Scales to assess psychosis in Parkinson's disease: Critique and

Task Force Report

Scales to Assess Psychosis in Parkinson’s Disease:

Critique and Recommendations

Hubert H. Fernandez, MD,

* Dag Aarsland, MD,

Gilles Fe´nelon, MD,

Joseph H. Friedman, MD,

Laura Marsh, MD,

Alexander I. Tro¨ster, PhD,

Werner Poewe, MD,

Olivier Rascol, MD,

Cristina Sampaio, MD,

Glenn T. Stebbins, PhD,

and Christopher G. Goetz, MD

Department of Neurology, McKnight Brain Institute/University of Florida, Gainesville, Florida, USA

Stavanger University Hospital, Centre for Clinical Neuroscience Research, Stavanger, Norway

Department of Neurology, Hoˆpital Henri-Mondor, Cre´til, France

Department of Clinical Neuroscience, Brown University Medical School, Providence, Rhode Island, USA

Department of Psychiatry, Johns Hopkins University, Baltimore, Maryland, USA

Department of Neurology, University of North Carolina, Chapel Hill, North Carolina, USA

Department of Neurology, Medical University of Innsbruck, Innsbruck, Austria

Laboratoire de Pharmacologie Medicale Et Clinique, Faculte de Medicine, Toulouse, France

Laboratory of Clinical Pharmacology and Therapeutics, Faculdade de Medicina de Lisboa, Lisboa, Portugal

Department of Neurological Sciences, Rush University Medical Center, Chicago, Illinois, USA

Abstract: Psychotic symptoms are a frequent occurrence in

Parkinson’s disease (PD), affecting up to 50% of patients. The

Movement Disorder Society established a Task Force on Rat-

ing Scales in PD, and this critique applies to published, peer-

reviewed rating psychosis scales used in PD psychosis studies.

Twelve psychosis scales/questionnaires were reviewed. None

of the reviewed scales adequately captured the entire phenom-

enology of PD psychosis. While the Task Force has labeled

some scales as “recommended” or “suggested” based on the

fulﬁlling-deﬁned criteria, none of the current scales contained

all the basic content, mechanistic and psychometric properties

needed to capture PD psychotic phenomena and to measure

clinical response over time. Different scales may be better for

some settings versus others. Since one scale may not be able to

serve all needs, a scale used to measure clinical response and

change over time [such as the Clinical Global Impression Scale

(CGIS)] may need to be combined with another scale better at

cataloging speciﬁc features [such as the Neuropsychiatric Inven-

tory (NPI) or Schedule for Assessment of Positive Symptoms

(SAPS)]. At the present time, for clinical trials on PD psychosis

assessing new treatments, the following are recommended primary

outcome scales: NPI (for the cognitively impaired PD population

or when a caregiver is required), SAPS, Positive and Negative

Syndrome Scale (PANSS), or Brief Psychiatric Rating Scale

(BPRS) (for the cognitively intact PD population or when the

patient is the sole informant). The CGIS is suggested as a second-

ary outcome scale to measure change and response to treatment

Key words: Parkinson; psychosis; scales; hallucination;

delusion.

Psychotic symptoms are a frequent occurrence in

Parkinson’s disease (PD), affecting up to 50% of pa-

tients.

1,2

Studies on psychosis have mostly focused on

visual hallucinations, the most common type of psy-

chotic symptom in PD.

3-6

However, hallucinations can

occur in all sensory domains and delusions of various

types are also relatively common.

4,7-9

Table 1 lists the

most relevant terms used in this review with their

deﬁnitions.

Over the course of PD, psychotic symptoms, once

present, tend to be persistent and progressive.

10-12

The

impact of psychosis is substantial in that it is associated

with dementia, depression, earlier mortality, greater care-

giver strain, and nursing home placement.

12-16

This article contains supplementary material available via the Inter-

net at http://www.interscience.wiley.com/jpages/0885-3185/suppmat.

*Correspondence to: Hubert Fernandez, Department of Neurology,

McKnight Brain Institute/University of Florida, 100 S. Newell Drive,

PO Box 100236, Gainesville, FL 32611.

E-mail: [email protected]ﬂ.edu

Received 3 May 2007; Revised 14 August 2007; Accepted 28

October 2007

Published online 3 January 2008 in Wiley InterScience (www.

interscience.wiley.com). DOI: 10.1002/mds.21875

Movement Disorders

Vol. 23, No. 4, 2008, pp. 484 –500

484

Until recently, there have been no standardized cri-

teria speciﬁcally designed to diagnose PD-related psy-

chosis. The Diagnostic and Statistical Manual of Men-

tal Disorders, Fourth Edition (DSM-IV-R)

and the

Structured Clinical Interview for DSM-IV-TR Axis I

disorders (SCID)

have been used, but they rely on

general categories like “psychotic disorder due to a

general medical condition” or “substance-induced

psychotic disorder.”

A NIH-sponsored workshop recently reviewed the PD

psychosis literature to provide criteria that distinguish

PD psychosis from other causes of psychosis.

On the

basis of these data, provisional criteria for PD psychosis

in the style of the Diagnostic and Statistical Manual of

Mental Disorders IV-R were proposed (see Table 2). The

criteria are inclusive and contain descriptions of the full

range of characteristic symptoms, chronology of onset,

duration of symptoms, exclusionary diagnoses, and as-

sociated features, such as dementia. They describe a

distinctive constellation of clinical features that are not

shared by other psychotic syndromes. These criteria re-

quire validation and perhaps reﬁnement, but form a use-

ful starting point for studies on PD psychosis.

Although the NIH-sponsored workshop focused on

diagnostic issues, this report concerns a critical evalua-

tion of scales that rate the severity of PD-related psycho-

sis. The Movement Disorder Society (MDS) has estab-

lished a Task Force on Rating Scales in PD, and this

critique applies the mission of this Task Force to rating

scales that address psychotic phenomena. Speciﬁcally,

the aims of this report were to (1) conduct a survey of

MDS members regarding their use of psychosis rating

scales, (2) review and critique available psychotic symp-

tom rating scales used in PD, and (3) make formal

recommendations to the society regarding the utility of

scales for research and clinical practice. The Task Force

charge did not include the design of a new scale, al-

though the recommendations could identify this need as

a possible conclusion.

PATIENTS AND METHODS

MDS Members Survey of Psychotic Symptom

Scale Use

As a background work to develop the critique and

recommendations for rating scales used in PD psychosis,

TABLE 1. Terms and deﬁnitions used in this critique

Term Deﬁnition

Psychosis A global term to encompass hallucinations, delusions and the “minor” phenomena of

illusions, “passage hallucinations” and “sense of presence”

Hallucinations Abnormal perceptions without a physical stimulus that can involve any sensory

modality and may be simple or complex in form.

