The Unified Parkinson's Disease Rating Scale (UPDRS): Status and ...

Movement Disorders
Vol. 18, No. 7, 2003, pp. 738 –750
© 2003 Movement Disorder Society
State of the Art Review
The Unified Parkinson’s Disease Rating Scale (UPDRS):
Status and Recommendations
Movement Disorder Society Task Force on Rating Scales for Parkinson’s Disease†*
Abstract: The Movement Disorder Society Task Force for
evant Incremental Difference, as well as testing its correlation
Rating Scales for Parkinson’s Disease prepared a critique of the
with the current UPDRS. If developed, the new scale should be
Unified Parkinson’s Disease Rating Scale (UPDRS). Strengths
culturally unbiased and be tested in different racial, gender, and
of the UPDRS include its wide utilization, its application across
age-groups. Future goals should include the definition of UP-
the clinical spectrum of PD, its nearly comprehensive coverage
DRS scores with confidence intervals that correlate with clin-
of motor symptoms, and its clinimetric properties, including
ically pertinent designations, “minimal,” “mild,” “moderate,”
reliability and validity. Weaknesses include several ambiguities
and “severe” PD. Whereas the presence of non-motor compo-
in the written text, inadequate instructions for raters, some
nents of PD can be identified with screening questions, a new
metric flaws, and the absence of screening questions on several
version of the UPDRS should include an official appendix that
important non-motor aspects of PD. The Task Force recom-
includes other, more detailed, and optionally used scales to
mends that the MDS sponsor the development of a new version
determine severity of these impairments. © 2003 Movement
of the UPDRS and encourage efforts to establish its clinimetric
Disorder Society
properties, especially addressing the need to define a Minimal
Key words: Unified Parkinson’s Disease Rating Scale
Clinically Relevant Difference and a Minimal Clinically Rel-
(UPDRS); Parkinson’s disease; rating scales; clinimetrics
The International Executive Council of the Movement
cerns the Unified Parkinson’s Disease Rating Scale
Disorder Society (MDS) has the authority and responsi-
[UPDRS].
bility to develop task forces that address special topics of
The UPDRS is a scale that was developed as an effort
Society interest. In 2001, based on the increasing reli-
to incorporate elements from existing scales to provide a
ance on standardized rating scales for Parkinson’s dis-
comprehensive but efficient and flexible means to mon-
ease (PD) and concerns that currently available scales
itor PD-related disability and impairment.1 Prior to its
may under-represent many elements of PD impairment
development, multiple scales, including the Webster,2
and disability, the society developed the Task Force for
Columbia,3 King’s College,4 Northwestern University
Rating Scales in PD. The mission of this Task Force is
Disability,5 New York University Parkinson’s Disease
three-fold: to critique existing scales, to identify clinical
Scale,6 and UCLA Rating Scales,7 were used in different
areas that are not adequately rated, and to make recom-
centers, making comparative assessments difficult. The
mendation on maintaining or modifying currently avail-
development of the UPDRS involved multiple trial ver-
able scales. Specifically, the mission does not include the
sions, and the final published scale is officially known as
enactment of any official changes in existing scales or
UPDRS version 3.0.1 The scale itself has four compo-
the development of new scales. The first critique con-
nents, largely derived from preexisting scales that were
reviewed and modified by a consortium of movement
disorders specialists (Part I, Mentation, Behavior and
†For a full list of contributors, see Appendix 1.
Mood; Part II, Activities of Daily Living; Part III, Motor;
*Correspondence to: Dr. Christopher G. Goetz, Rush-Presbyterian-
St. Luke’s Medical Center, 1725 W. Harrison Street, Chicago, IL
Part IV, Complications). The original concept of the
60612. E-mail: cgoetz@rush.edu
scale was to provide a core assessment tool that could be
Received 10 January 2003; Accepted 17 January 2003
accompanied by additional measures to focus on global
impairment or specific elements in more detail (Stanley
738

