Clinical Hub,UW Health Clinical Tool Search,UW Health Clinical Tool Search,Questionnaires,Related

Reliability and Responsiveness of the Barry-Albright Dystonia Scale

Reliability and Responsiveness of the Barry-Albright Dystonia Scale - Clinical Hub, UW Health Clinical Tool Search, UW Health Clinical Tool Search, Questionnaires, Related

Reliability and
responsiveness of the
Dystonia Scale
Margaret J Barry* MS PT PCS, Research Physical Therapist,
Department of Neurosurgery, Children’s Hospital of
Jessie M VanSwearingen PhD PT, Associate Professor,
Department of Physical Therapy, School of Health and
Rehabilitation Sciences,University of Pittsburgh;
A Leland Albright MD, Chief, Department of Neurosurgery,
Children’s Hospital of Pittsburgh, PA, USA.
*Correspondence to first author at Children’s Hospital,
Department of Neurosurgery, 3705 Fifth Avenue
Pittsburgh, PA 15213.
E-mail: mjbst95+@pitt.edu
The reliability and responsiveness of the Barry–Albright
Dystonia (BAD) Scale, a 5-point ordinal severity scale for
secondary dystonia, was assessed. For interrater reliability, 13
raters scored 10 videotaped patients; for intrarater reliability,
two raters rated the videotape again. For test–retest
reliability, patients were rated on two occasions. Four
inexperienced raters scored patients, received training, then
scored additional patients. To assess responsiveness, we
compared patient and physician global ratings of change
(better, same, and worse) with BAD Scale score changes for
18 patients on intrathecal baclofen (ITB) trials. We assessed
reliability with the intraclass correlation coefficient (ICC).
The mean ICC for total BAD Scale scores were as follows:
interrater reliability 0.866, intrarater reliability 0.967 and
0.978, test–retest reliability 0.978 (before training) and 0.967
(after training). We found the BAD Scale responsive to
change, with most improved scores in patients rated by the
patient, family, and neurosurgeon as ‘better’. The total scores
were reliable for experienced raters. We recommend training
for clinicians interested in using the scale.
Dystonia is characterized by sustained muscle contractions
causing abnormal postures or twisting and repetitive move-
ments (Fahn et al. 1987). It is typically classified by distribu-
tion, cause, or age of onset. In terms of distribution, dystonia
may be focal, as in writer’s cramp, or generalized, affecting
all or most of the body. Its cause may be a primary disorder of
unknown etiology, or secondary to brain disorders, such as
cerebral palsy (CP); it reportedly affects 15% of patients with
CP (Treves and Korczyn 1986). Dystonia may not be present
until the child reaches 3 to 5 years of age, and often progress-
es into late childhood or adolescence (Burke et al. 1980,
Albright 1996a). Secondary dystonia may also be present in
patients with acquired brain injuries (Krauss et al. 1992), but
the incidence is not well documented.
A relatively new treatment for generalized dystonia is
intrathecal baclofen (ITB), infused by an implanted program-
mable pump (Synchromed
, Medtronic, Minneapolis, MN,
USA) (Narayan et al. 1991; Penn et al. 1995; Albright et al.
1996, 1998). The pump continuously delivers small doses of
baclofen directly into the CSF, typically at the mid-thoracic
spinal level. Side effects such as lethargy and dizziness, often
seen with oral baclofen, are not as common with intrathecal
administration. As the treatment is somewhat invasive and
expensive, it is imperative to document its effects accurately.
ITB therapy is approved by the Food and Drug Administration
(FDA) for treatment of cerebral and spinal-origin spasticity,
but dystonia is not an indication for use on the current drug
labeling. Permission was granted from the FDA to study the
effects of ITB therapy on secondary generalized dystonia.
Documenting the degree of generalized dystonia is diffi-
cult, and so, therefore, is defining successful treatment inter-
vention. Studies describe dystonia seen in patients with CP
and traumatic brain injuries (TBI) (Burke et al. 1980,
Kyllerman et al. 1982, Saint-Hilaire et al. 1991, Ferraz et al.
1992, Krauss et al. 1992), but there is no universally accepted
measurement scale for dystonia in these patient populations.
In evaluating our initial cohort of patients with generalized
dystonia due to CP, it was difficult to determine the most
appropriate method of measuring change in dystonia
(Albright et al. 1996). In an attempt to quantify dystonic move-
ments, we modified the Fahn–Marsden Movement Scale
(Burke et al. 1985), which assesses movement primarily relat-
ed to function. In patients with certain types of dystonia, such
as spasmodic torticollis, an effective intervention results in
normal movement and function. However, some patients with
CP do not have voluntary control of their movements and have
few functional abilities. Furthermore, many of these patients
have significant cognitive impairments, making it difficult to
know whether they are attempting to move in response to the
examiner’s directions or whether they are unable to move.
Many scales, including the Fahn–Marsden Movement and
Disability Scales, lack even face validity in patients who are
severely involved because rating severity is based on function.
Rather than basing scoring criteria solely on functional abili-
ties, our revision of the scale included severity of posturing
and involuntary dystonic movements (Albright 1996b). For
patients who are not expected to gain function, the goals of
treatment may be easier care, improved comfort, and a reduc-
tion in dystonia (Table I).
