George A.

Koumantakis, PhD1
Julie Winstanley, PhD2
Jacqueline A. Oldham, PhD1

Study Design: Repeated measures design of active spinal position sense in individuals with and
without low back pain (LBP).
Objectives: Reproducibility and validity evaluation of thoracolumbar proprioception measurement.
Background: Proprioception studies in peripheral joints and the spine suggest that there may be
proprioception deficits due to injury, pain, or degeneration. Kinesthetic retraining may be useful in
rehabilitation of patients with LBP, but appropriate measures are required to objectively quantify
spinal proprioception.
Methods and Measures: Active-target reproduction in the sagittal, horizontal, and coronal planes
was assessed (3 separate occasions for 18 asymptomatic volunteers and 2 occasions for 62
patients with LBP). Repositioning accuracy was expressed as absolute errors (AE) and variable
errors (VE). Reliability was analyzed with intraclass correlation coefficient (ICC) and precision with
standard error of measurement (SEM) and calculation of the smallest detectable difference (SDD)
index. Repeated measures ANOVA and correlations were used for within-group comparisons and
discriminant analysis for between-group comparisons.
Results: Reproducibility was better for the asymptomatic group, with AE for flexion and rotation
being the most reliable (ICC = 0.760.80, SEM = 0.911.34). SDDs were high for all tests,
suggesting limited clinical applicability. Reproducibility for the same tests was poor-moderate
(ICC = 0.310.64, SEM = 0.453.90) for the patient group. AE for right-side rotation could
discriminate between subject groups with 83.3% specificity but only 54.8% sensitivity.
Conclusions: Proprioception testing, with the methods employed, did not demonstrate good
measurement properties in a sample of patients with recurrent LBP. Neither could it sufficiently
discriminate between individuals with and without LBP. Possible reasons for these findings are
discussed. J Orthop Sports Phys Ther 2002;32:327335.

Key Words: electrogoniometer, muscle spindle, neuromuscular dysfunction,

position sense

Centre for Rehabilitation Science, University of Manchester, Central Manchester Healthcare Trust,
Manchester, UK.
2
Rehabilitation Studies Unit, University of Sydney, Royal Rehabilitation Centre, Ryde, NSW, Australia.
Ethical permission for the recruitment of asymptomatic volunteers and patients was granted by the Central
Manchester Ethical Committee.
This study was supported by a grant from the State Scholarships Foundation (IKY) of Greece.
Please send correspondence to George A. Koumantakis, Centre for Rehabilitation Science, University of
Manchester, Central Manchester Healthcare Trust, Oxford Road, Manchester, M13 9WL UK. E-mail:
gak4@otenet.gr
Journal of Orthopaedic & Sports Physical Therapy

327

REPORT

roprioception is a term
frequently used in rehabilitation and has been
defined as the afferent
neural information
originating from joints, muscles,
tendons, and associated deep tissue mechanoreceptors.24 These
inputs are processed at 3 different
motor control centers of the
central nervous system (CNS)
spinal, lower brain (brainstem,
cerebellum), and cortical levels13,14,17,24,41resulting in the
appropriate neuromuscular output
for the seamless and accurate execution of movement. Proprioceptive information is processed at a
subconscious level by the first 2
CNS centers and consciously at
the cerebral cortex.24
A variety of methods assessing
the conscious component of
proprioception have been described. These involve appreciation of joint position either during
passive movement of a body segment by an external device34 or
during active movement initiated
by the individual. Several investigators have argued that it may be
more functionally relevant to assess proprioception deficits actively
during the performance of normal self-paced movements.5,14,16,46

RESEARCH

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Thoracolumbar Proprioception in
Individuals With and Without Low Back
Pain: Intratester Reliability, Clinical
Applicability, and Validity

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Any proprioception deficits present in midrange active movements are more likely to represent abnormal or insufficient afferent information emanating
from muscle receptors, as they provide the majority
of proprioceptive information regarding position and
movement of peripheral joints under these conditions.7,13,14 Altered muscular activation due to damage in nearby ligaments and facet capsules cannot be
discounted, but these structures are thought to provide a much larger afferent input towards end range
joint positions.13
Poor proprioception has been hypothesized to be
a significant contributing factor for chronic low back
pain (LBP). A few studies of patients with LBP have
identified significant proprioception and motor control deficits as being associated with the number of
past lumbar spine injuries,34 the presence of chronic
pain,16 mechanical pain,4 and muscle fatigue.35,44,47 Other studies, however, have failed to
establish such differences between individuals with
and without LBP,23,30 and have been unable to provide any definitive answers as to whether proprioceptive deficits existed in individuals with LBP or
whether the tests used were not sensitive enough.
These results reflect a lack of consensus regarding
the effects of spinal pathology on proprioception;
however, further research in the area is recommended to identify the best discriminative test for
back proprioception.23
Gill and Callaghan16 used the lumbar motion
monitor (LMM) electrogoniometer to assess differences in proprioception of the thoracolumbar spine
complex in the sagittal plane between asymptomatic
subjects and individuals with LBP. They showed that
there were small differences between the groups and
that the group with LBP demonstrated poorer
proprioception than the asymptomatic group. However, from the review of the literature in the area of
low back proprioception and the relatively few repeated measures studies presented in Table 1, a
number of important questions regarding intratester
reliability, clinical applicability, and validity were
raised. These questions need to be addressed before
the contribution of proprioceptive deficit to LBP can
be understood. The aims of the study therefore were
twofold: (1) to establish the degree of between-day
reliability of thoracolumbar position sense testing
with the LMM in asymptomatic subjects and those
with recurrent LBP and (2) to compare the accuracy
of position sense between the 2 groups.