---

UPDRS: STATUS AND RECOMMENDATIONS
739
Fahn, personal communication). For example, whereas
other data sets (Appendix 3). All unpublished data that
the UPDRS is often accompanied by and reported with
were used for the critique were shared with the entire
such scales as the Schwab and England and Hoehn and
Task Force membership. An audio-taped meeting assem-
Yahr scales, these latter scales are not part of the UPDRS
bled all participants for open discussion of the UPDRS.
per se.8,9
A draft of the report was prepared by the Writing Com-
As part of the background work to develop the Task
mittee and Secretariat staff and circulated to the Expert
Force on Rating Scales for PD, the MDS secretariat staff
Consultants. The final document was assembled by the
(Caley Kleczka, Director) sent a questionnaire on UPDRS
chairperson and approved by all Task Force members
utilization patterns to all MDS members (see Appendix
before submission to the MDS International Executive
2). Of 1,593 (1,405 e-mails and 188 fax communica-
Committee. This Task Force document has been ap-
tions), 185 members from all continents responded
proved by the MDS Scientific Issues Committee prior to
(12%). Ninety-six percent of responders had personal
submission
for
journal
peer-review
in
Movement
experience with the UPDRS, 87% using it in clinical
Disorders.
trials, 70% in clinical practice, and 69% in other research
venues. The sub-components of the UPDRS were used at
RESULTS
different frequencies, 60% of responders regularly using
Part I (Mentation), 80% Part II (Activities of Daily
Utilization of the UPDRS
Living), 98% Part III (Motor), 65% Part IV (Complica-
One of the core advantages of the UPDRS is that it
tions). Combination scores based on sums of different
was developed as a compound scale to capture multiple
parts were often used, I II III (37%) and II III (41%).
aspects of PD. It assesses both motor disability (Part II:
The questionnaire results demonstrated that responding
Activities of daily living) and motor impairment (Part
MDS members use the UPDRS frequently and for mul-
III: Motor section). In addition, Part I addresses mental
tiple purposes, including research and clinical practice.
dysfunction and mood, and Part IV assesses treatment-
Parts II and III are most widely used for both clinical and
related motor and non-motor complications. Of all avail-
research purposes. These results confirmed the decision
able clinical scales for the assessment of parkinsonian
to prioritize the UPDRS as the first Task Force
motor impairment and disability, the UPDRS is currently
assessment.
the most commonly used.10 Sixty-nine percent of 1994 –
1998 articles using a PD-rating scale relied on the UP-
METHODS
DRS as the standard tool.11 This trend is an international
one, and the UPDRS predominates as the primary scale
Administrative Organization
in published studies from both US and other geograph-
Under the Chairperson’s direction, a Writing Commit-
ical regions. It has been applied with equal frequency in
tee was identified to draft the UPDRS critique and to
studies of early and late PD.11 The wide usage and global
remain on the Task Force in ongoing manner for future
acceptance of the scale has resulted in its use for numer-
scale assessments. In addition, a series of movement
ous multicenter studies. Further, the standardized ratings
disorder or statistical specialists with specific experience
allow for summary scores used to communicate global
using the UPDRS participated as Expert Consultants.
severity of impairment and disability.
These specialists were recruited to serve on the Task
Another unique feature of the UPDRS is the availabil-
Force for this critique only, with plans to rotate Expert
ity of a teaching-videotape standardising the practical
Consultants for each future scale critique. The third
application of the scale and thereby serving as an impor-
group was the administrative staff of the MDS secretar-
tant asset to enhance inter-rater reliability.12 This feature
iat, assigned the organization of the review process,
is particularly relevant to the training of new raters and to
integration of materials, and editorial review.
the conduct of multicenter therapeutic trials in PD.
Despite its multidimensional approach with 4 different
Critique Process
sections, the UPDRS has proven an easy-to-use instru-
Through Medline search and familiarity with the
ment in clinical practice with an average time require-
UPDRS literature, the chairperson supplied each Task
ment for administration of the full scale between 10 and
Force member with a series of articles related to the
20 minutes.13 This time can be further shortened by
UPDRS. Questions compiled by the Writing Committee
self-administration of the Mentation and ADL sections
were addressed to all Expert Consultants with the request
by patients in the waiting room. There is good inter-rater
to furnish a written document to respond to each point
reliability between patients who self-complete the histor-
with suitable references from the articles provided or
ical sections (Part I, Mentation, and Part II, ADL) of the
Movement Disorders, Vol. 18, No. 7, 2003

