1562

Reliability of the Two-Minute Walk Test in Individuals With

Transtibial Amputation
Dina Brooks, PhD, Judith P. Hunter, MSc, Janet Parsons, MSc, Emma Livsey, BSc, PT, Janice Quirt, BSc,
Michael Devlin, MD
ABSTRACT. Brooks D, Hunter JP, Parsons J, Livsey E,
Quirt J, Devlin M. Reliability of the two-minute walk test in
individuals with transtibial amputation. Arch Phys Med
Rehabil 2002;83:1562-5.
Objective: To determine inter- and intrarater reliability of
the two-minute walk test (2MWT) in individuals with transtibial amputation.
Design: Prospective; test-retest method by a pair of trained
physical therapists.
Setting: Two regional amputee rehabilitation centers in Canada.
Participants: Thirty-three subjects (23 men, 10 women;
mean age standard error, 63.62.0y) with transtibial amputation; 6 in outpatient rehabilitation, 27 in inpatient rehabilitation. The most common primary diagnoses were peripheral
vascular disease (n15) and diabetes (n11).
Interventions: Each subject performed a total of four
2MWTs, 1 test for each rater, on 2 consecutive days at approximately the same time of day. Subjects were given at least a
20-minute rest between tests. The order of raters was randomized on the first day and reversed for the next day. The walk
tests were performed in the same enclosed corridors with the
same starting point for all tests. The subjects were familiar with
the test or were given 1 or more practice tests at least 1 day
before testing. Subjects were allowed to walk with a mobility
aid of their choice. Raters used a digital stopwatch to time the
tests and a calibrated wheel with a counter to measure the
distance walked in meters. The raters were blinded to each
others scores.
Main Outcome Measure: Distance walked in 2 minutes (in
meters).
Results: Within-rater reliability was high (intraclass correlation coefficient [ICC], .90 .96). Between rater reliability was
also high (ICC .98 .99). Analysis of variance (ANOVA)
showed a significant effect for day of test (P.001) in the
inpatient group but no effect for therapist (P.098) or for
interaction of day and therapist (P.710). Similarly, in the
outpatient group, ANOVA showed a significant effect for day
(P.013) but no effect for therapist (P.259) or interaction of
day and therapist (P.923).

From the Clinical Evaluation and Research Unit, West Park Healthcare Centre
(Brooks, Hunter, Parsons); St. Johns Rehabilitation Hospital (Hunter); Toronto
Rehabilitation Institute (Parsons); and Department of Physical Therapy (Brooks,
Hunter, Parsons, Livsey, Quirt) and Division of Physiatry (Devlin), Faculty of
Medicine, University of Toronto, Toronto, Ont, Canada.
Preliminary results presented at the Ontario Association of Amputee Care Conference, May 2001, Kingston, Ont, Canada.
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 author(s) or upon any
organization with which the author(s) is/are associated.
Reprint requests to Dina Brooks, PhD, Clinical Evaluation and Research Unit, West
Park Healthcare Centre, 82 Buttonwood Ave, Toronto, Ont M6M 2J5, Canada,
e-mail: dina.brooks@utoronto.ca.
0003-9993/02/8311-6979$35.00/0
doi:10.1053/apmr.2002.34600