Illusions Misperceptions of real stimuli that are often visual in nature

Delusions False, ﬁxed, idiosyncratic beliefs that are maintained despite evidence to the contrary

Sense of presence Experience that someone is present when nobody is actually there

Passage hallucinations Fleeting, vague imaging in the peripheral vision

TABLE 2. Proposed diagnostic criteria for PD associated psychosis

A. Characteristic symptoms [Presence of at least one of the following symptoms (specify which of the symptoms fulﬁll the criteria)]

Illusions

False sense of presence

Hallucinations

Delusions

B. Primary diagnosis

UK brain bank criteria for PD

C. Chronology of the onset of symptoms of psychosis

The symptoms in criterion A occur after the onset of PD

D. Duration

The symptom(s) in criterion A are recurrent or continuous for 1 mo

E. Exclusion of other causes

The symptoms in criterion A are not better accounted for by another cause of Parkinsonism such as dementia with Lewy bodies,

psychiatric disorders such as schizophrenia, schizoaffective disorder, delusional disorder, or mood disorder with psychotic

features, or a general medical condition including delirium

F. Associated features (Specify if associated)

With/without insight

With/without dementia

With/without treatment for PD (specify drug, surgical, other)

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 485

Movement Disorders, Vol. 23, No. 4, 2008

the Writing Committee reviewed and adapted the tem-

plate used by the MDS Rating Scale Task Force to

survey MDS members on their use of speciﬁc psychosis

rating scales in the PD population (see Appendix A,

available online at http://www.interscience.wiley.com/

jpages/0885-3185/suppmat).

Critique Process

PubMed searches (from 1950 to September 2005)

were conducted using Parkinson’s disease and psychosis/

hallucinations/delusions in combination with each of the

following terms: clinical course, functional outcome,

clinical features, antipsychotic drugs, neuroleptics, diag-

nosis, diagnostic criteria, rating scales, and clinical trials.

The resulting articles were then screened by the Chair of

the Writing Committee (HHF) before distribution to the

Task Force members to determine that they dealt specif-

ically with PD and that they were original contributions.

While several other psychosis scales exist, this critique

was limited to the scales used in published, peer-re-

viewed PD psychosis studies. Each of the Committee

members reviewed the primary reference for the psycho-

sis scale along with all articles in which the scales were

used in patients with PD. The review began with a

“primer on psychometric issues” given by the commit-

tee’s psychologist (AIT). Appendix B (supplementary

material) brieﬂy deﬁnes the clinimetric terms used in

evaluating the properties

21-23

of each scale. Each scale’s

strengths, weaknesses, and psychometric properties were

then determined. The features were then summarized,

and speciﬁc recommendations concerning the recom-

mended use of each scale in PD were made according to

the following deﬁnitions: (1) “recommended”: a scale

that has been applied to PD populations; there are data on

its use in clinical studies beyond the group that devel-

oped the scale; and, it has been studied clinimetrically

and considered valid, reliable, and sensitive to the given

behavior being assessed. Ideally this latter criterion is

met for PD psychosis speciﬁcally, but can be met if

strong clinimetric results are available for hallucinations

and psychosis in other contexts. (2) “suggested”: the

scale has been applied to PD populations, but only one of

the other criteria is fulﬁlled; (3) “listed”: the scale has

been applied to PD populations, but neither of the other

criteria is fulﬁlled. These designations are also being

used to develop the Appendix of ancillary scales to

complement the Movement Disorder Society-sponsored

revision of the Uniﬁed Parkinson Disease Rating Scale

(MDS-UPDRS) still in development.

The terms recom-

mended, suggested, and listed are designations based on

the evidence criteria listed above, but do not represent an

ofﬁcial stand by the MDS.

RESULTS

MDS Members Survey Results

Of over 2000 surveys sent to the entire MDS mem-

bership, 58 responses were received and tabulated. The

response rate was considered too meager to draw deﬁn-

itive conclusions on membership utilization of PD psy-

chosis scales. Nonetheless, 91% (N ⫽ 52) of the respon-

dents felt that there was a need for a psychosis rating

scale for PD. Seventy-two percent (N ⫽ 41) have used a

formal psychosis rating scale in PD, the majority of them

in clinical practice and clinical trials. Of the rating scales,

the Neuropsychiatric Inventory (NPI) was the most com-

monly used (N ⫽ 23), followed by the Brief Psychiatric

Rating Scale (BPRS) (N ⫽ 16), Parkinson Psychosis

Rating Scales (PPRS) (N ⫽ 8), Positive and Negative

Symptom Scale (PANSS), and Scale for the Assessment

of Positive Symptoms (SAPS) (N ⫽ 5 respondents each).

The majority of the respondents felt that none of the exist-

ing scales were ideal in clinical practice, but that some may

be useful in clinical trials or PD psychosis research.

If a new scale for PD were to be created, the majority

of respondents felt that it was “important” or “very

important” that the scale should take less than 10 minutes

to administer (86%), that a reliable caregiver/observer

also contributes to the history/information obtained

(86%), that the scale also captures illusions (67%) and

“sense of presence ” hallucinations (64%), that it was

able to differentiate drug-induced psychosis from delir-

ium (77%), that a nonphysician was able to administer

the scale (80%), and that it reﬂected the severity of

psychosis over the last 7 to 14 days (92%).

Psychosis Scales Used in Parkinson’s Disease

Each scale fulﬁlling the Recommended or Suggested

criteria is discussed in the printed report with expanded

materials available as supplementary materials on the

journal website. For Listed scales that did not meet the

higher criteria ratings, only a brief text is provided, and

readers are referred to the supplementary materials on

the journal website for their full discussion.

Table 3 summarizes characteristics of each scale and

Table 4 shows how each scale fulﬁlled the criteria for use

recommendations as a scale to assess PD psychosis.

Psychosis Scales Speciﬁc for Parkinson’s Disease.

Parkinson Psychosis Rating Scale. This scale is

designed to “rate the content, quality, severity, and

frequency of six domains (of psychotic phenomenol-

ogy in PD), and their functional impact based on

family report.”

The six domains are as follows:

visual hallucinations, illusions/misidentiﬁcation, para-

486 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

noid ideation, sleep disturbance, confusion, and sexual

preoccupation.

The original publication was based on 29 generally

elderly and demented PD patients with psychosis. The

six items are scored on a four-point scale (1 ⫽ absent to

4 ⫽ severe symptoms), anchor points are provided, with

total scoring guidelines: mild ⫽ 8 to 12; moderate ⫽ 13

to 18, severe ⫽ 19 to 24. An additional item with the

same scoring system rates the overall functional impact

of psychosis based on family report.

TABLE 3. Summary of the general properties of psychosis scales/inventories reviewed in this article

Scale

Time required to

administer the

scale (in min)

Are the items asked

in a structured

manner? (Y/N)

No special training

required to

administer the

scale. (Y/N)

Has a validation

study been

reported in PD?