---

740
C.G. GOETZ ET AL.
UPDRS and the treating neurologist who interviews
and averages of subscale scores obtained during on and
them.14 When patients self-complete Parts I and II, the
off scores among fluctuators have been documented in
physician’s time investment can focus primarily on Parts
comparison with placebo.34 UPDRS improvements have
III and IV. When physicians are not available, the
been seen in patients with dose-finding studies new treat-
UPDRS has been effectively taught to other medical
ments of advanced disease as well as in studies focusing
personnel for such applications as community-based
on mildly disabled patients.28,35 Published reports using
neurological examinations.15
the UPDRS, however, have focused almost exclusively
The UPDRS is increasingly used as a gold standard
on Caucasians, and the UPDRS characteristics have not
reference scale. The motor section has repeatedly been
been extensively investigated in different racial or ethnic
employed in attempts to develop surrogate markers for
minorities.36 Furthermore, the effects of gender and age
disease progression like beta-CIT-SPECT or 18-F-Dopa-
on UPDRS ratings during treatment interventions have
PET.16,17 The UPDRS is also the common reference
not been specifically examined.
scale in studies of instrument development for rating
Insufficient information is available on the ability of
specific aspects of PD.18–20 US and European regulatory
the UPDRS to discriminate between disease categories
agencies rely on the scale for new drug approvals,21 and
of clinical pertinence. To date, operative definitions of
the UPDRS has also been used to define the placebo
“minimal,” “mild,” “moderate” and “severe” stages of
response in PD.22 Almost all recent trials of surgical
PD have not been explicitly defined. UPDRS scores,
interventions for PD, both related to intracerebral trans-
however, correlate with the Hoehn and Yahr scale and
plantation and deep brain surgery, have employed the
the Schwab and England scale.13 Furthermore, within the
UPDRS. It is a key component of the Core Assessment
UPDRS, the objective, physician-derived Motor section
Programs for Intracerebral Transplantation and Surgical
(Part III) correlates well with the subjective, patient-
Interventional Therapies for PD (CAPIT/CAPSIT).23,24
derived Activities of Daily Living (Part II) section.13 If a
Although specifically developed to assess PD, the UPDRS
measure such as the Schwab and England scale were
has been utilized to rate parkinsonian features of other
used to define “minimal,” “mild,” “moderate,” and “se-
conditions, including normal aging, progressive supranu-
vere” PD, UPDRS scores could be tested to see how
clear palsy, and Lewy body dementia.15,25,26
consistently they increase as the disease advances clini-
cally. With this analysis, UPDRS scores could be devel-
Scale Application Across the Spectrum of PD
oped to define numerically these clinical categories, ex-
The UPDRS has been used in studies of early, mild
pressed as ranges with 95% confidence intervals.
PD, moderate but stable PD, and severe disease and
motor fluctuations.27,28 Prior studies have demonstrated
Clinimetric Issues
that the scale favors the assessment of moderate and
Of all available PD rating scales, the UPDRS has the
severe impairments, and may not be ideally configured to
additional advantage that it is the most thoroughly tested
assess very mild disease-related signs and symptoms.29
instrument from a clinimetric point of view. Almost
Although the UPDRS has been extensively applied to
one-third of all studies assessing clinimetric properties of
clinical trials of early PD to test the concept of neuro-
impairment and disability scales for PD identified in a
protection or the impact of therapies on reducing the
recent systematic review were targeted on the UPDRS.10
need to start levodopa (L-dopa) therapy, “floor” effects
Clinimetric scale evaluation usually assesses a scale’s
potentially limit the scale’s utility in the early stages of
reliability and validity. Reliability evaluations assess the
the illness where impairment is subtle. To address this
amount of measurement error in a scale, while validity
issue, some studies have permitted the inclusion of 0.5
evaluations assess the degree to which a scale measures
ratings or other designations based on such anchors as
what it is purported to measure. Reliability and validity
“may be normal for healthy elderly subjects.”30 Modifi-
are not independent: a scale cannot be valid if it is not
cation of the UPDRS with such new wording or rating
reliable, but it can be reliable without being valid.
options, however, has not been validated.
Reliability can be divided into two major domains:
Several longitudinal studies of PD have demonstrated
internal consistency (the degree to which scale items
that the UPDRS increases over time and scores are
measure similar constructs) and rater consistency or sta-
higher at key clinical decision-making points like the
bility (the level of rating agreement across multiple raters
need to introduce symptomatic therapy.27,31–33 Numerous
or in a single rater across time). The UPDRS has shown
studies indicate that the UPDRS is responsive to thera-
excellent internal consistency across multiple stud-
peutic interventions. Significant improvements in total
ies13,37,38 and retains this consistency across stages of
UPDRS scores, individual subscales (Parts II and III),
disease severity as measured by the Hoehn and Yahr
Movement Disorders, Vol. 18, No. 7, 2003