Arch Phys Med Rehabil Vol 83, November 2002

Conclusion: Although the 2MWT showed evidence of interand intrarater reliability in individuals with unilateral belowknee amputation, the distance walked in 2 minutes continued to
improve over time. This improvement was not solely the result
of a training and learning effect.
Key Words: Amputation; Exercise; Exercise test; Rehabilitation; Walking.
2002 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation
ALK TESTS ARE QUANTITATIVE measures of speed
and distance that provide information about functional
W
exercise capacity. In cardiorespiratory populations, some of the
timed walk tests, such as the 12-, 6-, and 2-minute walk tests,
are valid measures of exercise capacity.1 These tests correlate
O2/kg),
in differing degrees with volume of oxygen uptake (V
ventilation, some measures of pulmonary function, and morbidity and mortality.1
An informal unpublished Canadian survey of amputee programs in 1998 reported the two-minute walk test (2MWT) as
the second most commonly used outcome measure after the
FIM instrument. Recently, we evaluated the construct validity and responsiveness of the 2MWT in subjects with lowerlimb amputation.2 We concluded that the 2MWT was responsive to change with rehabilitation and that it correlated with
measures of physical function and prosthetic use in this population. To date, no investigators have reported the reliability of
the 2MWT in amputee populations.
The reliability of walk tests has been reported in other
populations. Larson et al3 reported high intrarater reliability of
the 12-minute walk (r.93.98) in individuals with chronic
obstructive pulmonary disease (COPD). Harada et al4 and
Montgomery and Gardner5 reported high test-retest reliability
of the 6-minute walk (r.95, r.94, respectively) in older
adults and individuals with peripheral arterial occlusive disease. Upton et al6 evaluated the 2MWTs test-retest reliability
in children with cystic fibrosis and reported no significant
difference with repeated testing as evidenced by a low mean
coefficient of variation (2.6%). Connelly et al7 reported the
2MWTs intra- and interrater reliability to be good to high
(intraclass correlation coefficient [ICC], .82.95) in a normal
frail elderly population. In addition, the test-retest reliability of
both comfortable and fast walking speeds of both amputees and
normal subjects has been reported to be high (ICC, .83.98).8
The purpose of the present study was to examine the interand intrarater reliability of the 2MWT in individuals with
transtibial amputation.
METHODS
The methods for the present study were based on the study
by Connelly7 who examined the reliability of the 2MWT in the
frail elderly. Our participants were physical therapists and
clients from 2 regional amputee rehabilitation programs in the
province of Ontario. Ethics approval was obtained from the
research ethics boards of the University of Toronto and the 2

1563

TWO-MINUTE WALK TEST IN AMPUTEES, Brooks

Table 1: Subjects Characteristics, Outpatients, Inpatients, and Total Sample

N
Age (y)
No. of prior walk tests
Days since prosthesis fitted
Days of rehabilitation
Days since amputation

Outpatients

Inpatients

All

6
55.34.4 (4469)
7.01.5 (110)
105.253.0 (28365)
134.046.7 (56365)
188.746.8 (63365)

27
65.42.1 (4280)
1.50.2 (16)
30.79.4 (7270)
31.53.5 (890)
89.915.0 (25365)

33
63.62.0 (4280)
2.60.5 (110)
44.212.8 (7365)
50.210.9 (8365)
107.816.1 (25365)

NOTE. Values are mean standard error (SE) and range.

centers involved. At each site, reliability was evaluated by

using the test-retest method with 2 physical therapists. Each
subject performed a total of 4 walks: on 2 consecutive treatment days, the subject completed 2 walks, 1 test for each rater,
at approximately the same time of day. Subjects were given at
least a 20-minute rest between each 2MWT. The order of raters
was randomized on the first day and reversed for the next day.
Participants
Study inclusion required each individual to have a limb
prosthesis, to have completed a minimum of 2 weeks of rehabilitation, to tolerate 2 minutes of walking, to have no prosthetic modifications planned, and to have no other medical
restrictions preventing them from participating in the test.
Subjects were excluded if they were cognitively impaired or
unable to give consent, poorly motivated to cooperate with the
procedure, or unable to participate on 2 consecutive treatment
days. All subjects gave informed consent before testing commenced.
A total of 33 individuals with transtibial amputation (23
men, 10 women) were recruited: 6 were undergoing outpatient
rehabilitation and 27 were enrolled in inpatient rehabilitation at
1 of 2 regional rehabilitation sites. Twelve of the subjects had
undergone amputation on the left lower extremity, and 21 had
undergone amputation on the right side. Sixteen subjects (10
inpatients, 6 outpatients) were from 1 regional rehabilitation
site, and 17 were from the other site. The majority of the
subjects (n31) required a mobility aid as follows: a walker or
rollator (n24), 1 or 2 canes (n5), or crutches (n2).
The subjects primary diagnoses were peripheral vascular
disease (n20), diabetes (n11), osteomyelitis (n1), and
sarcoma (n1). Thirty subjects presented with at least 1 additional comorbid diagnosis that included either diabetes, coronary artery disease, congestive heart failure, or chronic lung
disease. The individuals typically used patellar tendon-bearing
plastic laminate sockets, suspended by waist belt and fork
strap, and a solid ankle cushioned heel (SACH) foot. Although
no prosthesis changes were made after the study commenced,
changes may have been made 1 day before the first day of
testing.