(Y/N)

Does the scale look

into the full

spectrum of PD

psychosis? (Y/N)

Parkinson psychosis rating

scale 5–15 N Y Y N

Parkinson psychosis

questionnaire 5–15 Y Y Y N

Rush hallucination

inventory ⬎30 Y Y N N

Baylor hallucination

questionnaire 5–15 Y Y N N

Neuropsychiatric

inventory 15–30 Y N N N

Behavioral pathology in

Alzheimer’s disease

rating scale 15–30 Y Y N N

Brief psychiatric rating

scale 15–30 N N N N

Positive and negative

syndrome scale ⬎30 N N N N

Schedule for assessment

of positive symptoms ⬎30 Y Y N N

Nurses’ observation scale

for inpatient evaluation 5–15 Y Y N N

Clinical global impression

scale ⬍5N YN N

Uniﬁed Parkinson disease

rating scale Part I ⬍5N YY N

TABLE 4. Summary of “use recommendations” of psychosis scales used in PD

Psychosis scale

Applied

in PD

Used in studies

beyond original article

Satisfactory clinimetric

assessment

Scale

designation*

Parkinson psychosis rating scale ⻫⻫Suggested

Parkinson psychosis questionnaire ⻫⻫Suggested

Rush hallucination inventory ⻫ Listed

Baylor hallucination questionnaire ⻫ Listed

Neuropsychiatric inventory ⻫⻫⻫Recommended

Behavioral pathology in Alzheimer’s disease

rating scale ⻫⻫ Suggested

Brief psychiatric rating scale ⻫⻫⻫Recommended

Positive and negative syndrome scale ⻫⻫⻫Recommended

Schedule for assessment of positive symptoms ⻫⻫⻫Recommended

Nurses’ observation scale for inpatient

evaluation ⻫ Listed

Clinical global impression scale ⻫⻫ Suggested

Uniﬁed Parkinson disease rating scale Part I ⻫ Listed

*These designations are based on the Appendix of ancillary scales to complement the Movement Disorder Society-sponsored revision of the Uniﬁed

Parkinson Disease Rating Scale (MDS-UPDRS) still in development.

The deﬁnitions of these designations are as follows: “recommended”: a scale

that has been applied to PD populations; there are data on its use in clinical studies beyond the group that developed the scale; and, it has been studied

clinimetrically and considered valid, reliable, and sensitive to the given behavior being assessed; “suggested”: the scale has been applied to PD

populations, but only one of the other criteria is fulﬁlled; “listed”: the scale has been applied to PD populations, but neither of the other criteria is

fulﬁlled.

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 487

Movement Disorders, Vol. 23, No. 4, 2008

In the original report, little information is given re-

garding administration. It appears to be based on an

interview by an experienced clinician. The authors de-

scribe the scale as “easily administered,” but the time

required for administration was not given.

Strengths: The PPRS was speciﬁcally designed to as-

sess psychotic symptoms in patients with PD. It is short

(six items, plus a global assessment item) and allows

measurement of change over time. The visual hallucina-

tions item takes into account the frequency of hallucina-

tory events and insight, two important characteristics of

hallucinations in PD.

Weaknesses: The PPRS fails to capture the heteroge-

neity of psychosis in PD, and the single items each for

hallucinations and delusions provide a narrow range of

scores for tracking clinical change. Further, the symp-

toms that accompany psychosis, include “confusion,”

“sexual preoccupancy,” and “sleep disturbances,” which

are not speciﬁcally felt to be part of the speciﬁc syn-

drome of psychosis. With only three items devoted to

psychosis and three other items devoted to the associated

features, the ﬁnal score risks being burdened with non-

psychotic confounds. Finally, the anchors have multiple

features collapsed together (i.e., “3 ⫽ frequent; absence

of full insight; can be convinced” leaving no options for

the frequent hallucinations with full insight retained).

Psychometric properties: Validity and reliability eval-

uation is limited to the original report on the scale.

Given the small sample size (n ⫽ 29), those data must be

viewed as preliminary. Further, responsiveness of the

PPRS to active or passive intervention in a longitudinal

setting has not been fully tested. Interrater reliability for

individual items and total score was good to very good

(␳⫽0.80 –0.99) (though it is not speciﬁed why this type

of correlation coefﬁcient was employed). Internal con-

sistency of items across three raters ranged from 0.31 to

0.80. Values for hallucination and paranoid ideation

items are fair (0.64 –0.75); others are quite weak. Test–

retest reliability is described by the authors as “high” for

6 weeks, but coefﬁcients ranged from quite poor to fair

(0.06 –0.70). Concurrent validity examining correlation

of scores with BPRS (presumably using total scores) was

high (0.92); however, the relationship was weak with

Nurses’ Observation Scale for Inpatient Evaluation

(NOSIE) Psychotic dimension score (0.48). The instru-

ment appears sensitive to treatment (i.e., on ondanse-

tron).

Final Assessment: The PPRS fulﬁlls criteria as a sug-

gested scale for rating PD psychosis.

Parkinson Psychosis Questionnaire. The Parkinson

Psychosis Questionnaire (PPQ)

was developed as a

screening instrument for early recognition of psychosis

in PD. The scale was reviewed by clinicians regarding its

appropriateness and underwent further analysis of its

internal validity by studying test results in 50 patients

with parkinsonism relative to results on the BPRS and

the Diagnostic and Statistical Manual of Mental Disor-

ders, Fourth Edition (DSM-IV).

The scale includes screening probes followed by de-

tailed questions regarding the presence or absence of

sleep disturbance, hallucinations/illusions, delusions, and

orientation. Any positive answers within a domain trig-

ger inquiries about frequency and severity. A subscore is

the product of the frequency multiplied by the severity

score for that symptom category. The total score is the

sum of the subscores.

Strengths: The PPQ is one of a few scales developed

speciﬁcally for PD and provides detailed anchors and

guidelines for rating items. It provides a mechanism for

cataloging the presence or absence of most discrete psy-

chotic phenomena. Unlike the PPRS, it scores both se-

verity and frequency, and two of the four categories are

devoted speciﬁcally to hallucinations/illusions and to

delusions.

Weaknesses: The PPQ does not include all hallucina-

tory phenomena (olfactory, tactile, kinesthetic, and sense

of presence). Insight into hallucinations is not taken into

consideration. Some forms of delusions, including so-

matic delusions, are not assessed, and there is no mech-

anism for cataloging psychotic phenomena outside the

speciﬁc questions. Delusions (ﬁve items) are more de-

tailed than hallucinations (four items), although they are

less common in PD psychosis. One item, delusional

ideas of control or inﬂuence on actions and thoughts, is

rare in PD. The sleep and orientation items reduce the

speciﬁcity of the scale.

Psychometric properties: There is limited psychomet-

ric evidence based on the original study of 50 PD pa-

tients. The scale has not been tested extensively, and has

not been subject to evaluation in clinical trials or larger

samples, and its sensitivity to change has not been tested.

Internal consistency (of presumably the total scale)

was moderate–good (␣ 0.68). Interrater reliability is

unknown. Divergent validity with UPDRS Part III (␳⫽

⫺0.13) and MMSE (␳⫽⫺0.27) is good. The scale has

excellent agreement with SCID (100% sensitivity, 92%

speciﬁcity). There is a signiﬁcant, but unspeciﬁed corre-

lation with the BPRS. The use of additive versus multi-

plicative scores is not justiﬁed adequately.

