30 Physiotherapy Research International

Physiother. Res. Int. 14(1): 3041 (2009)

Published online 22 July 2008 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/pri.414

Factors with independent influence on

the timed up and go test in patients
with hip fracture
MORTEN TANGE KRISTENSEN Lund University, Department of Health Sciences,
Division of Physiotherapy, Lund University Hospital, Lasarettsgatan 7, SE-221 85
Lund, Sweden. Department of Physiotherapy, Orthopaedic Surgery, Hvidovre University
Hospital, Copenhagen, Denmark
NICOLAI BANG FOSS Department of Anaesthesiology, Orthopaedic Surgery, Hvidovre
University Hospital, Copenhagen, Denmark
HENRIK KEHLET Section of Surgical Pathophysiology 4074, Rigshospitalet University,
Copenhagen, Denmark
ABSTRACT Background and Purpose. Data on performance times for the timed up and
go (TUG) test with analyses of factors, that eventually could affect the result in patients
with hip fracture, have not been published to date. The aims of the present study, therefore,
were to assess normative reference values of TUG performances and determine the influence of individual and clinical factors on TUG-test scores in patients with hip
fracture. Method. In this prospective, descriptive study, a total of 196 consecutive patients
over the age of 60, and able to perform the TUG when discharged directly to their own
homes from a specialized orthopaedic hip fracture unit, were evaluated. The association
between TUG scores and categorical variables were examined, and linear regression was
used to investigate the factors influencing performance times. Results. Univariate analysis showed significant differences between all categorical variables, except gender, but
multivariate linear regression analyses showed that only a high pre-fracture function level,
evaluated by the New Mobility Score (B = 11), was independently associated with having
a good TUG score, while older age (B = 0.49), having an intertrochanteric fracture (B =
7), performing TUG with a walker (B = 15), and performing TUG in the later postoperative period (B = 0.39) were independently associated with having a poorer TUG
score. Conclusions. These preliminary normative reference values of TUG performances
in patients with hip fracture can be used as references, to which individuals can expect to
perform. Multivariate testing suggests that clinicians should use age, pre-fracture function,
fracture type and walking-aid specific data when interpreting the TUG test results. Physiotherapists should be aware of this if TUG scores are to be used predictively or as an
outcome measure in patients with hip fracture, especially in research. Copyright 2008
John Wiley & Sons, Ltd.

Copyright 2008 John Wiley & Sons, Ltd

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

Timed up and go test in patients with hip fracture 31

Key words: interpretation, patients with hip fracture, timed up and go (TUG) test

INTRODUCTION
A variety of measures are used by physiotherapists to assess the level of functional
mobility and as an outcome measure for
patients during post-hip fracture rehabilitation (Parker and Palmer, 1993; Freter and
Fruchter, 2000; Hall et al., 2000).
The timed up and go (TUG) test is a
score that measures functional mobility as
the time in seconds that it takes an individual to rise from an armchair, walk 3 m to a
line drawn on the floor and return to the
chair (Podsiadlo and Richardson, 1991). The
test has been used worldwide in several
studies, and mean (standard deviation [SD])
performance times from 7.7 (2.3) to 23.6 (16.7)
seconds for the TUG have previously been
reported for different groups of communitydwelling elderly according to gender, age,
chair seat height and use of walking aids
(Thompson and Medley, 1995; Siggeirsdottir
et al., 2002; Steffen et al., 2002; Bischoff
et al., 2003; Lusardi et al., 2003; Lin et al.,
2004). Higher mean (SD) performance times
of 28.2 (23.0) and 30.0 (17.4) seconds have
been reported for institutionalized elderly
(Bischoff et al., 2003; Nordin et al., 2006).
Very high intra- and inter-rater reliability
(Podsiadlo and Richardson, 1991; Noren
et al., 2001) and testretest reliability (Steffen
et al., 2002; Nilsagard et al., 2007) have been
found for the TUG. Also, significant correlations between the TUG score and scores
from the Barthel Index of activities of daily
living, the Berg Balance Scale and gait speed
testing (Podsiadlo and Richardson, 1991),
the General Motor Function Assessment
Scale (Gustafsson and Grahn, 2007), the
Four Square Step Test (Dite and Temple,
2002) and the Tinetti Balance (Lin et al.,