sample of 18 asymptomatic subjects and 62 consecutive patients with LBP, subsequently recruited for the
purposes of a randomized controlled trial, took part
in the study after informed consent had been obtained. The rights of the participants were protected
at all times. The anthropometric characteristics of all
participants are detailed in Table 2. All subjects apart
from one patient were right-hand dominant.
Asymptomatic participants were required not to
have any previous history of LBP and no other trunk
or lower-limb pathology, deformity, or conditions that
may affect motor control. Patients were recruited if
they had recurrent mechanical nonspecific LBP9 (at
least 2 episodes of LBP within the past year) with
pain duration of less than half the days in the past
year,48 were still working, and had no previous neurological condition. Patients were tested after the
acute stage of their current episode (greater than 6
weeks after onset) if their symptoms persisted. Before
testing, patients were asked to fill in the Short Form
McGill Pain Questionnaire (SFMPQ),29 a visual analog scale (VAS) of average pain intensity over the
past week,20 the Roland-Morris Disability Questionnaire (RMDQ),40 and to report their symptoms duration (current episode and first onset) (Table 3).

METHODS

Procedure

Subjects

All participants were required to perform 5 distinct movements up to certain target angles, first
identified with eyes open. After practicing each task
with visual feedback, they were then blindfolded and
instructed to identify again as accurately as possible

The Central Manchester Ethical Committee

granted ethical permission for the recruitment of
asymptomatic volunteers and patients. A convenience
328

Apparatus
A triaxial spinal electrogoniometer, a lumbar motion monitor (LMM) (Chattecx, Chattanooga, Ltd.,
USA), was used for angular measurement and subject
feedback. The device had a mass of 1 kg and was
attached to the body by harnesses at the thorax and
pelvis. It is designed to concurrently measure range
of motion (ROM), velocity, and acceleration of the
thoracolumbar spine as a unit in 3-dimensional (3D)
space relative to the pelvis. The LMM has been validated against video-based motion analysis and has
been shown to be more accurate than video-based
analysis when placed in multiple configurations on a
3D-calibration frame (0.96 accuracy in the sagittal,
1.71 in the frontal, and 0.50 in the transverse
planes).27 Also, the reproducibility of the LMM for
in vivo ROM measurement in the 3 primary planes is
suitably high (ICC = 0.820.96) for use in research
and clinical settings.15 The LMM was attached onto
the subjects according to previous specifications.16
Care was taken that both harnesses were appropriately secured on the subjects to avoid any movement
between the device and the subjects skin.

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

TABLE 1. Survey of the literature on between-day measurement of active low back proprioception.
Authors

Assessment Method

Sample Size and

Design

Results

Maffey-Ward et al

Active position sense in

sitting
Upright position
reproduction

Fastrak
AE

10 subjects without LBP

Test-retest (max 4 d
apart)

Flexion: AE = 2.6
Rotation: AE = 1.31.6
Side-flexion:
AE = 0.60.8
No significant differences
between days

Gill and Callaghan16

Active position sense in

standing and 4-point
kneeling
20 flexion angle
reproduction test

LMM
AE

10 subjects (5 with
chronic LBP and 5
without LBP)
test-retest (12 wk apart)

Flexion (standing):
AE = 5.486.53
Flexion (4-point
kneeling):
AE = 6.137.01
ICC (standing): 0.85
ICC (4-point kneeling):
0.86

Swinkels and Dolan45

Active position sense in Fastrak

AE
standing
Upright and halfway
positions reproduction in
flexion and side-flexion

20 subjects without LBP

Test-retest (within-day
and 2 wk apart)

Flexion:
AE = 1.675.27
Side-flexion:
AE = 0.403.70
Between-day (upright):
ICC = 0.110.78
Between-day (halfway):
ICC = 0.340.90

Brumagne et al5

Active position sense in

standing
Random targets in
anterior or posterior
pelvic tilt zone

Piezoresistive
electrogoniometer
AE

14 subjects without LBP

Test-retest (1 d apart)

