Clinical Rehabilitation 2004; 18: 326]/334

The Trunk Impairment Scale: a new tool to measure

motor impairment of the trunk after stroke
G Verheyden, A Nieuwboer Department of Rehabilitation Sciences, Faculty of Physical Education and Physiotherapy,
Katholieke Universiteit Leuven, Belgium, J Mertin Neurologic Rehabilitation Unit of the Kiliani-Klinik, Bad Windsheim,
R Preger Neurologic Rehabilitation Unit of the Klinik Kipfenberg, Germany, C Kiekens Physical Medicine and Rehabilitation
Unit of the University Hospital Pellenberg, Katholieke Universiteit Leuven and W De Weerdt Department of Rehabilitation
Sciences, Faculty of Physical Education and Physiotherapy, Katholieke Universiteit Leuven, Belgium

Received 17th December 2002; returned for revisions 2nd April 2003; revised manuscript accepted 15th June 2003.

Objective: To examine the clinimetric characteristics of the Trunk Impairment Scale

(TIS). This newly developed scale evaluates motor impairment of the trunk after
stroke. The TIS scores, on a range from 0 to 23, static and dynamic sitting balance as
well as trunk co-ordination. It also aims to score the quality of trunk movement and to
be a guide for treatment.
Design: Two physiotherapists observed each patient simultaneously, but scored
independently. Each patient was re-examined by one of the therapists.
Subjects: Twenty-eight patients in a rehabilitation setting.
Results: Kappa and weighted kappa values for item per item reliability ranged for all
but two, from 0.62 to 1. All percentages of agreement exceeded 81%. Intraclass
correlations (ICC) for the summed scores of the different subscales were between
0.85 and 0.99. Test]/retest and interobserver reliability for the TIS total score (ICC)
was 0.96 and 0.99, respectively. The 95% limits of agreement for the test ]/retest and
interexaminer measurement error were 2/2.90, 3.68 and 2/1.84, 1.84, respectively.
Cronbach alpha coefficients for internal consistency ranged from 0.65 to 0.89.
Content validity was defined. Spearman rank correlations with the Barthel Index
(r 5 /0.86) and the Trunk Control Test (r 5 /0.83) was used to examine construct and
concurrent validity, respectively.
Conclusions: Analysis of different clinimetric parameters support the use of the TIS
in both clinical use and future stroke research. Guidelines for treatment and level of
quality of trunk activity can be derived from the assessment.

Introduction little attention. D avies4 associates the loss of

M ost literatu re concerning motor rehabilitatio n
after stroke focuses on the upper and lower
extremity.1] 3 Trunk rehabilitation receives only
Address for correspondence: G eert Verheyden, K atholieke
U niversiteit Leuven, F aculty of Physical Education and
Physiotherapy, D epartment of Rehabilitation Sciences, Tervuur-
sevest 101, B-3000 Leuven, Belgium. e-mail: geert.verheyden@
flok.kuleuven.ac.be
# Arnold 2004
selective control in the trunk with problems of
breathing, speech, balance, gait, arm and hand
function. Sittin g balance is also reported as a
predictor of motor and functio nal recovery after
stroke.5] 8 The importance of recovery of trunk
function, although often stated by clinical experts,
needs to be confirm ed by scientific research.
In the litera ture, the use of a clinical tool to
measure trunk function is poorly documented.
Sometimes a 3-, 4- or 5-point ordinal scale is
10.1191/0269215504cr733oa