In a pilot study, 12 patients with dystonic CP were
assessed using our revision of the Fahn–Marsden Movement
Scale. That study attempted to determine if ITB was effective
404 Developmental Medicine & Child Neurology 1999, 41: 404–411

in reducing dystonia. A neurologist and a physical therapist
graded dystonia, and reliability between the two raters was
demonstrated. The scale quantified change after interven-
tion (Albright et al. 1998), and it was clear that the pilot data
warranted further study of the scale.
The criteria for several items on our scale, based on feed-
back from the neurologist, other physical therapists and our
own clinical experience with the scale, were clarified. The
initial criteria were not consistent across body regions, there-
fore criteria that followed similar guidelines for each region
were developed to simplify the scoring process. The modi-
fied version was termed the Barry–Albright Dystonia (BAD)
scale (Appendix I). The purpose of the present study was to
assess the reliability of the BAD Scale, specifically interrater,
intrarater, and test–retest reliability, and the effects of train-
ing on reliability. We also studied the construct validity of the
scale by determining responsiveness of the scale to clinically
meaningful changes in dystonia with intervention. The
intent was to define the basic psychometric properties of the
revised scale to determine the generalizability of its use by
clinicians in a variety of settings.
Eighteen consecutive patients (six females and 12 males)
with secondary dystonia, treated with ITB therapy, partici-
pated in this study. Their mean age was 16 years (range 3 to
42 years). All patients had generalized dystonia except for
one man who had hemidystonia (see Table I). Thirteen
patients had a primary diagnosis of CP; the others had an
acquired brain injury. An interdisciplinary team assessed
patients in the spasticity clinic at Children’s Hospital of
Pittsburgh between May 1994 and May 1996. Two patients
referred directly to the neurosurgeon (ALA) did not attend
the clinic. The recommended intervention for all patients
was ITB therapy. In all cases, the team, including patients and
families, felt that the severity of dystonia warranted this inva-
sive intervention. Dystonia substantially interfered with
functional activities, ease of care, or comfort. Most patients
had undergone trials of oral medication without success. All
patients/guardians signed informed consent and agreed to
participate in studies of ITB therapy. The studies were
approved by the Institutional Review Board of the Children’s
Hospital of Pittsburgh.
The authors developed the BAD Scale to address the need for
a clinical assessment tool that allowed for the varied cogni-
tive and physical impairments of persons with CP and TBI.
The BAD Scale is a 5-point, criterion-based, ordinal scale
designed to assess dystonia in eight body regions: eyes,
mouth, neck, trunk, and the four extremities (see Appendix
I). Raters score dystonia as none (0), slight (1), mild (2),
moderate (3), or severe (4). Individual scores for each region
are added for a total score. Each region has specific descrip-
tors for scoring, but the following general rules apply: slight
dystonia is present less than 10% of the time; mild dystonia
does not interfere with function or care; moderate dystonia
is characterized as interfering with a functional activity; and
severe dystonia prevents the performance of the activity – for
example, if dystonia prevents lower-extremity weight bear-
ing. If a rater is unable to assess even simple functional tasks,
such as sitting in a chair, the severity of the muscle contrac-
tions or abnormal posturing determines the score. If a rater
is unable to assign a score to a particular body region, this
region is excluded and a total score for the other regions may
be recorded. However, in reassessing the patient, the same
body regions scored at baseline should be included to allow
determination of change in dystonia.
The BAD Scale excludes two components of the
Fahn–Marsden Disability Scale: provoking factors and disabili-
ty. The assessment of provoking factors was eliminated as dys-
tonia was present both at rest and with activity in our patient
population. The Fahn–Marsden Disability Scale is limited in its
scope, rating tasks such as eating and hygiene on a 4-point
scale. Instead, functional assessment tools, such as the
Pediatric Evaluation of Disability Index (PEDI) (Feldman et al.
1990, Haley et al. 1992), the Functional Independence
Measure (Uniform Data System 1993, Ottenbacher et al. 1996),
and the Gross Motor Function Measure (GMFM) (Russell et al.
1989, 1993) are routinely used as outcome measures related to
the dimensions of functional limitation and disability due to
their demonstrated reliability and validity for our patient popu-
lation. The Canadian Occupational Performance Measure (Law
et al. 1994) was chosen to allow patients and their families to
identify treatment goals important to them.
Barry–Albright Dystonia Scale Margaret J Barry et al. 405
Table I: Patient characteristics
Patient Age (y)/ Primary Patient/family
sex diagnosis goals
1 7/F Encephalitis as Prevent deformity
2 42/F Cerebral palsy Ease of care,
improved function
3 15/M Cerebral palsy Ease of care, positioning
4 35/F Hypoxic episode Comfort, improved
at 5 y function
5 11/M Cerebral palsy Ease of care, comfort
6 6/M Cerebral palsy Ease of care, positioning
7 28/M Cerebral palsy Improved nutritional
8 33/M Brain injury at 6 y Improved ambulation,
9 9/M Cerebral palsy Improved
10 12/F Cerebral palsy Comfort, improved
11 14/M Cerebral palsy Improved ambulation
12 29/M Cerebral palsy Comfort, improved
13 3/M Cerebral palsy Prevent deformity,
14 14/F Cerebral palsy Comfort, improved
15 5/F Cerebral palsy Prevent deformity,
ease of care
16 14/M Near drowning at 7 y Ease of care
17 7/M Cerebral palsy Comfort, improved
18 6/M Holoprosencephaly Ease of care
M, male; F, female.