Copyright 2008 John Wiley & Sons, Ltd

2004) have been found. The TUG has also

been found to be able to discriminate multiple fallers from non-multiple fallers with
medians (2575% quartiles) of 16.7 (14.9
22.1%) and 12.3 (9.814.7%; Dite and
Temple, 2002). Shumway-Cook et al. (2000)
reported means (SD) of 22.2 (9.3) and 8.4
(1.7) for fallers and non-fallers, respectively,
while Okumiya et al. (1998) reported a cutoff point of 16.0 seconds in predicting falls
in community-dwelling older active.
In patients with hip fracture, the TUG
has been used in several studies (Freter
and Fruchter, 2000; Crotty et al., 2003;
Ingemarsson et al., 2003; Mendelsohn et al.,
2003; Kristensen et al., 2007), reporting
performance times at different time points
during the post-hip fracture rehabilitation.
At discharge, Ingemarsson et al. (2003)
found mean (SD) values from 37.7 (18.7) to
45.7 (23.4), while Crotty et al. (2003) showed
median (2575% quartiles) values of 37.0
(24.049.5%) for patients in an accelerated
discharge and home rehabilitation program.
Mendelsohn et al. (2003) and Freter and
Fruchter (2000), who used the TUG to assess
patients with hip fracture admitted to an
inpatient rehabilitation program, reported
higher means (SD) of TUG scores at 48.6
(25.6) and 55.1 (26.9) upon admission, compared with patients discharged to their own
homes (Crotty et al., 2003). In patients with
hip fracture, the TUG has been found useful
as an outcome measure (Freter and Fruchter,
2000; Crotty et al., 2003; Ingemarsson et al.,
2003; Mendelsohn et al., 2003) in predicting
falls with a cut-off point of 24.0 seconds
(Kristensen et al., 2007), and was found to
correlate well with the Berg Balance Scale
(Hall et al., 2000) and gait time (Freter and
Fruchter, 2000).

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

32 Kristensen et al.
However, data on performance times
with analyses of factors that eventually could
affect the TUG-performance in patients with
hip fracture remain unpublished. Understanding the factors that influence TUG
performance and determining normative
reference values for patients with hip fracture may be useful in screening patients
upon discharge for fall risk, and in interpreting change in TUG with clinical practice.
The aims of this study, therefore, were to
examine the influence of individual and
clinical factors on the TUG performance,
and to establish preliminary normative
reference values for the TUG in patients
with hip fracture when discharged from an
acute orthopaedic ward.
METHODS
Recruitment
Patients admitted from their own homes to
the special hip fracture unit at Hvidovre
University Hospital between September
2002 and August 2004 were included consecutively in this prospective, descriptive
study.
Inclusion and exclusion criteria
The inclusion criterion was that patients
should be able to perform the TUG at the
day before discharge from the hip fracture
unit. Patients, who were not allowed full
weight bearing post-operatively, and those
without pre-fracture indoor walking ability
(walking aids allowed, but without assistance), were excluded from the study.
Rehabilitation
Rehabilitation took place in the orthopaedic
ward, and patients were discharged when

Copyright 2008 John Wiley & Sons, Ltd

they were safely able to perform basic

mobility skills, defined as independence in
getting in and out of bed, sitting down and
standing up from a chair or toilet, and
walking with the aid to be used at home.
Only patients, who after 23 weeks of intensive rehabilitation at the orthopaedic ward
still required additional rehabilitation, were
transferred to a secondary rehabilitation
facility. During their hospital stay, all
patients followed a well-defined care plan
with multimodal fast-track rehabilitation
(Foss et al., 2005), including early surgery
within 24 hours of admission. Patients were
mobilized on the day of surgery, and an
intensive physiotherapy programme comprising two daily sessions was initiated on
the first day after surgery. The content of
this programme was focused on improvement in transferring, walking and balance.
The performance of the TUG is part of
normal discharge routines in the unit.
Ethical considerations
This study is part of Hvidovre University
Hospitals hip fracture project, which was
approved by the local ethics committee. The
study was approved by the Danish data
protection agency.
Measurements
Data were gathered prospectively in a database for all patients admitted to the unit.
Upon admission, information on age, sex,
cerebrovascular, cardiovascular and pulmonary diseases, type of hip fracture and
mental status evaluated by a validated 9point Danish version of the abbreviated
mental test (Qureshi and Hodkinson, 1974)
were recorded by physicians. The mental
score has a cut-off value of five or less
for cognitive impairment, which has been