Pelvic tilt:
AE = 1.731.81
ICC: 0.51
SEM: 0.5

Brumagne et al6

Active position sense in

standing
Random targets in
anterior or posterior
pelvic tilt zone

Piezoresistive
electrogoniometer and
3D Video Analysis
AE

11 subjects without LBP

Test-retest (1 d apart)
Correlation between the
2 testing methods

Pelvic tilt
(electrogoniometer):
AE = 1.851.99
Pelvic tilt (3D video
analysis):
AE = 1.691.93
ICC (electrogoniometer):
0.72
ICC (3D video analysis):
0.64
Pearsons r = 0.840.97

Lam et al23

Active position sense in

sitting
Upright position
reproduction

Fastrak
AE

20 patients with LBP

Test-retest (15 d apart)

Flexion:
AE = 2.252.32
Rotation:
AE = 1.251.31
Side-flexion:
AE = 0.810.85
No significant difference
between days

AE = absolute error; ICC = intraclass correlation coefficient; LBP = low back pain; LMM = lumbar motion monitor.

RESEARCH

TABLE 2. Anthropometric characteristics (mean SD) of all participants with and without low back pain (LBP).
Subjects Without LBP

Subjects With LBP

Variables

Women
(n = 10)

Men
(n = 8)

Total
(n = 18)

Women
(n = 32)

Men
(n = 30)

Total
(n = 62)

Age (y)
Height (cm)
Body Mass (kg)
Body Mass Index (kg/m2)

23.6 4.2
165.8 7.9
65.8 7.9
24.0 3.1

25.7 3.6
176.6 8.1
74.7 9.1
24.0 2.7

24.6 4.0*
170.6 9.5
69.8 9.4*
24.0 2.8*

39.8 10.9
165.9 6.1
74.4 13.8
26.9 4.4

35.8 9.9
177.5 6.7
81.4 11.1
25.8 2.7

38.2 10.7
171.6 8.6
77.8 12.8
26.4 3.7

* Significantly different than subjects with LBP P 0.002.

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

329

REPORT

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

26

Equipment and Error Index

TABLE 3. Characteristics of patients (n = 62) with recurrent low

back pain (LBP).
Variables
Pain
LBP 1st onset (mo)
McGill Questionnaire
(pain descriptors)
Average pain-VAS (0100)
Functional Status
Roland-Morris questionnaire

Mean SD (range)
53.6 50.6 (2240)
15.9 5.8 (628)
34.7 23.7 (089)
10.1 5.0 (022)

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

VAS = visual analog scale

the target angles by moving 3 times up to their point

in range. The 5 target angles were set as follows: 1 in
forward flexion, 2 for rotation (right and left), and 2
for side-flexion (right and left), covering the 3 primary planes of movement (Figure 1). A modification
of the protocol used by Gill and Callaghan16 was attempted in this study by limiting the repetitions for
each test to 3 instead of 10, along with the addition
of proprioception assessment in the other 2 primary
planes of movement. Angles were set at 20 of flexion, 15 rotation bilaterally, and 15 side-flexion bilaterally, as measured by the LMM. These were clearly
displayed on screen for subjects visual feedback and
have been used in previous publications.16,28 It was
initially ensured that the subjects ROM in each direction exceeded the target angles, although the sub-

jects full range of spinal motion was not recorded.

Movement toward each target angle was performed
at the subjects preferred speed 3 times within 30
seconds (angle reproduction test). The 5 tasks were
randomized to avoid order effects.18 The same examiner conducted all examinations and tests. Participants were asked to wear a loose t-shirt (if preferred)
and shorts or loose trousers for the tests in an attempt to negate any direct proprioceptive input from
clothing. Subjects were tested either barefoot or
wearing flat, comfortable shoes. The goniometer was
calibrated on every subject immediately before the
beginning of each of the tasks. Hip movement was
blocked for the rotation tests by asking the subjects
to rest the anterior aspect of their pelvis against a
plinth at the anterior superior iliac spines level. Subjects without LBP were tested on 3 occasions with
each test 1 week apart to negate any learning effects.
Patients with LBP were tested on 2 occasions set at
least 5 days apart. Tests were performed at the same
time of day to minimize any effect of diurnal variation. Results were withheld from the subjects until all
tests were completed in an attempt to control the
degree of effort from each participant on all occasions.10
Measurement of target accuracy was recorded and
2 different error indices of target estimation were
calculated: the absolute error (AE) and variable error (VE). AE is the average absolute deviation, with-

FIGURE 1. The active position sense tests employed.

330

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

out regard to direction, between the subjects responses and the target. It is regarded as a measure
of overall accuracy in performance, encompassing
constant error (bias), and variable error. The AE index has been used as a primary outcome measure in
spinal proprioception assessment and was calculated
to provide direct comparison with other studies. The
VE represents the variability of the subjects performance around their mean response, thus being sensitive to inconsistency in response. VE is the average
deviation between the subjects score on each trial
and their own average score. VE is not dependent on
whether or not subjects were close to the set target.42

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

RESULTS
Demographic data
Between-group comparisons (Table 2) indicated
that the patients with LBP were significantly heavier
and older than the control group, with a much
greater age range. However, further analysis did not
indicate any significant associations between any of
the proprioception error indices with age or gender
for either group (r = 0.070.21, P 0.09), therefore
each groups position sense data were pooled together for all further analyses.