Patients were videotaped at the time of initial evaluation to
document their baseline dystonia. Clinically, routine video-
taping of patients is performed to provide a basis for compar-
ison during the ITB trial with an external infusion pump:
initial videotape is helpful in documenting the change in
severity of the dystonia to justify the implant. In some cases,
patients elect to have a pump implanted for continuous infu-
sion of ITB during the same hospital admission. The use of
videotapes in the reliability study ensured that all raters
viewed exactly the same movements and postures.
Videotape sessions lasted 20 to 45 minutes depending on the
level of cooperation, mobility, and functional capabilities of
the patient. We asked patients to remain still and then to per-
form a variety of functional tasks, depending on their capa-
bilities. After observing several patients, we developed a
script for videotaping to promote more consistent proce-
dures. The standardized testing mentioned above was also
performed at this time, appropriate to the patient’s age,
functional, and developmental level.
After the initial assessments were completed, a physician
placed an intrathecal catheter and connected it to an exter-
nal microinfusion pump for delivery of baclofen. Trials of ITB
lasted 3 to 9 days. At the end of the trial period, the videotap-
ing procedure was repeated. One patient was videotaped at a
midpoint in the trial, before her mother noted a change in
dystonia, because she was receiving a high dose, and we
wanted to document her status. Patients and their families
rated the change in dystonia with ITB as better, same, or
worse. The neurosurgeon also rated the change. Agreement
between the surgeon and families was unanimous. These
global ratings of change were used in our study of respon-
Brief videotaped segments of our patients at the time of
initial evaluation and during the trial of ITB were edited. One
patient did not undergo an external pump trial; she was the
first patient with dystonia treated with ITB at our center, and
the neurosurgeon (ALA) implanted a programmable pump
after a slight improvement with high oral doses of baclofen
was found. Her first videotape session demonstrated her
functional abilities with the pump turned on, whilst the sec-
ond session, conducted a few days later, showed her
attempting the same activities with the pump turned off.
Videotape segments of all patients were then randomized
and used to assess the responsiveness of the BAD Scale to
clinical change with ITB compared with the patient, family,
and physician global rating of change. A subset of the video-
tapes (the first 10 subjects) was used to study the interrater,
intrarater, and test–retest reliability of our scale and to assess
the effects of training on reliability.
For the scale to fulfill its intended use, several forms of
reliability and validity needed to be established. For the scale
to be used widely, it was necessary to determine interrater
reliability (Portney and Watkins 1993). To use the scale
repeatedly, to identify change before and after intervention
or over time, intrarater and test–retest reliability were neces-
sary. Finally, to see whether the scale actually measures the
phenomenon of dystonia, it was imperative to measure valid-
ity. We did not feel that it was possible to assess concurrent
validity due to the lack of a ‘gold standard’ for the assessment
of dystonia in this patient population. The aim was to deter-
mine whether the scale identified clinically meaningful
change, and therefore the responsiveness of the BAD Scale to
global change after treatment, identified by physicians and
patients as better, worse, or the same (Stratford et al. 1996),
was examined.
A total of 17 raters participated in this study: seven physical
therapists, seven physicians, two nurses, and one occupation-
al therapist (see Table II). Raters had varying levels of clinical
experience with the scale and the patient populations of CP
and acquired brain injuries. Thirteen of the raters (raters 1 to
13) participated in the determination of interrater reliability,
and the other four raters (raters 14 to 17) were involved in an
investigation of the effects of training on reliability. Two raters
participated in the study of responsiveness.
Interrater reliability
We assessed the reliability between raters by comparing the
BAD Scale scores of various raters who scored the same
videotape of 10 patients. After reading the BAD Scale for 5
minutes, raters viewed videotaped segments of 10 patients
and rated the patients’ dystonia using the scale for all eight
body regions.
The group of raters included 10 physicians and therapists,
from several different medical centers across the country,
recruited at a continuing education course (raters 1 to 10,
see Table II). None of these raters had previous experience
with the BAD Scale. The other three raters (raters 11 to13,
see Table II) were clinicians from our center. Raters 11 and 12
had extensive clinical experience using the scale and rater 13
had no previous experience with the scale.
Intrarater reliability
To determine the consistency of individual raters, two raters
scored the videotape used for the interrater study, and then
scored the same videotape again 2 weeks later. A nurse and a
physical therapist (Raters 11, 12), both experienced with clini-
cal use of the scale, served as raters for this portion of the study.
