TUG A and TUG B for analyses within raters and TUG A and TUG C for analyses between raters. We chose to present the results of both relative reliability and absolute reliability.18–20 The relative reliability was analyzed by use of 2 different models of the ICC with a 95% confidence interval (CI).30 With ICC(1,1), it is assumed that all error is random measurement error. With ICC(3,1), it is assumed hat systematic error is not part of the measurement error. No systematic error is present when ICC(1,1) equals ICC(3,1).31 We calculated ICC(1,1) with a 1-way random-effects model for a single measure, and ICC(3,1) was calculated with a 2-way mixed model for a single measure. The absolute reliability was analyzed by use of the within-subject standard deviation (Sw) as described by Bland and Altman.32,33 The first step was to verify the distribution of the data by plotting the difference between 2 TUG measures against their mean. We found a relationship between the measurement error and the magnitude of the TUG scores; that is, heteroscedasticity was present (Fig. 1). In such a case, log-transformed data are used in the analyses.20,32,33 The Sw then was obtained from the square root of the within-subject residual mean square of a 1-way analysis of variance. The next step was to back-transform (antilog) this Sw value to the natural scale. The result was a ratio (antilogSw) that was used to calculate the variability of the TUG scores. Multiplying on the natural scale is equivalent to adding on the log scale, whereas dividing is equivalent to subtracting.33 The difference between a subject’s measurement and the true value would be expected to be less than antilogSw1.96 for 95% of the observations. The difference between 2 measurements for the same subject would be expected to be less than antilogSw2.77 for 95% of pairs of observations. Finally, to estimate the true value from the expected variability on either side of any observed value, the observed value was divided (lower bound) and multiplied (upper bound) by antilogSw1.96. Likewise, to estimate repeatability (test-retest), the observed value was divided and multiplied by antilogSw2.77.32,33
|
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
The time needed to perform the TUG ranged from 9.8 to 109.2 seconds. The distribution was skewed, with a mean and a median of approximately 30 and 25 seconds, respectively. The results are shown in Table 2.
|
Verbal cuing during testing was given to 77% of the participants in TUG A, to 68% in TUG B, and to 60% in TUG C. In the analyses without adjustment for possible confounding by the time needed to perform TUG, the MMSE scores were significantly related to participants who were cued in TUG B and TUG C (P=.038 and P=.002, respectively) but not in TUG A (P=.101). When the analyses were adjusted for the time needed to perform the TUG, it was shown that a slow performer was more likely to be cued and that the cognitive state was of importance only on day 3 (day 1: TUG A P=.010, MMSE P=.217; day 2: TUG B P=.055, MMSE P=.090; day 3: TUG C P=.023, MMSE P=.006).
Reliability
The median and interquartile range (Q1–Q3) of the difference between test sessions were 0.2 and –2.1 to 3.2 seconds for intrarater measurements, respectively; these values were 0.6 and –3.1 to 3.4 seconds for interrater measurements, respectively (Tab. 2). Relative reliability between test sessions reached ICC(1.1) values of .92 (95% CI=.86–.95) for intrarater measurements and .91 (95% CI=.86–.94) for interrater measurements. The corresponding values for ICC(3.1) were .91 (95% CI=.87–.94) and .91 (95% CI=.86–.94). Thus, no systematic error was present. Figure 1 shows a plot of the mean difference (in seconds) between TUG A scores and TUG B scores before log transformation.
The antilogSw1.96 within raters was 1.3507, and that between raters was 1.3488; the antilogSw2.77 within raters was 1.5164, and that between raters was 1.5135. These values were used to calculate the upper and lower bounds of the expected variability for any observed value. For example, the expected true value for a person who performs the TUG in 20 seconds is (20/1.3507)=14.8 seconds for the lower bound and (20x1.3507)=27 seconds for the upper bound; for a TUG performed in 30 seconds, the lower bound is 22.2 seconds and the upper bound is 40.5 seconds, and for a TUG performed in 60 seconds, the lower bound is 44.4 seconds and the upper bound is 81 seconds. Consequently, the expected ranges of the variability for repeated measurements are between (20/1.5164)=13.2 seconds and (20x1.5164)=30.3 seconds for a TUG performed in 20 seconds, between 19.8 and 45.5 seconds for a TUG performed in 30 seconds, and between 39.6 and 91 seconds for a TUG performed in 60 seconds. The results of analyses of the variabilities expected for single and repeated measurements are shown in Figure 2 and Table 3. The variabilities within raters and between raters were similar (interrater data not shown in Fig. 2).
