Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 2 , Pages 296-301, February 2009

Timed Walking Tests Correlate With Daily Step Activity In Persons With Stroke

Presented in part to the Australian Physiotherapy Association Conference, October 4–6, 2007, Cairns, Australia, and the 5th World Congress for NeuroRehabilitation, September 24–27, 2008, Brasilia, Brazil.

  • Suzie Mudge, MHSc

      Affiliations

    • Corresponding Author InformationReprint requests to Suzie Mudge, MHSc, Department of Surgery, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand
    • ,
  • N. Susan Stott, PhD

Department of Surgery, University of Auckland, Auckland, New Zealand

Article Outline

Abstract 

Mudge S, Stott NS. Timed walking tests correlate with daily step activity in persons with stroke.

Objectives

To examine the relationship among 4 clinical measures of walking ability and the outputs of the StepWatch Activity Monitor in participants with stroke.

Design

Correlational study.

Setting

Clinic and participants' usual environments.

Participants

Fifty participants more than 6 months after stroke were recruited. All participants were able to walk independently, but with some residual difficulty.

Interventions

Not applicable.

Main Outcome Measures

Rivermead Mobility Index (RMI), Rivermead Motor Assessment (RMA), six-minute walk test (6MWT), ten-meter walk test (10MWT), StepWatch outputs (based on daily step counts and stepping rates).

Results

The correlations between the RMA and all StepWatch outputs were low (ρ=0.36–0.48; P<.05), as were most for the RMI (ρ=0.31–0.52; P<.05). The 10MWT and 6MWT had moderate to high correlations (ρ=0.51–0.73; P<.01) with most StepWatch outputs. Multiple regression showed that the 6MWT was the only significant predictor for most StepWatch outputs, accounting for between 38% and 54% of the variance. Age and the RMI were further significant predictors of 1 and 2 outputs, respectively.

Conclusions

The 6MWT has the strongest relationship with the StepWatch outputs and may be a better test than the 10MWT to predict usual walking performance. However, it should be remembered that the 6MWT explains only half the variability in usual walking performance. Thus, activity monitoring captures aspects of walking performance not captured by other clinical tests and should be considered as an additional outcome measure in stroke rehabilitation.

Key Words: Motor activity, Walking, Rehabilitation, Stroke

List of Abbreviations: 6MWT, six-minute walk test, 10MWT, ten-meter walk test, RMA, Rivermead motor assessment, RMI, Rivermead mobility index, SF-36, 36-item Short Form Health Survey

 

STROKE IS THE MOST common cause of severe disability in adults,1 with persistent physical disability reported in 50% to 65% of persons who survive stroke.1, 2, 3 Although as many as 70% are able to walk independently after rehabilitation,3, 4 it appears that only a small percentage of these persons are able to walk functionally in the community.5, 6 This difference may reflect a discrepancy between testing walking in a clinical environment and monitoring usual walking in natural environments as has been suggested by the International Classification of Functioning, Disability and Health.7

There is a range of clinical tests available to assess walking after stroke, many of which have good psychometric properties and assess wider aspects of gait thought to relate to walking in community environments.8 Some tests involve direct therapist observation of walking, of which an aspect is then graded or measured. Examples include the 10MWT,9 the 6MWT,10 and the RMA.11 Other outcome measures rely on patient self-report of usual function, such as the RMI12 and the Functional Ambulation Categories.13

The advantage of the directly observed tests is their standardized nature, but they may be more reflective of best performance than usual performance. For example, self-selected gait speed (measured by the 10MWT) is a global indicator of physical functioning14 and can discriminate between different categories of community ambulation.6 However, community ambulation can be achieved by persons with stroke who have low gait velocities, suggesting that gait velocity alone is not sufficient as a measure of community ambulation.6 Self-report measures, on the other hand, may ask about usual performance; however, they depend on the accuracy of a patient's perception, cognition, and communication.15, 16 Indeed, a recent study has shown that persons with stroke have a higher subjective report of physical activity and exercise than is found on objective testing.17