406 Developmental Medicine & Child Neurology 1999, 41: 404–411
Table II: Rater characteristics
Rater Profession Experience Patient populations
1 Physical therapist 10 Pediatrics: CP, BI
2 Physical therapist 3 Pediatrics and adults: CP
3 Physical therapist 20 Pediatrics: CP
4 Orthopedic surgeon 25 Pediatrics: CP
5 Occupational therapist 13 Pediatrics and adults: CP
6 Physiatrist 2 Pediatrics and adults: CP, BI
7 Neurologist 20 Pediatrics: CP, BI
8 Neurologist 30 Pediatrics: CP
9 Neurologist 15 Adults: multiple sclerosis,
10 Neurologist 16 Pediatrics: CP, BI
11 Nurse 11 Pediatrics and adults: CP, BI
12 Physical therapist 7 Pediatrics and adults: CP, BI
13 Neurosurgeon Senior Pediatrics and adults: CP, BI
14 Nurse 16 Pediatrics: CP, BI
15 Physical therapist 11 Pediatrics: CP, BI
16 Physical therapist 1 Pediatrics: CP, BI
17 Physical therapist 1 Pediatrics: CP, BI
CP, cerebral palsy; BI, brain injury.

Test–retest reliability
Test–retest reliability is important in determining that the
scores remain constant when there is no change in the prop-
erty that is being measured. This form of reliability ensures
that if there is a change in scores, it is not due to instability of
the scale (Portney and Watkins 1993). To determine
test–retest reliability, a physical therapist (rater 12) scored
videotapes for six patients at initial evaluation. The same
patients were then rated performing the same activities,
without any intervention, in another videotape session
where no change was expected. For example, a session in
clinic upon initial evaluation followed by a second session
upon admission to the hospital 1 week later, before the ITB
trial, were rated.
To determine if training in the use of the scale would
improve reliability, one nurse and three physical therapists
(raters 14 to 17, see Table II), all with no previous experience
with the scale, participated in training sessions. All raters
scored five patients from the videotape, then the first author
(rater 12) discussed the scoring criteria for rating those
patients. The training session lasted about 30 minutes. After
training, raters scored five more patients. Interrater reliabili-
ty before and after the training were compared to determine
its effects. Improvement in reliability after training was
Validity of the BAD Scale was examined in the context of its
responsiveness to clinically meaningful change. We com-
pared change in BAD Scale scores with outcomes estab-
lished by family, patient, and physician’s global ratings of
better, same, or worse. Two raters assessed videotapes of 18
patients before and after ITB, viewed in random order. A
neurosurgeon (rater 13) blinded to treatment status and a
physical therapist (rater 12), who was not blinded, partici-
pated in the responsiveness study, scoring all items of the
BAD Scale demonstrated on the videotape. We expected the
dystonia scale score to decrease for patients whose clinical
status was better, to stay the same for patients whose clini-
cal status was unchanged, and to increase for patients rated
Interrater, intrarater and test–retest reliability were deter-
mined by the intraclass correlation coefficient (ICC [2,1]) for
the individual items and the total scores. The ICC 2,1 is an
appropriate reliability coefficient to demonstrate the extent
of agreement in assigning rank order scores (Shrout and
Fleiss 1979, Lahey et al. 1983) and was chosen because the
the scale is intended to have broad clinical use. The random
effects model of the ICC allows one to generalize the results
of the reliability study to clinicians who are similar to those
who participated in the study. The ICC typically ranges from
0 to 1, with scores closer to 1 representing greater reliability
(Portney and Watkins 1993).
To analyse the results of the responsiveness portion of our
study, we used descriptive statistics to indicate the differ-
ences in BAD Scale scores before and after intervention by
outcome group (better, same, or worse).
Interrater reliability for the eight individual BAD Scale item
scores by 13 raters for 10 patients using the ICC 2,1 ranged
from 0.061 to 0.866 (Table III), yielding a mean value of
0.422. Overall, the reliability for the total score was excellent,
Comparison of the eight individual BAD Scale item scores by
the same rater for the videotape of 10 patients rated at two
sessions yielded mean ICC values of 0.869 for rater 11 and
0.802 for rater 12. The ICC values for the total scores were
0.967 and 0.978 respectively (Table IV). The mean ICC for
the individual items for rater 12, excluding the left upper
extremity, was 0.873.
For test–retest reliability, the mean ICC value for the eight
individual items was 0.730 and for the total scores for six
patients was 0.978 (rater 12, Table V). The ICC of 0 for the left
upper extremity resulted from the extremely low variability
in the scores, despite actual score differences of only one
level for three patients between test and retest and complete
agreement for the other three patients.
Barry–Albright Dystonia Scale Margaret J Barry et al. 407
Table III: Interrater reliability of scoring individual items and
total score of the BAD Scale for 13 raters
Item ICC
Eyes 0.423
Mouth 0.288
Neck 0.558
Trunk 0.717
Left upper extremity 0.133
Right upper extremity 0.650
Left lower extremity 0.061
Right lower extremity 0.542
Total BAD Scale scores 0.866
ICC, intraclass correlation coefficient.
Table IV: Intrarater reliability of scoring individual items and
total score of the BAD Scale for two raters
Item Rater 1 Rater 2
Eyes 0.960 0.814
Mouth 0.622 0.625
Neck 0.870 0.881
Trunk 0.958 0.926
Left upper extremity 0.963 0.308
Right upper extremity 0.936 0.905
Left lower extremity 0.824 1.000
Right lower extremity 0.816 0.960
Total BAD Scale scores 0.967 0.978
ICC, intraclass correlation coefficient.