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

Timed up and go test in patients with hip fracture 33

described as an independent predictor of
long-term mortality (Parker and Palmer,
1993). The pre-fracture functional level,
evaluated by the New Mobility Score (NMS;
Parker and Palmer, 1993) and use of walking
aids were recorded by physiotherapists at the
orthopaedic ward. Physiotherapists recorded
the performance time in seconds, the walking
aid used when performing and the postoperative day for the TUG on the day before
discharge to the community. Also, the reason
for patients not being able to perform the
TUG and their discharge residential status
were recorded at that time.
The fracture type was classified as
cervical according to Garden (Caviglia
et al., 2002), as intertrochanteric according
to Evans (Andersen et al., 1990) or as
subtrochanteric.
The NMS is a composite score of the
patients ability to perform indoor walking,
outdoor walking and shopping before the hip
fracture, providing a score between zero and
three (0 = not at all, 1 = with help from
another person, 2 = with an aid and 3 = no
difficulty) for each function, resulting in a
total score from 0 (no walking ability at all)
to 9 (fully independent). Patients were asked
about their walking ability in the last few
weeks before admittance with hip fracture
and, if necessary, relatives or caregivers
were consulted for verification. The NMS
has a high inter-tester reliability (Kristensen
et al., 2008) and a score of more than 5 is a
valid predictor of a good long-term outcome
in patients with hip fracture (Parker and
Palmer, 1993; Kristensen et al., 2005).
The TUG measures the time (in seconds)
that it takes an individual (as quickly and
safely as possible) to rise from an armchair
(chair seat height = 45 cm), walk 3 m to a
line drawn on the floor and return to the
chair. The time was measured from a seated
position (back against the backrest) with a

Copyright 2008 John Wiley & Sons, Ltd

stopwatch started on the command ready

go and stopped again when the seated
position was regained. No individual physical assistance was allowed; however, verbal
cuing during the test was allowed, if necessary. The patient was given a practice trial
followed by one timed trial, both performed
on the day before discharge. Patients performed the TUG in accordance with national
guidelines previously developed and examined for reliability in the study unit, with
intra-tester and inter-tester reliability (intraclass correlation) of 0.95 and 0.93, respectively. All TUG-tests were conducted by one
of six trained orthopaedic physiotherapists
after participation in one training session.
The use of walking aids was also assessed,
with reports of the walking aid used during
the TUG test.
Statistics
Descriptive statistics, means, standard deviations (SD) and 95% confidence intervals
(CI) were calculated for the TUG score by
gender, age group, NMS, type of walking
aid used pre-fracture, and when performing
the TUG at discharge, fracture type and
post-operative day of performing the TUG.
Only four patients had subtrochanteric fractures and no differences were found compared with patients with intertrochanteric
fractures. Accordingly, all were classified
as intertrochanteric in the analyses. If data
for the TUG scores, evaluated using the
KolmogorovSmirnov test, were normally
distributed, the Students t-test or the oneway analysis of variance (ANOVA) test was
used; otherwise the MannWhitney test or
the KruskallWallis test was used. The cutoff points for the individual variables were
as follows: gender (male female), age
(6069 to 7079 to 8089 to 90 years of age,
according to previous studies; Steffen et al.,

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

34 Kristensen et al.
2002; Lusardi et al., 2003; Bohannon, 2006),
NMS (05 [low pre-fracture level] to 69
[high], according to previous studies; Parker
and Palmer, 1993; Kristensen et al., 2005),
pre-fracture walking aids (nonestick
or elbow crutchesrollator), fracture type
(cervicalintertrochanteric), walking aids at
discharge (1 or 2 elbow crutchesrollator
walker) and post-operative day of performing TUG (day 0 to 10later, due to overall
median day 10 of performing the TUG). For
categorical data with significant differences
between groups, multivariate linear regression analysis was carried out to identify
independent factors associated with TUG
performance. Age and day of performing the
TUG were entered as continuous variables;
fracture type was entered as categorical,
while the pre-fracture walking aid and the
walking aid used when performing the TUG
were entered in the analysis and were coded
as dummy variables. A dummy variable is
one of the variables in a group (like crutches
in type of walking aids), which is excluded
from the regression, making this the base
category against which the other walking
aids used when performing the TUG are
compared with. No pre-fracture walking aid
and crutches when performing the TUG
were chosen as indicators due to the fastest
(best) TUG performance times in these
groups. Individuals under 60 years of age
and those discharged to further inpatient
rehabilitation in the community, were
excluded from all analyses in order to provide
normative values and analyses of factors, as
similar as possible, to those of previous
studies using the TUG in older adults with
or without hip fracture.
Data are presented as means (SD) when
normally distributed, otherwise presented as
medians (2575% quartiles). The level of
significance was set at p < 0.05. All analyses
were performed with SPSS version 12.0.