Reliability and Clinical Applicability

There were no significant differences between
means obtained for days 1, 2, and 3 for the
proprioception error indices in each group, with the
exception of the right rotation VE in the individuals
without LBP and the flexion AE in the patients with
LBP (Table 4). Data were significantly different for
right rotation VE due to a reduction in the error
recorded on subsequent occasions, whereas the op331

REPORT

Anthropometric characteristics, proprioception

acuity data, and patients health status parameters
were initially tested for normality of distribution using the Kolmogorov-Smirnov test. Most data conformed to a normal distribution, so parametric statistical analyses were employed. The only parameter
that was not normally distributed was current episode duration (median = 12 weeks, interquartile
range (IQR) = 816) for the participants with LBP.
Anthropometry was tested for between-group differences using independent samples t-tests. The effect of age (Pearsons correlation) and gender (independent samples t-tests) on proprioception acuity was
examined separately for each group. Participants
ages were of particular interest, as age-related
proprioception and general motor control have been
found to decline in subjects over 60 years of
age.1,22,43 Also, a possible effect of gender was considered important, as there are several studies describing differences in the performance of several
tasks by subjects of different gender, either due to
structural or motor control muscle physiology differences.2 Proprioception acuity data (AE and VE) are
presented as means SD. Paired samples t-tests were
used to identify between-side differences for rotation
and side flexion for each day and for both error indices in participants with and without LBP. This last
analysis was performed because a positive relationship between proprioception asymmetry betweensides and back pain has been previously described.34
For reliability, repeated measures analysis of variance (ANOVA) were performed for each of the tests
to assess any significant between-day differences for
the individuals without LBP (days 13) and the individuals with LBP (days 12). Two reliability indices
were computed from the ANOVA components, as
recommended by Rankin and Stokes:38 the intraclass
correlation coefficient (ICC) model 3, ( = 3)12
and the standard error of the measurement (SEM).
The ICC is a unitless index of reliability based on
the ratio of the within-to-between subject differences
and values between 0.75 and 1.00 are considered acceptable for research.12 The SEM indicates the preci-

RESEARCH

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Statistical Analysis

sion of measurement (magnitude of error present)

in the same units as the original measurement (degrees) and therefore can be directly compared
against subjects values.11 The SEM can be calculated
as the square root of the within-subject mean
squared (WMS) error from the repeated measures
ANOVA.12 Also, a clinical applicability index, the
smallest detectable difference (SDD) was derived
from the SEM (SDD = 1.96(2) SEM) and expressed
as a percentage of the parameters grand mean. It is
a useful index for diagnostic tests, indicating the
level of change in a parameter attributed with 95%
certainty to a true change in a subjects condition,
instead of being caused by test-retest errors.39 A
small SDD associated with repeated test application
renders a measurement more responsive to change.
The diagnostic validity (a form of construct-related
validity) of the tests was examined, as there is no
gold-standard measure of proprioception acuity to
test for criterion validity of our test.19 The ability of
the tests to differentiate between the 2 groups was
analyzed using a stepwise discriminant analysis
(simple MANOVA) by merging between-days data
only for the measures that were adequately stable in
both groups.
All statistical analyses were performed using the
Statistical Package for the Social Sciences (SPSS for
Windows, Chicago, IL). Statistical significance was set
at 0.05 or lower and the Bonferroni correction
method was used to avoid the probability of a type I
error in cases where the number of comparisons was
extensive.

posite trends were recorded for flexion AE. The reliability coefficients were 0.69 and 0.41 for those 2
variables, respectively (Table 5). No differences between the right and left sides for rotation (P = 0.13
0.78) and side flexion (P = 0.600.97) tests were
identified.
In subjects without LBP, AE for flexion and rotation had the highest ICC (0.760.80). In this same
group, ICC ranged from 0.20 to 0.69 for all other
parameters. SEM was large (0.451.34) for all pa-

rameters measured, indicating lack of precision. For

patients with LBP the measurements had ICC ranging from 0.24 to 0.64 and precision was generally
lower than the group without LBP (SEM = 0.45
3.90).
The clinical applicability of measures was limited,
as indicated by the magnitude of the SDD in both
groups of participants (Table 5). Generally, the smallest detectable difference that can be confidently
(95% confidence interval) attributed to nontest-

TABLE 4. Means SD for both error indices on all testing occasions and repeated measures ANOVA P values for significant between-day
differences.
Group