Table 2: Distance Walked in 2 Minutes and Inter- and Intrarater

Reliability (ICCs) of 2 Raters Recording 2MWTs (m) of Inpatient
Transtibial Amputees on 2 Days

Day 1
Day 2
ICC

Rater 1

Rater 2

ICC*

50.14.6
57.05.0
.90

51.54.2
57.55.2
.94

.98
.98

NOTE. Values are mean SE.

* The reliability on any single day, reflecting interrater reliability.
The reliability of any single rater, reflecting intrarater reliability.

Study Protocol
At each site, the walk tests were performed in an enclosed
corridor. The same corridor was used for each test. The corridors were level ground, relatively free from distractions, and
longer than 40m. The starting point was the same for all tests
and was clearly marked. The subjects were instructed to walk
as far as they could in the 2 minutes. To control for learning
and practice effects, the subjects were familiar with the test or
were given 1 or more practice tests at least 1 day before testing,
usually in a different corridor than that used for testing. Subjects were allowed to walk with a mobility aid of their choice
and were given a rolling start of 2 or 3 steps. No talking was
permitted by raters or subjects during the tests. Raters used a
digital stopwatch to time each test and a calibrated wheel with
a counter to measure the distance walked in meters. The raters
were blinded to each others score.
Statistical Analysis
Reliability was determined by calculating the ICC. The ICC
provides a measure for evaluating reliability because it takes
into account both the between- and within-subjects components
of variance as well as the heterogeneity of the sample.9 Values
greater than .60 are considered acceptable reliability.10
We examined the consistency of the test over time by repeated-measures 2-way analysis of variance (ANOVA) to determine the difference over the 2 days and between therapists.
Post hoc analysis was performed by using the Student-Newman-Keuls procedure.
RESULTS
The characteristics of the subjects in the total sample and in
the inpatient and outpatient subgroups are in table 1. Individuals undergoing outpatient rehabilitation were generally
younger. They had undergone significantly more walk tests
before testing than the individuals undergoing inpatient rehabilitation (P.05). Furthermore, days since amputation, days
of rehabilitation, and days since prosthetic fitting were greater
in the outpatient group (all P.03).
Tables 2 and 3 portray the mean distance walked on the 2

Table 3: Distance Walked in 2 Minutes and Inter- and Intrarater

Reliability (ICCs) of 2 Raters Recording 2MWTs (m) of Outpatient
Transtibial Amputees on 2 Days

Day 1
Day 2
ICC

Rater 1

Rater 2

ICC*

121.118.6
137.914.7
.95

123.215.5
140.715.9
.96

.98
.99

NOTE. Values are mean SE.

* The reliability on any single day.
The reliability of any single rater.

Arch Phys Med Rehabil Vol 83, November 2002

1564

TWO-MINUTE WALK TEST IN AMPUTEES, Brooks

days for the 2 raters in the inpatient and outpatient subgroups,

respectively. Within-rater reliability was high, with ICCs ranging from .90 to .96. Between-rater reliability was also high,
with ICCs ranging from .98 to .99. A scatterplot of the distance
walked for both days is in figure 1.
In the inpatient group, a 2-way repeated-measures ANOVA
showed a significant effect of day of test (P.001) but no
effect for therapist (P.098) or interaction of day and therapist
(P.710). Similarly, in the outpatient group, a 2-way repeatedmeasures ANOVA showed a significant effect of day (P.013)
but no effect for therapist (P.259) or interaction of day and
therapist (P.923). Regardless of the tester, distances walked
on day 2 were greater than on day 1 for both groups. Figure 2
shows the changes in the distance walked for tests 1 and 2 (on
day 1) and tests 3 and 4 (on day 2).
DISCUSSION
The 2MWT exhibits good within- and between-rater reliability in individuals with transtibial amputation. However, the
distance walked in 2 minutes was not constant over time, but
increased over the 2 days in individuals undergoing either outor inpatient rehabilitation.
Reliability is a fundamental measurement property that is
relatively easy to determine. It is quantified in terms of degree
of consistency and repeatability when properly administered
under similar circumstances. Clinically, this property is important because it allows the clinician to determine the amount of
noise or random error in the tool.6 The clinician, when inter-

Fig 2. Mean distance SE walked by (A) inpatients and (B) outpatients in 2 minutes for each test regardless of the rater on 2 consecutive days.