Copyright 2008 John Wiley & Sons, Ltd

RESULTS
A total of 436 patients with hip fracture were
admitted from their own homes to the unit,
with 266 being able to perform the TUG
upon discharge (Figure 1). Among these,
196 patients, aged 60 years or older, performed the TUG at a median of 10 (714)
days after surgery, with a mean test time of
35.9 (18.9) seconds when discharged to their
own homes, directly from the acute orthopaedic ward. The pre-fracture status for the
196 patients included in analysis, and the 131
patients unable to perform the TUG upon
discharge from the unit are shown in Table
1. The 131 patients, who were not able to
perform the TUG, were significantly older,
had lower pre-fracture functional level and
mental status upon admission, and were,
discharged later (p < 0.001).
Thirty-three patients were excluded, as
they were younger than 60 years of age.
They performed the TUG at a mean of 29.0
(16.4) seconds at median day 7 (510) after
surgery. A further number of 37 patients
with a median age of 84 (8090) years, who
performed the TUG at a mean time of 79.1
(62.5) seconds, at median day 19 (1427)
after surgery, were discharged to further
inpatient rehabilitation, and therefore excluded. The association between postoperative day of TUG and TUG performance
on categorical variables with descriptive
data for mean (SD) and 95% CI of mean for
the TUG-test score are shown in Table 2.
All variables, except gender, were significant in univariate testing (Table 2) and were
included in the multivariate analysis, which
showed that only a high pre-fracture functional level (p < 0.001) was independently
associated with having a fast TUG time,
while older age (p = 0.002), having a intertrochanteric fracture (p = 0.006), performing TUG with a walker (p < 0.001) and
Physiother. Res. Int. 14: 3041 (2009)
DOI: 10.1002/pri

Timed up and go test in patients with hip fracture 35

436 patients admitted
from their own homes
Patients excluded due to:
- Low prefracture level
- Surgical restrictions of
mobilisation
Patients not performing
the TUG due to:
- Medical complications
- Death during admittance
- Low functional level
at discharge
- Mental status
- Logistic
Total

39 patients < 60 years of age

(33 performed the TUG at
discharge)

8
25

33

22
35
57
9
8 131
164

397 patients 60 years of

age

37 patients performed the

TUG when discharged to
further inpatient rehabilitation

196 patients performed the TUG when

discharged directly to their own homes.

FIGURE 1: Flowchart of 436 consecutive patients with hip fracture admitted to an acute orthopaedic ward
from their own homes, with 196 performing the timed up and go (TUG) test when discharged directly to
their own homes.

TABLE 1: Baseline characteristics and results for patients with hip fracture able to perform the timed up
and go (TUG) test (n = 196) and potential patients (n = 131) not able to perform the test when discharged
from an acute orthopaedic ward
Characteristics

Not included
n = 131

Included
n = 196

Age (years)
Women
Low pre-fracture functional level (New Mobility Score < 6)
Pre-fracture use of walking aids
Cerebrovascular disease
Cardiovascular disease
Pulmonary disease
Dementia on admission
Fracture type:
Cervical
Trochanteric
Subtrochanteric
Low functional mobility at discharge (TUG of 30 seconds)
Post-operative day of discharge

83 (7990)
107 (82)
95 (73)
101 (77)
47 (36)
46 (35)
16 (12)
38 (29)

80 (7485)
156 (80)
48 (24)
87 (44)
26 (13)
75 (38)
18 (9)
9 (6)

63 (48)
60 (46)
8 (6)

27 (1751)

116 (59)
76 (39)
4 (2)
112 (57)
11 (815)

Data are presented as number (percentage) for categorical data and as median (2575 quartiles) for numerical data.

Copyright 2008 John Wiley & Sons, Ltd

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

36 Kristensen et al.
TABLE 2: Associations between the timed up and go (TUG) test in seconds performed at time of discharge
from an acute orthopaedic ward and categorical variables in patients with hip fracture (n = 196)
Variables

n (%)

Post-operative
day of TUG

TUG in
seconds

95% CI
of mean

Univariate
association (p)

Overall
Male
Female
Age
1: Age 6069
2: Age 7079
3: Age 8089
4: Age 90
Pre-fracture functional level
NMS 05 (low)
NMS 69 (high)
Pre-fracture walking aid
1: None
2: Stick or elbow crutches
3: Rollator
Fracture type
Cervical
Intertrochanteric
Walking aid when TUG tested
1: Elbow crutches
2: Rollator
3: Walker
Post-operative day of TUG
Performed at day 010
Performed after day 10

196
39 (20)
156 (80)

10 (714)
12 (715)
10 (713)

35.9 (18.9)
32.4 (16.0)
36.7 (19.5)

