837

Mobility-Related Function in Older Adults:

Assessment With a 6-Minute Walk Test
Nancy D. Harada, PhD, PT, Vicki Chiu, MS, Anita L. Stewart, PhD
ABSTRACT. Harada ND, Chiu V, Stewart AL. Mobility- distance of 50 feet. It is often used by clinicians to measure
related function in older adults: assessment with a 6-minute walking velocity over a fixed distance.’ The 2-, 6-, and
walk test. Arch Phys Med Rehabil 1999;80:837-41. 12-minute walking tests measure the distance walked within
their respective time frames.2-6
Objective: To determine the usefulness of the 6-minute walk Most studies of the 6-minute walk have found it to be a
test as an integrated measure of mobility in older adults. valuable measure of cardiovascular exercise capacity in elderly
Design: Observational study. patients with congestive heart failure and chronic lung dis-
Setting: Community centers and retirement homes in the Los ease.3,6-10 In these patients, the distance walked in 6 minutes has
Angeles area. a moderate to high correlation with cardiovascular respiratory
Patients: Eighty-six older adults without significant disease. capacity as measured by maximum oxygen uptake during a
Interventions: None treadmill or bicycle exercise test6-10 The advantage of the
Main Outcome Measures: Assessments included the 6-minute walk over treadmill tests of exercise capacity is that it
6-minute walk, chair stands, standing balance, gait speed, body is better tolerated by elderly patients.@ Peetersand Mets6 found
mass index, and self-reported physical functioning and general
that 22% of elderly patients with congestive heart failure were
health perceptions. unable to perform treadmill testing because of the speed of the
Results: One-week test-retest reliability of the B-minute walk
treadmill and consequent fear of falling, whereas all patients
was .95. As hypothesized, the B-minute walk distance was
significantly greater for active than for inactive older adults were able to complete the 6-minute walk.
The 6-minute walk has also been applied as a tool for disease
(p < .OOOl),moderately correlated with chair stands (r = .67),
standing balance (r = .52), and gait speed (Y = - .73). It had a severity classification and assessmentof prognosis in patients
low correlation with body mass index (r = - .07). The correla- with chronic heart failure and as an outcome measure following
tion of the 6-minute walk with self-reported physical function- an intervention.4Jj,10-13The 6-minute walk can differentiate
ing was .55, and its correlation with general health perceptions severely ill from less severely ill patients with chronic heart
was .39. Self-report and performance measures explained 69% failure, according to the New York Heart Association functional
of the variance in 6-minute walk scores. classification.4x6,7It is a strong predictor of morbidity and
Conclusions: The 6-minute walk test is reliable and is valid mortality in patients with left ventricular dysfunction and is
in relation to the performance and self-reported indicators of therefore useful as a prognostic tool in identifying patients with
physical functioning tested in this study. It could serve as a a higher likelihood of morbidity and mortality events.4J0It has
useful integrated measure of mobility. been applied as an outcome measure to determine the effective-