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
Older people living in residential care facilities are often frail. They have multiple diagnoses, their intake of medications is high, they lack general strength, and they are susceptible to sickness and disease. This multisystem involvement most certainly provides sources of variability and can be attributed to a condition of unstable disability or frailty. Campbell and Buchner described frailty as "a condition or syndrome which results from a multisystem reduction in reserve capacity to the extent that a number of physiological systems are close to, or past, the threshold of symptomatic clinical failure."34(p315) Frailty is considered to be the root cause of unstable disability.34 Movement involves an interaction between multiple systems, such as sensory input, neuromuscular function, and cognitive processes, and requires an ongoing adaptation to tasks and environment, anticipation, and planning. Despite the fact that the TUG consists of basic everyday movements, its various components can be complicated.
First, to be able to rise from a sitting position to a standing position requires both strength and technique. Walking a path for 3 m includes both acceleration and deceleration as participants prepare for a turn. The turning sequence is challenging for older people with balance disorders.35 Finally, turning around to sit down challenges both balance and orientation in adapting the body position to the chair. It has been suggested that a within-subject variation of 15% may be considered acceptable when repeated measurements of physical performance are obtained for young subjects who are healthy36 as well as middle-aged people with neck or back pain symptoms and middle-aged people who are healthy.37,38 An individual patient variability of up to 25% has been found for gait speed in patients 2 to 6 years after stroke.39 Our participants had an even higher degree of variability. This variability could be a consequence of the participants’ high degree of unstable disability and the combination of numerous sequential movements in the TUG.
The ICC is strongly affected by the range of scores used to calculate the coefficient; ICC is high when the difference in scores between measurements is small in comparison with the range of scores between the studied participants.18–20 There is no universal agreement on the interpretation of a correlation coefficient. A variety of guidelines are suggested in the literature: >.75 equals "excellent reliability,"30 .80 is "very reliable,"40 and >.75 indicates "good reliability."41 Rockwood et al13 reported ICCs of .56 for people with cognitive impairment and .50 for people without cognitive impairment. However, several factors may have contributed to these results, as there was an average of 112±72 days between test and retest, a design not well suited for evaluating the repeatability of a measure. Thomas and Hageman14 reported an ICC of .87 for their small sample (n=9) of subjects with cognitive impairment. Our analyses showed high ICC(1.1) values of .91 and .92, with the lowest bound of the 95% CI being .86; there was, however, a wide range of variability in absolute figures. We wanted to provide clinicians with means to interpret the results of TUG performance at an individual level. Our results for a single measure, antilogSw1.96, and for repeated measures, antilogSw2.77, can be used to calculate the expected variability of any observed value, as described in the "Results" section, for a frail older person dependent in ADL and living in a residential care facility. The exact level of acceptance for measurement errors seems to be impossible to state. It depends on the purpose of the measurement; therefore, determination of the reliability of a measurement is a clinical decision. Conclusions regarding assessments cannot be made without knowledge of the expected range of variability of scores for the TUG or any other instrument.
For clinical individual-based interpretation of an intervention, the results should exceed the boundaries of variability found in the present study to be interpreted as a true change. As an example, we can apply our findings to an individual patient who performs the TUG in 50 seconds. We can expect this person’s true value to be somewhere between 37 and 67.5 seconds for a single measurement. If our intention is to use the TUG measure before treatment and after treatment for this subject, the final assessment would have to be under 33 seconds in order to be interpreted as a true positive change or above 75.8 seconds to be interpreted as a true negative change. The size of this range is so wide that it would seem impossible to expect such an improvement for a person who requires this amount of time to perform the TUG. Thus, the use of the TUG for the evaluation of treatment of an individual patient, regardless of the cognitive level, is questionable when the performance is slow. However, the TUG still could be used to provide a description of the patient’s ability, indicated by the time needed to perform the TUG.
The TUG is easily applicable even for a person with cognitive impairment. We allowed and recorded additional verbal cuing during TUG performance, as we assumed that some subjects with cognitive dysfunctions would need such cuing. Instead, we found that it was participants with slower TUG performance who were more likely to be cued at the first 2 sessions. It was not until the third test session that cognitive function, as measured with the MMSE, was related independently to the presence of verbal cuing. Therefore, despite the fact that participants were tested for the third time within 1 week, those with lower cognitive functions lacked recognition of the situation, whereas participants with higher cognitive levels, at that time, most likely gave an impression of confidence in performing the TUG.
The TUG has been recommended as a screening tool for identifying older people who are at risk for falling.2,3 The TUG times from 13 to 16 seconds have been reported as cutoff values for predicting falls in community-dwelling older people.42,43 With regard to our findings of variability, it seems to be difficult to state a fixed time limit as a cutoff between a person who is at risk for falling and a person who is not at risk for falling when the TUG is used at an individual level in a population of frail older people dependent in ADL and living in residential care facilities. Even with a fast TUG performance of 10 seconds, the expected true value could be within a range of 7.4 to 13.5 seconds. The causes of falls are a complex phenomenon, and the characteristics of our participants reveal that, in many ways, they are at high risk of sustaining falls. Therefore, the usefulness of a certain cutoff for the TUG could be questioned, but this notion warrants further investigation for this population.
