| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
GL Creel, PT, MHS, NCS, is Lecturer, Department of Physical Therapy, College of Health Professions, University of Florida, PO Box 100154, Gainesville, FL 32610 (USA) (gcreel{at}hp.ufl.edu). Address all correspondence to Ms Creel
KE Light, PT, PhD, is Associate Professor, Department of Physical Therapy, College of Health Professions, University of Florida
MT Thigpen, PT, MHS, NCS, is Doctoral Student, Departments of Physical Therapy and Exercise and Sports Science, University of Florida
Ms Creel and Ms Thigpen were graduate students in the Department of Physical Therapy at the University of Florida during this study, which was undertaken in partial fulfillment of the requirements for their Master of Health Sciences degrees
Submitted August 23, 1999;
Accepted August 4, 2000
 | Abstract |
|---|
Key Words: Activities of daily living • Elderly • Mobility • Timed movement
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussionConclusionAppendixReferences |
|---|
The Timed Movement Battery (TMB) is a new assessment tool designed to measure the mobility of older people.13 It was developed in an effort to provide a measure of mobility that has validity, reliability, and sensitivity for elderly individuals with a broad spectrum of mobility. The TMB incorporates common, everyday movements that are associated with the performance of basic activities of daily living (BADL) and instrumental activities of daily living (IADL), such as supine-to-sit, sit-to-stand, and forward walking.1,2,4–7,9,13–15 The remaining movements in the TMB were chosen because it has been reported that elderly patients with declining independence in ambulation (eg, turning, stepping over objects, backward walking) have problems performing them.13,16 Basic activities of daily living include moving from a bed to a chair and performing self-care activities (eg, bathing, toileting, dressing, feeding, locomotion).1,2,4,17 Instrumental activities of daily living include those tasks necessary for higher levels of independent function that involve interaction with the environment, such as using the telephone, managing money and medications, cooking and cleaning, and using public transportation.4,6
The TMB consists of 11 movements performed at 2 speeds. The movements are supine-to-sit, sit-to-supine, sit-to-stand, forward ambulation for 6 m (20 ft), backward ambulation for 3 m (10 ft), walking and stepping over a 5.1-cm (2-in) obstacle, walking and stepping over a 15.2-cm (6-in) obstacle, a figure-8 walk, ascending 4 steps, descending 4 steps, and moving from a supine-on-the-floor position to a standing position. Originally, the battery also included side-stepping for 3 m to the left and 3 m to the right.13,14 The TMB has since been modified, and side-stepping has been replaced with stepping over 5.1- and 15.2-cm obstacles.
For each movement, subjects are instructed to move at their normal or customary pace (ie, self-selected speed) for one trial. Subjects are then asked to perform each movement as quickly as they can safely perform the task (ie, their "maximum-movement" speed). Table 1 describes the starting and stopping points for the TMB. Subjects assume the start position and are verbally cued to begin the movement. Timing of the movement is initiated simultaneously with the start signal and stopped as described in Table 1.
|
.97 for all movements except the self-selected speed of the supine-to-sit movement.13 The TMB was also effective in differentiating among 3 groups of elderly subjects according to their diagnoses in the balance clinic and history of falls.14 The purpose of our study was to further explore the construct and concurrent validity of scores obtained with the TMB as a measure of mobility in elderly individuals. Construct validity is the degree to which an instrument measures the theoretical construct it was designed to measure.18 We first defined the underlying traits of mobility in elderly individuals, because these traits cannot be measured directly. We were interested in those aspects of mobility that allow elderly people to maintain their independence in BADL and IADL. We defined the construct of mobility in elderly individuals as having at least 3 dimensions: (1) adequate mobility to perform the movements required for BADL, (2) the ability to perform these movements at speeds that are conducive to independence in their environment (timed mobility), and (3) the ability to maintain stability in an upright position under a variety of conditions (balance).
Concurrent validity is the degree of agreement between a new measure and another measure for which validity has already been established.18 We selected at least one criterion measure for each of our 3 defined dimensions of mobility in elderly people (ADL, timed mobility, and balance). Because we lack a "gold standard" for any of these dimensions, we chose 4 measures with some published evidence of validity and reliability for elderly subjects (aged 60 years or older), but we realize that these measures have some error associated with them, which limits inferential use.