Error

Day 1

Day 2

Day 3

3.24 2.22
1.43 1.19
7.08 5.34
2.29 1.59

2.80 1.70
1.65 1.04

0.16
0.67
0.03*
0.62

2.05 1.68
0.76 0.46

0.22
0.014*
0.52
0.18

1.91 1.47
0.89 0.67

0.16
0.98
0.73
0.55

1.78 1.19
0.72 0.42

0.21
0.08
0.42
0.25

1.98 1.04
1.05 0.84

0.60
0.24
0.62
0.86

20 Flexion
Control

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Patients

Absolute
Variable
Absolute
Variable

3.67 1.82
1.69 1.02
5.46 3.54
2.18 1.55

15 Right Rotation
Control
Patients

Absolute
Variable
Absolute
Variable

2.46 2.00
1.16 0.45
3.43 2.24
1.19 0.73

1.80 1.11
0.98 0.58
3.66 2.60
1.05 0.54

15 Left Rotation
Control
Patients

Absolute
Variable
Absolute
Variable

2.44 1.45
0.90 0.63
3.36 2.47
1.15 0.64

1.96 1.26
0.93 0.46
3.39 2.39
1.08 0.89

15 Right-Side Flexion
Control
Patients

Absolute
Variable
Absolute
Variable

2.33 1.44
0.96 0.47
2.58 1.56
1.05 0.61

2.44 1.05
1.07 0.63
2.38 1.42
0.93 0.61

15 Left-Side Flexion
Control
Patients

Absolute
Variable
Absolute
Variable

2.26 1.37
0.77 0.50
2.48 1.84
0.89 0.62

2.37 1.55
0.81 0.50
2.31 1.74
0.88 0.60

*Significant at P 0.05.
TABLE 5. Reliability-clinical applicability results for various error indices based on the 3 separate testing occasions for individuals without
LBP and on 2 separate occasions for the patients with LBP.
Control

20 flexion
15 right rotation
15 left rotation
15 right-side flexion
15 left-side flexion

Patients

Error

ICC3,3
(95% CI)

SEM ()

SDD (%)

ICC3,3
(95% CI)

SEM ()

SDD (%)

Absolute
Variable
Absolute
Variable
Absolute
Variable
Absolute
Variable
Absolute
Variable

0.76 (0.470.90)
0.45 (0.000.78)
0.76 (0.480.90)
0.69 (0.330.88)
0.80 (0.570.92)
0.20 (0.000.68)
0.22 (0.000.68)
0.47 (0.000.78)
0.48 (0.000.79)
0.54 (0.000.81)

1.34
0.96
1.13
0.37
0.91
0.56
1.18
0.45
1.17
0.53

114.2
169.8
147.0
103.4
119.0
164.8
151.0
130.4
145.4
170.4

0.41 (0.020.64)
0.53 (0.220.72)
0.58 (0.290.75)
0.33 (0.000.60)
0.51 (0.170.71)
0.37 (0.000.63)
0.42 (0.040.65)
0.31 (0.000.58)
0.24 (0.000.54)
0.64 (0.400.78)

3.90
1.25
1.98
0.58
1.96
0.69
1.28
0.55
1.68
0.45

171.7
157.6
160.8
142.8
164.6
106.2
141.7
151.5
193.3
134.8

ICC = intraclass correlation coefficient; CI = confidence interval; SEM = standard error of measurement; SDD = smallest detectable difference; LBP =
low back pain.
332

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

degrees
Mean (SD)

retest error exceeded the grand mean values in the

tests, with even the lowest SDD being of equal magnitude to the mean values (all SDD 103%).
Further exploration of the data set was warranted,
as reliability and clinical applicability are not the sole
criteria for good measurement properties of diagnostic testing in cases, for instance, where tests can adequately discriminate between populations. For the
following section, only data that did not demonstrate
systematic between-day differences were explored.
This excluded right rotation VE data for the group
without LBP and flexion AE data for the patients
with LBP (Table 4).

GROUP
2

DISCUSSION

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

18 62

Flex VE

18 62

18 62

L Rot AE

R Rot AE

18 62

18 62

R SBend AE

L Rot VE

18 62

18 62

L Sbend AE

R Sbend VE

L Sbend VE

Error index

FIGURE 2. Means SD of absolute error (AE) and variable error

(VE) of both groups tested (R = right; L = left; Flex = flexion; Rot =
rotation; Sbend = side-flexion; LBP = low back pain).

LBP and data collection was performed on more

than 3 occasions to identify the presence of any
learning effects. As no major learning effects were
identified, apart from right-rotation VE, patients with
LBP subsequently recruited were tested on 2 occasions only. Again, in general, the measurements
made on occasion 2 were not systematically different
from those made on occasion 1 (apart from flexion
AE).
The mean AE data for flexion and associated variability in subjects without LBP were lower than those
from the Gill and Callaghan study.16 This may be
attributed to the fact that the task in the current experiment was easier to perform, as only 3 repetitions
were required, whereas 10 repetitions were required
in the previous study. As 30 seconds were given to
subjects to complete each test, it was thought more
appropriate to limit the number of repetitions and
increase the subjects concentration on the task.
The low discriminatory power (discriminating between individuals with or without LBP) of the position sense tests employed is in agreement with several previous studies. Gill and Callaghan16
recognized that with 20 flexion there was a significant overlap of the data for subjects without LBP
and for subjects with chronic LBP, despite the significant between-population differences identified. These
results are not limited to studies using the LMM as a
measuring device. Luoto et al25 have also reported a
54% discriminant power using a one-footed postural
control task and Taimela et al,47 using a threshold to
detection of movement proprioception task, reported
90% specificity but merely 25% sensitivity in patients
with chronic or recurrent LBP. Furthermore, Lam et
al23 did not find any significant differences between
the patients with LBP in their study and
333