used without mentioning origin or basic statistica l
characteristics. 5,7,9 The Sittin g Balance Scale devel-
oped by N ieuwboer et al.10 showed poor reliability,
especially for the items evaluating the quality of
trunk activity. The Trunk Control Test by Collin
and Wade is a quick and reliable measure with
predictive validity.11,12 Limitatio ns of the test are
that it does not take the quality of movement into
account 11 as well as the moderate correlation with
trunk strength, 13 measured using a hand dyna-
mometer. The latter was explained by Bohannon
because it needs more than trunk muscle strength
to complete the tasks of the Trunk Control Test. 13
In several existing scales, (isolated) trunk activity is
scored as one component. Examples are the `leg
and trunk’ scale of the R ivermead M otor Assess-
ment, 14 the `balanced sittin g’ item of the M otor
Assessment Scale, 15 the `postural control’ part of
the Chedoke-M cM aster Stroke Assessment, 16 the
`lying and sittin g’ tests of the Postural Assessment
Scale for Stroke Patients17 and the `verticality’ and
`abdominal manual muscle testing’ items of the
Stroke Impairment Assessment Set. 18
The aim of this study was to develop a
comprehensive tool to measure motor impairment
of the trunk after stroke, the Trunk Impairment
Scale (TIS), and to examine reliability, internal
consistency and validity of the TIS. The scale
should include the observation of quality of trunk
movement and be a guide for the treatment of the
trunk in stroke patients.
Methods
The TIS consists of three subscales: static sittin g
balance, dynamic sittin g balance and co-ordina-
tion. Each subscale contains between three and
ten items. The TIS score ranges from a minimum
of 0 to a maximum of 23.
In a preliminary study, scoring the symmetry
and manual lengthening of the hemiplegic and
nonhemiplegic side of the trunk were considered as
well. D ue to poor reliability, these items were
removed from the scoring system. Other parts of
the scale were redefined. The current scale is given
in the Appendix.
Stroke patients were recruited in rehabilitatio n
centres and were included if they were allowed to
T he T runk Impairment S cale 327
sit upright. Exclusion criteria were a hip prosthesis
at the nonhemiplegic side or a score of 0 on the
comprehension or speech item of the European
Stroke Scale.19 F urther data collection to define
the population consisted of patient’s age, sex,
hemiplegic side, type of stroke, days since stroke,
Barthel Index,20 BrunnstroÈ m-F ugl-M eyer test 21 for
upper and lower extremity and Trunk Control
Test.11
Twenty-eight patients (14 female and 14 male)
were included in the study. Eleven had a right
hemiplegia, 15 a left hemiplegia and two were
bilaterally affected. Eight patients suffered from
a haemorrhagic stroke, 20 had an ischaemic
accident. M edian age was 63 years, media n days
since stroke 61 (Table 1). The majorit y of patients
had moderate AD L function and limited motor
recovery (Table 1).
Ethical approval was obtained from the Ethical
Commission, M edical Faculty, K .U.Leuven, Bel-
gium. Informed consent was obtained from all
patients.
For the reliability study, each patient was
examined twice. On one occasion, two phy-
siotherapists scored the TIS concurrently but
independently. On another occasion, one of the
therapists assessed the patient alone. The therapist
who examined the patient alone also instructed the
patient when both observers were scoring simulta-
neously. The two observations were always on the
same day, separated by 1 or 2 hours of recovery
time. D uring that time no treatment was offered.
The observations were planned every half hour, so
at least two different patients were evaluated before
seeing the same patient again. To further minimize
recall bias, the observers filled in the score sheet
but did not add up the scores. Allocation of the
Table 1 Patient characteristics
Median (Q1,Q3)a Range
Age (years) 63 (47,71) 32 ]/87
Days since stroke 61 (46,94) 21 ]/2341
Barthel Index 60 (33,80) 5 ]/100
BrunnstroÈm-Fugl-Meyer 9 (4,54) 0 ]/66
Test (arm)