---

UPDRS: STATUS AND RECOMMENDATIONS
741
staging system.38,39 This high degree of internal consis-
posed of six factors, axial/gait bradykinesia, right brady-
tency may be artificially inflated due to redundancy in
kinesia, left bradykinesia, rigidity, rest tremor, and
the large number of items in Parts II and III of the
postural tremor, has been shown to be stable across on
UPDRS.
and off stages.38,39 and to have a similar factor structure
Assessments of rater consistency have examined both
when used in other movement disorders.25
inter-rater reliability and intra-rater reliability. Inter-rater
Additional validity studies have been conducted to
reliability appears adequate for the total UPDRS13,18 as
assess the ability of the UPDRS to detect changes in
well as the Activities of Daily Living14 and the Motor
function in either untreated or treated states. In general,
Examination30,40 sections. There are two reports of un-
these studies have demonstrated that the UPDRS is sen-
acceptably low inter-rater reliability for selected items
sitive to change in clinical status.
assessing speech and facial expression on the Motor
Examination section of the UPDRS.30,40 Other studies,
Minimal Clinically Relevant Difference and
however, reported acceptable inter-rater reliability esti-
Minimal Clinically Relevant Incremental Difference
mates for these items.13,18 There are three published
Implicit to the strength and utility of a rating scale is
reports examining intra-rater reliability. One study used
the determination of increases or decreases that represent
the UPDRS and the other used a modified version of the
clinically relevant changes in the disease under consid-
scale applied to elderly community subjects without the
eration. Identifying the threshold or smallest difference
specific diagnosis of PD.15,40 Both of these studies
between two assessments that has an impact on disability
showed low to medium intra-rater reliability. Among 400
or handicap in a disease is known clinimetrically as the
early-stage PD subjects examined on two occasions,
Minimal Clinically Relevant Difference (MCRD). How-
separated by approximately 2 weeks, the intraclass cor-
ever simple and straightforward the MCRD may be con-
relation coefficients were very high: total score, 0.92;
ceptually, its operational definition is difficult to estab-
Mentation, 0.74; Activities of Daily Living, 0.85; Motor,
lish, and very few scales are associated with a well-
0.90.41
defined
MCRD.
Several
factors
complicate
the
Validity can be divided into three major domains: face
establishment of a MCRD for the UPDRS. First, PD
or content validity, criterion validity, and construct va-
signs vary throughout the day in parkinsonian patients
lidity. The UPDRS has adequate face validity and sam-
even without motor fluctuations. The natural moment-to-
ples important and typical domains associated with PD.
moment or visit-to-visit variation in the UPDRS among
In addition, its construction was guided by experts in the
patients considered to be stable in overall function has
field and based on previous scales. Criterion validity has
not been extensively studied.41 Second, because the four
not been established because there is no absolute “gold-
subscales of the UPDRS measure different aspects of PD
standard” that can be used for this assessment. The
and rely on physician-based and patient-based assess-
majority of validation studies have assessed the construct
ments, a single MCRD may not exist. Third, for a disease
validity of the UPDRS. These studies have generally
like PD, different MCRD values may apply at different
found satisfactory results regarding convergent validity
disease severities. Specifically, a smaller MCRD may
with other instruments assessing PD, such as the Hoehn
likely apply to groups of patients with mild disease,
and Yahr or Schwab and England scales or timed motor
whereas a larger differential value would be expected for
tests.20,38,39,42 Divergent validity, or the degree to which
groups with more severe illness. MCRD has been par-
the scale does not measure domains unrelated to PD, has
ticularly well studied for pain assessment scales and to a
not been well established. However, one study found a
lesser degree in assessment measures for asthma and
significant correlation between the UPDRS and measures
chronic obstructive pulmonary disease.43–46 To date, no
of mental status and depression.13 This finding may not
MCRD has been established for the UPDRS, but appro-
indicate poor divergent validity, but rather the associa-
priate studies based on methods extrapolated from the
tion of mental status and mood changes associated with
pain literature could be performed with the aim of estab-
PD.
lishing MCRD score ranges with confidence intervals.46
Multiple studies have examined construct validity of
The MCRD concept is applicable in two settings,
the UPDRS through factor analysis. These studies have
individual and group. At the individual level, though a
found between three and six factors that account for a
MCRD is not statistically enumerated, intervention de-
significant proportion of the total scale variance.13,18,37–39
cisions within each physician/patient relationship are
The resultant factors form rational groupings of the
guided by this concept of a minimal change from the
items, and suggest that the scale has a valid multidimen-
prior visit that warrants clinical recognition. For the
sional assessment format. One factor structure, com-
design of clinical trials involving groups of patients, a
Movement Disorders, Vol. 18, No. 7, 2003