33.238.6
27.337.6
33.639.8

=0.22d

27 (14)
55 (28)
94 (48)
20 (10)

9 (614)
8 (612)
11 (814)
14 (1025)

27.4 (9.3)
30.7 (14.6)
38.4 (17.6)
50.1 (31.3)

23.731.1
26.734.6
34.842.0
35.464.7

<0.001c
1,2 vs. 3,4

48 (24)
148 (76)

14 (1018)
9 (712)

45.0 (23.8)
33.0 (16.0)

38.151.9
30.435.6

<0.001b

109 (56)
43 (22)
44 (22)

8 (612)
12 (915)
13 (1018)

33.0 (16.7)
42.3 (25.8)
37.0 (14.2)

29.836.1
32.741.3
34.450.2

=0.006a
1 vs. 2

116 (59)
80 (41)

10 (714)
13 (1119)

31.7 (14.8)
42.0 (22.2)

29.034.4
37.147.0

=0.006b

69 (35)
59 (30)
68 (35)

7 (610)
11 (815)
13 (919)

25.2 (9.8)
35.1 (14.0)
47.6 (22.6)

22.827.5
31.438.7
42.153.0

<0.001a
1 vs. 2,3
2 vs. 3

103 (53)
93 (47)

7 (69)
14 (1219)

30.9 (14.6)
41.5 (21.3)

29.033.7
37.145.9

<0.001b

Values are presented as number of patients (percentage), as median (2575 quartiles), as mean (SD) and
95% CI of mean. The following tests were used as the statistical test for significance: a ANOVA, b Students
t-test, c KruskallWallis Test, d Mann-Whitney Test.
* Significant differences between groups.
CI = confidence intervals; NMS = New Mobility Score; SD = standard deviation.

performing TUG in the later post-operative

period (p = 0.021), were independently associated with performing the TUG slowly
(Table 3). The R square showed that the
regression model, which included age, prefracture functional level, fracture type, type
of walking aid and day of TUG performance
was able to explain 39% of the variation in
TUG scores. All residuals were normally
distributed and no multi-collinearity was
found.

Copyright 2008 John Wiley & Sons, Ltd

DISCUSSION
In the present study, we have established
mean (SD) and 95% CI of mean TUGperformance times for categorical patient
variables by gender, age group, pre-fracture
functional level, type of walking aid, fracture type and post-operative day of TUG test
that can serve as initial normative reference
values for people during the post-hip
fracture rehabilitation period. Univariate

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

Timed up and go test in patients with hip fracture 37

TABLE 3: Multivariate linear regression indicating the factors independently influencing the time in seconds
for the timed up and go (TUG) test (n = 196)
Variables

Age
High pre-fracture level, New Mobility Score 69
Use of crutches pre-fracture*
Use of rollator pre-fracture*
Intertrochanteric fracture
Use of rollator when TUG**
Use of walker when TUG**
Post-operative day of TUG performance

0.486
10.746
0.192
2.477
6.506
0.955
15.096
0.392

95 % CI for B
Lower

Upper

0.185
16.488
5.807
8.975
1.899
7.786
9.428
0.012

0.788
5.003
6.190
4.020
11.114
5.875
20.763
0.772

0.002
<0.001
0.950
0.453
0.006
0.783
<0.001
0.043

Dummy parameters, indicator set as no use of walking aids * and crutches **.

analyses showed significant differences

between all categorical variables, except
gender. Finally, age, pre-fracture functional
level, fracture type, post-operative day of test
and the walking aid used when performing
the TUG were shown to be independently
associated with TUG performance times in
multivariate linear regression analyses.
Gender
Previous studies have reported that women
have reduced walking speed (Bohannon,
1997) and TUG scores (Thompson and
Medley, 1995) compared with men, but like
others (Medley and Thompson, 1997; Hall
et al., 2000; Lusardi et al., 2003), we found
no differences between genders on TUG
scores in our study. One explanation for
these disparate findings may be that
Bohannon (1997) and Thompson and Medley
(1995) studied patients younger than 80
years of age, while ours and other studies
(Medley and Thompson, 1997; Hall et al.,
2000; Lusardi et al., 2003), included patients
older than 80 years of age. It seems likely
that differences due to gender, also in patients
with hip fracture, are diminished when age