A comparison of the eight individual item scores for four
raters before and after a training session yielded a mean ICC
of 0.087 before training, and a substantially improved mean
ICC of 0.587 after training (Table VI). The ICC for the total
score after training improved to 0.967 from a value of 0.549
before training.
The BAD Scale scores distinguished between the patients
with clinical outcomes considered better and the patients
with outcomes considered the same or worse. Thirteen
patients were better, three the same, and two worse, accord-
ing to the perceptions of patients, families, and their physi-
cian (see Table VII). The mean reduction in the scores for the
patients who were better, same, and worse were 12.1, 3.3,
and 5.5 respectively (see Fig. 1). Total scores did not increase
for any outcome group, even for the patients rated worse. We
did not perform statistical analysis of the differences in the
scores due to the small sample size and the unequal distribu-
tion of patients among outcome groups.
Documentation of the wide range of functional limitations
and disabilities due to dystonia is challenging. The involun-
tary dystonic movements of the arm may knock items out of
someone else’s hand, interfere with getting through a door-
way, and may cause embarrassment. Patients with dystonic
CP also tend to be greatly influenced by primitive reflexes
(Kyllerman et al. 1982) which further complicates the task of
motor assessment.
We believe that the BAD Scale provides a reasonable
assessment of the severity of dystonia for patients with CP
and TBI. It has the potential to be a useful tool for healthcare
professionals with divergent clinical backgrounds who
408 Developmental Medicine & Child Neurology 1999, 41: 404–411
Table V: Test–retest reliability of scoring individual items and
total score of the BAD Scale for one rater
Item ICC
Eyes 0.925
Mouth 0.762
Neck 0.954
Trunk 0.636
Left upper extremity 0.000
Right upper extremity 0.931
Left lower extremity 0.706
Right lower extremity 0.922
Total BAD Scale scores 0.978
Table VI: Interrater reliability of scoring individual items and
total score of the BAD Scale before and after training for four
before training after training
Eyes 0.175 0.281
Mouth 0.119 0.282
Neck 0.282 0.791
Trunk –0.036 0.892
Left upper extremity –0.011 0.244
Right upper extremity –0.029 0.872
Left lower extremity 0.000 0.417
Right lower extremity 0.197 0.913
Total BAD Scale scores 0.549 0.967
Table VII: Patient outcomes
Patient Global Total BAD Scale Scores
Change in total
Before ITB After ITB BAD Scale scores
1 Same 24 21 3
3 Same 25 22 3
7 Same 26 22 4
8 Worse 12 9 3
11 Worse 20 12 8
2 Better 27 14 13
4 Better 28 15 13
5 Better 32 6 26
6 Better 25 10 15
9 Better 21 18 3
10 Better 22 6 16
12 Better 21 13 8
13 Better 29 20 9
14 Better 23 7 16
15 Better 23 16 7
16 Better 26 15 11
17 Better 23 11 12
18 Better 19 4 15
Mean 23.7 13.4 10.3
SD 4.3 5.6 6
SEM 1 4.2 4.1
ITB, intrathecal baclofen.
Rated by patient, family, and physician.
Assessed by rater 12.
Figure 1: Mean Change in BAD Scale scores by global
ratings of change.
Better Worse Same
Global ratings




Barry–Albright Dystonia Scale Margaret J Barry et al. 409
assess these patients at different centers. In this study, the
total scores demonstrated excellent reliability, but individual
items did not. Clinicians using this scale need to consider
this issue.
The ICC for the total scores for the 13 raters who participat-
ed in the interrater reliability portion of the study was an excel-
lent 0.866. However, the ICC for the individual items varied,
and for some items was poor. One problem influencing the
ICC was the limited variability of scores for the left upper and
lower extremities. The scores for test–retest reliability for the
left upper extremity for rater 12 illustrates the problem. The
rater scored three subjects exactly the same on test and retest
(4 points), and the other three subjects scored 4 the first time
and 3 the second time. This difference of only one point for
half the subjects resulted in an ICC of 0. The limited variability
of scores means a greater penalty for even slight (one level),
infrequent differences in scores between raters resulting in a
lower mean ICC than an observation of the raw data may sug-
gest (Portney and Watkins 1993). To define the influence of
limited variability on the reliability of the extremity item
scores, it will be important in future studies to include a wider
range of severity of dystonia. In addition, the small sample size
contributed to the statistical problems.
Eleven raters had no previous experience with the scale,
and only a few minutes to review the scale before rating
patients. Many of the raters, especially the physical therapists
in the group, reported that they could not assess dystonia
without handling the patients. The orthopedic surgeon and
neurologists did not report difficulty scoring from the video-
tapes, and they rated patients more quickly. The reliability of
the physicians was not higher than the therapists’, but the
physicians did not seem as apprehensive as the therapists in
assigning scores.