Activity monitors are one way of monitoring usual walking performance in natural environments because they can be worn during everyday activities over extended periods. The typical output is counts with respect to time, which can give information about amount, rate, and patterns of activity. An activity monitor that has been used to investigate ambulatory activity after stroke is the StepWatch Activity Monitor.18, 19, 20a The monitor contains a custom sensor that uses a combination of acceleration, position, and timing to determine the number and rate of steps taken. The StepWatch has been shown to have criterion validity18, 21 and is reliable19, 22 for step counting in persons with stroke. The output of the StepWatch is based on the number of steps taken on 1 leg, which is doubled to represent steps taken on both legs.19, 20, 23, 24 The most commonly reported output of the StepWatch, mean steps a day,19, 20, 23, 25 correlates moderately with self-selected gait speed (r=0.55)20 and scores on the Functional Independence Measure (r=0.62)23 and Berg Balance Scale (r=0.58)20 in patients with stroke. Recent research has also shown that mean steps a day shows a low correlation to peak exercise capacity (r=0.316)24 but is not related to self-reported fatigue severity24, 26 or economy of gait.24

Many other outputs of the StepWatch are available, which include calculations based on rate of stepping. The peak activity index is the average step rate of the fastest 30 minutes over 24 hours, regardless of when they occurred. Sustained activity measures are also available for 1, 5, 20, 30, and 60 minutes and are calculated by scanning the accumulated 24-hour data to determine the maximum number of steps taken during continuous intervals of 1, 5, 20, 30, and 60 minutes. The number of steps at high (above 60 steps/min), medium (between 30 and 60 steps/min), and low (below 30 steps/min) step rates can also be calculated. We have recently shown good test-retest reliability for a number of these additional outputs in persons with stroke, particularly peak activity index and maximum number of steps in 5 and 1 minutes.22 However, the relationship between commonly used clinical measures of walking ability and these additional StepWatch outputs has not been studied.

Thus, the aims of this study were to determine the strength of the relationship between commonly used clinical tests of walking ability and the available StepWatch outputs, and in particular to determine how well clinical walking tests predict ambulatory activity in natural environments as measured by the StepWatch. Self-selected gait speed was measured by the 10MWT, and gait endurance was measured by the 6MWT, both of which are used commonly8 and have good psychometric properties.10

We chose the RMI to capture self-reported mobility because 6 of the 15 items report on walking situations and it has good psychometric properties.8 The RMA was also selected because 5 of the 13 items directly test walking conditions.8 Both the RMI and RMA reflect a breadth of walking conditions, such as walking over uneven surfaces and walking outside, that are not evaluated by the commonly used timed walking tests. We hypothesized that performance during these common walking conditions may have a stronger relationship to usual walking activity in natural settings than do the timed walking tests.

Back to Article Outline

Methods 

Participants 

A power calculation based on mean steps a day (SD of 4390 steps a day) and the 6MWT (SD of 124m) from pilot data (n=16) suggested that a sample size of 24 would achieve 99% power (α=0.05)27 for a single correlation. To ensure adequate power for a multiple regression analysis, a convenience sample of 50 persons with chronic stroke was recruited based on formulae by Green28 (minimum of 46 participants for 5 predictors and estimated multiple correlation of 0.50). Participants were recruited from the hospital stroke service and local and newspaper advertising and were eligible for inclusion if they were at least 6 months poststroke and were able to walk independently, but with some residual difficulty, confirmed by a score of less than 2 on at least 1 of the walking items (a, d, e, g, h, or i) of the physical functioning scale of the SF-36.29 Participants also had to walk in the community at least once a week, determined by the response to the question, “How many times do you walk past your letterbox, on average in 1 week?” Persons were excluded if they had fallen more than twice in the previous 6 months, had another serious health problem affecting walking (eg, musculoskeletal or cardiovascular condition), or were unable to complete the testing for another reason (eg, inability to follow instructions).