Copyright 2008 John Wiley & Sons, Ltd

groups over 80 are included in the analyses,

as in those groups, other factors may be
more predictive of TUG performances.
Age
Normative reference values from 8.1 to 11.3
seconds with significant differences between
three age groups (6069, 7079 and 8099)
in TUG performance among primarily
healthy elderly have been established in a
descriptive meta-analysis including 4,395
patients from 21 studies (Bohannon, 2006).
One study by Thompson and Medley (1995;
not included in the meta-analysis) found no
difference between three age groups (65
79); Steffen et al. (2002) found a trend
towards age-related declines in three age
groups (6089), while Lusardi et al. (2003)
in four age groups 60101 found age to be
an important predictor of TUG performance
times. We studied four age groups (6099)
similar to Lusardi et al. (2003) and found
that older age in multivariate analysis
was also significantly associated with
slower TUG performances, but similar to
Thompson and Medley (1995), we did not
find any significant differences between our

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

38 Kristensen et al.
two similar age groups (6079) in the univariate analyses. Accordingly, the importance of the use of age-specific reference
values, as stated by Lusardi et al. (2003)
when attempting to characterize functional
performance in older adults, seems to be
relevant in the interpretation of TUG performances in older patients with hip fracture.
Pre-fracture functional level
A large number of studies in patients with
hip fracture have previously established that
the pre-fracture functional level and ambulatory ability are important predictors of
rehabilitation outcome in patients with hip
fracture (Guccione et al., 1996; Koval et al.,
1998; Cree and Nade, 1999). Accordingly,
we found that a patient with a high prefracture functional level, evaluated by the
NMS, requires on the average 11 seconds
less to perform the TUG compared with a
patient with a low pre-fracture level to
perform the TUG compared with a patient
with a low pre-fracture level. This shows
that the TUG, at this early stage of rehabilitation, has the ability to reflect the prefracture functional level. This fact should
also be taken into consideration when interpreting the TUG test results in patients with
hip fracture.
Fracture type
Patients with an intertrochanteric fracture
required on the average 7 seconds more to
perform the TUG, compared with patients
with cervical fractures in our study. This is
in accordance with a previous study that
showed elderly women with an intertrochanteric fracture to be more functionally
impaired, compared with one with a femoral
neck fracture, at the time of discharge
(Haentjens et al., 2007).

Copyright 2008 John Wiley & Sons, Ltd

Type of walking aid

The use of walking aids compared with none
has previously been found to be associated
with slower TUG performances among
community-dwelling elderly (Podsiadlo and
Richardson, 1991; Thompson and Medley,
1995; Lusardi et al., 2003). As all patients in
our study needed a walking aid to perform
the TUG, we could not make this comparison. We found, however, that those who used
a stick or elbow crutches before the hip fracture, compared with independent ambulators were also significantly slower when
performing the TUG in univariate, but not
in multivariate analyses. In contrast to this,
we found significant differences in the univariate analyses between all three walking
aids (crutches, rollator and walker) used
when performing the TUG. Performing the
TUG with a walker was found to be
the independent variable with the strongest
influence on TUG performances. Thus, a
patient performing the TUG with a walker
requires, on average, 15 seconds more to
perform the TUG compared with a patient
performing the TUG with crutches. These
results are equal to findings by Medley and
Thompson (1997), who examined the effect
of these three walking aids on TUG performance times in 187 community-dwelling
elderly and independent ambulators, with no
history of lower extremity fracture or neurological disease. The choice of walking aid
upon discharge and when performing the
TUG, is a bit more complex in patients with
hip fracture. Crutches are primarily chosen
in patients having a high post-operative
functional level, while a walker instead
of a rollator among those with a lower postoperative level was chosen most often due
to indoor conditions at their discharge residence, such as doorsteps. A total of 35% (68
out of 196) of patients in our study performed

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

Timed up and go test in patients with hip fracture 39

the TUG with a walker at a mean of 47.6
(22.6) seconds. Our results might have been
different if these patients had been given the
opportunity to perform the TUG with a rollator instead, as the 59 patients, who actually
did this despite higher age and a lower prefracture level, performed the TUG significantly faster at a mean of 35.1 (14.0) seconds
(p < 0.001). Therefore, the effect of different
walking aids on TUG performances should
be investigated in future studies. Physiotherapists should for the time being
report the walking aid used when performing
the TUG, and if one is to compare an individuals test results over time, the same aid
should be used for the retest. Also, performances between individuals using different
walking aids should not be compared.
Our study showed that those performing
the TUG late in the post-operative period
also had significantly slower mean scores
reflecting a lower ambulatory level, in accordance with the need of a longer recovery
period before fulfilling discharge criteria.
No previous studies have examined the
TUG in patients with hip fracture as detailed
in this study, but several studies have used
the test as an outcome measure, providing
TUG performance times at different time
points during post-hip fracture rehabilitation (Freter and Fruchter, 2000; Crotty et al.,
2003; Ingemarsson et al., 2003; Mendelsohn
et al., 2003; Kristensen et al., 2007). Our
performance values are within the range
(SD) of comparable studies (Crotty et al.,
2003; Ingemarsson et al., 2003), giving more
strength and relevance to the values and conclusions presented in this study. On the other
hand, scores of patients discharged to further
inpatient rehabilitation in our study had
slower TUG times than those reported on
admission in studies by Mendelsohn et al.
(2003) and Freter and Fruchter (2000).
It may be due to the different standards