The therapists and the nurse who participated in the train-
ing exercise had very low initial reliability scores, but these
improved markedly after a brief training session with an
experienced rater. Possibly the raters in the interrater relia-
bility study followed the rules more closely knowing that
they would be rating the videotapes only once. In the train-
ing group, however, the raters knew that they would rate, be
trained, then rate other videotapes after training. Therefore,
the training group may not have followed the rules as closely
in the initial session. In any case, based on this small sample
of raters, training in the use of the BAD Scale appears to be
beneficial for clinicians intending to use it.
Intrarater reliability for two raters familiar with the scale
demonstrated very good reliability, suggesting that the scale
is a reliable tool for experienced professionals using it with
their patients. Despite the influence of the extremely limited
variability of the left upper extremity, test–retest reliability
was also acceptable, suggesting that the tool is a consistent
measure of dystonia. Given the results, we feel confident that
changes in BAD Scale scores reflect change in dystonia and
not inconsistencies in measurement. However, the rater was
familiar with the patients and scoring may thus have been
inadvertently biased. To our knowledge, aside from the
Fahn–Marsden Movement Scale which we feel is inappropri-
ate for our patient population, there are no other accepted
tools for assessing the severity of dystonia (Albright et al.
1996). As the BAD Scale addresses the issues related to a
severely involved patient population, the scale demonstrates
face validity.
The BAD Scale seems to be responsive to change as identi-
fied by the patient, family, and physician. The largest change
in scores before and after ITB therapy occurred in the
patients reported to be better. Much less change was noted
for patients who were reportedly worse or unchanged after
intervention. The small number of subjects in the same and
worse categories make it difficult to draw any definitive con-
The patients who had the highest scores initially, reflect-
ing the most severe dystonia, improved the most after ITB.
Patients with milder dystonia and lower initial scores did
not perceive improvement, although the scores would
suggest that some reduction in dystonia occurred in those
patients as well. It is possible that the scores are accurate
before and after treatment, and other that factors may
impact the perceptions of the patients and their families,
for instance, the magnitude of change may not have been
clinically significant. Another explanation would be that
patients with lower levels of impairment and disability
expected to eliminate their dystonia and have no residual
disability after intervention. In patients with goals of func-
tional gains, it is sometimes difficult to assess improved
function during an ITB trial. The medication and CSF leaks
may cause headaches, nausea, and vomiting when upright.
Patients may become weak after lying in bed for several
days. Furthermore, some patients seem to require time
and physical and occupational therapy to change their
motor patterns before they experience improved function
with ITB therapy. In our experience, all of these factors
play a role in limiting the amount of functional change
seen during an ITB trial. For patients with goals of greater
comfort and ease of care, the changes are often seen within
days. For instance, positioning in a wheelchair is immedi-
ately easier if severe trunk dystonia is eliminated. The rela-
tion between an impairment, like dystonia, and disability
is not always linear. In fact, for some patients, reducing
dystonia results in decreased function because they rely on
dystonic posturing to keep them upright.
Another limitation of this study is the limited range of
levels of severity in our videotapes of patients. Most
patients had moderate to severe generalized dystonia,
limiting our assessment of reliability and validity of the
BAD Scale for milder dystonia. The level of severity in our
population may be due to the fact that many patients came
to our center specifically for treatment of dystonia and
were appropriate candidates for an invasive intervention.
Furthermore, to simplify this study, we did not use patients
with movement disorders other than dystonia, so we do
not know if our scale distinguishes between dystonia and
athetosis or spasticity.
Future studies to address the issue of differentiating
between movement disorders are important because the
effectiveness of treatments, such as ITB therapy, varies
with the type of movement disorder. ITB therapy is effec-
tive in the treatment of spasticity and dystonia (Albright
1996b), but not for athetosis nor ataxia. The development
of other clinical assessment tools would be helpful in eval-
uating patients with movements such as athetosis, chorea,
and ataxia. Assessing a milder level of dystonia is another
area of study, as well as increasing the data to support or
negate the need for training for reliable use of the scale.
Also, this study considered only videotaped assessments.

In the clinical assessment of patients, there may be greater
variability in the examination eliciting dystonia, and the
scale may be less reliable. The use of raters not familiar
with the patients may increase confidence in the
test–retest stability of the measure. In addition, a power
analysis to determine the appropriate sample size would
strengthen the statistical conclusions regarding the psy-
chometric properties. Perhaps most importantly, we need
to correlate the outcomes of tests, such as the Pediatric
Evaluation of Disability, with the BAD Scale outcomes and
determine concurrent validity in patients where functional
gains are the goal of intervention.
The BAD Scale is reliable for the assessment of dystonia
secondary to CP and TBI in the hands of experienced
users. It seems that if raters followed the scale’s rules, the
total score was reliable. Reliability for individual body
regions was not acceptable. Training improved reliability
and is recommended for clinicians planning to use the
scale. Preliminary analysis supports the responsiveness of
the scale: patients with a rating of better showed the great-
est improvement, those with ratings same and worse
showed less change in scale scores. This scale may be use-
ful as part of a multidimensional assessment, and further
studies of training, reliability, validity, and responsiveness
of the BAD Scale are warranted.
Accepted for publication 29th January 1999.
We thank the patients and raters who participated in this study.
