The Reliability of the Star-Excursion Test in

Assessing ~ynami c Balance

Stephen ). Kinzey, PhD '
Charles W. Armstrong, PhD *
D
ynamic balance is re-
quired for normal daily
activities, such as walk-
ing, running, and stair
climbing. Sports activi-
ties also require proper balance con-
trol. The visual, somatosensory, and
vestibular systems all contribute to
the maintenance of balance (10) and
may be adversely affected by muscu-
loskeletal injury, head trauma, dis-
ease, or aging. These influences on
the visual, somatosensory, and vestib-
ular systems might decrease a per-
son's ability to perform dynamic ac-
tivities and, thus, impede normal
daily functioning (2,6,17,23). Quanti-
fication of balance, or postural con-
trol, is often necessary to assess the
level of injury or ability to function in
order to initiate an appropriate plan
of care (15,17).
A valid and reliable technique to
measure balance is stabilometry (1 7).
This method uses a force plate or
other similar device to measure the
displacement of an individual's cen-
ter-of-pressure while standing in a
stationary position (1 7). Center-of-
pressure represents a weighted aver-
age of all the pressures over the sur-
face area in contact with the ground
(20). Quan tification of center-of-pres-
sure movement may be used to evalu-
ate numerous parameters (ie., mean
position of the center-of-pressure, velo-
city of center-of-pressure movement,
and total distance traveled by the cen-
ter-of-pressure) influenced by the con-
trol mechanism affecting balance (16).
Many activities of daily living and
sports are classified as dynamic activi-
Quantification of dynamic balance is often necessary to assess a patient's level of injury or
ability to function in order to initiate an appropriate plan of care. Some therapists use the star-
excursion test in an attempt to quantiQ dynamic balance. This test requires the patient to balance
on one leg while reaching with the other leg. For the purpose of this study, the reach was
performed in four directions. No previous researchers have attempted to evaluate the reliability of
this test. Twenty healthy subjects between the ages of 18 and 35 years participated in this study.
During two testing sessions, each subject was required to perform five reaching trials in four
directions. Reliability estimates, calculated using the intraclass correlation coefficient (2/ I ) , ranged
from 0.67 to 0.87. Six duplicate practice sessions were suggested to increase this range above
0.86. Task complexity may account for the moderate reliability estimates. Subjects should engage
in a learning period before being evaluated on the star-excursion test.
Key Words: dynamic balance, lower extremity, reaching, reliability
' Assistant Professor, Department of Exercise Science and Leisure Management, Applied Biomechanics and
Motor Performance Laboratory, The University of Mississippi, University, MS 38677
Professor, Department of Health Promotion and Human Performance, Applied Biomechanics Laboratory,
The University of Toledo, Toledo, OH
ties (9). Dynamic activities are those
that cause the center of gravity to
move in response to muscular activ-
ity. This muscular activity may arise
from any source of external or inter-
nal disturbance. During dynamic ac-
tivity, the center-of-pressure travels
between the base of support bound-
aries and sometimes outside the base
of support (9). Because stabilometry
is performed under static conditions,
the results may not be directly appli-
cable to assessing a patient's capabil-
ity for dynamic activities (21).
Unfortunately, few practical
methods exist for evaluating dynamic
balance (7). Methods have been re-
ported to assess dynamic balance in-
volving perturbation of stationary
subjects, through translation of a
force plate, or by using an external
force to perturb the subject and a
nonmoving force plate (5,18). These
methods are costly, impractical, and,
under some conditions, may pose a
risk to subjects. For these reasons,
new and practical techniques or pro-
tocols to evaluate dynamic balance
need to be investigated.
The starexcursion test is used by
some therapists to assess dynamic bal-
ance. It has been proposed that the
test meamres dynamic balance, re-
quiring patients to maintain balance
on a single limb, while manipulating
the other limb. This test requires the
subject to reach along a previously
marked line with one leg while stand-
ing on the other leg. This reaching
task is done along four different diag-
onal lines in four different directions.
The distance reached in each direc-
tion is recorded separately and the
four individual scores are interpreted
as a representation of dynamic bal-
ance and offer clinicians a practical
alternative for assessing dynamic bal-
ance. The purpose of this study was
Volume 27 Number 5 May 1W8 JOSPT
R E S E A R C H S T U D Y
FIGURE. Layout of the star-excursion test showing
how the tape was applied to the floor. Subjeas stood
within the square box and reached out in the four
diagonal directions. LA = Left-anterior, LP = Left-
posterior, RA = Right-anterior, RP = Right-posterior.
to evaluate the reliability of the star-
excursion test.