---

742
C.G. GOETZ ET AL.
uniform MCRD definition is desirable both for analyses
randomised trial that compared L-dopa to pergolide for 3
of efficacy and futility. In some studies using the
years, an a priori stopping rule for established superiority
UPDRS, a 30% improvement in the Part III score has
of one treatment over the other was set as a between-
been applied to define “responders.”35 This empirically
group difference greater than 4 points in the Part III
determined figure is often used in clinical medicine,
UPDRS score at one year.52
based largely on the erroneous assumption that placebo
effects occur in 30% of patients, regardless of disease,
Ambiguities of the UPDRS
scale, study duration, or impairment under consider-
In the context of marked strengths and wide usage of
ation.47,48 The 30% UPDRS change from baseline used
the UPDRS, a number of limitations nonetheless exist.
in clinical studies has not been experimentally derived,
First, as a composite scale, the UPDRS is uneven in the
and furthermore, does not specifically presume to repre-
type of information it gathers. For example, Section I is
sent a minimal change of clinical significance.
conceptually different from parts II and III, and as a
One method to establish a MCRD would be to follow
screening assessment for the presence of depression,
patients with both sequential UPDRS scores and a global
dementia or psychosis, it cannot be used as an adequate
estimate of change, for example, the Clinician Interview-
severity measure of any of these behaviours. In cases of
based Impresssion of Change scale.49,50 This scale ranges
interventions, targeting such non-motor problems of PD,
from 1 (“very much improved”) to 4 (“no change”) to 7
specific additional scales are generally used.53–55 An
(“very much worse”) relative to either a prior visit or a
appendix to the UPDRS with a series of recommended
determined baseline. The key anchors, 3 (“minimally
improved”) or 2 (“moderately improved”) could poten-
scales for more detailed measurement of all screening
tially be examined relative to the corresponding UPDRS
questions would enhance consistency of data collection
scores to calculate a MCRD expressed as appropriate
among researchers. At present, such appendices do not
UPDRS ranges and confidence intervals. If such ratings
exist. Likewise, section IV is constructed differently than
are obtained from investigator and patient, these values
the rest of the UPDRS with a mixture of 5-point options
could be examined against the total UPDRS as well as
and dichotomous (yes/no) ratings that are difficult to
the specific scale sections.
analyse together. As such, though this portion is some-
Allied to the concept of MCRD is the Minimal Clin-
times used in clinical trials, most intervention studies for
ically Relevant Incremental Difference (MCRID). Rather
dyskinesias or motor-fluctuations currently rely primar-
than comparing two assessments within a patient or
ily on other scales. A number of additional dyskinesia
group (pre- vs. post-treatment), this terms refers to the
scales have been proposed to supplement the UPDRS56–58
difference between two groups at the end of a compara-
and most studies of patients with motor fluctuations have
ble period. In the case of a clinical trial, the MCRID
used self-scoring on-off diaries.34 Similarly dichotomous,
would determine the relevant expected difference at the
yes/no questions for the presence of gastrointestinal com-
end of a treatment between a placebo-group and the
plaints, orthostatic hypotension, or sleep problems (Part IV,
patients receiving the treatment in question. Knowing the
items 40 – 42) can only be used as screening items to assess
MCRID would allow clinicians to determine the thresh-
presence or absence of select clinical problems. The Task
old UPDRS value that would discriminate two treat-
Force members considered these items as insufficient to
ments. So far, there is no experimentally generated or
assess severity of impairment or disability related to non-
systematically analysed data on a MCRID for the total or
motor domains of PD.
sub-component scores of the UPDRS. There is, however,
Some items of the motor section have relatively poor
limited experience with this concept based on expert
inter-rater reliability, including speech, facial expression,
opinion or reliance on differences found in previous
posture, body bradykinesia, action tremor, and rigidi-
trials. In these cases, opinions or data widely accepted by
ty.13,30,59 Although the UPDRS teaching tape for Part III
the scientific community are used to determine estimates
provides visual anchors to improve inter-rater reliability,
of an end of treatment score associated with clinically
such tapes for the rest of the UDPRS have not been
important differences in patient function between two
developed.12 A specific example of a key testing problem
interventions and thereby to estimate a MCRID. Among
is the assessment of postural stability in Part III. Because
the few examples of an empirically used MCRID based
the response of the patient and the assigned rating de-
on experience and literature reviews, in one pallidotomy
pend directly on the force of the postural threat, stan-
trial that enrolled advanced PD patients with high pre-
dardized instructions and application of the test are es-
operative UPDRS motor scores, a MCRID for Part III of
sential for consistent ratings. These instructions are not
the UPDRS motor was established at 10 points.51 In a
part of the UPDRS.
Movement Disorders, Vol. 18, No. 7, 2003