Final Assessment: The PPQ fulﬁlls criteria as a sug-

gested scale for rating PD psychosis.

Rush Hallucination Inventory. This 53-item inven-

tory allows four answers to most questions, but also

includes informational questions (how long to fall asleep,

488 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

what is bedtime?) and yes/no answers to others.

The

questions with four answers allow for frequency and

duration ratings. These questions are not given weights.

Illusions and hallucinations as well as emotional colora-

tion (fear) are considered, and visual, auditory, tactile,

and olfactory domains are separately assessed.

The ﬁrst set of questions addresses sleep disorders.

There are 11 in number including informational ques-

tions such as quality of sleep, frequency questions on

vivid dreams, nightmares, excessive daytime sedation,

and use of sedative medications.

The second section is devoted to visual illusions and

consists of seven questions, the ﬁrst being a screening

question, asking if illusions occur. If not, one skips to the

next section. The third section pertains to auditory illu-

sions and has a maximum frequency of three times per

week. The six further questions are the same as for the

visual illusions. The fourth section is on hallucinations,

with an initial focus on visual hallucinations. The initial

screening question deﬁnes maximal frequency as three

times per week. This section asks how long the halluci-

nation persists, what time of day and environmental

circumstances, and whether they are frightening. Similar

questions are then asked about auditory hallucinations.

Then, a section on presence hallucinations, phrased

somewhat ambiguously, asks, “During the past month,

have you had the experience of feeling something or

someone out of nowhere. . . that is, have you had a

sensation when nothing was there?” The last section asks

about olfactory hallucinations.

Strengths: This scale speciﬁcally outlines that it covers

the past 1 month. It comprehensively probes into the

major and minor forms of hallucinations and includes

detailed questioning of associated sleep disturbances.

The questions take into account the frequency and dura-

tion of each type of hallucination.

Weaknesses: This inventory assumes that severity is

based on frequency and negative emotional association.

There is no overall rating for any section. Further, al-

though written as a questionnaire, there are no speciﬁc

instructions. Delusions are not included in the inventory.

None of the information is solicited from an observer.

Contentwise, there is no mechanism for applying the

scale to patients with dementia.

In the visual illusions section, there is no question on

how clearly the vision is seen, how persistent or anything

about the content other than whether it is frightening.

The most severe rating for visual illusion frequency is

three times per week (which may not be unlikely for a

normal person with impaired vision). There are questions

of interest, such as, on what situations (day, night, dark,

light) are most likely to stimulate visual illusions, but not

useful for either treatment studies or diagnosis. There is

no question on visual acuity. Similarly, in the auditory

illusions section, no questions on hearing impairment are

asked. Duration of the illusion is not ascertained nor are

any questions on content asked (e.g., are the sounds

isolated such as a bark or a name being called, or is it

sustained music, conversations, clearly heard, heard at a

distance, etc). For olfactory hallucinations, there is no

question on smell or taste impairment and no question on

taste. The lack of description of the hallucinated content

or whether there is an emotional reaction (other than

fear) is another drawback.

Psychometric properties: The inventory has unknown

psychometric properties. The correlation between the

scale and MMSE (measuring cognitive state) and the

UPDRS motor score at baseline is difﬁcult to interpret.

That is, is the Rush Hallucinations Inventory Score sim-

ply a proxy for disease severity? Or, does it indicate, as

one might expect, that persons with more marked motor

problems and cognitive compromise are also more likely

to have hallucinations?

Final Assessment: The Rush Hallucination Inventory

fulﬁlls criteria as a listed scale for rating PD psychosis.

Baylor Hallucinations Questionnaire. This six-

item, four-point scale questionnaire assesses only hallu-

cinations

: (1) visual hallucinations, (2) auditory hallu-

cinations, (3) presence hallucinations, (4) insight: “can

you tell that the hallucinations are not real?,” (5) “do you

attempt to communicate with the hallucinations?,” and

(6) “how upset is your family by the hallucinations?.”

The answers to the questions 1 to 3 are: 0 ⫽ I do not have

this problem; 1 ⫽ rare; 2 ⫽ occasionally (about once/

week); 3 ⫽ frequently (more than three times per week);

4 ⫽ all the time (more than once each day).

Strengths: The strengths of this scale are its ease and

rapidity of administration, characterization of the distress

caused to the family, and delineation of symptom fre-

quency. It focuses speciﬁcally on hallucinations in PD,

and is anchored in frequency and insight, two clinically

important dimensions of PD hallucinations.

Weaknesses: The frequency anchors are not clear and

the equivalence of daily and “all the time” is likely to be

a problem for many patients and caregivers. The spec-

trum used to rate symptom frequency seems dispropor-

tionate. For example, patients or raters may not view

hallucinations experienced more than once a day as oc-

curring all the time if they are very ﬂeeting. A person

with frequent presence hallucinations may have them

three or four times daily without being terribly bothered.

Despite focusing on hallucinations only, not all halluci-

nations are included. It does not assess tactile hallucina-

tions or other rare modalities such as olfactory halluci-

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 489

Movement Disorders, Vol. 23, No. 4, 2008

nations. Delusions are not taken into account. The last

item concerning the reactions of “the family” is interest-

ing but debatable as reactions of the family may not

reﬂect the severity of the patient’s hallucinations.

Psychometric properties: The psychometric properties

are unknown. The score reportedly correlates with global

impression but the correlation coefﬁcient was not re-

ported.

The scale has been used in one small study

only.

Final Assessment: The Baylor Hallucinations Ques-

tionnaire fulﬁlls criteria as a listed scale for rating PD

psychosis.

Scales Developed to Assess Psychosis in Patients

with Dementia.

Neuropsychiatric Inventory. The NPI is a 12-item

scale developed primarily for the assessment of psycho-

pathology in patients with dementia.

While the NPI has

also been used in studies of neuropsychiatric distur-

bances with nondemented patients, its standard adminis-

tration assumes that the subject has dementia and that the

interview is conducted by trained rater with a knowl-

edgeable caregiver. There are now other validated and

widely used versions: the nursing home NPI and the

questionnaire version (NPIQ), but informant report re-

mains the source of information about the patient.

The twelve items covered by the NPI are delusions,

hallucinations, agitation/aggression, depression/dyspho-

ria, anxiety/elation/euphoria, apathy/indifference, disin-

hibition, irritability, aberrant motor behavior, nighttime

behaviors, and appetite/eating behaviors. To facilitate

detection of behavioral changes in these domains and

minimize administration time, the NPI uses screening

questions about symptoms and behaviors for each of the

12 items. More speciﬁc questions are asked only if the

screening probe is positively endorsed. For each positive

screening probe, the severity and frequency of the related

symptoms are rated and a product of these two ratings is

multiplied to obtain the score for each domain. A sepa-

rate rating is attached to the “distress caused to the

caregiver,” or the “occupational disruption” at the nurs-

ing home. The NPI is copyrighted.