METHODS
Subjects
Twenty subjects (nine males and
11 females) volunteered to partici-
pate in this study and provided in-
formed consent as approved by the
Institutional Review Board, University
of Toledo, Toledo, OH. To qualify
for participation, the subjects could
not have: 1) incurred any ankle
trauma requiring medical attention
within the past 2 years, 2) a history of
any dizziness, 3) any inner ear disor-
ders, 4) any nervous system problems,
5) any bone or joint abnormalities,
6) a history of loss of consciousness,
7) any uncorrected problems with
vision, or 8) any other problem that
might adversely affect the control of
balance. All subjects were between
the ages of 18 and 35 years.
Test Description
The starexcursion test layout
consisted of four lines, applied to the
floor with athletic tape: two forming
vertical and horizontal lines and two
positioned perpendicular to each
other and at 45O with respect to the
vertical and horizontal lines (Figure).
A rectangle representing the starting
position of the feet wa. placed at the
center point. This box wa. large
enough to fit the subject's feet while
straddling the vertical line. A standard
tape measure wa. used to quantify the
distance (cm) from the center point to
the point that each subject reached
along each diagonal using the distal
part of the foot. Calculators were then
used to reduce the trial data into direc-
tional averages.
Protocol
The subject. completed two test-
ing sessions (pretest and posttest) 7
days apart. For each session, subjects
were instructed to stand with both
feet positioned inside the boundaries
of the starting box. A trial was initi-
ated when the subject began to reach
in one of the following four diagonal
directions: right-anterior (RA), left-
anterior (LA), right-posterior (RP) ,
and left-posterior (LP). When reach-
ing along the lines to the subject's
right, the subject used his/her right
leg to reach while using his/her left
leg as the support limb and vice
versa. Subject. were not allowed to
touch the ground with the reaching
leg at any time during the reach. The
maximal reach distance was the fur-
thest point along the directional line.
In accordance with the usual test pro-
tocol, the maximal reach distance was
measured visually. The visual mark
was then replaced by a finger, and
the distance was then measured using
a tape measure. The test administra-
tor was positioned on his knees along
the directional line. The trial was
complete after the subject returned
to the starting point by placing the
reaching leg in the starting box with
the support leg. After completion of
a single trial, the subject was given
the time necessary to regain the
proper starting position before start-
ing the next trial. Five consecutive
trials in each direction were com-
pleted before the subject began to
reach in the next direction.
During the starexcursion test,
the subject. were instructed to move
in any way possible to achieve a maxi-
mum reach distance without moving
the support foot. These movements
included but were not limited to
knee flexion of the support limb and
dorsiflexion of the support limb; hip
flexion in the reaching limb, knee
extension in the reaching limb, and
plantar flexion in the reaching limb;
and slight hyperextension of the
trunk. Typically, clinicians also use a
subjective description of what move-
ment strategies are used by the pa-
tient when performing the star-excur-
sion test. This subjective description
is then used to identify possible dif-
ferences between limbs.
After performing a local survey of
clinicians that used this test, it was
determined that the star-excursion
test is usually performed with shoes
on. Furthermore, no recommenda-
tions concerning the height of the
reaching limb are made when the
test is administered. In an attempt to
keep our protocol similar to a clinical
application, no effort was made to
control for these items.
Experimental Design and
Statistical Analysis
A pretest-posttest design was used
to assess the four dependent mea-
sures corresponding to the average
of the three best reaches for each
diagonal direction: right-anterior
(RA), left-anterior (LA), right-poste-
rior (RP), and left-posterior (LP) .
The direction of reach was deter-
mined by a balanced Latin square
to reduce the possibility of an order
effect.
Intraclass correlation coefficient
(2,l) (ICC 2, l) (13) and the Spear-
man Brown prophecy (1) were used
to estimate the reliability (ie., agree-
ment between scores) of the starex-
cursion test. The ICC(2.1) was cho-
sen as the reliability estimate, since it
provides an estimate that includes
the variability of measurement. taken
by any investigator on any subject
JOSFT Volume 27 Number 5 May 1998
3.57
RE S E ARCH S T U D Y
Left-
Righ
Left-
Righ
iredion ICC(2,l) SEM
anterior 0.87 3.43
t-anterior 0.67 4.78
posterior 0.87 3.48
1-posterior 0.82 3.99
TABLE 1. Summary oi the ICC(2,l) estimates o i
reliability and standard error of measurement (SEM).