Copyright 2008 John Wiley & Sons, Ltd

regarding transfer to further inpatient

rehabilitation.
A potential limitation of our study and
the use of the TUG test in patients with hip
fracture is that not all patients were able to
perform the test. Nevertheless, a total of
72% (266 out of 368 potential patients)
performed the TUG test upon discharge.
Implications
Initial normative reference values of TUG
performances have been established in
patients with hip fracture, but physiotherapists should take factors such as age, the
pre-fracture functional level, the fracture
type and the walking aid used when performing the TUG into consideration when
interpreting the TUG test results. TUG
performances using different walking aids,
especially those that have not been investigated, should not be compared.
ACKNOWLEDGEMENTS
This project was funded by a grant from the IMK
Fonden and the Danish Physiotherapy Organization,
Copenhagen, Denmark. The authors thank Derek
Curtis, MSc, PT and Thomas Bandholm, MSc, PT for
comments on the manuscript.

REFERENCES
Andersen E, Jorgensen LG, Hededam LT. Evans
classification of trochanteric fractures: an assessment of the interobserver and intraobserver reliability. Injury 1990; 21: 377378.
Bohannon RW. Comfortable and maximum walking speed of adults aged 2079 years: reference
values and determinants. Age Ageing 1997; 26:
1519.
Bohannon RW. Reference values for the timed up and
go test: a descriptive meta-analysis. Journal of
Geriatric Physical Therapy 2006; 29: 6468.
Caviglia HA, Osorio PQ, Comando D. Classification
and diagnosis of intracapsular fractures of the
proximal femur. Clinical Orthopaedics and
Related Research 2002; 399: 1727.

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

40 Kristensen et al.
Cree AK, Nade S. How to predict return to the community after fractured proximal femur in the
elderly. The Australian New Zealand Journal of
Surgery 1999; 69: 723725.
Crotty M, Whitehead C, Miller M, Gray S. Patient
and caregiver outcomes 12 months after homebased therapy for hip fracture: a randomized controlled trial. Archives of Physical Medicine and
Rehabilitation 2003; 84: 12371239.
Dite W, Temple VA. A clinical test of stepping and
change of direction to identify multiple falling
older adults. Archives of Physical Medicine and
Rehabilitation 2002; 83: 15661571.
Foss NB, Kristensen MT, Kristensen BB, Jensen PS,
Kehlet H. Effect of postoperative epidural analgesia on rehabilitation and pain after hip fracture
surgery: a randomized, double-blind, placebocontrolled trial. Anesthesiology 2005; 102:
11971204.
Freter SH, Fruchter N. Relationship between timed
up and go and gait time in an elderly orthopaedic
rehabilitation population. Clinical Rehabilitation
2000; 14: 96101.
Guccione AA, Fagerson TL, Anderson JJ. Regaining
functional independence in the acute care setting
following hip fracture. Physical Therapy 1996;
76: 818826.
Gustafsson U, Grahn B. Validation of the General
Motor Function Assessment Scale An instrument for the elderly. Disability and Rehabilitation
2007; 18.
Haentjens P, Autier P, Barette M, Venken K,
Vanderschueren D, Boonen S. Survival and functional outcome according to hip fracture type:
a one-year prospective cohort study in elderly
women with an intertrochanteric or femoral neck
fracture. Bone 2007; 41: 958964.
Hall SE, Williams JA, Senior JA, Goldswain PR,
Criddle RA. Hip fracture outcomes: quality of life
and functional status in older adults living in the
community. Australian and New Zealand Journal
of Medicine 2000; 30: 327332.
Ingemarsson AH, Frandin K, Mellstrom D, Moller M.
Walking ability and activity level after hip
fracture in the elderly a follow-up. Journal of
Rehabilitation Medicine 2003; 35: 7683.
Koval KJ, Skovron ML, Aharonoff GB, Zuckerman
JD. Predictors of functional recovery after hip
fracture in the elderly. Clinical Orthopaedics and
Related Research 1998; 348: 2228.
Kristensen MT, Bandholm T, Foss NB, Ekdahl C,
Kehlet H. High inter-tester reliability of the New

Copyright 2008 John Wiley & Sons, Ltd

Mobility Score in patients with hip fracture.