REPORT

A summary of between-day measurement of active

lower spinal proprioception in recent studies is presented in Table 1. Nearly all the studies assessed dayto-day variability using some form of correlation coefficient and only one of the studies23 presented data
on individuals with LBP. Gill and Callaghans16 study
analyzed reliability in a mixed population. However,
establishing adequate reliability solely on the basis of
high ICC can, in some instances, mask aspects of
measurement errors. Complementary analysis using
SEM and SDD has therefore been strongly recommended.11,38,39 Conversely, a relatively low ICC may
be of little concern if a small SEM suggests that the
inconsistency of measurement is occurring in an acceptably small range.11 In this study, the relative high
magnitudes of SEM and SDD in comparison with the
raw data may support the low ICC findings. For the
group of patients with recurrent LBP, these indices
suggest that the test-retest error is excessively high
for a measure of this type to be sensitive to change
attributed to clinical interventions.
For the purposes of the present study, testing was
initially performed in a group of individuals without

18 62

RESEARCH

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Mean error values from subjects without LBP (days

13) and patients with LBP (days 12) were calculated (Figure 2) and introduced to a stepwise
discriminant analysis to identify whether the tests
could discriminate between the 2 groups. Simple
comparisons of mean error values between the 2
groups with independent-samples t-tests showed significant differences for flexion VE (P = 0.01), rightside rotation AE (P = 0.002), and left-side rotation
AE (P = 0.003). Discriminant analysis indicated that
the most discriminative data set between the 2
groups was the right-side rotation AE, with 83.3%
specificity (correct identification of 15 out of 18 subjects without LBP), but only 54.8% sensitivity (correct identification of just 34 out of 62 patients with
LBP). Overall correct subject identification was
61.3%.

LBP

0
N=

Diagnostic Validity

Non-LBP

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

asymptomatic volunteers from a previous study performed with the exact methodology.26
The general conclusion that can be drawn from
these concurrent findings is that patient identification on the basis of proprioception and motor control tests has low accuracy. Large normative databases
of more reliable proprioceptive testing methods will
need to be established in the future to identify appropriate cut-off scores of repositioning errors,
which, if exceeded, would suggest impaired
proprioceptive performance.
Alternatively, the presence of pain in the lumbar
spine may have a facilitatory effect in proprioception
acuity through at least 2 mechanisms. It has been
shown that chemical products of inflammation in the
facet joints8 or intramuscular injection of bradykinin
to the paraspinals36 increases the excitatory state of
the fusimotor system innervating the muscle spindles
and possibly providing better proprioceptive awareness. Also, certain psychological states like fear can
lead to increased fusimotor drive, a phenomenon
described as fusimotor set.37 In chronic LBP, the
presence of increased paraspinal electromyographic
activity in full forward flexion (loss of the flexionrelaxation phenomenon21,49) or even during the
swing phase of walking,3 might be related to these
excitatory mechanisms.
The LMM goniometer spans the area between the
T7 and S1 vertebrae, and therefore, testing
proprioception with this instrument could be criticized as not being selective enough of the lumbar
spine. However, it could also be argued that analyzing repositioning accuracy of an area as a whole, covering several vertebrae, can readily identify spinal
proprioception deficits. Results from a recent sagittal
plane movement analysis with videofluoroscopy in
asymptomatic subjects and patients with L4-level
spondylolisthesis suggested that several patients had a
generalized movement pattern classified as being abnormal.31 This may signify a departure from the normal neuromuscular programming, as has been hypothesized to happen due to injury,33,32 with a
probable effect on proprioception acuity of patients
with LBP.
This study assessed proprioception acuity with 1
methodological approach (an active movement reproduction task), as opposed to other methods using
passive movement reproduction.34 As performance
under 1 of those testing conditions has not been
shown to predict performance in another,34 findings
from this study are limited in that respect.

CONCLUSION
The reliability of the measurements obtained using
an active thoracolumbar position sense protocol was
established as low in a group of patients with subacute LBP. More than one method of results analysis
and presentation provided a better understanding of
334

the measurement error magnitude. For this group of

patients with recurrent LBP, no proprioception deficits could be clearly identified with detailed multipleplane testing methods, compared to an asymptomatic
control group. These results agree with other studies
suggesting that patients with LBP experience no
proprioception deficits.16,23,25,47 The lumbar motion
monitor, using the methods described in this study,
cannot be recommended for the identification of
patients with recurrent LBP and may be insufficiently
sensitive to measure the effects of treatment.