0 1999 by the American Congress of Rehabilitation Medi- ness of exercise interventions for elderly patients.“-l3 These
cine and the American Academy of Physical Medicine and studies have found the 6-minute walk to be sensitive to change
Rehabilitation after an exercise intervention.
In the clinical setting, the 6-minute walk could be useful for
rehabilitation practitioners as an integrated measure of mobility
T IMED WALKING TESTS are often used in clinical
practice and research to assessaspects of physical function.
Several types of walking tests have been developed. They are
function. It measures a patient’s level of mobility, taking into
account any limitations imposed by major body systems such as
the musculoskeletal, cardiovascular, sensory, and neurologic
administered by having the patient walk for a specified distance
or a specified time. For example, the 50-foot walk test is systems.15-18These impairments may have an additive or
conducted by timing how long it takes a patient to walk a interactive effect on an individual’s level of mobility. During
clinical evaluation, the rehabilitation practitioner conducts a
personal interview and performs objective teststo determine the
From the Geriatric Research, Education, and Clinical Center of the Department of specific impairments that cause the patient’s difficulty with
Veterans Affairs Medical Center, West Los Angeles (Dr. Handa); the University of functional activities, and appropriate treatment strategies are
California Los Angeles School of Medicine (Dr. Harada, Ms. Chiu) and the
UCLA/VA/RAND MEDTEP Center for Asians and Pacific Islanders (Dr. Harada, Ms.
subsequently devised.15 For example, an elderly patient may
Chiu), Los Angeles; and the Institute for Health and Aging of the University of have difficulty walking to the store. Further clinical testing may
California San Francisco, San Francisco (Dr. Stewart), CA. reveal impairments in strength and endurance, for which
Submitted for publication September 1, 1998. Accepted in revised form December specific exercises might be prescribed. A patient’s score on the
23, 1998.
Suooorted bv National Institute on A&e grant AGO9931 and Aaencv
6-minute walk test (either distance or gait speed) could
- _ for Health
Care’dolicy ani Research grant HS07376. - - summarize the effects of the strength and endurance impair-
Presented in part at the 50th Annual Scientific Meeting of the Gerontological ments on walking and could also provide information about the
Society ofAmerica, November 14-18,1997, Cincinnati, OH. functional ability to walk by comparison with established
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
norms. Thus, by having the patient walk for 6 minutes, the
organization with which the authors are associated. rehabilitation practitioner can obtain a global assessmentof the
Reprint requests to Nancy Harada, PhD, PT, GRECC (llG), Department of Veterans extent to which impairments affect mobility.
Affairs Medical Center-West Los Angeles, 11301 Wilshire Boulevard, Los Angeles, The aims of this study were to explore the usefulness of the
CA 90073.
0 1999 by the American Congress of Rehabilitation Medicine and the American
6-minute walk as an integrated performance-based measure of
Academy of Physical Medicine and Rehabilitation mobility by (1) determining the l-week test-retest reliability,
0003-9993/99/8007-5183$3.00/O known-groups validity, and convergent validity of the 6-minute