---

UPDRS: STATUS AND RECOMMENDATIONS
743
Additionally, there is some redundancy of items in
and it has not been examined extensively in other ethnic
both the ADL and motor sections. While duplication of
or racial groups.36
material enhances the internal consistency of the scale,
some critics consider such enhancement a spurious in-
Comorbidities and the UPDRS
flation.13 Redundancy also increases the time required to
PD is more prevalent in subjects over age 50 years,
administer the scale. Efforts to reduce redundancy have
and the co-existence of other diseases like diabetes,
led to the Short Parkinson’s Evaluation Scale (SPES),
stroke, and arthritis can confound the evaluation of PD-
based directly on the UPDRS, but with fewer items and
related impairment and disability. Furthermore, short-
reduced rating options of 0 –3.18 Whereas the elimination
term disabilities resulting from a fracture or the exacer-
of redundancy and the enhancement of inter-rater reli-
bation of rheumatic disease may increase overall
ability are overall positive goals for scales, the shrinkage
disability without altering PD severity itself. Conversely,
of numerical options clinimetrically diminishes the ca-
correction of cataracts may improve overall patient func-
pacity to discriminate change. The majority of the Task
tion and facilitate the execution of activities of daily
Force considered 0 – 4 ratings to be preferred over 0 –3.
living without directly affecting PD. Finally, common
Allied to duplication of information is the concern that
co-existent disorders like depression can potentially af-
aspects of parkinsonian motor impairment in Part III are
fect the speed of a patient’s movement, alter motivation,
not necessarily reflective of the impact of each cardinal
and enhance perceptions of disability even when PD
feature on overall function. For example, bradykinesia-
itself is stable. The question of how the UPDRS should
related items are overrepresented in terms of the number
accommodate these various issues of co-morbidities is
not specifically detailed in the scale instructions. Two
of assessment items in comparison to tremor and postural
different prototypic paradigms could be used: the first, a
stability.
concerted attempt to disregard all components of impair-
The allocation of items to specific sections of the
ment or disability due to conditions unrelated to PD,
UPDRS is not altogether consistent, leading to potential
using the UPDRS as a scale of PD-related dysfunction in
ambiguity of interpretation. Part II, titled “Activities of
its purest sense; a second strategy involves a “rate-as-
Daily Living,” includes a mixture of items that directly
you-see” approach, using the UPDRS to describe a pa-
relate to daily activities (e.g., dressing, eating), but also
tient’s functional impairment regardless of the direct
examine patient perceptions of primary disease manifes-
relationship to PD. The first approach has the advantage
tations (e.g., tremor, salivation). Items that overlap these
of focusing on PD itself, but it is highly subject to
two categories include the gait items that assess primary
investigator and patient bias. For accurate assessment of
parkinsonian features (freezing, falls), but impact on
Part III, the rater would need to have a list of co-
walking as an activity of daily living. Renaming Part II
morbidities to maximize appropriate interpretation of
as “Historical” or “Patient Perceptions” would semanti-
signs. The second strategy deals with the reality of the
cally, but not conceptually, resolve this ambiguity. Items
patient status without interpretation of underlying cause
assessing function outside activities of daily living could
for impairment, but will likely inflate ratings that have
alternatively be reassigned to another section of the
minimal or no direct relationship to PD. Standardized
scale.
instructions for rating PD in the context of co-morbidi-
The UPDRS Part II is culturally biased, and the an-
ties do not exist in the current UPDRS, and the Task
choring descriptions for some item ratings are not appli-
Force members agreed that clarification of data handling
cable to all ethnic environments. For example, Dressing
for co-morbidities would be an important asset of a
(Item 10) describes difficulty with buttons, even though
future scale modification. The rate-as-you-see method
many traditional cultures do not use them; Cutting Food/
for Part III (Motor Examination) would likely reduce
Handling Utensils (Item 9) presumes that food is regu-
inter-rater variability, but the inclusion of “open fields”
larly cut for eating and that utensils are used, although
in the margins of the scale document for raters to note the
some cultures serve food in bite-size portions and some
contribution of other medical conditions would be
do not use eating utensils. Although the scale was con-
needed for full interpretation. Lang60 suggested open-
sidered applicable to most international urban settings,
field marginal denotations to indicate when dyskinesias
the UPDRS may be limited by ambiguities when applied
(D), excessive parkinsonian tremor (T), or apraxia (A)
in epidemiological research efforts that involve field
confound the execution of rated tasks. Similar denota-
work to rural and geographically isolated cultures. Even
tions could be included in the margins of the scale
within Western cultures, the UPDRS has been primarily
for confounding co-morbidities that cause weakness, or-
used in studying Caucasians with Parkinson’s disease,
thopedic, or non-parkinsonian coordination deficits.
Movement Disorders, Vol. 18, No. 7, 2003