Fig 1. Distance walked by (A) inpatients and (B) outpatients in 2

minutes for 2 raters on 2 different days. The solid line indicates the
line of identity.

Arch Phys Med Rehabil Vol 83, November 2002

preting the results of outcome measures, must be able to

determine how much of the measured change is due to real
change in the clients health status and how much is due to
measurement error. Reliability is population specific11 and
should not be seen as a property that a particular test does or
does not possess. Instead, any measure will have certain reliability when applied by certain clinicians, to a specific population, and under specific conditions. In individuals with transtibial amputation, the 2MWT is a reliable measure.
Some of the other psychometric properties of the 2MWT,
mainly validity and responsiveness, have been reported in other
populations including the elderly, individuals with respiratory
disease, and individuals with amputation. Butland et al12 and
Bernstein et al13 both evaluated concurrent validity in individuals with respiratory disease and reported that the distance
walked in the 2MWT strongly correlated with the distances
walked in both the 6- and the 12-minute walk tests (r.892
.955). Bernstein13 examined construct validity in individuals
with COPD and reported that the distance walked in 2 minutes

1565

TWO-MINUTE WALK TEST IN AMPUTEES, Brooks

moderately correlated to maximal and strongly correlated to

submaximal oxygen consumption (r.45, r.55, respectively). Furthermore, Upton et al6 concluded that, in children
with cystic fibrosis who had near-normal respiratory function,
the 2MWT was a more discriminative measure than peak
expiratory flow rate. In individuals with lower-extremity amputation, the 2MWT distance correlated with measures of
physical function and prosthetic use.2
With respect to responsiveness, Guyatt et al14 showed that
responsiveness to treatment, measured by the within-individual
standard deviation (SD) of distance walked in 2 minutes, was
less in the 2-minute test than in the 6-minute test (SDs, .90 vs
.74) in individuals with chronic airflow limitation, chronic
heart failure, or both.14 However, in a sample of elderly men
with COPD, changes in 2MWTs were strongly correlated with
O2/kg (r.53) and maximal oxygen uptake
changes in V
O2max) (r.53).13 In individuals with lower-extremity am(V
putation, we1 have recently shown that the 2MWT was responsive to change with rehabilitation.2
One surprising finding from the present study was that the
distance walked in 2 minutes improved over the 2 days of
testing. This improvement was statistically significant in each
of the 2 subgroups (inpatients, outpatients). One possible explanation is that subjects experienced a training or learning
effect. Guyatt14 established the presence of a learning and
training effect by performing 6 repeated tests on the 2-minute
walk. The distance walked improved on the first 2 walks
compared with the last 4 walks in adults with chronic airflow
limitation or chronic heart failure or both. Furthermore, training and learning effects with repeated testing have been identified in several studies on the 6-minute walk test (6MWT).1
For instance, 1 study15 showed that 6MWT distances could
improve 60m after 3 repeated walk tests in individuals with
COPD. However, in these studies, this effect was attenuated by
the third test, which was not the case in our present study.
Although a learning and training effect may have influenced
our findings, we do not believe that this was the main reason
for the improvement observed. The change in distance walked
over the 2 days was highly significant in both subgroups of
individuals (inpatients, outpatients); however, the majority
(5/7) of outpatient subjects had undergone more than 5 training
walks before testing. All subjects had at least 1 training walk,
and the majority had at least 2. We speculate that 2 other
factors may have contributed to the improvement over time.
First, the same corridor and starting point were used for all
tests. Considering that the distances walked by the subjects
were relatively short, it is possible that they remembered their
performance and were encouraged through external cues to try
harder on subsequent tests. Most other studies on the 2-minute
walk looked at performance in the first 2 minutes of a 6- or a
12-minute walk test, thus minimizing the effect of memory on
performance. Figure 2 shows that, regardless of the tester,
performance improved between tests on the first and second
day. This continued improvement supports the notion of a
potential effect of memory and the individual desire to show
improved performance. Second, because these subjects were
undergoing rehabilitation, it is possible that their walking capacity may have improved overnight. Alternatively, repeated
walking as performed during the tests may be therapeutic, and
therefore the changes may reflect a treatment effect of the test
itself. Future studies could examine this effect by studying
individuals several months after termination of rehabilitation or
by using a different corridor or different starting point for
repeated testing.
As mentioned, most other studies evaluating the 2MWT
have evaluated different populations, and McDowell and New-