ACKNOWLEDGEMENTS
The authors thank Mr. MJ Callaghan and Mr. P
Doherty for helpful comments in the design of the
study, Mr. PJ Watson for reviewing the manuscript,
and the Royal Liverpool Hospital (UK) for loan of
the equipment used in this study.

REFERENCES
1. Alaranta H, Moffroid M, Elmqvist L, Held J, Pope M,
Renstrom P. Postural control of adults with
musculoskeletal impairment. Crit Rev Phys Rehabil
Med. 1994;6:337370.
2. Alexander M. Gender differences in biomechanical
aspects of performance. Crit Rev Phys Rehabil Med.
1998;10:1536.
3. Arendt-Nielsen L, Graven-Nielsen T, Svarrer H, Svensson P. The influence of low back pain on muscle
activity and coordination during gait: a clinical and
experimental study. Pain. 1996;64:231240.
4. Brumagne S, Cordo P, Lysens R, Verschueren S, Swinnen S. The role of paraspinal muscle spindles in
lumbosacral position sense in individuals with and
without low back pain. Spine. 2000;25:989994.
5. Brumagne S, Lysens R, Spaepen A. Lumbosacral position sense during pelvic tilting in men and women
without low back pain: test development and reliability
assessment. J Orthop Sports Phys Ther. 1999;29:345
351.
6. Brumagne S, Lysens R, Spaepen A. Lumbosacral repositioning accuracy in standing posture: a combined
electrogoniometric and videographic evaluation. Clin
Biomech. 1999;14:361363.
7. Brumagne S, Lysens R, Swinnen S, Verschueren S. Effect
of paraspinal muscle vibration on position sense of the
lumbosacral spine. Spine. 1999;24:13281331.
8. Cavanaugh JM, Ozaktay AC, Yamashita T, Avramov A,
Getchell TV, King AI. Mechanisms of low back pain: a
neurophysiologic and neuroanatomic study. Clin
Orthop. 1997;166180.
9. Clinical Standards Advisory Group. Back Pain: Report
of a CSAG Committee on Back Pain. London, UK:
HMSO; 1994.
10. Currier D. Elements of Research in Physical Therapy.
3rd ed. Baltimore, MD: Williams & Wilkins; 1990.
11. Denegar C, Ball D. Assessing reliability and precision of
measurement: an introduction to intraclass correlation
and standard error of measurement. J Sport Rehab.
1993;2:3542.
12. Fleiss J. The Design and Analysis of Clinical Experiments. New York, NY: John Wiley and Sons; 1986.
J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

335

Journal of Orthopaedic & Sports Physical Therapy

Downloaded from www.jospt.org at NTU Medical Library on September 22, 2015. For personal use only. No other uses without permission.
Copyright 2002 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

REPORT

J Orthop Sports Phys Ther Volume 32 Number 7 July 2002

31. Okawa A, Shinomiya K, Komori H, Muneta T, Arai Y,

Nakai O. Dynamic motion study of the whole lumbar
spine by videofluoroscopy. Spine. 1998;23:17431749.
32. Panjabi MM. The stabilizing system of the spine. Part I.
Function, dysfunction, adaptation, and enhancement. J
Spinal Disord. 1992;5:383389; discussion 397.
33. Panjabi MM. The stabilizing system of the spine. Part II.
Neutral zone and instability hypothesis. J Spinal Disord.
1992;5:390396; discussion 397.
34. Parkhurst TM, Burnett CN. Injury and proprioception in
the lower back. J Orthop Sports Phys Ther.
1994;19:282295.
35. Parnianpour M, Nordin M, Kahanovitz N, Frankel V.
1988 Volvo award in biomechanics. The triaxial coupling of torque generation of trunk muscles during
isometric exertions and the effect of fatiguing isoinertial
movements on the motor output and movement patterns. Spine. 1988;13:982992.
36. Pedersen J, Sjolander P, Wenngren BI, Johansson H.
Increased intramuscular concentration of bradykinin
increases the static fusimotor drive to muscle spindles
in neck muscles of the cat. Pain. 1997;70:8391.
37. Prochazka A, Hulliger M. The continuing debate about
CNS control of proprioception. J Physiol.
1998;513:315.
38. Rankin G, Stokes M. Reliability of assessment tools in
rehabilitation: an illustration of appropriate statistical
analyses. Clin Rehabil. 1998;12:187199.
39. Roebroeck ME, Harlaar J, Lankhorst GJ. The application
of generalizability theory to reliability assessment: an
illustration using isometric force measurements. Phys
Ther. 1993;73:386395; discussion 396401.
40. Roland M, Morris R. A study of the natural history of
back pain. Part I: Development of a reliable and
sensitive measure of disability in low-back pain. Spine.
1983;8:141144.
41. Rothwell J. Proprioceptors in muscles, joints and skin.
In: Rothwell J, ed. Control of Human Voluntary Movement. London, UK: Chapman & Hall; 1994:86126.
42. Schmidt R. Methodology for studying motor behavior.
In: Schmidt R, ed. Motor Control and Learning. A
Behavioral Emphasis. Champaign, IL: Human Kinetics;
1988:5662.
43. Skinner HB, Barrack RL, Cook SD. Age-related decline
in proprioception. Clin Orthop. 1984;208211.
44. Sparto PJ, Parnianpour M, Reinsel TE, Simon S. The
effect of fatigue on multijoint kinematics and load
sharing during a repetitive lifting test. Spine.
1997;22:26472654.
45. Swinkels A, Dolan P. Regional assessment of joint
position sense in the spine. Spine. 1998;23:590597.
46. Swinkels A, Dolan P. Spinal position sense is independent of the magnitude of movement. Spine.
2000;25:98104; discussion 105.
47. Taimela S, Kankaanpaa M, Luoto S. The effect of
lumbar fatigue on the ability to sense a change in
lumbar position. A controlled study. Spine.
1999;24:13221327.
48. Von Korff M. Studying the natural history of back pain.
Spine. 1994;19:2041S2046S.
49. Watson PJ, Booker CK, Main CJ, Chen AC. Surface
electromyography in the identification of chronic low
back pain patients: the development of the flexion
relaxation ratio. Clin Biomech. 1997;12:165171.