Strengths: The NPI has a number of strengths. Admin-

istration of this scale is relatively efﬁcient with screening

probes that capture delusions and hallucinations as well

as the range of most psychiatric symptoms in patients

with PD. The structured interview questions potentially

enable administration of the NPI by less clinically expe-

rienced professionals without reducing scale validity or

reliability. Open-ended questions for each item also al-

low recording of behaviors not listed for a particular

domain. Separating symptom frequency from symptom

severity provides a means to track the frequency, inci-

dence, prevalence, and the dynamics of various psychi-

atric phenomena over time. As such, the NPI allows the

rater to capture mild but very frequent phenomena or

moderate but less frequent phenomena. The scale also

provides some questions to characterize speciﬁc psy-

chotic phenomena. Ratings of other symptoms, e.g., ag-

itation and anxiety allow characterization of additional

psychiatric phenomena that may occur with psychosis

and improve when the psychotic symptoms improve.

Weaknesses: While the NPI is applicable to a range of

neuropsychiatric conditions, its development as an in-

strument to evaluate patients with dementia potentially

limits its application in PD patients who are not de-

mented. Accordingly, if the NPI is to be used in clinical

studies of PD patients, the scale needs to be modiﬁed so

that informant- and patient-derived information is ob-

tained in a standardized fashion. The instrument inquires

about most psychotic phenomena, but it does not provide

a systematic way of capturing the presence or character

of the “minor ” forms of psychosis (i.e., illusions and

passage and sense of presence hallucinations). The total

score does not provide a speciﬁc index of psychosis,

because other behaviors are included in the ﬁnal outcome

Psychometric properties: In Alzheimer’s disease, the

NPI showed good internal consistency (Cronbach’s ␣

0.87– 0.88). The test–retest reliability over 2 to 3 weeks

for the 10 constituent scales and the total score of the NPI

ranged from 0.51 to 0.97 for frequency of occurrence of

symptoms and from 0.51 to 1.00 for ratings of the

severity of symptoms. The interrater agreement was “90

to 100%.” Concurrent validity was established with Be-

have-AD: the Behave-AD score and NPI frequency score

correlation was 0.66, while the correlation with the NPI

severity rating of symptoms was 0.71. Correlations

among corresponding (similar) subscales of the Be-

have-AD and NPI tended to be weaker: 0.54 to 0.78 for

frequency of symptoms and 0.47 to 0.80 for severity of

symptoms.

The NPI has been used in several studies

related to PD psychosis. Although the NPI may be useful

for tracking the incidence and presence of psychosis,

some antipsychotic treatment studies suggest that the

NPI may not be as sensitive to change in the PD popu-

lation

30-32

relative to the Brief Psychiatric Rating Scale

(BPRS). This may be related to the multiplicative scoring

metric, which results in noncontinuous scores as symp-

tom frequency and severity increase. In addition, there

probably is a nonlinear relationship between symptom

severity (intensity) and frequency and these constructs

may have differential sensitivity to treatment. Clinimet-

ric testing has been performed on the total score and not

490 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

the speciﬁc subscores related to hallucinations and psy-

chotic behaviors.

Final Assessment: The NPI fulﬁlls criteria as a rec-

ommended scale for rating PD psychosis, especially in

the cognitively impaired population.

Behavioral Pathology in Alzheimer’s Disease Rat-

ing Scale. The Behave-AD

was designed to measure

behavioral disturbances in dementia, especially those

occurring in AD, by excluding symptoms that primarily

result from cognitive and functional impairments. Hal-

lucinations (ﬁve items) and delusions (seven items) are

assessed, and therefore the scale has been applied to PD

psychosis. The Behave-AD is a two-part scale. Part 1

(symptomatology) includes 25 items that measure behav-

ioral disturbances classiﬁed in seven categories: paranoid

and delusional ideation (7 items), hallucinations (5

items), activity disturbances (3 items), aggressiveness (3

items), diurnal rhythm disturbances (1 item), affective

disturbance (2 items), and anxieties and phobias (4

items). Each symptom is scored on four-point scale of

severity, where 0 means “not present” and 3 means

“present, generally with an emotional and physical com-

ponent.” Part 2 (global rating) is a four-point global

assessment of the overall magnitude of the behavioral

symptoms in terms of disturbance to the caregiver and/or

dangerousness to the patient. Ratings are based on care-

giver reports of symptoms occurring in the preceding 2

weeks. The scale takes ⬃20 min or less to administer. A

shorter, 12-item observer-rated derived scale has been

developed.

Strengths: In the setting of AD, this relatively brief

scale is sensitive to therapeutic interventions, such as the

use of antipsychotic drugs.

Its administration is opera-

tionalized to provide low variability. The emphasis of the

scale is on delusions (seven items) and hallucinations

(ﬁve items). In particular, the detailed rating of delusions

is well ﬁtted to the PD population. It is a caregiver-based

scale, which is an advantage in demented populations.

Weaknesses: The scale was speciﬁcally constructed

for the assessment of behavioral disorders of AD not in

patients with PD, in which the proﬁle of behavioral

disorders is different. In particular, hallucinations are

more frequent than delusions in PD and should receive

heavier representation. Of the ﬁve items on hallucina-

tions, only one item is on visual hallucinations. Equal

weight is then placed on auditory, olfactory, haptic, and

“other” hallucinations that are less prevalent in PD. “Mi-

nor” forms, such as illusions, passage hallucinations, and

sense of presence are not taken into account. Retained or

lost insight and frequency, both clinically considered as

key features of PD hallucinations and delusions are not

considered. The frequency of hallucinations and delu-

sions is also not considered. On the other hand, the scale

includes “diurnal rhythm disturbances” that are common

in PD, such as in response to anti-PD medications. They

should, therefore, not be considered a priori as a psychi-

atric symptom. Basing symptom severity on the caregiv-

er’s response carries the risk of indirectly measuring

other features such as “agitation impression” rather than

the actual delusion or psychosis. Some items are con-

founded by motor, cognitive, and behavioral features of

PD, such as the items on depression and anxiety. Since

the scale is caregiver-based, nonobservable intrapsychic

symptoms may not be directly measured.

Psychometric properties: In AD, interrater agreement

for total scores and for Paranoia and Delusions subscales

was very high (greater than 0.90). Strong internal con-

sistency and validity evaluations have been established

for AD. Its internal consistency appears to be excellent

for the scale overall (0.96 total).

Regarding validity, the

items are grouped to “measure” seven areas; recent fac-

tor analysis with 151 AD patients supports ﬁve factors,

only one measuring psychosis. A study on its concurrent

validity shows a 0.92 correlation between Behave-AD

Paranoid and Delusional Ideation category and Dementia

Signs and Symptoms Scale (DSS); 0.93 between Be-

have-AD Hallucinations and DSS. Paranoid and Delu-

sional Ideation scores are sensitive to serotonergic and

antipsychotic treatment.

In spite of these positive eval-

uations in AD, however, the scale has only rarely been

used in PD psychosis.

Final Assessment: The Behave-AD fulﬁlls criteria as a

suggested scale for rating PD psychosis.

Scales Developed to Assess Psychosis in

Schizophrenia.