---

744
C.G. GOETZ ET AL.
TABLE 1. Items not covered by the Unified Parkinson’s
impractical if the UPDRS is to remain a standard scale to
Disease Rating Scale
be applied in both clinical and research domains. The
concept that screening questions could cover these topics
Item
was favored, but an appendix should be added to the
Anhedonia
Bradyphrenia
UPDRS that would include officially recognized scales
Anxiety
to assess each of the screened areas in further depth.61
Hypersexuality
Another option would be the development of multiple
Sleep disorders (insomnia, excessive daytime sleepiness)
Fatigue
UPDRS versions of different lengths and different levels
Dysautonomia (urinary dysfunction, constipation,
of comprehensiveness, leaving the choice of scale to the
impotence, sweating)
physician (daily practice, in-depth evaluation, clinical
Dysregulation
Health-related quality of life
trials).62 Multiple UPDRS versions, however, would
have the potential to cause reporting ambiguity and un-
dermine the “unified” concept that anchors the UPDRS.
Whereas these denotations cannot be handled in a simple
The Task Force members favored the maintenance of a
statistical manner, they would be potentially useful in
single UPDRS that has sufficient screening questions to
clinical practice.
capture problems related to all aspects of PD with official
In contrast, Part II (Activities of Daily Living) has
appendices that recommend scales to assess each of the
instructions asking patients specifically to rate disability
screened areas in further depth when needed.
that they attribute to PD.1 The Task Force members
acknowledged that most individual patients are comfort-
CONCLUSIONS AND RECOMMENDATIONS
able with this introspective process, although patient-bias
The UPDRS is the most widely used clinical rating
may be unstable over time and subject to educational
scale for PD. The data cited in this critique highlight the
efforts and patient experience that clarify the contribu-
well-established and respected status of the UPDRS
tion of other illnesses to overall disability.
within the movement disorders, scientific, and regulatory
In clinical trials, the problems related to co-morbidi-
communities. The Task Force members unanimously
ties and their impact on UPDRS scores can be minimized
considered the concept of a single clinical rating scale to
by excluding patients with medical conditions that con-
be an important tool for clear and consistent communi-
found interpretation of the primary rating measure. Al-
cation among movement disorder colleagues. The
ternatively, some studies permit co-morbid conditions,
strengths of the UPDRS are many, and the scale provides
but only when they are stable and of long duration. If
a relatively comprehensive assessment of motoric as-
these conditions do not overly elevate the baseline scores
pects of PD. Extensive clinimetric analyses have already
and introduce concerns of “ceiling effects,” the UPDRS
been conducted on the UPDRS, providing it both scien-
can still measure change from the intervention under
tific and clinical credibility. Some items, however, have
study.
poor inter-rater reliability, and individual text anchors or
instructions for data acquisition are ambiguous. The
Important Elements Not Covered
UPDRS is less comprehensive in its assessment of non-
Several key elements of PD are not covered by the
motor features of the disease.
UPDRS. When the scale was formulated in the mid-
The Task Force members agreed that the identified
1980s, the developers were well aware of this limitation,
weaknesses were substantive, but amenable to correc-
but they made choices to delete questions on some par-
tion. They could be reduced or eliminated if the scale
kinsonian impairments, mainly to create a scale that was
was retained in its basic structure and core elements, but
reasonably simple and short (S. Fahn, personal commu-
modified and expanded to reflect the growing knowledge
nication). After more than a decade of scale utilization,
of PD-related impairments. Modifications should focus
however, the Task Force members considered that these
on clarity, resolution of ambiguity, and also provide
initial choices should be re-considered. In view of the
ratings for the very mild impairments seen in early PD.
anatomical, neurochemical, physiological, and concep-
These changes were considered important for optimising
tual evolution in thinking about PD, clinical neuroscien-
PD evaluation across the spectrum of clinical severities.
tists may currently need to have a scale that adequately
Expansion, however, should be limited to essential items
reflects the multifaceted elements of PD and assesses
that are currently not covered by the UPDRS, so that the
additional non-motor symptoms and signs that contribute
scale would retain its ready utility in both clinical and
to disability and quality of life. Several areas of concern
research settings. It was emphasized that any change in
exist (Table I), though changes in all items may be
the place of a given existing item, any modification in the
Movement Disorders, Vol. 18, No. 7, 2003