For the dimension of ADL, we chose 2 measures, the Barthel Index (BI)19 and an 18-item ADL/IADL scale.5 The BI was developed to monitor functional independence in people who require long-term rehabilitation and has been used extensively in research as a measure of ADL. The BI is simple to administer, and the reliability and validity of data obtained with the BI have been supported in several studies with a variety of patient populations.20–24 The BI is frequently used as a criterion measure in the validation of new ADL measures.24–27
Limitations of the BI include a ceiling effect and insensitivity to change at the upper end of the scale in patients with neurologic disorders.28,29 More specific to the aging population, Yohannes et al25 found that the BI failed to distinguish clinically between elderly subjects without disabilities and subjects who have a disability resulting from chronic airflow limitation. To address this problem, the authors recommended a scale that incorporates IADL as well as BADL to identify difficulties in individuals who have higher functioning. The 18-item ADL/IADL scale recommended by Kempen and Suurmeijer5 is such a scale. Because of the expected ceiling effect of the BI in our sample population, we chose to use this ADL/IADL scale as an additional measure that would more fully address the abilities of our subjects.
This 18-item ADL/IADL scale was developed for use with noninstitutionalized elderly people in an effort to increase the range of measurement with regard to the needs of this population (Appendix). This scale is based on the work of Kempen and Suurmeijer,5 who identified the ADL and IADL items that measure different levels of difficulty and provide an accurate measure of disability in this population. One variation they proposed is to use 3 responses to the questions about the subject's abilities (ie, can perform independently and easily, independent with difficulty, requires assistance or supervision) rather than only 2 responses (able to perform independently, not able to perform independently). This variation has recently been recognized as a potentially important step in identifying those individuals who have begun to demonstrate a gradual decline in their ability to independently accomplish their daily tasks.29,30
We chose the Berg Balance Scale (BBS) as a criterion measure for the dimension of balance based on indications that it has some validity as a measure of balance. Reliability in populations similar to our sample has also been documented.31 The BBS consists of 14 activities that are commonly performed in everyday life. The intrarater and interrater reliability of the BBS scores was very high (ICC [2,3]=.98) in their sample of 31 subjects (mean age=83 years, age range not reported).31 Initial validation studies of elderly individuals (N=113) indicated that the BBS scores differentiated between those who required assistive devices for gait and those who did not require an assistive device.32 These studies also indicated that the BBS scores differentiated among the living situations (hospital, rehabilitation hospital, or home) of subjects 12 weeks after stroke. Shumway-Cook et al33 found the BBS to be the best single predictor of falls in a cohort of community-dwelling elderly adults (aged 65 years or older). Newton34 recommended the BBS for screening this population despite the possibility of a ceiling effect in those groups that are not homebound.
We chose the Timed "Up & Go" Test (TUGT) as a measure of timed mobility. This test requires the subject to stand up from a standard armchair (seat height=41 cm), walk 3 m, turn around, walk back to the chair, and be seated. The subject is timed from the command "go" until he or she returns to a seated position in the chair.35 Podsiadlo and Richardson36 modified the "Up & Go Test"35 by changing the scoring system from a scale of 1 to 5 based on the evaluator's assessment of the subject's gait stability to the time it took to complete the activity. The validity and reliability of scores obtained with this measure were investigated with a cohort of elderly people with poor-to-fair balance abilities (mean age=73.8 years, range=52–94). The results showed good correlation of the TUGT with the BBS (r=–.81) and the BI (r=–.78). Intrarater and interrater reliability of the TUGT scores were reported as very high (ICC=.99) in this same cohort. Many recently developed measures of physical ability use timing of tasks to improve objectivity of the scoring.31,33,36–38 Timing of task performance has been found to provide greater discrimination of differences in ability, especially in elderly subjects with higher functional abilities.38
By examining the subjects using these measures, we wanted to address the following questions:
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
Subjects were 17 women with a mean age of 79.4 years (SD=6.3, range=69–92) and 13 men with a mean age of 75 years (SD=7.2, range=65–92). Further descriptive data on the subjects are presented in Table 2. All subjects completed all aspects of the testing with the exception of 2 movements. Eight of the subjects refused or were unable to go from being supine on the floor to standing in the TMB. One of these subjects also refused to ascend and descend the 4 steps at maximum-movement speed. Three subjects refused to go from being supine on the floor to standing at maximum-movement speed. These refusals occurred because the subjects feared injuring themselves or were unable to perform the movement. There were no incidences of injury during testing, and subjects reported only minimal to moderate fatigue following completion of data collection.