Brief Psychiatric Rating Scale. The BPRS was de-

signed to measure clinical change in patients with

schizophrenia.

Developers of the BPRS intended for it

to be administered by experienced psychiatrists and psy-

chologists and that the 20 to 30 minute scale required

staff training and monitoring to ensure adherence to item

deﬁnitions. The BPRS includes 18 items with one item

devoted to hallucinatory behavior, one to suspiciousness,

and one to unusual thought content. Ratings are to be

based solely on clinician-observed symptom severity.

Each item is scored on a seven-point scale ranging

from “not present” to “extremely severe.” The total

BPRS score is the sum of the scores for each of the 18

items and can be used as a global measure of psychopa-

thology.

Strengths: The BPRS has been used more often than

any other symptom rating scale in clinical trials of anti-

psychotic agents in patients with PD.

10,30-32,40-53

It is

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 491

Movement Disorders, Vol. 23, No. 4, 2008

relatively brief to administer. It is a good measure of

overall psychopathology in a wide range of patient

groups. This is relevant in studies of PD psychosis be-

cause patients frequently experience other psychiatric

symptoms (e.g., depression and anxiety) or behaviors

(e.g., uncooperativeness) that cut across diagnostic cate-

gories and may be affected by antipsychotic treatment. In

addition, empirically derived subscores, based on pa-

tients with schizophrenia, provide an index of the sever-

ity of related psychotic phenomena as well as indepen-

dent symptom areas (e.g., mood phenomena). The BPRS

contains items that enable characterization of different

delusional phenomena that can occur in PD, such as

bizarre delusions, somatic delusions (or concerns), and

grandiosity. The seven-point scoring system allows a

large gradation of measures including minor (i.e., “very

mild”) hallucinations when scoring.

Weaknesses: The scale does not provide adequate

detailed characterization of the various psychotic phe-

nomena that occur in PD. The scale does not permit

differential scoring of the intensity and frequency of

different types of hallucinatory phenomena (e.g., fre-

quent formed visual hallucinations and rare auditory

hallucinations). Anchors for the item on hallucinations

exclude “vivid mental imagery,” which could represent

illusions in PD patients.

Raters for the BPRS need training. Some items, e.g.,

“physical tension,” “mannerisms,” “blunted affect,” and

“motor retardation” may be confounded by the motor

aspects of PD, dementia, or apathy. It may also be

unclear how to evaluate the items on “conceptual disor-

ganization,” “guilt feelings,” “grandiosity,” and “excite-

ment” in patients with PD and psychosis. In cognitively

impaired patients, the restricted use of patient’s report

during the interview and direct observation of the patient

may limit the validity of the BPRS.

Psychometric properties: In general, the BPRS total

score and the BPRS “psychosis subscore” appear to be

sensitive to change in overall psychopathology in pla-

cebo-controlled and open-label clinical studies on the

treatment of psychosis in PD. Despite its extensive use in

PD psychosis studies, however, the scale has yet to

undergo formal psychometric evaluation in this popula-

tion. It might be difﬁcult to replicate the tight psycho-

metric properties originally reported in schizophrenia

Hedlund and Vieweg

reported item interrater reliabili-

ties (Pearson coefﬁcients of 0.63– 0.83) unless signiﬁcant

effort is spent in training observers, especially on obser-

vational versus patient-report items. One study reported

needing more than 30 training sessions to achieve intra-

class correlation coefﬁcients ⬎0.80 using seven psychi-

atrists.

Modiﬁed descriptive anchors may help improve

reliability. Positive and negative symptom items have

good internal consistency, with ␣⬎0.81.

Regarding internal consistency, a recent meta-analysis

of 26-factor analytic studies

showed good support for

the core four-factor model (meaning that the instrument

measures four domains relevant to psychosis); however,

a ﬁve-factor solution is also supported by meta-analysis:

affect, positive symptoms, negative symptoms, resis-

tance, and activation (the ﬁfth best emerges in studies of

schizophrenia). Most validity studies examined conver-

gent validity: BPRS positive and negative symptom

scores correlate well with same scales from PANSS

(0.92 and 0.82). Construct validity in terms of sensitivity

to change with treatment is supported in many studies.

The BPRS may be less useful in patients with mild

symptoms.

Final Assessment: The BPRS fulﬁlls criteria as a rec-

ommended scale for rating PD psychosis, especially in

the cognitively intact population and as a means to

tracking response to treatment or other interventions.

Positive and Negative Syndrome Scale. The Posi-

tive and Negative Syndrome Scale (PANSS) is a 30-item

scale with 7 positive symptom items, 7 negative symp-

tom items, and 16 general psychopathology symptom

items. The scale is completed by the physician, on the

basis of “verbal report and manifestations during the

course of the interview as well as reports of behavior by

primary care workers or family.”

Each item is scored

on a seven-point severity scale from 1 (symptom absent,

deﬁnition does not apply) to 7 (extreme). The positive

and negative symptom scales are often reported sepa-

rately. The PANSS was based originally on the BPRS

and on the Psychopathology Rating Schedule. It was

designed to measure symptoms in schizophrenia and is

commonly used in that setting in trials of antipsychotic

agents. The PANSS or its positive symptom subscore has

been used in several studies of the treatment of drug-

induced psychosis in PD.

11,60-63

Strengths: Detailed deﬁnitions and speciﬁc criteria for

all rating points are provided in the scale. The positive

scale also includes behavioral phenomena (hostility, sus-

piciousness, excitement) that can accompany hallucina-

tions or delusions. It provides a “minimal” rating and

therefore permits rating the presence of minor forms of

hallucinations in a standardized fashion.

The general psychopathology subscale includes other

psychiatric phenomena that are frequently, though not

invariably present in PD patients with psychosis. And,

the scale has a sufﬁcient number of items to permit

conduction of factor analyses for psychosis in PD and to

determine whether the syndrome of PD psychosis is

associated with speciﬁc patterns of psychotic symptoms.

492 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

Weaknesses: The scale was speciﬁcally constructed

for the assessment of psychopathology of schizophrenia,

not PD. The duration of administration is relatively long

(recommended interview time is 30 – 40 minutes) and

includes assessment of complex phenomena. The exam-

iner should have experience in psychiatry, but applica-

bility in PD samples may be limited. The positive scale

of the PANSS includes a single item devoted to “hallu-

cinatory behavior,” and the operational criteria for sever-

ity scoring are based on the hallucinatory syndrome of

schizophrenia (which are mostly verbal and distressing).

Among the other positive symptoms subscale of the

PANSS, delusions may apply to patients with PD but the

other items (conceptual disorganization, excitement,

grandiosity, suspiciousness/persecution, and hostility)

are not well adapted to PD psychopathology. The nega-

tive symptoms subscale has many items that can overlap

with presence of dementia or apathy.

Psychometric properties: The positive symptom scale

and the individual items for “hallucinations” and “delu-

sions” have been sensitive to changes in therapeutic trials

in PD.

In studies among persons with schizophrenia,

interrater reliability is good with intraclass correlation

coefﬁcients of 0.80 or above.