Testing Protocol 

The study was approved by the Northern Regional Ethics Committee of New Zealand. All participants attended a rehabilitation clinic for initial testing and gave written informed consent. The clinical tests were administered by 1 examiner. The RMI is a self-report of ability to perform up to 15 mobility items, with answers given as either yes or no. The highest score of 15 indicates an ability to climb up and down 4 steps with no rail and run 10 meters. The RMA was tested in a clinic and outside environment, and patients were scored on each of the 13 items based on their ability to perform the mobility task. The maximum score of 13 indicates an ability to run 10 meters and hop on the affected leg 5 times. Self-selected gait speed was measured at a comfortable pace over 10 meters (10MWT), and gait endurance was tested by the 6MWT, both following standardized protocols.30

A StepWatch was calibrated and attached to the lateral side of the ankle of the nonparetic leg with a strap or cuff. The monitor has an infrared light that flashes with every step, which was matched to a manual count of steps during walking 5 meters at each of 3 walking speeds (fast, slow, and self-selected). The sensitivity and cadence settings were adjusted, if necessary, until the flashes corresponded exactly with the manual count during the 3 walking speeds.

Participants were instructed to wear the monitor for the next 3 days, removing it for sleeping and showering. Participants were given an instruction sheet with details about the care of the StepWatch, and a follow-up appointment was made to pick up the monitor. Data were exported to Excel,b where the number of steps detected over a 24-hour period was doubled to obtain steps a day for both legs. A subgroup of these patients (n=37) also agreed to participate in further data collection for a larger study of reliability testing.22

Statistical Analyses 

Variables were tested for normality using the Shapiro-Wilk statistic. The level of association between the variables was assessed using the Pearson correlation coefficient for normally distributed variables or the Spearman rank correlation coefficient for variables without a normal distribution, with significance accepted at the .05 level. A correlation above 0.90 was interpreted as very high, 0.70 to 0.89 as high, 0.50 to 0.69 as moderate, 0.30 to 0.49 as low, and less than 0.29 as little if any correlation.31 Age and sex were also tested for correlation with StepWatch outputs because they were potentially confounding factors. A forward linear multiple regression analysis was performed for each of the significant variables from the correlation entered as independent variables and the StepWatch outputs as the dependent variables. All calculations were performed using SPSS.c

Back to Article Outline

Results 

Fifty participants enrolled in the study. Forty-nine of the 50 participants, mean ± SD age of 67.4±12.5 years and 6 to 219 months after stroke, completed the study (table 1). The remaining participant did not have 3 complete days of data and therefore was excluded from the analysis. There were 29 men and 20 women. Eighteen participants had right-sided paresis. The median score on the physical functioning index of the SF-36 was 18 (range, 10–29), where the maximum score of 30 indicates no limitations with all items, including walking more than a mile, climbing several flights of stairs, and running, and a score of 10 indicates significant limitations with all items. All participants walked independently with an assistive device, if necessary. However, median scores on the RMI and RMA indicated that the participants had difficulty with higher-level mobility tasks such as running, hopping, and climbing up and down steps without a handrail. The mean steps a day showed a wide variation between participants from a low of 1225 steps a day to a high of 21,273 steps a day (see table 1). However, the median of 4765 steps a day in this study was lower than the 6565 steps a day reported by Bohannon32 for apparently healthy adults over 65 years.

Table 1. Study Sample Characteristics
Mean ± SDMedianRange
Demographics
Age (y)67.4±12.5 38–89
Months since stroke66±61 6–219
Physical functioning index (to/from SF-36) 1810–29
Clinical test
10MWT (m/s)0.67±0.32 0.12–1.42
6MWT (m)230±121 42–568
RMA 105–13
RMI 136–15
StepWatch outputs
Mean steps/d 47651225–21,273
Percentage of time with no steps 83%53–96
Number of steps at low rate (<30 steps/min)2334±565 493–5331
Number of steps at high rate (>60 steps/min) 6550–10,590
Peak activity index (steps/min)58.7±10.6 17–112
Highest step rate in 60 minutes (max 60) (steps/min) 18.75–89
Highest step rate in 1 minute (max 1) (steps/min)81.5±11.1 23–128

Only 2 clinical tests (10MWT and 6MWT) and 3 StepWatch outputs (number of steps at a low rate, peak activity index, and highest step rate in 1 minute) were distributed normally. Thus, most correlations shown in table 2 use the Spearman correlation coefficient. Sex showed no correlation with any of the StepWatch outputs but age showed a significant but low correlation (ρ=−0.33; P<.05) with number of steps at a high rate and highest step rate in 60 minutes. The correlations between the RMA and all the StepWatch outputs were less than 0.50, as were most for the RMI. There were 2 moderate correlations between StepWatch outputs and the RMI; mean steps a day was positively correlated (ρ=0.51; P<.01), and percentage of time with no steps was negatively correlated (ρ=−0.52; P<.01). The 10MWT had moderate correlations with most StepWatch outputs, with the highest step rate in 1 minute reaching a high level of correlation (r=0.71; P<.01). The 6MWT reached at least a moderate level of correlation with all StepWatch outputs, with peak activity index (r=0.72; P<.01) and highest step rate in 1 minute (r=0.73; P<.01) reaching a high level of correlation.