|
Raters
Two physical therapists collected all of the data. Both raters had at least 10 years of clinical experience treating geriatric clients. They had previous experience using the BBS and TUGT in the clinic; however, neither rater had prior experience using the TMB. Interrater reliability on the BBS was measured on a pilot sample of 11 subjects using the kappa statistic (
), a chance-corrected measure of agreement for reliability between raters when the data are ordinal.41 Our coefficient of agreement () was .80, which we believe represents excellent agreement for this type of measure. Our interrater reliability on the TUGT had an intraclass correlation coefficient (ICC [3,2]) of .99 and an intrarater ICC (3,2) of .98. Following training in the use of the TMB, the raters demonstrated ICCs (3,2) of .99 for interrater reliability and .99 for intrarater reliability testing for the battery of timed movements. The BI was completed as a self-report by our subjects. The interrater reliability (r) of data obtained with this method compared with observation is .88 (P<.001).23 The interrater and intrarater reliability (r) of data obtained with the Folstein Mini Mental Exam has been reported as .80 to .95 in a variety of sample groups.39,40
Procedure
Informed consent was obtained from each participant. We first completed the Folstein Mini-Mental Exam to screen volunteers for eligibility. We then collected demographic data and measured the subjects' initial heart rate and blood pressure to establish a baseline of physiologic function prior to testing. Each participant then completed a self-report version of the BI.25
After completing the BI, the subjects were asked to complete the 3 performance-based tests (BBS, TUGT, and TMB) in 1 of 6 randomly assigned sequences. Movements of the TMB were performed in random order to control for fatigue. Subjects completed 3 trials of the TUGT and the supine-to-sit movement at the self-selected speed in the TMB. The means of these trials were used in our analysis. Because all other movements in the TMB demonstrated high test-retest reliability,13 they were performed once at self-selected speed and once at maximum-movement speed. All timed activities were scored to the nearest hundredth of a second.
After each movement series, subjects were allowed to rest for 1 to 3 minutes, as needed. There was a mandatory rest period of at least 3 minutes between the performance measures to allow for recovery from possible fatigue.
Following completion of all the physical performance measures, the subjects were administered the 18-item ADL/IADL scale (questionnaire) by interview. We monitored the subjects' heart rate and blood pressure at various intervals during the testing and after the final rest period to ensure safety of the subjects and their tolerance to activity. Completion of all data collection required between 60 and 120 minutes per subject.
Data Analysis
The Spearman rank correlation coefficient is a nonparametric analog of the Pearson product moment correlation coefficient and is used for ordinal data.42 Spearman rank correlations were calculated using the subjects' scores on the BI, the BBS, the 18-item ADL/IADL scale, and the individual and composite scores on the TMB. A composite TMB score represents the sum of 10 individual movement times: sit-to-supine and supine-to-sit, sit-to-stand, stepping over 5-cm and 15-cm (2-in and 6-in) obstacles, backward and forward ambulation, up and down stairs, and the figure-8 walk. Scores on the floor-to-stand movement were not included in the composite TMB score because all subjects were not able to complete this movement. Separate correlations were calculated on the self-selected and maximum-movement speeds of the TMB.
We analyzed the relationship between the TUGT and the individual and composite scores of the TMB using the Pearson product moment correlation coefficient. In this case, both sets of data were ratios as opposed to a combination of ratios and ordinal data as in the BBS, the BI, and the 18-item ADL/IADL scale.
An independent-samples t test was used to determine whether the composite scores on the TMB between those subjects who reported difficulty with ADL/IADL and those who reported no difficulty with ADL/IADL were different. We used a significance level of P<.01 for our statistical analyses.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
|
|
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
In our study, we targeted elderly people who had moderate or no difficulties in performing ADL. Eighteen of our subjects achieved the maximum score of 100 on the BI. The lower correlations between the TMB scores and the BI scores may be related to the lack of sensitivity of the BI and the relatively high level of ability in our subjects. These findings are in agreement with those of Seeman et al,38 who reported improved sensitivity in testing elderly subjects with minimal or no functional deficits when standardized performance measures rather than self-report indexes were used. The third criterion test, the BBS, addresses the balance component of our construct of mobility in the elderly. The composite TMB scores show what we would argue are good correlations with our subjects' scores on the BBS (Tab. 3).
The correlations of the individual movements of the TMB with the criterion tests are not as robust as the correlations between the composite TMB scores and the scores on the criterion tests (Tabs. 3–5). The higher correlations between the composite TMB scores and the scores on the criterion tests is indicative of the need to perform the full battery of movements in order to get the most complete picture of the individual's abilities. The correlations between the individual movements of the TMB and the BI were consistently lower than the correlations between the individual movements and the other criterion tests, with the exception of 2 movements. The sit-to-supine and supine-to-sit movements, when performed at self-selected speeds, demonstrated higher correlations with the BI than with the BBS or the 18-item ADL/IADL scale. Including these 2 movements in the TMB may be very important in achieving the goal of having a tool that is applicable to people with lower levels of ability as well as those with higher levels of ability.