---

UPDRS: STATUS AND RECOMMENDATIONS
745
content or definition of a given item, or any addition of
questions identify the presence of deficits in a spe-
a new item would necessarily require new validation
cific function.
testing. Although a new official version of the UPDRS
j. Be tested by factor analysis to determine core im-
would compromise already operative longitudinal stud-
pairments that potentially could lead to a reduced
ies based on the current UPDRS version, statistical meth-
number of items that would serve to form a short
ods readily exist to compare the current version and a
version of the scale for everyday clinical practice.
new version and could be conducted for total, section,
3. The MDS should encourage efforts to conduct com-
and individual item scores to address this dilemma. If
prehensive clinimetric testing of the new scale to in-
developed, the new version could be tested in the context
clude not only standard assessments of reliability and
of two particular scale issues that have not been studied
validity, but also:
with the current UPDRS version: 1) the definition of
a. Analysis of the new scale’s correlation with the
UPDRS scores that define clinically pertinent categories,
current UPDRS.
“minimal,” “mild,” “moderate,” and “severe” PD; and 2)
b. Tests to establish UPDRS score ranges with confi-
the definition of UPDRS scores for a MCRD and MC-
dence intervals that define clinically pertinent cate-
RID. Based on these considerations, the Task Force on
gories, such as “minimal,” “mild,” “moderate,” and
Rating Scales in PD made the following recommenda-
“severe” PD.
tions:
c. Evaluations of MCRD and MCRID.
d. Assessments of the scale’s validity across race, gen-
1. The Movement Disorder Society should organize and
der, and age.
sponsor an official new version of the UPDRS.
4. If the International Executive Committee designates a
2. The new version should:
panel to develop a new version of the UPDRS, the
a. Retain the basic structure of the current UPDRS,
panel of experts selected for this project should in-
with sections that include both physician-derived
clude movement disorder specialists that represent ac-
and patient-derived data and retain at least 0–4
ademic and practice settings and statistical experts. All
ratings for motor items assessing severity of impair-
current members of the Task Force would be willing
ment or disability.
to serve on this panel. The development and testing of
b. Provide specific instructions for the acquisition of
proposed changes should include input from the entire
data on each item.
MDS membership.
c. Eliminate ambiguities in all text anchors.
5. Until a new UPDRS version is developed, clinical and
d. Cover the full spectrum of PD, especially providing
research efforts should retain primary reliance on the
ratings for mild parkinsonism.
UPDRS version 3.0, which is still considered an over-
e. Be responsive to concerns of cultural bias.
all strong and useful assessment tool for evaluating
f. Have specific instructions on a uniform manner of
impairment and disability in PD.
dealing with co-morbidities.
g. Better characterize and rate dyskinesias, capturing
APPENDIX 1
peak-dose and end of dose dyskinesia, both choreic
and dystonic forms.
Movement Disorder Society Task Force on Rating
h. Include additional screening questions on uncov-
Scales for Parkinson’s Disease
ered non-motor items that impact on the overall
Chairperson: Christopher G. Goetz, MD
function of PD subjects.
Writing Committee: Werner Poewe, MD, Olivier Rascol,
i. Incorporate an appendix that lists officially recom-
MD, Cristina Sampaio, MD, and Glenn T. Stebbins, PhD
mended additional scales applicable to each screen-
Expert Consultants: Stanley Fahn, MD, Anthony E. Lang,
ing question. These optionally used scales would
MD, Pablo Martinez-Martin, MD, PhD, Barbara Tilley, PhD,
assess severity of impairment related to non-motor
and Bob Van Hilten, MD, PhD
elements of PD and be applied when the screening
MDS Secretariat staff: Caley Kleczka, Lisa Seidl
Movement Disorders, Vol. 18, No. 7, 2003

---

746
C.G. GOETZ ET AL.
APPENDIX 2
Movement Disorders, Vol. 18, No. 7, 2003

---

UPDRS: STATUS AND RECOMMENDATIONS
747
APPENDIX 3
Movement Disorders, Vol. 18, No. 7, 2003

---