Special thanks to Barb Shultz RN BSN and Jeff Campbell MD for
assistance in data collection. We also appreciate support for this
work provided by Grant No. 5M01R-R00084 from National
Institutes of Health, General Clinical Research Center, Bethesda,
MD and a grant from Medtronic Inc., Minneapolis, MN.
Barry–Albright Dystonia Scale
Patient’s Name: _____________________________ Date: _________
Directions: Assess the patient for dystonia in each of the following
regions: eyes, mouth, neck, trunk, each upper and lower extremity
(8 body regions). Write the scores on the lines provided. Rate severi-
ty based only on dystonia as evidenced by abnormal movements or
postures. When assessing functional limitations, do not score dysto-
nia based on other factors, such as weakness, lack of motor control,
cognitive deficits, primitive reflexes, and/or other movement disor-
ders as defined below.
Dystonia: Sustained muscle contractions causing twisting and
repetitive movements or abnormal postures
Spasticity: Velocity-dependent resistance to passive stretch
Athetosis: Distal writhing or contorting movements
Chorea: Brief, rapid, unsustained, irregular movements
Ataxia: Incoordination of movement characterized by wide-
based unsteady gait, flailing movements.
Eyes: signs of dystonia of the eyes include: prolonged eyelid spasms,
and/or forced eye deviations
0 – Absence of eye dystonia
1 – Slight. Dystonia less then 10% of the time and does not interfere
with tracking
2 – Mild. Frequent blinking without prolonged spasms of eye closure,
and/or eye movements less than 50% of the time.
3 – Moderate. Prolonged spasms of eyelid closure, but eyes open
most of the time, and/or eye movements more than 50% of the
time that interfere with tracking, but able to resume tracking
4 – Severe. Prolonged spasms of eyelid closure, with eyes closed at
least 30% of the time, and/or eye movements more than 50% of
the time that prevent tracking
* – Unable to assess eye movements
Eyes: _____
Mouth: signs of dystonia of the mouth include: grimacing, clenched
or deviated jaw, forced open mouth, and/or forceful tongue thrusting
0 – Absence of mouth dystonia
1 – Slight. Dystonia less than 10% of the time and does not interfere
with speech and/or feeding
2 – Mild. Dystonia less than 50% of the time and does not interfere
with speech and/or feeding
3 – Moderate. Dystonia more than 50% of the time, and/or dystonia
that interferes with speech and/or feeding
4 – Severe. Dystonia more than 50% of the time, and/or dystonia that
prevents speech and/or feeding
* – Unable to assess mouth movements
Mouth: _____
Neck: signs of dystonia of the neck include: pulling of the neck into
any plane of motion: extension, flexion, lateral flexion or rotation
0 – Absence of neck dystonia
1 – Slight. Pulling less than 10% of time and does not interfere with
lying, sitting, standing and/or walking
2 – Mild. Pulling less than 50% of the time and does not interfere
with lying, sitting, standing and/or walking
3 – Moderate. Pulling more than 50% of the time and/or dystonia
that interferes with lying, sitting, standing and/or walking
4 – Severe. Pulling more than 50% of the time and/or dystonia that
prevents sitting in standard wheelchair, standing and/or walking
410 Developmental Medicine & Child Neurology 1999, 41: 404–411

(e.g. requires more than standard head rest for seating)
* – Unable to assess neck movements
Neck: _____
Trunk: signs of dystonia of the trunk include: pulling of the trunk into
any plane of motion: extension, flexion, lateral flexion or rotation
0 – Absence of trunk dystonia
1 – Slight. Pulling less than 10% of the time and does not interfere
with lying, sitting, standing and/or walking
2 – Mild. Pulling less than 50% of the time and does not interfere
with lying, sitting, standing and/or walking
3 – Moderate. Pulling more than 50% of the time, and/or dystonia
that interferes with lying, sitting, standing and/or walking
4 – Severe. Pulling more than 50% of the time, and/or dystonia that
prevents positioning in standard wheelchair, standing and/or
walking (e.g. requires adapted seating system to control postur-
ing, such as ASIS bar)
* – Unable to assess trunk movements
Trunk: _____
Upper extremities: signs of dystonia of the upper extremities
include: sustained muscle contractions causing abnormal posturing
of the upper extremities
0 – Absence of upper extremity dystonia
1 – Slight. Dystonia less than 10% of the time and does not interfere
with normal positioning and/or functional activities
2 – Mild. Dystonia less than 50% of the time and does not interfere
with normal positioning and/or functional activities
3 – Moderate. Dystonia more than 50% of the time and/or dystonia
that interferes with normal positioning and/or upper extremity
4 – Severe. Dystonia more than 50% of the time and/or dystonia that
prevents normal positioning and/or upper extremity function
(e.g. arms restrained in wheelchair to prevent injury)
* – Unable to assess upper extremity movements
Left upper extremity: _____
Right upper extremity: _____
Lower extremities: signs of dystonia of the lower extremities
include: sustained muscle contractions causing abnormal posturing
of the lower extremities
0 – Absence of lower extremity dystonia
1 – Slight. Dystonia less than 10% of the time and does not interfere
with normal positioning and/or functional activities
2 – Mild. Dystonia less than 50% of the time and does not interfere
with normal positioning and/or functional activities
3 – Moderate dystonia more than 50% of the time and/or dystonia
that interferes with normal positioning and/or lower extremity
weight bearing or function
4 – Severe dystonia more than 50% of the time and/or dystonia that
prevents normal positioning and/or lower extremity weight bear
ing and/or function (e.g. cannot maintain standing due to severe
dystonia at ankles)
* – Unable to assess lower extremity movements
Left lower extremity: _____
Right lower extremity: _____
Total score: _____ Rater’s initials: _____
Albright AL. (1996a) Intrathecal baclofen in cerebral palsy
movement disorders. Journal of Child Neurology 11: (Suppl.)