A way in which to minimize or stabilize expected variability in physical performance assessments may be to use the mean of several repeated measurements. The mean of 3 repeated measurements of manually timed gait, assessed at the same test session, was reported to represent the highest level of reliability in a frail older population.44 However, this conclusion was based on ICC statistics. Therefore, the degree of variability remains unknown to the reader, and further studies are warranted.
The methodological framework was considered in order to create optimal conditions for the evaluation. Instructions on how to perform the TUG were thorough, and all physical therapists involved in testing procedures were trained before the assessments. We managed to conduct the retest at approximately the same time as the preceding test. Multiple pairs of therapists made the assessments, and the physical therapists were unaware of the previous results. This strategy acknowledged that the rater is a component of the test and allowed the subject-rater source of variability to influence the results. This strategy reflects the use of the tool in the clinical setting. Thus, our results reflect a combination of instrument errors, tester errors, and true subject variability. We consider these design conditions to be important for the clinical interpretation of TUG performance.
We allowed verbal cuing during TUG performance because, on the basis of our clinical experience, we assumed that a person with cognitive impairment might need such cuing to accomplish the task. Consequently, cuing could be considered to be compensatory for cognitive impairment and thus possibly necessary for the valid use of the TUG for people with cognitive deficits. However, our results revealed that cuing also was given to slow performers and thus may have been used as an encouragement to proceed in the test. The use of "subtle verbal guidance" during TUG assessments in people with dementia has been mentioned by Thomas and Hageman.14(p22) In contrast, the occurrence of cuing in TUG assessments has not been reported elsewhere; therefore, we do not know whether and to what extent cuing is used in reality. In the present study, we simply recorded whether cuing was present or not; therefore, we cannot provide an exact description of the cues that were actually used. We can conclude only that verbal cuing was sometimes required for the assessment of TUG scores in this population. Whether such cuing affects the validity of TUG scores warrants further evaluation.
We believe that our findings are representative for a population of frail older people with multiple impairments, dependent in ADL, and living in residential care facilities. However, the present study did not include participants with MMSE scores of <10. Thus, our findings are not applicable to people with MMSE scores of <10.
| Conclusion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
| Footnotes |
|---|
All authors provided concept/idea/research design data collection and analysis, project management, and facilities/equipment. Ms Nordin provided writing. Dr Lundin-Olsson provided subjects and institutional liaisons. Mr Rosendahl and Dr Lundin-Olsson provided consultation (including review of manuscript before submission). The authors thank their colleagues who participated in the testing procedure for their valuable contribution to this study.
This study was approved by the Ethics Committee of the Medical Faculty of Umea University.
This investigation was supported by grants from the Aldrecentrum Vasterbotten, the Minnesfond of the Registered Physiotherapy Association, the Erik and Anne-Marie Detlof’s Foundation of Umea University, the Swedish Council for Working Life and Social Research, and the Osterman Research Foundation.
This research was presented as a poster presentation at the International Congress of The World Confederation for Physical Therapy, June 7–12, 2003, Barcelona, Spain, and as an oral presentation at a national conference for physical therapists arranged by the Swedish Association of Registered Physiotherapists, October 15–17, 2003, Stockholm, Sweden.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Nordin, N. Lindelof, E. Rosendahl, J. Jensen, and L. Lundin-olsson Prognostic validity of the Timed Up-and-Go test, a modified Get-Up-and-Go test, staff's global judgement and fall history in evaluating fall risk in residential care facilities Age Ageing, July 1, 2008; 37(4): 442 - 448. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E Busse, H. Khalil, L. Quinn, and A. E Rosser Physical Therapy Intervention for People With Huntington Disease Physical Therapy, July 1, 2008; 88(7): 820 - 831. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Steffen and M. Seney Test-Retest Reliability and Minimal Detectable Change on Balance and Ambulation Tests, the 36-Item Short-Form Health Survey, and the Unified Parkinson Disease Rating Scale in People With Parkinsonism Physical Therapy, June 1, 2008; 88(6): 733 - 746. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Netz, S. Axelrad, and E. Argov Group physical activity for demented older adults feasibility and effectiveness Clinical Rehabilitation, November 1, 2007; 21(11): 977 - 986. [Abstract] [PDF]
|
|
|
|
 |
|
 D. Sumukadas MBBS MD, M. D. Witham MBChB PhD, A. D. Struthers MBChB MD, and M. E.T. McMurdo MBChB MD Effect of perindopril on physical function in elderly people with functional impairment: a randomized controlled trial Can. Med. Assoc. J., October 9, 2007; 177(8): 867 - 874. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Conradsson, L. Lundin-Olsson, N. Lindelof, H. Littbrand, L. Malmqvist, Y. Gustafson, and E. Rosendahl Berg Balance Scale: Intrarater Test-Retest Reliability Among Older People Dependent in Activities of Daily Living and Living in Residential Care Facilities Physical Therapy, September 1, 2007; 87(9): 1155 - 1163. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