Table 2. Correlation Coefficient⁎ for StepWatch Outputs and Clinical Gait Tests and Age
StepWatch OutputRMIRMA10MWT6MWTAge
Mean steps/d0.510.470.550.67−0.29†
Percentage of time with no steps−0.52−0.47−0.41−0.57NS
Number of steps at low rate (<30 steps/min)0.470.440.46⁎0.58⁎NS
Number of steps at high rate (>60 steps/min)0.31†0.420.540.60−0.33†
Peak activity index0.370.400.64⁎0.72⁎−0.28†
Highest step rate in 60 minutes (maximum 60)0.460.480.510.59−0.33†
Highest step rate in 1 minute (maximum 1)0.36†0.410.71⁎0.73⁎NS

Abbreviation: NS, not significant.

⁎Use of Pearson correlation coefficient. All other correlations use Spearman correlation coefficient.

†Correlation is significant at the .05 level. All other correlations are significant at the .01 level.

Regression analysis using StepWatch outputs as the dependent variables and age, RMI, RMA, 10MWT, and 6MWT as the independent variables showed that for most StepWatch outputs, the 6MWT was the single most significant predictor (table 3). The 6MWT accounted for between 30% (for number of steps at a low rate) and 54% (for mean steps a day) of the variance in the StepWatch outputs. For 3 outputs (highest step rate in 60 minutes, percentage of time with no steps, and number of steps at a low rate), other variables made an independent contribution to the variance. Age made a significant contribution to the variance in highest step rate in 60 minutes over and above that of the 6MWT, increasing the explained variance from 44% to 49%. The 6MWT and the RMI independently contributed to both the percentage of time with no steps and the number of steps at a low rate. For the percentage of time with no steps, the addition of RMI increased the explained variance from 40% to 47%. For the number of steps at a low rate, the addition of the RMI increased the explained variance from 30% to 36%.

Table 3. Stepwise Linear Regression Models of Selected StepWatch Outputs
StepWatch Output/Predictors⁎Regression CoefficientsR2R2 ChangePAdjusted R2Constant
Mean steps/d
6MWT26.20.54 0.0000.53159.7
Percentage of time with no steps
6MWT0.0000.40 0.0000.380.92
6MWT and RMI0.000/−0.0140.460.060.0000.441.07
Number of steps at low rate (<30 steps/min)
6MWT5.390.33 0.0000.321092
6MWT and RMI3.92/186.50.410.080.0000.39−908.4
Number of steps at high rate (>60 steps/min)
6MWT12.20.46 0.0000.45−625.5
Peak activity index (steps/min)
6MWT0.1260.51 0.0000.5029.7
Highest step rate in 60 minutes (max 60; steps/min)
6MWT0.0900.44 0.0000.434.05
6MWT and Age0.082/−0.3120.490.050.0000.4726.6
Highest step rate in 1 minute (max 1; steps/min)
6MWT0.1360.54 0.0000.5350.3

⁎Predictors indicated in italics.

Back to Article Outline

Discussion 

The aims of this study were to determine the strength of the relationship between commonly used clinical tests of walking ability and the available StepWatch outputs, and in particular to determine how well clinical walking tests predict ambulatory activity in natural environments. We found that both the 10MWT and the 6MWT were, in general, more highly correlated with the StepWatch outputs than was either the RMI or the RMA. However, on regression analysis, the 6MWT was the only significant predictor for all but 3 of the StepWatch outputs, with the 10MWT making no further independent contribution to the variance.