Arch Phys Med Rehabil Vol 80, July 1999

838 6-MINUTE WALK, Harada

walk in a sample of community-dwelling older adults without correlates well with published data on the strength of knee
significant disease, and (2) determining the extent to which the flexor and extensor muscles in groups of men and women of
6-minute walk summarizes information on performance mea- various ages.22,23
sures and self-reported physical functioning. The hypotheses Several methods of administering this test have been re-
were: (1) the 6-minute walk is highly reliable over a l-week ported.19,20This study followed the method documented by
period; (2) distance covered over 6 minutes is greater for older Guralnik and colleagues2o for the Short Physical Performance
adults living in community centers than for those living in Battery in the Established Populations for Epidemiologic
retirement homes (known-groups validity); (3) distance cov- Studies of the Elderly (EPESE). The subject stood up and sat
ered in 6 minutes is moderately correlated with chair stands, down from a chair five times as quickly as possible with arms
balance, gait speed, and self-reported physical functioning and crossed over the chest. A tester used a stopwatch to time the
general health perceptions (convergent validity); (4) distance maneuver. The stopwatch was started when the subject began
walked in 6 minutes is negatively and moderately correlated the first rise from the chair, and stopped when the subject stood
with body mass index (convergent validity); and (5) a large on the fifth repetition and all body movement had ceased.
proportion of the variance in the 6-minute walk measure is Scoring was done according to the method used by Guralnik.20
determined by self-reported physical functioning, strength, The number of seconds to complete the test was divided into 4
balance, and gait speed. categories based on a population-based quartile algorithm;
those who were unable to do the test or had to use their arms
METHODS were assigned a score of zero. A higher score thus indicated
better strength. Guralnik20 found chair stand performance to
Subjects differ as hypothesized by age group and gender.
A convenience sample of 86 older adults was recruited from Tandem balance. The test of standing balance was also
two community centers and three retirement homes in Los selected from the battery of physical performance measures
Angeles. One community center and one retirement home developed by GuralnikzO for the EPESE studies. Guralnik
served primarily the Japanese-American community. The size describes a battery of three balance tests with the subject
of the total sample was based on a statistical power analysis of standing with feet side-by-side, semi-tandem, and tandem. The
the number of cases needed to detect correlations of 0.3 with side-by-side position was first demonstrated by the tester, and
80% power. The community centers and retirement homes were the subject was then asked to stand in the position while being
selected by design to obtain subjects who were known to be timed with a stopwatch. The timing stopped when subjects
more and less active. The sample was stratified by age group moved their feet or lost their balance, or when 10 seconds had
(65 to 74yrs, and 75yrs and older) and gender. Older adults were elapsed. This process was repeated for the semi-tandem and
recruited for the study through announcements, flyers, and word tandem positions. Although subjects performed all three tests,
of mouth. only the tandem balance test was used in this analysis. The
score used in these analyses was the number of secondsup to 10
Performance-Based, Clinical, and Self-Report Measures that the subject could hold the tandem position.