. For individual items,

interrater correlations (with patients with schizophrenia)

have ranged from 0.54 to 0.93

or 0.23 to 0.88.

Ade-

quate interrater reliability can be achieved in three train-

ing sessions.

For the positive and negative scales, in-

terrater reliability has been reported at 0.82 and 0.86.

Internal consistency is good with Cronbach’s ␣ from

0.73 to 0.87 for the three subscales.

The fact that PANSS shows correlation with BPRS

(convergent validity) is not surprising, since the items

were derived in part from the BPRS. Correlations be-

tween positive scale with SAPS and between negative

scale with SANS are 0.77.

Although it has three orig-

inal scales (positive symptoms, negative symptoms, gen-

eral psychopathology), almost all-factor analytic studies

with schizophrenic and mood disordered patients have

shown this to be not supported. Instead, most studies

support a ﬁve- to six-factor model.

Final Assessment: The PANSS fulﬁlls criteria as a

recommended scale for rating PD psychosis especially

for tracking treatment response. The scale can be used in

cognitively intact or impaired populations as it relies not

only on patient report and clinician observation during

the interview but also from the reports of primary care-

givers and family.

Schedule for Assessment of Positive Symptoms

(SAPS). The SAPS was developed to assess and pro-

vide qualitative information about speciﬁc features of

hallucinations delusions, behavioral changes associated

with psychosis, and thought disorder.

The instructions

state that the interview should be administered as part of

a standardized interview with additional information ob-

tained from nursing staff or others who have observed

the patient. The scale is designed to include single items

as well as global ratings for each symptom cluster. The

rater is instructed to take detailed notes when the patient

describes the symptoms. The scale was not developed as

a tool for measuring change, although it has been used

this way in treatment trials for PD psychosis. The hallu-

cinations section includes seven items: one each on vi-

sual hallucinations, olfactory hallucinations, and somatic

or tactile hallucinations; three items on auditory halluci-

nations, of which two rate certain “ﬁrst rank ” symptoms

(such as voices conversing and voices commenting,

which should be rated independent of the more typical

auditory hallucinations); and a global rating. The rater is

instructed not to rate illusions or hallucinations that oc-

cur when the person is falling off or waking up from

sleep or in the context of an illness or medication expo-

sure that might be associated with occurrence of hallu-

cinations. The individual hallucinations items are rated

on a continuum based on their frequency (occasional to

daily, with the latter being most severe). However, the

global hallucinations item scoring is based on both the

frequency and the extent to which the hallucinations are

disruptive. The delusions section has 13 items, including

12 individual items reﬂecting various types of delusions

and 1 global delusions score. The types of delusions

rated include persecutory, grandiose, jealousy, guilt, re-

ligious, somatic, and referential, as well as ﬁrst rank

symptoms of being controlled, mind reading, thought

broadcasting, thought insertion, and thought withdrawal.

These items are rated according to degree of conviction

about the idea, the frequency to which the idea is con-

sidered, and whether it affects behavior. The global rat-

ing also takes into account bizarre delusions, but is

otherwise rated in a similar fashion to the individual

items. The next section rates “bizarre behavior” and

includes four items plus a global score that reﬂect a range

of phenomena, including abnormal dress, manneristic

behavior that is socially inappropriate behavior, aggres-

sive behavior, and repetitive stereotyped behaviors. The

ﬁnal section rates formal “thought disorder,” which is

characterized by a disruption in how ideas are linked to

one another in the context of communication. This sec-

tion includes eight items. There are no speciﬁc instruc-

tions for scoring the SAPS. Some studies in PD patients

use the sum of all the items as well as the global scores.

Others look only at global scores or just the hallucina-

tions and delusions scores.

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 493

Movement Disorders, Vol. 23, No. 4, 2008

Strengths: The SAPS is easy to administer with a

structured interview and clear anchors provided as part

of the scale. It assesses the range of various subtypes of

hallucinations and delusions, and this may provide a tool

for cataloging the range of hallucinatory and delusional

phenomena in PD.

The global severity rating for each subsection provides

a useful measure of overall symptom severity. In partic-

ular, the global rating of hallucinations is a good ques-

tion, as it rates severity by its impact, with “mild” (pa-

tient unsure if they are real), “moderate” (vivid and

mildly bothersome), “marked” (vivid, frequent, and “per-

vade his life”) and “severe” (very vivid and extremely

troubling). The global rating for delusions is similarly

useful.

Weaknesses: Like other scales, the SAPS was devel-

oped for use in patients with schizophrenia, not PD, so

items do not rate the more common types of hallucina-

tions or delusions in PD or capture the range of severity

of those symptoms and vice versa (covering many symp-

toms that are uncommon in PD). The scale speciﬁcally

excludes illusions. It does not provide a systematic way

of capturing the presence or character of all psychotic

phenomena, including other minor forms. The hallucina-

tions items are weighted toward auditory hallucinations.

The presence of insight is not taken into account with

scoring.

The scale was not intended for use in patients with

dementia or cognitive impairment that limits awareness

that symptoms are present. Furthermore, the anchors for

scoring hallucinations are confusing to apply in PD and

may not reﬂect the overall severity of the phenomena

because frequency and severity are dissociated in the

scoring metric. For example, vivid visual hallucinations

with insight that occur daily and do not disrupt behavior

would score a “5” (severe) in this item, but it is unclear

where they would be rated for the global item.

The behavioral section includes disparate items that

are poorly deﬁned and cover too varied dimensions in a

single rating. Phenomena in the thought disorder section

overlap with features of aphasia and it would be impos-

sible to distinguish the etiology of the language distur-

bance in a patient with psychosis. An inclusive rating

approach, however, would inﬂate the overall SAPS score

of a patient with dementia and aphasia who has minimal

hallucinations.

Raters without experience interviewing patients with

schizophrenia may have less familiarity with many of the

constructs within each subsection, especially in eliciting

the various types of delusions. This could have a signif-

icant effect on validity and reliability of the instrument.

The SAPS provides a set of structured interview ques-

tions, but experienced raters are aware of the need to

probe in greater depth to clarify the presence or absence

of a given phenomena based on information provided by

the patient or others as well as observations of the

patient. To that end, the structured interview could yield

a scale that is administered reliably but lacks validity.

Psychometric properties: Studies using the SAPS in

clinical trials of PD psychosis (especially the subsection

on delusions and hallucinations) show that it is sensitive

to change in response to effective treatment.

34,50

How-

ever, SAPS has not been subject to careful psychometric

analysis in PD.

Nonetheless, interrater reliability for SAPS summary

score in psychotic patients is good (0.84).

The intra-

class coefﬁcient (ICC) is 0.94.

For the global domain,

intraclass correlations ranged from 0.50 to 0.91

Test–

retest reliability is weak–moderate (0.54).

70,71

Internal

consistency is weaker for the overall instrument (Cron-

bach ␣ 0.48) than for the four global domain scores (␣

ranging from 0.66 to 0.79).