(13). The ICC(2.1) is represented by
the following equation (13):
where BMS = between mean square,
EMS = residual mean square, JMS =
between judges' mean square, k =
the number of sets of scores, and
n = the number of persons observed.
Mean square terms were obtained
using the F statistic calculated under
the reliability procedure in SPSSm for
WindowsTM, Version 6.1 (SPSS, Inc.,
Chicago, IL).
The Spearman Brown prophecy
estimates the reliability of a test. Us-
ing only two sets of obsenations, it
allows the researcher to estimate how
many observations might be neces-
sary to achieve a desired estimate of
reliability. The Spearman Brown
prophecy is represented by the fol-
lowing formula:
where k = a factor by which the orig-
inal set of data is lengthened, r,,. =
the value of ICC(2,1), and r,,. = the
new estimate of reliability. The factor
k can be altered to decide how many
sets of observations are necessary to
achieve a suitable estimate of reli-
ability.
RESULTS
The reliability estimates obtained
for each of the four reaches are pre-
sented in Table 1. Reaches that were
in the left diagonal directions, per-
formed while the subjects stood on
their right foot, produced the highest
estimates of reliability: left-anterior,
ICC(2,l) = 0.87; left-posterior,
ICC(2,l) = 0.87. The ICC(2,l) esti-
mates for reaches along the right an-
terior and posterior diagonal direc-
tions were 0.67 and 0.82, respectively.
Table 2 contains the results from
the Spearman Brown prophecy. The
Spearman Brown prophecy provides
an indication that at least six practice
sessions of five trials per direction
per session, taking the average of the
best three reaches, may be necessary
to achieve a reliability measure rang-
ing from 0.86 to 0.95. A minimum
reliability estimate of 0.95 would be
achieved if the subject engaged in 18
separate practice sessions, consisting
of five trials per direction per session.
DISCUSSION
The task of standing in normal
adults can be described as a position
of quasi-static equilibrium. Standing
is quasi-static because the center-of-
pressure travels within the base of
support, indicating overall sway with-
out a corresponding translation of
the base of support (20).
The goal of the starexcursion
test is to force subjects to disturb
their equilibrium to a near maximum
(ie., the reach is maximum but does
not cause a fall which would be indic-
ative of a maximal disturbance) and
then return back to the state of equi-
librium (starting point). Normally,
adults do not challenge their state of
equilibrium to the extent required in
this test. Additionally, the movement
patterns employed in the test are not
common to normal functional activi-
Direction k
Number of Three
Trial Averages
6
18
6
10
18
6
18
6
18
TABLE 2. Summary of the Spearman Brown
prophecy reliability estimates.
The goal of the star-
excursion test is to
force subjects to
disturb their
equilibrium to a
near maximum.
ties or those involved in sports. Thus,
the task involved in the starexcursion
test would appear to be a novel one.
Table 1 lists the moderate reli-
ability estimates that were found for
the four components of the test.
Moderate estimates of reliability indi-
cate that a subject mav exhibit a
change in scores due to some unmea-
surable circumstance, a random
movement pattern, or any other pos-
sible influence, including mental
state.
One possible explanation for
these moderate estimates is that the
subjects may have chosen different
movement trajectories in an attempt
to obtain a maximal reach distance.
These random movement trajectories
could contribute to random low or
high reaches instead of a consistent
reach distance. Also, the type of
movement required during the star-
excursion test is both multilimb and
multiarticular. For instance, distance
reached in an anterior direction
would be greatly affected by the
amount of knee flexion and ankle
dorsiflexion of the support limb and
hip flexion, knee extension, and an-
kle plantar flexion of the reaching
limb. These movements inherently
have more variability associated with
them than .single limb and uniarticu-
lar movements (19). To decrease the
variability involved in a complex ta..k
requires practice. Both task complex-
ity and motor ability influence how
many practice sessions are necessary
to achieve consistent result.. (12).
The human organism uses the
redundancy within the sensorimotor
358
Volume 27 Number 5 Mav 1998 JOSPT
R E S E A R C H S T U D Y
system to reduce this variability when
realizing the solution to a given task
(1 1 ). Three systems, visual, vestibular,
and the somatosensory, are used to
provide the information to the senso-
rimotor system (10). This redun-
dancy is demonstrated when subjects
are able to maintain an upright posi-
tion under the commonly used foam
and dome conditions (14), which
confuse or eliminate portions of the
sensorimotor system used in postural
control. Although the task of main-
taining upright posture becomes in-
creasingly difficult, subjects are still
able to perform under the test condi-
tions (eg., maintenance of posture
when blindfolded) or when confused.