Each subject was assessedon several performance, clinical, d-Foot waZk. Gait speed was measured by having the
and self-report measures. These included (1) 6-minute walk, subject walk a distance of 8 feet. An S-foot walking line was
(2) chair stands, (3) tandem standing balance, (4) 8-foot walk, marked off with an additional 2 feet at either end. Each subject
(5) weight and height, and (6) SF-36 physical functioning and was instructed to walk along the line at their usual speed and
general health perception scales. timed for 2 trials. Scoring was the number of seconds required
6-Minute walk. The subject walked around a series of to walk 8 feet. Guralnik20 divides the number of seconds into
traffic cones, which were placed to mark off a circular walking quintiles, with those people who are unable to perform the test
area of about 40 feet in diameter that was measured before the receiving a zero. Since all our subjects were able to perform the
test. Subjects were instructed to attempt to walk for 6 minutes, test, the number of seconds was used as a continuous measure.
covering as much ground as possible at a work effort that The average of the 2 trials was used in the analysis.
allowed the person to talk without becoming short of breath. Body mass index. Body mass index was calculated using
The tester walked alongside the subject, and timed the walk the following formula: Body mass index = weight in kilograms/
with a stop watch. Subjects were not prompted by the tester height in meters squared (kg/m2).24 The subject was weighed
because previous studies have found that encouragement pro- using a portable digital scale. A tape measure attached to the
vided by the tester affects performance, with patients receiving wall was used to measure height. This measurement was
encouragement walking greater distances than those who do not obtained by alignment of a straight edge from the top of the
(p < .02).14The subject was permitted to stop and rest, but was head.
instructed to resume walking as soon as possible (if able). A Self-reported physical functioning and general health per-
counter was used to count the number of laps completed by the ceptions (SF-36 Health Survey). The SF-36 health status
subject. After 6 minutes, the subject was instructed to stop survey assessesseveral domains of health-related quality of
walking, a marker was placed on the ground, and the distance life, including physical functioning and general health percep-
walked during the last lap was measured by a rolling tape tions.25 Physical functioning refers to limitations in a variety of
measure. The total distance was derived by multiplying the activities such as bathing, dressing, walking, bending, climbing
number of laps by the circumference of the walking circle, and stairs, and running. General health perceptions are personal
adding the distance covered on the last lap. beliefs about general health status. The reliability of these two
Chair stands. Chair stands are a performance-based test measures for adults older than 65 years of age using internal
that measures quadriceps and lower body strength.19 Chair consistency methods has been determined to be .92 for physical
stands have been significantly correlated with walking speed functioning and .78 for general health perceptionsz6
and standing balance (Spearman coefficients = .48 and .39,
p < .OOl,respectively). 2o Other studies of chair stands indicate Procedure
good reliability. In a study of frail hospitalized elders, the Participants were scheduled for two visits, spaced 1 week
percent agreement among raters was 93.3.21 The test also apart. All visits were at either the community center or