BrunnstroÈm-Fugl-Meyer 19 (6,25) 0 ]/31
Test (leg)
Trunk Control Test 75 (43,94) 0 ]/100
a
Q1 means quartile one; Q3 means quartile three.
328 G Verheyden et al.
patients to the observers as well as the order of the
two observations were randomized. Each observer
examined a group of 14 patients twice, amounting
to 28 patients in total.
In this study, every item of the scale was
performed three times to avoid a possible scoring
bias if a patient reached the maximum score after
one or two attempts.
Test ]/retest reliability was measured by compar-
ing the results of the therapists who examined the
patient twice. To determine inter-rater reliability,
the results of both therapists who observed the
patient simultaneously were compared.
Test ]/retest and interobserver reliability were
determined for all scale items. K appa and weighted
kappa values were calculated for dichotomous and
ordinal scales, respectively. Percentage of agree-
ment was also determined for all items. Test ]/retest
and interobserver reliability for the subtotals and
total score was examined by means of intraclass
correlation (IC C). The 95% test ]/retest and inter-
examiner measurement error interval according to
H aas22 was determined. Cronbach’s alpha was
calculated to check for internal consistency of the
subscales and total scale. Content validity was
evaluated. The TIS total score was compared with
the Barthel Index by means of Spearman rank
correlations for determining constru ct validity and
with the Trunk Control Test for concurrent
validity.
Item per item reliability was established when
the kappa or weighted kappa statistic exceeded
0.6023 or when more then 80% agreement was
observed. Test ]/retest and interobserver reliability
of the subtotals and TIS total was reached when
the ICC was 0.80 or higher. Cronbach’s alpha
should exceed 0.70, which is suggested as a value of
scale reliability and indicates underlying con-
struct. 24
Results
As seen in Table 2, most kappa or weighted kappa
values ranged from 0.62 to 1. Test ]/retest agree-
ment of item 2 of static sittin g balance (0.51) and
item 2 of co-ordination (0.46) were insufficient.
For these items, a high percentage of agreement,
89% and 93% respectively, was observed. N o
kappa or weighted kappa values could be calcu-
lated for item 3 (test ]/retest reliability) and item 4
(test ]/retest and inter-rater reliability) of co-ordi-
nation because of a skewed distribution of the
data. Percentage of agreement for these items
ranged from 86% to 100%.
ICCs for static and dynamic sittin g balance,
co-ordination and the total TIS are presented in
Table 3 and were between 0.85 and 0.99. The 95%
test ]/retest and interexaminer measurement error
interval on the total TIS score was 2/2.90, 3.68
and 2/1.84, 1.84, respectively.
Internal consistency by means of Cronbach’s
alpha was calculated for the subscales static sittin g
balance (0.79), dynamic sittin g balance (0.86) and
co-ordination (0.65). Cronbach’s alpha for the
total Trunk Impairment Scale was 0.89.
Content validity of the TIS was achieved
through literatu re review, observing stroke pa-
tients, clinical experience of the authors and
discussing the content of the scale with specialists
within the field of stroke rehabilita tion. Spearman
rank correlation between the TIS and the Barthel
Index (construct validity) was 0.86, between the
TIS and the Trunk Control Test (concurrent
validity) 0.83.
The time needed to complete the TIS ranged
from 2 to 18 minutes. All obtained scores ranged
between 0 and 21. The median score (quartile 1,
quartile 3) was 14 (10, 16).
Discussion
The aim of this study was to develop a measure-
ment tool to evaluate the impairment of the trunk
Clinical messages
. The Trunk Impairment Scale (TIS) is a new
tool to measure motor impairment of the
trunk after stroke.
. The TIS evaluates static and dynamic sittin g
balance as well as co-ordination of trunk
movement.
. The TIS has sufficient reliability, internal
consistency and validity for use in clinical
practice and stroke research.
 T he T runk Impairment S cale 329