Furthermore, in experiments in-
volving the analysis of upper-limb-
reaching trajectories, visual feedback
has been shown to be important in
achieving consistency (22). In the
absence of visual feedback, it has
been proposed that an individual
may use internal models of the limb
to aid in the positioning during
reaching tasks (3). Additional infor-
mation regarding the position of the
lower limb may be arrived at using
proprioceptive mechanisms. Proprio-
ceptive feedback during reaching
provides information regarding the
limb position and has been shown to
be important in the development of
an internal coordinate system (4). It
is important to note that because hu-
mans rarely engage in reaching tasks
using their lower limbs, these neural
pathways which are readily available
during upper extremity reaches may
not be as readily available when de-
scribing lower limb movement.
Finally, the strategy chosen by
the subjects in performing the test
involved a controlled lowering of the
body to maximize the reach. This was
accomplished through eccentric and
isometric contractions of the knee
extensors of the supporting limb.
The extent to which this can be done
is clearly dependent on the strength
of these muscles. It is possible then
that the ability to control the support
limb in a partial squat using an iso-
metric contraction could contribute
to the lack of consistency in scores.
An optimal reach during the star-
excursion test might require precise
integration of the nervous system
function and musculoskeletal system.
The absolute strength of the support
limb musculature may also influence
the outcome of the test. Because the
star-excursion test is novel, the devel-
opment of this integration might be
difficult and require practice. The
practice would allow for the proper
neural circuitry to become activated
and the resulting coordination of
movement to be developed. This may
explain the linear relationship be-
tween the amount of practice sessions
and the estimated reliability shown in
Table 2.
The speed at which the move-
ment is performed may also affect
the consistency of results. During tar-
geted reaching tasks involving upper
limb movement, speed has a direct
Both task complexity
and motor ability
influence how many
practice sessions are
necessary to achieve
consistent results.
relationship with accuracy and pro-
vides an experimental example of
Fitt's Law (8). As the limb speed in-
creases, the accuracy of the limb
movement decreases. It is probable
that the same relationship in the
lower limb exists during the starex-
cursion test. The starexcursion test
requires a manual measurement
when the subject achieves maximum
reach distance. Therefore, the s ub
jects did not use a strategy that in-
volved a rapid reach with the lower
limb when performing the starexcur-
sion test. However, in upper extrem-
ity testing, a slight decrease in move-
ment time (100 msec) caused a large
decrease (20%) in movement accu-
racy (8). Although no movement
time was recorded, subjects might
have made adjustments of this mapi -
tude, causing a resultant change in
accuracy and reach distance.
CONCLUSIONS
The utility of clinical diagnostic
testing is dependent on the reliability
and validity of the testing procedure.
Because exact or near exact repeat
performances were not exhibited in
this investigation, the starexcursion
test might not be an appropriate test
of dynamic balance. While practice
may improve the reliability of the
test, clinicians typically do not have
an unlimited amount of time to reha-
bilitate an individual's injury. There-
fore, the time necessary to practice
the starexcursion test is not always
available.
Without reliability, there can be
no validity; therefore, we must ques-
tion what exactly is measured during
the starexcursion test. Although the
starexcursion test may not provide
an appropriate method for evaluating
a patient's dynamic balance, it may
play an important part in rehabilita-
tion. Because of the strength, move-
ment complexity, and neural control
associated with the starexcursion test,
it may be a useful activity for progres-
sively stimulating the involved systems
in selected patients.
In light of the present findings, it
seems that a test of dynamic balance
should involve activities that are
more common and in accordance
with activities of daily living. For in-
stance, we might have typical move-
ment patterns that occur when we
descend stairs. Therefore, it might be
possible that dynamic activities of this
type could be structured in such a
way that reliable measurements could
be made. Typical movements that are
normally performed should probably
JOSPT Volume 27 Number 5 May 1998
R E S E A R C H S TUDY
be the basis for evaluation methods Methodology and effects of midazolam lnt J Sports Med 6(3): 180- 182, 1 985
instead of using novel movements
sedation. Acta ~tol aryngol 113(3):245-
17. Tropp H, Odenrick P, Sandlund B, Od-
like the starexcursion test.
248, 1993
kvist LM: Stabilometry for studying pos-
8. Keele SW, Posner MI: Processing of vi-
tural control and compensation in ver-
sual feedback in rapid movements. 1
tigo of central and beripheral origin.