The 6MWT is seen as a measure of submaximal exercise performance.33 Thus, the ability of the 6MWT to predict variations in walking performance in a natural environment is perhaps not unexpected. It is possible that distance on the 6MWT could be used as a quick test to estimate usual walking activity. From our data, the 95% confidence interval for the regression equation for an individual who achieved a distance of 153 meters would suggest that the individual might average between 3078 and 5231 steps a day.

Self-selected gait speed measured over a short distance (eg, 10MWT) is the most commonly used test to assess walking ability in a clinical situation.8 It is extremely quick and easy to administer, and from both this study and others20, 34 is moderately correlated to mean steps a day, in participants with both stroke (r=0.55)20 and neurological disorders (r=0.58).34 However, our data suggest that the 6MWT may be a better clinical test to use to predict usual walking performance. The 10MWT nevertheless is very highly correlated with the 6MWT and still has a role, particularly if it is not possible to test walking for 6 minutes.

Both the RMI and RMA showed a low correlation with most StepWatch outputs. These data are similar to a previous study of participants with neurologic disorders that showed a low correlation between mean steps a day and the RMI (r=0.49).34 One explanation for this finding is that both the RMI and the RMA assess mobility, rather than walking per se. For example, they both assess bed mobility and transfer skills. They also assess wider aspects of walking, such as stair climbing, walking outside, and walking over uneven surfaces, which are thought to be important aspects of usual walking performance.35 Although the StepWatch accurately identifies steps under these walking conditions,21 it does not distinguish between these different aspects of walking, which might explain the lower correlation.

However, the RMI, which measures self-reported mobility, was an independent predictor of 2 StepWatch outputs (percentage of time with no steps and number of steps at a low stepping rate). Both of these outputs reflect reduced levels of walking activity. This result suggests that patients' perception of reduced mobility may be able to predict aspects of usual walking performance. Although self-reported measures of physical activity have been shown to be inflated compared with mean steps a day,17 it is still possible that some persons with stroke voluntarily restrict activity if they have a low perception of their functional ability.36 However, whether the perception of reduced mobility is a causative factor in the low levels of activity or a consequence of it is not certain.

In addition to the 6MWT, age was an independent predictor of, and inversely related to, the highest step rate in 60 minutes. This StepWatch output measures the highest step rate in a continuous 60-minute period and might be expected to decrease with reduced exercise performance, as measured by the 6MWT.37 However the finding that age also makes an independent contribution was unexpected because age has not been shown to relate to walking speed in adults with chronic stroke.38 This finding suggests that the level of sustained activity decreases with age in people with chronic stroke over and above that which can be attributed to decreased endurance.

Half of the variability in StepWatch outputs of usual walking performance is not accounted for by the clinical walking tests. Because community walking is related to other physical characteristics in addition to gait speed,6, 39 it is also possible that physical factors such as balance,40 fitness,41 use of assistive devices, and motor function may also affect usual walking performance. It is also likely that behavioral, personal, environmental, and social factors will have some impact on walking performance in natural environments,14, 42 but there is little research in this area. Until these factors are identified, there would seem to be a place for the inclusion of activity monitoring as an outcome measure during stroke rehabilitation.

Study Limitations 

A limitation of this study is the selected nature of the participants, which may not generalize to the entire stroke population. Furthermore, participants may have changed their walking activity in their own environment as a result of the monitoring, thus not giving completely accurate data on usual performance.

In addition, this study was adequately powered to detect a correlation coefficient r>0.5 in the regression analysis, but more subjects would have been needed to detect a smaller effect size,28 such as shown by the lower correlations between both the RMI and the RMA and the StepWatch outputs. However, the question remains whether such a level of correlation should be considered clinically significant.

It should be acknowledged that while the StepWatch is an objective measure of usual walking, the information gained is limited to the amount and rate of walking and patterns of activity. The StepWatch cannot, for instance, give information about functional goals achieved or effectiveness and energy cost of walking.

Back to Article Outline

Conclusions 

The 6MWT is the clinical test with the strongest relationship with the StepWatch outputs. Thus, the 6MWT may be a better test than the 10MWT to predict usual walking performance; however, it should be remembered that half of the variability in usual walking performance is not explained by either clinical walking test. Activity monitoring detects aspects of usual walking performance in participants with stroke not captured by clinical tests and should be considered as an additional outcome measure for rehabilitation programs.