RESEARCH

13. Gandevia S, Burke D. Does the nervous system depend

on kinesthetic information to control natural limb
movements? Behav Brain Sci. 1992;15:614632.
14. Gandevia S, McCloskey DI, Burke D. Kinaesthetic
signals and muscle contraction. Trends Neurosci.
1992;15:6265.
15. Gill KP, Callaghan MJ. Intratester and intertester reproducibility of the lumbar motion monitor as a measure of
range, velocity and acceleration of the thoracolumbar
spine. Clin Biomech. 1996;11:418421.
16. Gill KP, Callaghan MJ. The measurement of lumbar
proprioception in individuals with and without low
back pain. Spine. 1998;23:371377.
17. Goodwin GM, McCloskey DI, Matthews PB. The contribution of muscle afferents to kinaesthesia shown by
vibration induced illusions of movement and by the
effects of paralysing joint afferents. Brain. 1972;95:705
748.
18. Hicks C. Research for Physiotherapists. Project Design
and Analysis. 2nd ed. Edinburgh, Scotland: Churchill
Livingstone; 1995.
19. Huck S, Cormier W. Reading Statistics and Research.
2nd ed. New York, NY: Harper Collins; 1996.
20. Jensen MP, Turner JA, Romano JM, Fisher LD. Comparative reliability and validity of chronic pain intensity
measures. Pain. 1999;83:157162.
21. Kaigle AM, Wessberg P, Hansson TH. Muscular and
kinematic behavior of the lumbar spine during flexionextension. J Spinal Disord. 1998;11:163174.
22. Kaplan FS, Nixon JE, Reitz M, Rindfleish L, Tucker J.
Age-related changes in proprioception and sensation of
joint position. Acta Orthop Scand. 1985;56:7274.
23. Lam SS, Jull G, Treleaven J. Lumbar spine kinesthesia in
patients with low back pain. J Orthop Sports Phys Ther.
1999;29:294299.
24. Lephart S, Riemann B, Fu F. Introduction to the
sensorimotor system. In: Lephart S, Fu F, eds.
Proprioception and Neuromuscular Control in Joint
Stability. Champaign, IL: Human Kinetics; 2000:xvii
xxiv.
25. Luoto S, Aalto H, Taimela S, Hurri H, Pyykko I,
Alaranta H. One-footed and externally disturbed twofooted postural control in patients with chronic low
back pain and healthy control subjects. A controlled
study with follow-up. Spine. 1998;23:20812089; discussion 20892090.
26. Maffey-Ward L, Jull G, Wellington L. Toward a clinical
test of lumbar spine kinesthesia. J Orthop Sports Phys
Ther. 1996;24:354358.
27. Marras W, Fathallah F, Miller R, Davis S, Mirka G.
Accuracy of a three-dimensional lumbar motion monitor for recording dynamic trunk motion characteristics.
Int J Indust Ergon. 1992;9:7587.
28. Marras WS, Parnianpour M, Ferguson SA, et al. The
classification of anatomic- and symptom-based low
back disorders using motion measure models. Spine.
1995;20:25312546.
29. Melzack R. The short-form McGill pain questionnaire.
Pain. 1987;30:191197.
30. Newcomer K, Laskowski ER, Yu B, Larson DR, An KN.
Repositioning error in low back pain. Comparing trunk
repositioning error in subjects with chronic low back
pain and control subjects. Spine. 2000;25:245250.