Table 2 Kappa or weighted kappa, lower value of the 90% con® dence limit of the kappa or weighted kappa and percentage of
agreement for test]/retest and inter-observer agreement

Item j /j wa Test ]/retest agreement Inter-observer agreement

Valueb 90%lclc %d Valueb 90%lclc %d

Static sitting balance

Item 1 j 1 1 100% 1 1 100%
Item 2 j 0.51 0.11 89% 1 1 100%
Item 3 j w 0.87 0.77 86% 0.97 0.92 96%

Dynamic sitting balance

Item 1 j 0.70 0.47 86% 1 1 100%
Item 2 j 0.78 0.59 89% 0.93 0.81 96%
Item 3 j 0.62 0.37 82% 0.84 0.66 93%
Item 4 j 1 1 100% 1 1 100%
Item 5 j 1 1 100% 1 1 100%
Item 6 j 0.78 0.59 89% 0.93 0.81 96%
Item 7 j 0.93 0.81 96% 1 1 100%
Item 8 j 0.73 0.49 89% 0.91 0.76 96%
Item 9 j 0.62 0.37 82% 0.84 0.67 93%
Item 10 j 0.71 0.40 93% 0.76 0.49 93%

Co-ordination
Item 1 j w 0.76 0.57 86% 0.71 0.48 86%
Item 2 j 0.46 2/0.07 93% 0.78 0.44 96%
Item 3 j w *e 86% 0.70 0.50 82%
Item 4 j *e 100% *e 96%

a
j /j w indicates the use of a kappa (j ) or weighted kappa (j w) for statistic analysis.
b
Value of the calculated kappa or weighted kappa.
c
Lower value of the 90% confidence limit of the kappa or weighted kappa.
d
Percentage of agreement.
e
No Kappa or weighted Kappa could be calculated because of the skewed distribution of the data.

after stroke and to investigate its reliability, inter- respectively, were found in combination with a
nal consistency and validity. high percentage of agreement. H aas22 points out
that if there is a large percentage of agreement but
most of that agreement is limited to one of the
Reliability possible scores, the kappa value is not an appro-
For item 2 of static sittin g balance and co- priate index of reliability. In this study this would
ordination, low kappa values of 0.51 and 0.46, suggest that these items are too easy. This conclu-
sion would not be sound because a number of
Table 3 ICC for test]/retest and inter-observer agreement patients scored 0 on the tests. Secondly, we expect
that the evaluation of the static sittin g balance will
Total Test]/retest Inter-observer be particularly relevant when examining an acute
agreement agreement stroke population. Limited variance was also
Static sitting 0.91 (0.83) 0.99 (0.99) found for items 3 and 4 of co-ordination as well.
balance These items evaluate the symmetry of the (timed)
Dynamic sitting 0.94 (0.89) 0.98 (0.96) co-ordination of the lower part of the trunk.
balance Symmetric trunk movements are apparently diffi-
Co-ordination 0.87 (0.76) 0.85 (0.74)
Trunk Impairment 0.96 (0.93) 0.99 (0.97) cult for stroke patients. Because the TIS should
Scale also be a guide for treatment, inclusion of these
items seems appropriate. Besides if fast, alternating
Values are presented as ICC (90% lower confidence limit). trunk movements are difficult to achieve, there will
330 G Verheyden et al.
be no ceiling effect of the scale. The level of
difficulty for the various items could be determined
by means of a Rasch analysis. H owever a very large
number of patients is needed for this analysis.
Although the percentage of observer agreement
is high and exceeds the proposed limit, several
authors point out that the agreement due to
chance alone is not taken into account when
percentage agreement is presented. 22,25 To establish
the reliability of items 1 and 2 of static sittin g
balance (easy items) and items 2, 3 and 4 of co-
ordination (difficult items), a new study should be
set up examining the TIS on an acute ward with
severely impaired stroke patients and on a nearly
fully recovered stroke population, respectively.
Measurement error
In this study, values for inter-rater reliability
exceeded those for test ]/retest agreement. This
could be explained by the fact that in the case of
the inter-rater reliability, both observers were
scoring the patient at the same time, so all possible
areas of bias and variability were minimized.
Variation within the patient’s performance can
also be a possible reason for lower test ]/retest
reliability. This is confirm ed by the higher test ]/
retest examiner measurement error, 3.68 (16% of
TIS total) in comparison with the 1.84 (8% of TIS
total) for the interexaminer measurement error.
The latter is comparable with the reported limits of
agreement of the Action Research Arm test
(between 10 and 10.88%) and the BrunnstroÈ m-
F ugl-M eyer assessment scale (between 7.58 and
10%).26 For the Sickness Impact Profile, Becker-
man et al .27 also found a minimum decrease of
9.26% before an unbiased improvement can be
considered. The test ]/retest examiner measurement
error of the TIS (3.68) was based on a method
suggested by H aas,22 an alternative to the method
of Altman and Bland.28 An increase of 4 points on
the TIS can be seen as an improvement without
reproducibility bias.
Internal consistency
Cronbach’s alpha for the subscale co-ordination
was under the suggested value of 0.70. This reflects
only moderate reliability according to N unnally.24
Still the ICC values for the subscale co-ordination
are well above the critical value. F urther, 0.70 is
only a rule of thumb. H atcher and Stepanski29
describe literatu re where useful alpha coefficients
are reported under 0.70 or 0.60. In this study
Cronbach’s alpha was mainly used to examine the
underlying construct of the subscales and total
TIS, which seems confirm ed by the presented data.
Comparison with other scales
K appa and weighted kappa values for the test ]/
retest and inter-rater agreement of the individual
items of the TIS ranged from 0.46 to 1 and from
0.70 to 1, respectively. K appa values reported
for the lying and sitting items of the Postural
Assessment Scale for Stroke Patients 17 are simila r.
They ranged between 0.45 and 1 for intra- and
between 0.64 and 1 for inter-rater reliability.
N ieuwboer et al.10 reported lower values from
0.20 to 1. Items regarding quality of movement
ranged from 0.20 to 0.64. The previous two studies
do not mention percentage of agreement if there
was a low kappa value. Total scale score of the TIS
is highly reliable. Reported Spearman rho correla-
tion coefficients for the inter-rater agreement of
the Trunk Control Test 11 was 0.76, for the sittin g
balance item of the M otor Assessment Scale30
0.99. Product moment and rank order correlation
coefficients only have a limited value as indices of
reliability. Systematic errors are not taken into
account. 22,25,31,32 Adding up dichotomous and
ordinal items is a widely used method in scale
development. The total score can thus be seen as a
continuous variable. Therefore ICC can be used as
an appropriate statistic for examining reliability.
F ranchignoni et al. 12 reported for the Trunk
Control Test Cronbach’s alpha coefficients of 0.83
and 0.86. Benaim et al. 17 found for the Postural
Assessment Scale for Stroke Patients 0.95 as
coefficient of internal consistency. These are in
line with the results for the TIS in this study (0.89).
Content, construct (r5 /0.86 with Barthel Index)
and concurrent validity (r 5 /0.83 with Trunk Con-
trol Test) of the TIS were established. Construct
validity of the Trunk Control Test was examined
by comparison with the F unctional Independence
M easure (F IM ). Correlation coefficients ranged
from 0.71 to 0.79 and from 0.82 to 0.86 for the
total F IM and for the motor part of the F IM ,
respectively.12 Benaim et al.17 found a correlation
of 0.73 between the Postural Assessment Scale for
Stroke Patients and the F IM . A correlation of 0.28
was found between the sittin g balance item of the