REFERENCES
Exp Psycho1 77(1): 155-1 58, 1968
Electromyogr Clin Neurophysiol 27:
9. Mann RA, Hagy JL, White V, Liddell D:
77- 82. 1 98.5
Crocker L, Algina J: Introduction to
Classical and Modern Test Theory, pp
1 17-1 18. Orlando, FL: Harcourt Brace
Jovanovich, Inc., 1986
Era P, Heikkinen E: Postural sway dur-
ing standing and unexpected distur-
bance of balance in random samples of
men of different ages. J Gerontol40(3):
287-295, 1 985
Ghez C, Gordon J, Ghilardi M F: Impair-
ments of reaching movements in pa-
tients without proprioception. ll. Effects
of visual information on accuracy.
I Neurophysiol 73(11:36 1-37.?, 1995
Gordon), Ghilardi MF, Ghez C: Impair-
ments of reaching movements in pa-
tients without proprioception. l. Spatial
errors. J Neurophysiol 73(1):347-360,
1995
Hi l l KM, Harburn KL, Kramer JF, Noh S,
Vandervoort AA, Matheson JM: Com-
parison of balance responses to an ex-
ternal perturbation test, with and with-
out an overhead harness safety system.
Gait Posture 2:27-3 1, 1994
lngersoll CD, Armstrong CW: The ef-
fects of closed-head injury on postural
sway. Med Sci Sports Exerc 24(7):73 9 -
743, 1992
Ledin T, Gupta A, Larsen L E, Odkvist LM:
Randomized perturbed posturography:
The initiation'of gait. J Bone Joint Surg
6 1A:232-239, 1979
10. Nashner LM, Shumway-Cook A, Martin
0 : Stance posture control in select
groups of children with cerebral palsy:
Deficits in sensory organization and
muscular coordination. Exp Brain Res
4 9:3 93- 409, 1 983
1 1. Newel1 KM, Corcos DM: Issues in vari-
ability and motor control. In: Newel1 KM,
Corcos DM (eds), Variability and Motor
Control, pp 1-12. Champaign, IL: Hu-
man Kinetics Publishers, Inc., 1993
12. Schmidt RA: Motor Control and Learn-
ing: A Behavioral Emphasis, Cham-
paign, IL: Human Kinetics Publishers,
Inc., 1988
13. Shrout PE, Fleiss JL: lntraclass correla-
tions: Uses in assessing rater reliability.
Psycho1 Bull 86(2):420-428, 1979
14. Shumway-Cook A, Horak FB: Assessing
the influence of sensory interaction on
balance. Ph ys Ther 66: 1548 - 1550, 1 986
15. Tropp H, Ekstrand J, Gillquist ): Sta-
bilometry in functional instability of the
ankle and its value in predicting injury.
Med Sci Sports Exerc 16(1):64-66,
1984
16. Tropp H, Odenrick P, Gillquist J: Sta-
bilometry recordings in functional and
mechanical instability of the ankle joint.
, . - -
18. Vrethem M, Ledin T, Emerudh J, Odkvist
L, Holmgren H, Moller C: Correlation
between dynamic posturography, clini-
cal investigation, and neuropphy in pa-
tients with polyneuropathy. ORL J Oto-
rhinolaryngol Relat Spec 53(5):294 -298,
1991
19. Walter CB, Swinnen SP, Franz EA: Sta-
bility of symmetric and asymmetric dis-
crete bimanual actions. In: Newel1 KM,
Corcos DM (eds), Variability and Motor
Control, pp 359-380. Champaign, IL:
Human Kinetics Publishers, Inc., 1993
20. Winter DA: A.B.C. (Anatomy, Biome-
chanics and Control) of Balance During
Standing and Walking, Waterloo, On-
tario: Waterloo Press, 1995
21. Winter DA, Patla AE, Frank IS, Walt SE:
Biomedical walking pattern changes in
the fit and healthy elderly. Phys Ther
70(6):340-347, 1990
22. Wolpert DM, Ghahramani Z, Jordan
MI: Are arm trajectories planned in ki-
nematic or dynamic coordinates? An
adaptation study. Exp Brain Res 103(3):
460-470, 1995
23. Woollacott MH, Shumway-Cook A:
Changes in postural control across the
life span: A systems approach. Phys
Ther 70(12):799-807, 1990
Volume 27 Numher 5 Mav 1W8 eJOSPT