Arch Phys Med Rehabil Vol 80, July 1999

 B-MINUTE WALK, Harada 839

retirement home from which the subject was recruited. Approxi- Table 2: I-Week Test-Retest Reliability of I-Minute Walk
mately 1 week before the first scheduled visit, participants were n r
mailed a packet that contained an informed consent form and a
questionnaire asking about demographic and medical informa- Total 83 .95
tion. Participants were instructed to read the informed consent By site
and to bring the completed questionnaire to their first visit. Retirement home 34 .91
At the first scheduled visit, the study was described to the Community center 49 .87
participant by the research staff and the informed consent By gender
obtained. Each participant was screened for cognitive function- Male 29 .94
ing using the Folstein Mini Mental State Examination.27 Female 54 .95
Individuals who received a score of 24 or higher then com- By age group
pleted the two self-reported scalesfrom the SF-36 health status 65-74yrs 41 .94
survey, after which the 6-minute walk test was administered. At z75yrs 42 .94
the end of the first visit, the participant received $10 for their
participation and was scheduled for the second visit 1 week
later. accounted for by self-reported physical functioning and perfor-
At the second visit, the participant performed the chair mance tests of impairments.
stands, gait speed, and standing balance tests. Weight and
height were recorded, and the second 6-minute walk test was RESULTS
administered. At the end of the second visit, participants again Demographic and medical characteristics of the sample are
received $10 for their participation. presented in table 1. The average age of the total sample was 75
years; the average age of the retirement home participants was 6
Data Analysis years older than community center participants. Sixty-three
Descriptive statistics (distributions and measures of central percent of the total sample were women. The racial breakdown
tendency) were examined for each measure using the SAS of the participants was 62% Japanese American, 33% white,
statistical software system.28 Pearson’s correlation coefficients and 5% African American. Thirty-three percent of the total
were calculated to assessthe one-week test-retest reliability. sample were married; however, there were large differences in
The mean score of the two 6-minute walk trials was used to marital status by site. Fifty-one percent of participants at the
assess convergent and known-groups validity. Convergent community centers were married; only 6% at the retirement
validity was determined by calculating Pearson’sor Spearman’s homes had spouses.The mean Folstein cognitive status score in
correlation coefficients between the 6-minute walk and other this sample that scored above 24 was 28, indicating good
performance-based or self-report measures. To examine known- cognitive functioning. The most common medical conditions
groups validity, a t test was conducted to determine whether the were arthritis (41%), hypertension (38%), visual deficits (27%),
mean walking distances between the two sites were signifi- and hearing deficits (26%).
cantly different. Ordinary least squaresregression with forward Table 2 reports the l-week test-retest reliability coefficients
selection was used to determine the amount of variance in the for the total sample, by site, by gender, and by age group. For
mean 6-minute walk score that was explained by self-reported the total sample, l-week test-retest reliability on the 6minute
physical functioning, chair stands, balance, and gait speed. walk was high (v = .95). Test-retest reliability was slightly
Self-reported general health was not included in the regression higher for retirement home (v = .91) than community center
model because we wanted to test the amount of variance (v = .87) participants. Women demonstrated slightly higher
l-week test-retest reliability than men (Y = .95 vs .94). Both
Table 1: Demographic and Medical Characteristics age groups demonstrated the same reliability of .94.
Known-groups validity was tested by comparing walking
Retirement Community distances on the 6minute walk between more and less active
Homes Centers
Inactive Active Total
older adults. More active older adults (from the community
(n = 35) (n = 51) (n = 86) centers) covered significantly greater distances (1,629 feet) than
less active older adults (living in retirement homes) (901 feet,
Age, vrs t = 10.1,~ < .OOOl)(table 3).
Mean (SD) 79 (‘3 73 (5) 75 (6) Convergent validity was tested by determining the correla-
Range 65-89 65-86 65-89
tion of 6-minute walk distances with performance-based,
Gender, % female 63 63 63 clinical, and self reported measures of physical functioning and
Race, %
general health perceptions (table 4). In the total sample, the
White 31 35 33
6minute walk distance was moderately correlated with lower
African-American 9 2 5
body strength (chair stands) (Y= .67), tandem balance (Y = .52),
Japanese-American 60 63 62 and gait speed (Y = - .73) as hypothesized. Higher correlations
Education, mean (SD) years 13 (3) 14 (3) 13 (3) were found for retirement home participants than community
Marital status, % married 6 51 33
Folstein score, mean (SD) 27 (2) 29 (1) 28 (2)
Self-reported medical conditions, % Table 3: Known-Groups Validity: Distance Walked During
6-Minute Walk
Visual deficits 34 22 27
Hearing deficits 29 24 26 Retirement Community
Arthritis 40 41 41 Homes Centers
Hypertension (n = 35) (n = 51) t Test
50 29 38
Asthma 14 4 8 Distance
Diabetes 29 12 19 Mean (SD), feet 901 (350) 1,629(313) t= lO.l,p< .OOOl
Falling in last 12 months 26 24 25 Range, feet 300-I ,665 777-2,353