Journal of Rehabilitation Medicine 2008; 40:
589591.
Kristensen MT, Foss NB, Kehlet H. Timed Up and
Go and New Mobility Score as predictors of function six months after hip fracture. Ugeskr Laeger
2005; 167: 32973300.
Kristensen MT, Foss NB, Kehlet H. Timed up & go
test as a predictor of falls within 6 months after
hip fracture surgery. Physical Therapy 2007; 87:
2430.
Lin MR, Hwang HF, Hu MH, Wu HD, Wang YW,
Huang FC. Psychometric comparisons of the
timed up and go, one-leg stand, functional reach,
and Tinetti balance measures in communitydwelling older people. Journal of the American
Geriatrics Society 2004; 52: 13431348.
Lusardi MM, Pellecchia GL, Schulman M. Functional performance in community living older
adults. Journal of Geriatric Physical Therapy
2003; 3: 1422.
Medley A, Thompson M. The Effect of Assistive
Devices on the Performance of Community
Dwelling Elderly on the Timed Up and Go Test.
Issues on Aging 1997; 20 (1:97): 37.
Mendelsohn ME, Leidl DS, Overend TJ, Petrella RJ.
Specificity of functional mobility measures in
older adults after hip fracture: a pilot study.
American Journal of Physical Medicine and
Rehabilitation 2003; 82: 766774.
Nilsagard Y, Lundholm C, Gunnarsson LG, Dcnison
E. Clinical relevance using timed walk tests and
timed up and go testing in persons with multiple
sclerosis. Physiotherapy Research International
2007; 12: 105114.
Nordin E, Rosendahl E, Lundin-Olsson L. Timed Up
& Go test: reliability in older people dependent
in activities of daily living focus on cognitive
state. Physical Therapy 2006; 86: 646655.
Noren AM, Bogren U, Bolin J, Stenstrom C. Balance
assessment in patients with peripheral arthritis:
applicability and reliability of some clinical
assessments. Physiotherapy Research International 2001; 6: 193204.
Okumiya K, Matsubayashi K, Nakamura T, Fujisawa
M, Osaki Y, Doi Y, Ozawa T. The timed up &
go test is a useful predictor of falls in community-dwelling older people. Journal of the
American Geriatrics Society 1998; 46: 928
930.
Parker MJ, Palmer CR. A new mobility score for
predicting mortality after hip fracture. Journal of

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri

Timed up and go test in patients with hip fracture 41

Bone and Joint Surgery. British Volume 1993; 75:
797798.
Podsiadlo D, Richardson S. The timed Up & Go: a
test of basic functional mobility for frail elderly
persons. Journal of the American Geriatrics
Society 1991; 39: 142148.
Qureshi KN, Hodkinson HM. Evaluation of a
ten-question mental test in the institutionalized
elderly. Age Ageing 1974; 3: 152157.
Shumway-Cook A, Brauer S, Woollacott M. Predicting the probability for falls in communitydwelling older adults using the Timed Up & Go
Test. Physical Therapy 2000; 80: 896903.
Siggeirsdottir K, Jonsson BY, Jonsson H, Jr.,
Iwarsson S. The timed Up & Go is dependent on
chair type. Clinical Rehabilitation 2002; 16:
609616.
Steffen TM, Hacker TA, Mollinger L. Age- and
gender-related test performance in communitydwelling elderly people: Six-Minute Walk Test,
Berg Balance Scale, Timed Up & Go Test, and

Copyright 2008 John Wiley & Sons, Ltd

gait speeds. Physical Therapy 2002; 82:

128137.
Thompson M, Medley A. Performance of community
dwelling elderly on the timed up and go test.
Physical and Occupational Therapy in Geriatrics
1995; 13: 1730.
Von Dechend M, Stahelin HB, Monsch AU, Iversen
MD, Weyh A, von DM, Akos R, Conzelmann M,
Dick W, Theiler R. Identifying a cut-off point for
normal mobility: a comparison of the timed up
and go test in community-dwelling and institutionalised elderly women. Age Ageing 2003; 32:
315320.
Address correspondence to: Morten Tange Kristensen, Department of Physiotherapy 236, Hvidovre
University Hospital, Kettegaard Alle 30, Copenhagen
DK-2650,
Denmark.
E-mail:
morten.tange.
kristensen@hvh.regionh.dk
(Submitted January 2008; accepted June 2008)

Physiother. Res. Int. 14: 3041 (2009)

DOI: 10.1002/pri