M otor Assessment Scale and the BrunnstroÈ m-
F ugl-M eyer assessment.30
An additional aim of the TIS was to score the
quality of trunk movement and to be a guide for
treatment. The four tasks of the subscale dynamic
sitting balance evaluate an appropriate shortening
or lengthening of both sides of the trunk as well
as the use of compensation strategies. In the
present study, this method was found more reliable
than scoring a symmetrical or asymmetrical per-
formance as index of quality of movement. 10 Static
and dynamic sittin g balance as well as the co-
ordination of the trunk are considered. Although
not all aspects of trunk function are covered by the
TIS, it can be used as a guideline for treatment.
Proprioception of the trunk and muscle strength
for instance are not evaluated in the TIS. Therefore
further validation of the TIS is required to evaluate
its value as a measure of total trunk function.
Conclusion
This study reports the development of the Trunk
Impairment Scale, a clinical test to measure motor
impairment of the trunk after stroke. The TIS
measures static and dynamic sittin g balance and
co-ordination of the trunk. Item per item relia-
bility, subtotal and scale total agreement were
established as well as the internal consistency of
the subscales and the total scale. Content, con-
struct and concurrent validity were evaluated.
Statistica l analysis of the different clinimetric
parameters of the TIS endorse its further use in
clinical practice as well as in research. The assess-
ment can be used as a guideline for treatment and
takes the quality of trunk movement into account.
To the best of our knowledge, the TIS is a unique
measurement of its kind in stroke literatu re.
Acknowledgements
The authors would like to thank Inge Taillieu
and Ilse Baert for helping to collect the data. We
are also grateful for the logistic support of L Lutter
(K iliani-K linik Bad Windsheim, G ermany),
F Roelandt (K linik K ipfenberg, G ermany) and
P Popelier (U niversitair Ziekenhuis Pellenberg,
Belgium).
T he T runk Impairment S cale 331
References
1 M oreland J, Thomson M A. Ef®cacy of
electromyographic biofeedback compared with
conventional physical therapy for upper-extremity
function in patients following stroke: a research
overview and meta-analysis. Phys T her 1994; 74:
534]/47.
2 M oreland JD, Thomson M A, Fuoco AR .
Electromyographic biofeedback to improve lower
extremity function after stroke: a meta-analysis.
A rch Phys M ed R ehabil 1998; 79: 134]/40.
3 van der Lee JH , Snels IAK , Becker man H,
Lankhorst G J, Wagenaar RC, Bouter LM . Exercise
therapy for arm function in stroke patients: a
systematic review of randomized controlled trials.
Clin R ehabil 2001; 15: 20 ]/31.
4 Davies PM . Problems associated with the loss of
selective trunk activity in hemiplegia. In: R ight in
the middle. Berlin, H eidelberg: Springer-Verlag,
1990: 31 ]/65.
5 Wade D T, Skilbeck CE, Langton Hewer R .
Predicting Barthel AD L score at 6 months after an
acute stroke. A rch Phys M ed R ehabil 1983; 64: 24 ]/
28.
6 Loewen SC, Anderson BA. Predictors of stroke
outcome using objective measurement scales.
Stroke 1990; 21: 78 ]/81.
7 Sandin K J, Smith BS. The measure of balance in
sitting in stroke rehabilitation prognosis. S troke
1990; 21: 82 ]/86.
8 Kwak kel G, Wagenaar RC, Kollen BJ, Lankhorst
GJ. Predicting disability in stroke ]/ a critical review
of the literature. A ge A geing 1996; 25: 479]/89.
9 Bohannon RW. Lateral trunk ¯exion strength:
impairment, measurement reliability and
implications following unilateral brain lesion. Int J
R ehabil R es 1992; 15: 249]/51.
10 Nieuwboer A, Feys H, De Weerdt W, N uyens G, De
Corte E. D eveloping a clinical tool to measure
sitting balance after stroke: a reliability study.
Physiotherapy 1995; 81: 439]/45.
11 Collin C, Wade D. Assessing motor impairment
after stroke: a pilot reliability study. J N eurol
Neurosurg Psychiatry 1990; 53: 576]/79.
12 Franchignoni F P, Tesio L, R icupero C, M artino
M T. Trunk control test as an early predictor of
stroke rehabilitation outcome. S troke 1997; 28:
1382]/85.
13 Bohannon RW. Recovery and correlates of trunk
muscle strength after stroke. Int J R ehabil R es 1995;
18: 162]/67.
14 Lincoln N, Leadbitter D. Assessment of motor
function in stroke patients. Physiotherapy 1979; 65:
48 ]/51.
332 G Verheyden et al.
15 Carr JH , Shepherd R B, N ordholm L, Lynne D.
Investigation of a new motor assessment scale for
stroke patients. Phys T her 1985; 2: 175]/80.
16 G owland CA. Staging motor impairment after
stroke. S troke 1990; 21: II-19]/II-21.
17 Benaim C, PeÂrennou DA, Villy J, Rousseaux M ,
Pelissier JY. Validation of a standardized
assessment of postural control in stroke patients.
The postural assessment scale for stroke patients
(PASS). Stroke 1999; 30: 1862]/68.
18 Tsuji T, Liu M , Sonoda S, D omen K , Chino N. The
stroke impairment assessment set: its internal
consistency and predictive validity. A rch Phys M ed
R ehabil 2000; 81: 863]/68.
19 H antson L, D e Weerdt W, D e Keyser J et al . The
European stroke scale. Stroke 1994; 25: 2215]/19.
20 M ahoney F, Barthel D. F unctional evaluation: the
Barthel index. M d M ed J 1965; 14: 61 ]/65.
21 F ugl-Meyer AR , JaÈ aÈ skoÈ L, Leyman I, Olsson S,
Steglind S. The post-stroke hemiplegic patient.
S cand J Rehabil M ed 1975; 7: 13 ]/31.
22 H aas M . Statistical methodology for reliability
studies. J M anip Physiol T her 1991; 14: 119]/32.
23 Landis R J, Koch G G. The measurement of observer
agreement for categorical data. Biometrics 1977; 33:
159]/74.
24 N unnally J. Psychometric theory. N ew York:
M cG raw-H ill, 1978.
25 Kramer M S, Feinstein AR . Clinical biostatistics.
LIV. The biostatistics of concordance. Clin
Pharmacol T her 1981; 29: 111]/23.
26 van der Lee JH , Beckerman H , Lankhorst G J,
Bouter LM . The responsiveness of the action
research arm test and the F ugl-Meyer assessment in
chronic stroke patients. J R ehabil M ed 2001; 33:
110]/13.
27 Beckerm an H , Roebroeck M E, Lankhorst G J,
Bech er JG, Bezemer PD, Verbeek AL. Smallest real
difference, a link between reproducibility and
responsiveness. Qual L ife R es 2001; 10: 571]/78.
28 Altman D G, Bland JM . M easurement in medicine:
the analysis of method comparison studies.
Statistician 1983; 32: 307]/17.
29 Hatcher L, Stepanski EJ. Assessing scale reliability
with coef®cient alpha. In: A step-by step approach to
using the SA S Ò system for univariate and
multivariate statistics. Cary, N C: SAS Institute Inc.,
1994: 505]/16.
30 Poole JL, Whitney SL. M otor assessment scale for
stroke patients: concurrent validity and interrater
reliability. Arch Phys M ed R ehabil 1988; 69: 195]/
97.
31 Haas M . The reliability of reliability. J M anip
Physiol Ther 1991; 14: 199]/208.
32 Bland JM , Altman DG. Statistical methods for
assessing agreement between two methods of
clinical measurement. L ancet 1986; 1: 307]/10.
 T he T runk Impairment S cale 333