Arch Phys Med Rehabil Vol80, July 1999

840 B-MINUTE WALK, Harada

Table 4: Construct Validity: Correlation of Performance-Based, likely to have their impairments captured by the performance
Clinical, and Self-Report Measures With 6-Minute Walk Distance tests used in this study.
Retirement Community Our findings that 69% of the variance in the 6-minute walk
Home CMXW Total was explained by self-reported physical functioning, lower
(n = 35) (n = 51) (n = 86) body strength, balance, and gait speed confirmed our hypoth-
Performance/clinical measures esis. These findings indicate that the 6minute walk could act as
Chair stands (+) .70 .53 .67 an integrated measure of self-reported physical functioning and
Tandem balance (+) .59 .39 .52 performance tests of impairments. Although a large proportion
8-foot walk (-1 -.68 -.45 -.73 of the 6-minute walk variance was accounted for in our model,
Body mass index (-) -.28 -.21 -.07 3 1% is still unexplained. This 3 1% may be partially explained
Self-report measures by other impairments not tested in this study, such as cardiovas-
Physical functioning (+) .46 .53 .55 cular capacity or range of motion limitations.
General health perceptions (+) .I0 .24 .39 Our results are encouraging because they indicate that the
6minute walk could be useful to rehabilitation practitioners.
A (+) indicates that a higher score is better; (-) indicates lower score There are many potential clinical applications of the 6minute
is better.
walk. The distance walked in 6 minutes can be used to assessan
individual’s endurance. Gait velocity can be averaged over the
center participants. In the total sample, low correlations were 6-minute interval and compared with functional gait speeds
found with body mass index, which did not confirm our established in the literature. For example, the literature has
hypothesis. The 6minute walk distance was moderately corre- defined independent community ambulation as the ability to
lated with the physical functioning subscale (r = .55) and the walk at least 332 meters at a gait velocity of 80m/min.29 In
general health perceptions subscale (u = .39) of the SF-36. For another study of older pedestrians, Hoxie and coworkers30
self-report measures, higher correlations were found in commu- found that 96% walked slower than a 1.22mfsec guideline used
nity center participants. by city traffic engineers to set traflic signal timing. Therefore,
The amount of variance in the 6-minute walk distance that an individual’s performance on the 6minute walk could be
was explained by the self-reported measure of physical function- used as an indicator of ability to ambulate in the community.
ing, lower body strength, standing balance, and gait speed was Other studies have used the g-minute walk as an outcome
69% (R2 = .69, p = .OOOl).The forward selection procedure is measure to determine a patient’s improvement following a
summarized in table 5. Gait speed entered into the model first, rehabilitation intervention. Gunnarsson and colleagues13found
followed by self-reported physical functioning, chair stands, an improvement of 11.6% from 373 meters to 421 meters in
and tandem standing balance. All variables, except tandem community dwelling older adults with a mean age of 78 years
standing balance, were significant at p < .05. engaged in a 1Zweek moderate-intensity walking program. In
contrast, other shorter timed walking tests may not demonstrate
DISCUSSION sensitivity to change. Grace and colleagues1 found that the
With the growth of the older adult population in the United 50-foot walk time was a poor outcome measure in rheumatic
States, there has been an increased demand for rehabilitation disease drug trials, based on their review of 5 1 published trials
practitioners to help these individuals maintain their highest using this measure. As compared with other outcomes also
level of functioning. Performance-based measures such as the employed in these 51 trials, the 50-foot walk time was
6-minute walk can be incorporated into the evaluation process statistically significant in only 21 trials (41% of the time) as
to assessthe mobility of these patients. compared with 81% for the best outcome measure, morning
The results of this study indicate that the 6minute walk test stiffness. In addition, the mean of the differences in the 50-foot
is highly reliable; it conformed to our hypotheses in five out of walk time in the 21 trials where statistically significant differ-
six validity tests of older adults without significant disease who ences were found was only 2 seconds.’
live in community centers or retirement homes. The only The subjects in our study represent a convenience sample of
measure that did not correlate as hypothesized with the older adults drawn by design from community centers and
6minute walk was body mass index. Of the self-reported retirement homes or the community at large. Therefore, the
measures, general health demonstrated a lower correlation. results cannot be generalized to older adults living in other
Unlike the other measures, general health is not a direct settings such as nursing homes or to those not attending
measure of physical function. The direction of the correlation, community centers. Since all of the Asian participants were
however, was in the hypothesized direction. English-speaking, it is improbable that ethnicity would limit the
Higher correlations between 6-minute walk and performance generalizability of these results.
measures were found for older adults in retirement homes than To improve the clinical usefulness of the 6minute walk,
in community centers. This could be partially explained by the further studies should be conducted to refine our understanding
ceiling effects of some of the performance tests in the commu- of the relationship of the 6minute walk with impairments
nity sample. Older adults in the retirement homes were more (limitations in range of motion, strength, proprioception, pain)
and functional limitations involving mobility (transfers, walk-
Table 5: Forward Regression of B-Minute Walk Distance on ing, shopping). By understanding these relationships, rehabilita-
Self-Reported and Clinical Measures (n = 86) tion practitioners will obtain a full understanding of how a
Model performance-based measure such as the 6minute walk can be
Step Variable w P used to evaluate the functional status of older adults.
1 8-foot walk .54 .OOOl
2 Self-reported physical functioning .64 .OOOl References
3 Chair stands .67 .0222 1. Grace EM, Gerecz EM, Kassam YB, Buchanan WW, Tugwell PS.
50-foot walking time: a critical assessment of an outcome measure
4 Tandem standing balance .69 .0683
in clinical therapeutic trials of antirheumatic drugs. Br J Rheuma-
Model R* = .69, p = .OOOl. to1 1988;27:372-4.