ell11 caution that reliability is population specific. We suspect

that the continuous improvement seen in our present study
reflects the specific activity limitations of individuals recovering from lower-extremity amputation. Further research on the
learning effect in this population is needed. Further validation
of the 2MWT in these individuals should be undertaken, including different time frames for testing, comparisons with
other measures of functional mobility (eg, the timed-up-andgo), balance tests, and formal gait analysis.
CONCLUSION
The 2MWT is practical, simple, quick, and easy to administer. In the present prospective study, we found evidence of
good inter- and intrarater reliability of the 2MWT in for the
population of persons with unilateral below-knee amputation.
However, the distance walked in 2 minutes was not constant,
and it improved over time. The improvement was not solely the
result of a training and learning effect and does not appear to
plateau with repetition. The explanation for the improved performance remains unclear.
Acknowledgments: We thank Janet Campbell, Debbie Albuquerque, Ian Lowe, Maria Laskowski, and Rosalie Chan for their assistance
with data collection.
References
1. Solway S, Brooks D, Lacasse Y, Thomas S. A qualitative systematic review of the measurement properties of functional walk tests
used in the cardiorespiratory domain. Chest 2001;119:256-70.
2. Brooks D, Parsons J, Hunter JP, Devlin M, Walker J. The
2-minute walk test as a measure of functional improvement in
persons with lower limb amputation. Arch Phys Med Rehabil
2001;82;1478-83.
3. Larson JL, Covey MK, Vitalo CA, Alex AG, Patel M, Kim MJ.
Reliability and validity of the 12-minute distance walk in patients
with chronic obstructive pulmonary disease. Nurs Res 1996;45:
203-10.
4. Harada ND, Chiu V, Stewart AL. Mobility-related function in
older adults: assessment with a 6-minute walk test. Arch Phys
Med Rehabil 1999;80:837-41.
5. Montgomery PS, Gardner AW. The clinical utility of a six-minute
walk test in peripheral arterial occlusive disease patients. J Am
Geriatr Soc 1998;46:706-11.
6. Upton CJ, Tyrrell JC, Hiller EJ. Two minute walking distance in
cystic fibrosis. Arch Dis Child 1988;63:1444-8.
7. Connelly DM, Stevenson TJ, Vandervoort AA. Between- and
within-rater reliability of walking tests in a frail elderly population. Physiother Can 1996;48(1):47-51.
8. Boonstra AM, Fidler V, Eisma WH. Walking speed of normal
subjects: aspects of validity of gait analysis. Prosthet Orthot Int
1993:17:78-82.
9. Norman GR, Streiner DL. Biostatistics: the bare essentials. Hamilton (Ont): BC Decker; 1998.
10. Masanic CA, Bayley MT. Interrater reliability of neurologic soft
signs in an acquired brain injury population. Arch Phys Med
Rehabil 1998;79:811-5.
11. McDowell I, Newell C. Measuring health: a guide to treating
scales and questionnaires. New York: Oxford Univ Pr; 1987.
12. Butland RJ, Pang J, Gross ER, Woodcock AA, Geddes DM. Two-,
six-, and 12-minute walking tests in respiratory disease. BMJ
(Clin Res Ed) 1982;284:1607-8.
13. Bernstein M, Despars J, Singh N. Reanalysis of the 12-minute
walk in patients with chronic obstructive pulmonary disease.
Chest 1994;105:163-7.
14. Guyatt GH, Pugsley SO, Sullivan MT, et al. Effect of encouragement on walking test performance. Thorax 1984;39:818-22.
15. Guyatt GH, Sullivan MJ, Thompson PJ, et al. The 6-minute walk:
a new measure of exercise capacity in patients with chronic heart
failure. CMAJ 1985;132:919-23.
Arch Phys Med Rehabil Vol 83, November 2002