Appendix ] Trunk Impairment Scale (TIS)

/

The starting positio n for each item is the same. The patient is sittin g on the edge of a bed or treatment
table without back and arm support. The thighs make full contact with the bed or table, the feet are hip
width apart and placed flat on the floor. The knee angle is 908. The arms rest on the legs. If hypertonia is
present the positio n of the hemiplegic arm is taken as the starting positio n. The head and trunk are in a
midline position.
If the patient scores 0 on the first item, the total score for the TIS is 0.
Each item of the test can be performed three times. The highest score counts. N o practice session is
allowed.
The patient can be corrected between the attempts.
The tests are verbally explained to the patient and can be demonstrated if needed.

Item

Static sitting balance

1 Starting position Patient falls or cannot maintain starting position for 10 seconds e 0
without arm support
Patient can maintain starting position for 10 seconds e 2
If score5 /0, then TIS total score5 /0

2 Starting position Patient falls or cannot maintain sitting e 0

Therapist crosses the unaffected leg over the hemiplegic leg position for 10 seconds without arm support
Patient can maintain sitting position for 10 seconds e 2

3 Starting position Patient falls e 0

Patient crosses the unaffected leg over the hemiplegic leg Patient cannot cross the legs without arm support on bed or table e 1
Patient crosses the legs but displaces the trunk more than 10 cm e 2
backwards or assists crossing with the hand
Patient crosses the legs without trunk displacement or assistance e 3
Total static sitting balance /7