Arch Phys Med Rehabil Vol 80, July 1999

 6-MINUTE WALK, Harada 841

2. McGavin CR, Gupta SP, McHardy. Twelve-minute walking test for 16. Bendall MJ, Bassey EJ, Pearson MB. Factors affecting walking
assessing disabilitv in chronic bronchitis. BMJ 1976:1:822-3. speed of elderly people. Age Ageing 1989;18:327-32.
3. Butland”RJA, Pang J, Gross ER, Woodcock AA,‘Geddes DM. 17. Woo J, Ho SC, Lau J, Chan SG, Yuen YK. Age-associated gait
Two-, six-, and twelve-minute walking tests in respiratory disease. changes in the elderly: pathological or physiological? Neuroepide-
BMJ 1982;284:1607-8. miology 1995;14:65-71.
4. Bittner V, Weiner DH, Yusuf S, Rogers WJ, McIntyre KM, 18. Giuliani CA. Adult hemiplegic gait. In: Smidt G, editor. Gait in
BangdiwaIa SI, et al. Prediction of mortality and morbidity with a rehabilitation. New York: Churchill Livingstone; 1990. p. 253-66.
6-minute walk test in patients with left ventricular dysfunction. 19. Csuka M, McCarty DJ. Simple method for measurement of lower
JAMA 1993;270:1702-7. extremity muscle strength. Am J Med 1985;78:77-81.
5. Bernstein ML, Despars JA, Singh NP, Avalos K, Stansbury DW, 20. Guralnik JM, Simonsick EM, Ferrucci L, Glynn RJ, Be&man LF,
Light RW. Reanalysis of the 12minute walk in patients with
chronic obstructive pulmonary disease. Chest 1994; 105: 163-7. Blazer DG, et al. A short physical performance battery assessing
6. Peeters P, Mets T. The 6-minute walk as an appropriate exercise lower extremity function: association with self-reported disability
test in elderly patients with chronic heart failure. J Gerontol Med and prediction of mortality and nursing home admission. J
Sci 1996;51A:M147-51. Gerontol Med Sci 1994;49:M85-94.
7. Guyatt GH, Sullivan MJ, Thompson PJ, Fallen EL, Pugsley SO, 21. Winograd C, Lemsky CM, Nevitt MC, Nordstrom TM, Stewart
Taylor DW, et al. The 6-minute walk: a new measure of exercise AL, Miller CJ, et al. Development of a physical performance and
capacity in patients with chronic heart failure. Can Med Assoc J mobility examination. J Am Geriatr Sot 1994;42:743-9.
1985;132:919-23. 22. Murray MP, Gardner GM, Mollinger LA, Sepic SB. Strength of
8. Lipkin DP, Striven AJ, Crake T, Poole-Wilson PA. Six minute isometric and isokinetic contractions. Phys Ther 1980;60:413-9.
walking test for assessing exercise capacity in chronic heart 23. Murray MP, Duthie EH Jr, Gambert SR, Sepic SB, Mollinger LA.
failure. BMJ 1986;292:653-5. Age-related decline in knee muscle strength in normal women. J
9. Roomi J, Johnson MM, Waters K, Yohannes A, Helm A, Connolly Gerontol 1985;40:275-80.
MJ. Respiratory rehabilitation, exercise capacity and quality of life 24. Garrow JS, Webster J. Quetelet’s Index (W/Hz) as a measure of
in chronic airways disease in old age. Age Ageing 1996;25: 12-6. fatness. Int J Obesity 1985;9:147-53.
10. Cahalin LP, Mathier MA, Semigran MJ, Dee GW, DiSalvo TG. 25. Ware JE. SF-36 Health Survey-Manual and interpretation guide.
The six-minute walk test predicts peak oxygen uptake and survival Boston: The Health Institute, New England Medical Center; 1993.
in patients with advanced heart failure. Chest 1996;110:325-32. 26. McHomey CA, Ware UE, Rachel Lu JF, Sherboume C. The MOS
11. Fitts SS, Guthrie MR. Six-minute walk by people with chronic 36-Item Short-Form Health Survey (SF-36): III. Tests of data
renal failure. Assessment of effort by perceived exertion. Am J quality, scaling assumptions, and reliability across diverse patient
Phys Med Rehabil 1995;74:54-8. groups. Med Care 1994;32:40-66.
12. Hogue CC, Belyea M, DeVellis BM, Riegger-Krugh C, Rosemond
C. Effects of an integrated exercises intervention on sedentary 27. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”-a
older adults with arthritis [abstract]. Gerontologist 1997;37:125. practical method for grading the cognitive state of patients for the
13. Gunnarsson OT, Judge JO, Earles DR, Marcella GR. A comparison clinician. J Psychiatr Res 1975;12:189-98.
of walking programs for older adults: effects on six minute 28. SAS Institute Inc. SAS/STAT User’s Guide, Version 6. 4th ed.
walking distance [abstract]. Gerontologist 1997;37: 126. Cary (NC): SAS Institute; 1993.
14. Guyatt GH, Pugsley SO, Sullivan MJ, Thompson PJ, Berman LB, 29. Menard-Rome K, Sobush DC, Bousamra M, Haasler GB, Lipchik
Jones NL, et al. Effect of encouragement on walking test perfor- RJ. Self-selected walking velocity for functional ambulation in
mance. Thorax 1984;39:818-22. patients with end-stage emphysema. J Cardiopulm Rehabil 1997;
15. Sullivan PE. Ambulation: an integrated framework to achieve a 1785-91.
functional outcome. In: Guccione AA, editor. Geriatric physical 30. Hoxie RE, Rubenstein LZ, Hoenig H, Gallagher BR. The older
therapy. St. Louis: Mosby; 1993. p. 253-68. pedestrian. J Am Geriatr Sot 1994;42:444-50.

Arch Phys Med Rehabil Vol 80, July 1999