Analysis of the TUGT and TMB scores allowed the use of the Pearson product moment correlation because both tests are scored by ratio data (time). Pearson product moment correlations between the TUGT scores and the composite TMB scores were, in our view, very good to excellent for self-selected speed (r=.89) and maximum-movement speed (r=.79). Fifty percent of the individual movements of the TMB performed at self-selected speed demonstrated very good to excellent correlation with the TUGT scores (r=.81–.92). Forty percent of the movements of the TMB performed at maximum-movement speed correlated highly with the TUGT scores (r=.77–.84). These high correlations provide further evidence for the validity of scores on the TMB as a measure of mobility in elderly people. The supine-to-sit and sit-to-supine movements demonstrated the lowest Pearson product moment correlations with the TUGT scores. We believe this result is consistent with the idea that these particular movements are appropriate for testing individuals with lower functional levels than those targeted for this study.
One goal established by the Research Plan for the National Center for Medical Rehabilitation in 1993 was to "develop more accurate and sensitive methods for quantifying slight to moderate functional impairments."43 Another goal was to "develop practical methods for the systematic evaluation of determinants of mobility in order to quantify the degree of functional impairment in the performance of routine daily tasks."43 We believe that the TMB begins to address these issues in a number of ways. First, the TMB involves movements that are required in the performance of daily activities. Second, these movements incorporate various aspects of the determinants of mobility such as balance and speed of movement. Third, measuring subjects' performance at their self-selected and maximum-movement speeds allows us to quantify their mobility and begin to perform examinations appropriate to their level of function. Whether the measure truly reflects function, however, awaits further research.
The time needed to complete physical tasks has most frequently been tested while the subject performs at preferred or customary speed. In few studies do subjects perform tasks at the fastest possible speed.33,38 In the TMB, we have added this requirement. Having individuals perform at their fastest possible speed, we believe, greatly diminishes the problem of a ceiling effect that is typical of measures of physical performance. Another important concept is "reserve speed" or the ability to perform these activities at speeds faster than a person's usual (self-selected) pace. The impact of "reserve speed" on an individual's ability to safely function is an area that needs further research.
Although we believe that our findings provide preliminary evidence regarding the construct and concurrent validity of scores on the TMB, we recognize there are limitations that must be considered when summarizing the overall implications. First, our data on one of the movements of the TMB (floor-to-stand) were incomplete and thus were not included in the composite scores. Although this is unfortunate and raises questions regarding how inclusion of these movement times might influence the results, it also indicates the inability or hesitancy of this population to perform this activity. The possibility exists that, upon further investigation, this movement may be highly discriminatory for identifying those individuals who are just beginning to have difficulty with ADL.
Another limitation of our study was the use of only 2 raters. Although this decision was necessary due to the logistics of our study, it limits our ability to make generalizations to a larger population of raters. Furthermore, as is always the case in the use of human subjects, the subjects' performance in the laboratory may have been influenced by their personality or motivation level. In addition, our choice of criterion tests may have limited our ability to make definitive conclusions. For example, although the 18-item ADL/IADL scale was more sensitive to varying levels of independence in ADL than the BI, there was still a ceiling effect with this test in our subjects. This example illustrates the problem we are trying to address: the lack of sensitivity in the tools that are available at present for assessing ADL in this population. Further studies are needed to investigate the TMB's performance in relation to other measures of mobility. Some suggested measures include the Physical Performance Test,44 the physical functioning section of the Medical Outcomes Study 36-Item Short Form Health Survey,45 and the Functional Assessment Screen.11 More rigorous analysis of the individual items of the TMB would be beneficial in determining whether the movements or movements with specific discriminatory attributes are redundant and to what extent the measure predicts dysfunction and disability.
| Conclusion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
| Appendix |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
|
| Footnotes |
|---|
This study was approved by the University of Florida Health Center Institutional Review Board.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. P.J. Huijbregts, G. F. Teare, C. McCullough, T. M. Kay, D. Streiner, S. K.C. Wong, S. E. McEwen, and I. Otten Standardization of the Continuing Care Activity Measure: A Multicenter Study to Assess Reliability, Validity, and Ability to Measure Change Physical Therapy, June 1, 2009; 89(6): 546 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L Behrman, A. R Lawless-Dixon, S. B Davis, M. G Bowden, P. Nair, C. Phadke, E. M Hannold, P. Plummer, and S. J Harkema Locomotor Training Progression and Outcomes After Incomplete Spinal Cord Injury Physical Therapy, December 1, 2005; 85(12): 1356 - 1371. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. V Bastille and K. M Gill-Body A Yoga-Based Exercise Program for People With Chronic Poststroke Hemiparesis Physical Therapy, January 1, 2004; 84(1): 33 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Carmeli, S. Bar-Chad, M. Lotan, J. Merrick, and R. Coleman Five Clinical Tests to Assess Balance Following Ball Exercises and Treadmill Training in Adult Persons With Intellectual Disability J Gerontol A Biol Sci Med Sci, August 1, 2003; 58(8): M767 - M772. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