Suppliers

Back to Article Outline

Acknowledgments 

We gratefully acknowledge the contributions of Pat Bennett, RN, and Alan Barber, MD, of the Auckland District Health Board for participant recruitment.

Back to Article Outline

References 

  1. Adamson J, Beswick A, Ebrahim S. Is stroke the most common cause of disability?. J Stroke Cerebrovasc Dis. 2004;13:171–177
  2. Bonita R, Solomon N, Broad JB. Prevalence of stroke and stroke-related disability: estimates from the Auckland stroke studies. Stroke. 1997;28:1898–1902
  3. Kelly-Hayes M, Beiser A, Kase CS, Scaramucci A, D'Agostino RB, Wolf PA. The influence of gender and age on disability following ischemic stroke: the Framingham study. J Stroke Cerebrovasc Dis. 2003;12:119–126
  4. Jorgensen HS, Nakayama H, Raaschou HO, Olsen TS. Recovery of walking function in stroke patients: the Copenhagen Stroke Study. Arch Phys Med Rehabil. 1995;76:27–32
  5. Goldie PA, Matyas TA, Evans OM. Deficit and change in gait velocity during rehabilitation after stroke. Arch Phys Med Rehabil. 1996;77:1074–1082
  6. Lord SE, McPherson K, McNaughton HK, Rochester L, Weatherall M. Community ambulation after stroke: how important and obtainable is it and what measures appear predictive?. Arch Phys Med Rehabil. 2004;85:234–239
  7. International Classification of Functioning, Disability and Health: ICF. Geneva: World Health Organization; 2001;
  8. Mudge S, Stott S. Outcome measures to assess walking ability following stroke—a systematic review of the literature. Physiotherapy. 2007;93:189–200
  9. Bohannon RW. Walking after stroke: comfortable versus maximum safe speed. Int J Rehabil Res. 1992;15:246–248
  10. Flansbjer U, Holmback AM, Downham D, Patten C, Lexell J. Reliability of gait performance tests in men and women with hemiparesis after stroke. J Rehabil Med. 2005;37:75–82
  11. Lincoln N, Leadbitter D. Assessment of motor function in stroke patients. Physiotherapy. 1979;65:48–51
  12. Collen FM, Wade DT, Robb GF, Bradshaw CM. The Rivermead Mobility Index: a further development of the Rivermead Motor Assessment. Int Disabil Stud. 1991;13:50–54
  13. Collen FM, Wade DT, Bradshaw CM. Mobility after stroke: reliability of measures of impairment and disability. Int Disabil Stud. 1990;12:6–9
  14. Buchner DM, Cress ME, Esselman PC, et al. Factors associated with changes in gait speed in older adults. J Gerontol A Biol Sci Med Sci. 1996;51:M297–M302
  15. Wade DT. Measurement in neurological rehabilitation. Oxford: Oxford University Press; 1992;
  16. Giantomaso T, Makowsky L, Ashworth NL, Sankaran R. The validity of patient and physician estimates of walking distance. Clin Rehabil. 2003;17:394–401
  17. Resnick B, Michael K, Shaughnessy M, et al. Inflated perceptions of physical activity after stroke: pairing self-report with physiologic measures. J Phys Act Health. 2008;5:308–318
  18. Macko RF, Haeuber E, Shaughnessy M, et al. Microprocessor-based ambulatory activity monitoring in stroke patients. Med Sci Sports Exerc. 2002;34:394–399
  19. Haeuber E, Shaughnessy M, Forrester LW, Coleman KL, Macko RF. Accelerometer monitoring of home- and community-based ambulatory activity after stroke. Arch Phys Med Rehabil. 2004;85:1997–2001
  20. Michael KM, Allen JK, Macko RF. Reduced ambulatory activity after stroke: the role of balance, gait, and cardiovascular fitness. Arch Phys Med Rehabil. 2005;86:1552–1556
  21. Mudge S, Stott NS, Walt SE. Criterion validity of the StepWatch Activity Monitor as a measure of walking activity in patients after stroke. Arch Phys Med Rehabil. 2007;88:1710–1715
  22. Mudge S, Stott NS. Test-retest reliability of the StepWatch Activity Monitor outputs in individuals with chronic stroke. Clin Rehabil. 2008;22:871–877