— (1996b) Spasticity and movement disorders in cerebral palsy.
Journal of Child Neurology 11: (Suppl.) 1–4. (Editorial.)
— Barry MJ, Fasick P, Barron W, Shultz B. (1996) Continuous
intrathecal baclofen infusion for symptomatic generalized
dystonia. Neurosurgery 38: 934–8.
— — Painter MJ, Shultz B. (1998) Infusion of intrathecal baclofen
for generalized dystonia in cerebral palsy. Journal of
Neurosurgery 88: 3–76.
Burke RE, Fahn S, Marsden CD, Bressman SB, Moskowitz C,
Friedman J. (1985) Validity and reliability of a rating scale for the
primary torsion dystonias. Neurology 35: 73–7.
— — Gold AP. (1980) Delayed-onset dystonia in patients with
“static” encephalopathy. Journal of Neurology, Neurosurgery,
and Psychiatry 43: 789–97.
Fahn S, Marsden CD, Calne DB. (1987) Classification and
investigation of dystonia. In: Marsden CD, Fahn S, editors.
Movement Disorders 2. London: Butterworth. p 332–58.
Feldman AB, Haley SM, Coryell J. (1990) Concurrent and construct
validity of the pediatric evaluation of disability inventory.
Physical Therapy 70: 602–10.
Ferraz HB, Andrade LAF. (1992) Symptomatic dystonia: clinical
profile of 46 Brazilian patients. The Canadian Journal of
Neurological Science 19: 504–7.
Haley SM, Coster WJ, Ludlow LH, Haltiwanger JT, Andrellos PJ.
(1992) Pediatric Evaluation of Disability (PEDI) Development,
Standardization and Administration Manual. Boston: New
England Medical Center Hospitals Inc.
Krauss JK, Mohadjer M, Braus DF, Wakhloo AK, Nobbe F, Mundinger F.
(1992) Dystonia following head trauma: a report of nine patients
and a review of the literature. Movement Disorders
7: 63–272.
Kyllerman M, Bager B, Bensch J, Bille B, Olow I, Voss H. (1982)
Dyskinetic cerebral palsy. Acta Paediatric Scandanavia 71: 543–50.
Lahey MA, Downey RG, Saal FE. (1983) Intraclass correlations:
there’s more there than meets the eye. Psychological Bulletin
93: 586–95.
Law M, Baptiste S, Carswell A, McColl MA, Polatajko H, Pollock N.
(1994) Canadian Occupational Performance Measure. 2nd edn.
Ottawa: CAOT Publications.
Narayan RK, Loubser PG, Jankovic J, Donovan WH, Bontke CF.
(1991) Intrathecal baclofen for intractable axial dystonia.
Neurology 41: 1141–2.
Ottenbacher KJ, Hsu Y, Granger CV, Fiedler RC. (1996) The
reliability of the functional independence measure: a
quantitative review. Archives of Physical Medicine and
Rehabilitation 77: 1226–32.
Penn RD, Gianino JM, York MM. (1995) Intrathecal baclofen for
motor disorders. Movement Disorders 10: 675–7.
Portney LG, Watkins MP. (1993) Foundations of Clinical Research
Applications to Practice. Norwalk, CT: Appleton and Lange.
Russell D, Rosenbaum P, Cadman DT, Gowland C, Hardy S, Jarvis S.
(1989) The gross motor function measure: a means to evaluate
the effects of physical therapy. Developmental Medicine and
Child Neurology 31: 341–52.
— — Gowland C, Hardy S, Lane M, Plews N, McGavin H, Cadman D,
Jarvis S. (1993) Gross Motor Function Measure Manual. 2nd edn.
Hamilton, Ontario: McMaster University.
Saint Hilaire MH, Burke RE, Bressman SB, Brin MF, Fahn S. (1991)
Delayed-onset dystonia due to perinatal or early childhood
asphyxia. Neurology 41: 216–22.
Shrout PE, Fleiss JL. (1979) Intraclass correlations: uses in assessing
rater reliability. Psychology Bulletin 86: 420–8.
Stratford PW, Binkley JM, Riddle DL. (1996) Health status measures:
strategies and analytic methods for assessing change scores.
Physical Therapy 76: 1109–23.
Treves T, Korczyn AD. (1986) Progressive dystonia and paraparesis
in cerebral palsy. European Neurology 25: 148–53.
Uniform Data System. (1993) Guide for the Uniform Data Set for
Medical Rehabilitation (Adult FIM), Version 4.0. Buffalo, NY:
State University of New York at Buffalo.
Barry–Albright Dystonia Scale Margaret J Barry et al. 411