Dynamic sitting balance

1 Starting position Patient falls, needs support from an upper extremity or the elbow e 0
Patient is instructed to touch the bed or table with the hemiplegic does not touch the bed or table
elbow (by shortening the hemiplegic side and lengthening the Patient moves actively without help, elbow touches bed or table e 1
unaffected side) and return to the starting position If score5 /0, then items 2 and 3 score 0

2 Repeat item 1 Patient demonstrates no or opposite e 0

shortening/lengthening
Patient demonstrates appropriate shortening/lengthening e 1
If score5 /0, then item 3 scores 0

3 Repeat item 1 Patient compensates. Possible compensations are: (1) use of upper e 0
extremity, (2)
contralateral hip abduction, (3) hip flexion
(if elbow touches bed or table further then proximal half of femur),
(4) knee flexion, (5) sliding of the feet
Patient moves without compensation e 1

4 Starting position Patient falls, needs support from an upper extremity or the elbow e 0
Patient is instructed to touch the bed or table with the unaffected does not touch the bed or table
elbow (by shortening the unaffected side and lengthening the Patient moves actively without help, elbow touches bed or table e 1
hemiplegic side) and return to the starting position If score5 /0, then items 5 and 6 score 0

5 Repeat item 4 Patient demonstrates no or opposite e 0

shortening/lengthening
Patient demonstrates appropriate shortening/lengthening e 1
If score5 /0, then item 6 scores 0
334 G Verheyden et al.

Item

6 Repeat item 4 Patient compensates. Possible compensations are: (1) use of upper e 0
extremity, (2)
contralateral hip abduction, (3) hip flexion
(if elbow touches bed or table further then proximal half of femur),
(4) knee flexion, (5) sliding of the feet
Patient moves without compensation e 1

7 Starting position Patient demonstrates no or opposite e 0

Patient is instructed to lift pelvis from bed or table at the shortening/lengthening
hemiplegic side (by shortening the hemiplegic side and lengthening Patient demonstrates appropriate shortening/lengthening e 1
the unaffected side) and return to the starting position If score5 /0, then item 8 scores 0

8 Repeat item 7 Patient compensates. Possible compensations are: (1) use of upper e 0
extremity, (2) pushing off with the ipsilateral foot (heel loses contact
with the floor)
Patient moves without compensation e 1

9 Starting position Patient demonstrates no or opposite shortening/lengthening e 0

Patient is instructed to lift pelvis from bed or table at the Patient demonstrates appropriate shortening/lengthening e 1
unaffected side (by shortening the unaffected side and lengthening If score5 /0, then item 10 scores 0
the hemiplegic side) and return to the starting position

10 Repeat item 9 Patient compensates. Possible compensations are: (1) use of upper e 0
extremities, (2) pushing off with the ipsilateral foot (heel loses
contact with the floor)
Patient moves without compensation e 1
Total dynamic sitting balance /10

Co-ordination
1 Starting position H emiplegic side is not moved three times e 0
Patient is instructed to rotate upper trunk 6 times (every shoulder Rotation is asymmetrical e 1
should be moved forward 3 times), first side that moves must be Rotation is symmetrical e 2
hemiplegic side, head should be fixated in starting position If score5 /0, then item 2 scores 0

2 Repeat item 1 within 6 seconds Rotation is asymmetrical e 0

Rotation is symmetrical e 1

3 Starting position H emiplegic side is not moved three times e 0

Patient is instructed to rotate lower trunk 6 times (every knee Rotation is asymmetrical e 1
should be moved forward 3 times), first side that moves must be Rotation is symmetrical e 2
hemiplegic side, upper trunk should be fixated in starting position If score5 /0, then item 4 scores 0

4 Repeat item 3 within 6 seconds Rotation is asymmetrical e 0

Rotation is symmetrical e 1
Total co-ordination /6

Total Trunk Impairment Scale /23

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.