  23. Shaughnessy M, Michael KM, Sorkin JD, Macko RF. Steps after stroke: capturing ambulatory recovery. Stroke. 2005;36:1305–1307
  24. Michael K, Macko RF. Ambulatory activity intensity profiles, fitness, and fatigue in chronic stroke. Top Stroke Rehabil. 2007;14:5–12
  25. McDonald CM, Widman LM, Walsh DD, Walsh SA, Abresch RT. Use of step activity monitoring for continuous physical activity assessment in boys with Duchenne muscular dystrophy. Arch Phys Med Rehabil. 2005;86:802–808
  26. Michael KM, Allen JK, Macko RF. Fatigue after stroke: relationship to mobility, fitness, ambulatory activity, social support, and falls efficacy. Rehabil Nurs. 2006;31:210–217
  27. Dupont WD, Plummer WD. Power and sample size calculations for studies involving linear regression. Control Clin Trials. 1998;19:589–601
  28. Green SB. How many subjects does it take to do a regression analysis?. Multivariate Behav Res. 1991;26:499–510
  29. Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36), I: conceptual framework and item selection. Med Care. 1992;30:473–483
  30. Hill K, Miller K, Denisenko S, Clements T, Batchelor F. Manual for clinical outcome measurement in adult neurological physiotherapy. 2nd ed.. St Kilda: Australian Physiotherapy Association; 2001;
  31. McDowell I. Measuring health: a guide to rating scales and questionnaires. 3rd ed.. New York: Oxford University Press; 2006;
  32. Bohannon RW. Number of pedometer-assessed steps taken per day by adults: a descriptive meta-analysis. Phys Ther. 2007;87:1642–1650
  33. Noonan V, Dean E. Submaximal exercise testing: clinical application and interpretation. Phys Ther. 2000;80:782–807
  34. Busse ME, Wiles CM, van Deursen RW. Community walking activity in neurological disorders with leg weakness. J Neurol Neurosurg Psychiatry. 2006;77:359–362
  35. Patla A, Shumway-Cook A. Dimensions of mobility: defining the complexity and difficulty associated with community walking. J Aging Phys Activ. 1999;7:7–19
  36. Stretton CM, Latham NK, Carter KN, Lee AC, Anderson CS. Determinants of physical health in frail older people: the importance of self-efficacy. Clin Rehabil. 2006;20:357–366
  37. Steffen TM, Hacker TA, Mollinger L. Age- and gender-related test performance in community-dwelling elderly people: six-minute walk test, Berg Balance Scale, Timed Up & Go Test, and gait speeds. Phys Ther. 2002;82:128–137
  38. Bohannon R. Gait performance of hemiparetic stroke patients: selected variables. Arch Phys Med Rehabil. 1987;68:777–781
  39. van de Port IG, Kwakkel G, Lindeman E. Community ambulation in patients with chronic stroke: how is it related to gait speed?. J Rehabil Med. 2008;40:23–27
  40. Patterson SL, Forrester LW, Rodgers MM, et al. Determinants of walking function after stroke: differences by deficit severity. Arch Phys Med Rehabil. 2007;88:115–119
  41. Pang MY, Eng JJ. Determinants of improvement in walking capacity among individuals with chronic stroke following a multi-dimensional exercise program. J Rehabil Med. 2008;40:284–290
  42. Patla AE. Mobility in complex environments: implications for clinical assessment and rehabilitation. Neurol Rep. 2001;25:82–90
  • a Orthocare Innovations, 6405 218th St SW, Suite 100, Mountlake Tce, WA 98043-2180.
  • b Excel 2003; Microsoft Corp, One Microsoft Way, Redmond, WA 98052-7329.
  • c Version 14.0; SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

 Supported by a Health Research Council Clinical Training Fellowship (Mudge). Research costs were funded by the Health Research Council (grant no. HRC 06/059) and the Wishbone Trust.

 No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.

PII: S0003-9993(08)01604-3

doi:10.1016/j.apmr.2008.07.025

Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 2 , Pages 296-301, February 2009

Article Tools