Correlations with PANSS

and BPRS are consistently high. For example, Norman et

al.

found correlation 0.91 between PANSS positive and

SAPS summary score. Nicholson et al. 1995 found

BPRS positive symptoms (various deﬁnitions) correlated

well with SAPS composite (0.89⫹). Single factor struc-

ture generally not supported.

Final Assessment: The SAPS fulﬁlls criteria as a rec-

ommended scale for rating PD psychosis. While not

intended to be used to track changes in treatment, it has

been used for this purpose in PD psychosis. It is best

used in nondemented populations.

Nurses’ Observation Scale for Inpatient Evalua-

tion. The Nurses’ Observation Scale for Inpatient Eval-

uation (NOSIE-30) is an inpatient ward behavior rating

scale. A ﬁrst version consisted of 80 items of ward

behavior completed by a pair of nursing service mem-

bers. A revised form, the NOSIE-30, is a selection of 30

items.

Each item is rated on a ﬁve-point scale of

frequency (0 ⫽ never; 1 ⫽ sometimes; 2 ⫽ often; 3 ⫽

usually; 4 ⫽ always). The behavior is rated as observed

in the last 3 days. Factor analysis has identiﬁed six

factors, three “positive” (social competence, social inter-

est, and personal neatness) and three “negative” (irrita-

bility, manifest psychosis, and retardation).

This scale has been designed to evaluate the behavior

of hospitalized schizophrenic patients. The original re-

port on the NOSIE-30 was based upon a sample of 630

schizophrenic men.

Strengths: The scale is brief (takes under 10 minutes to

complete) with simple scoring, focusing on frequency.

Raters require little training and simply ﬁll in the re-

sponses at the end of the 3-day period without having to

494 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

ask the patient questions. No interpretation of behavior is

required (e.g., whether a patient is uncommunicative

because of depression, delusional thinking, confusion,

etc) making this a reliable and simple tool for nurses.

It has been used for over 40 years in a large number of

published studies on inpatient schizophrenics and shown

to be sensitive to changes in therapeutic trials with a high

interrater reliability.

Weaknesses: The items in the NOSIE-30 were chosen

to describe behavior of severe (men) schizophrenic in-

patients (and therefore not applicable to the outpatient

setting). A number of items do not ﬁt well with PD

patients (e.g., “tries to be friendly with others”). It scores

frequency, but not severity. Some of the items can be

confounded by the motor state or apathy (e.g., “is slow

moving or sluggish,” “sits, unless directed into activity,”

etc).

The NOSIE-30 has been used in only a few studies of

PD patients with psychosis

75,76

and often together with

other psychosis rating scales.

It contains several items focused on social integration

and maintaining social norms, assuming normal physical

function. It has no items related to delusions, apathy or

anxiety.

Pyschometric properties: Original reports by Honig-

feld et al.

do not provide adequate psychometric data.

There is a weak correlation between the NOSIE psycho-

sis score with Friedberg et al.’s PPRS (0.48). Internal

consistency of positive symptom items is fair (KR 0.68),

but weak for negative symptoms (KR 0.10).

Interrater

reliability is adequate for global score but not for sub-

scores.

Final Assessment: The NOSIE-30 fulﬁlls criteria as a

listed scale for rating PD psychosis.

Other Scales Used in Parkinson’s Psychosis Studies.

Clinical Global Impression Scale. The CGIS is

comprised of three subscales: CGI-severity; CGI-im-

provement; and CGI-therapeutic effect.

Each subscale

is a single item rating that is the closest thing to a clinical

“gestalt.” In CGI-severity, the clinician simply catego-

rizes the severity of the patient’s problem using a nom-

inal scoring system with: 0 ⫽ not assessed; 1 ⫽ normal;

2 ⫽ borderline ill; 3 ⫽ mildly ill; 4 ⫽ moderately ill; 5 ⫽

markedly ill; 6 ⫽ severely ill; 7 ⫽ among the most

extremely ill patients. Similarly, in CGI-improvement,

the clinician categorizes the improvement using a similar

nominal scoring system from very much improved to

very much worse. The only instruction is to rate this

patient within the spectrum of “similar patients.” This

guideline would be interpreted in the context of PD

psychosis as meaning other PD patients with psychosis,

excluding, for example Alzheimer’s disease or schizo-

phrenia. The scale does not include a battery of questions

and interpretation of severity encompasses whatever the

rater considers important. It may be applied to any ill-

ness.

Strengths: This scale takes only a few seconds to

complete once the history and examination are con-

cluded. The formulation of the score involves no extra

effort on the clinician’s behalf, as there are no particular

questions to ask and the clinician should have obtained

sufﬁcient information in any encounter to be able to

complete this question. The scale provides a measure of

clinical relevance, which may be a useful addition to the

more sophisticated scales with high reliability and sen-

sitivity, but where the clinical relevance of a change may

be limited. The scale is in the public domain, making it

cost-free to use.

Weaknesses: Whereas the CGIS is a “ﬂexible ” and

open scale that permits the rater to consider the “whole

picture,” assessment of single items can be easily con-

fusing. For example, hallucinations of similar degree in

different patients may cause markedly different re-

sponses in the patient, making it a major problem in one

case and a minor problem in the other. Raters for the CGI

may also confound more than one problem, such that

depression, anxiety, or motor dysfunction itself may alter

the rater’s interpretation of the severity of the illness (i.e.,

a patient with mild psychosis may have severe anxiety,

causing the global impression to have an uncertain in-

terpretation). Importantly, item 3 (minimally improved)

and 5 (minimally worse) do not indicate whether these

refer to minimal in the sense of the minimal change that

is still clinically signiﬁcant or to “minimal” in the sense

of “inconsequential.”

Psychometric properties: While there is limited psy-

chometric information available, the scale has been used

in several large randomized PD psychosis clinical trial

and appears to be sensitive to change.

50,61,80

Guy

notes

that the limited studies available on psychometrics are

quite critical of the scale but that some studies may not

have adequately evaluated psychometric properties of

the scale by virtue of having included heterogeneous

samples, e.g., patients with schizophrenia, depression

and anxiety. Interrater reliability for the severity scale is

low (0.41– 0.66). Nonetheless, it has been used in a

several PD psychosis studies, either as the primary or

secondary outcome measure and has been found to al-

ways correlate with the other more in depth psychosis

scale.

50,61,80

Some guidelines for the use and training are

probably needed to establish adequate interrater reliabil-

ity, at least in multicenter studies in PD samples.

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 495

Movement Disorders, Vol. 23, No. 4, 2008

Final Assessment: The CGIS fulﬁlls criteria as a sug-

gested scale for rating PD psychosis. It is best used as an

additional outcome measure to complement a more de-

tailed psychosis scale.

Uniﬁed Parkinson Disease Rating Scale, Part I.

Part 1 of the UPDRS has four items (“mentation”;

“thought disorders,” “depression” and “motivation/initia-

tive), scored from 0 (normal) to 4 (most severe), based on

the week prior to assessment.

It was developed as a

subscale of the larger UPDRS with the goal of tapping

nonmotor phenomena. The ratings for Part I are based on

a clinical interview of the patient, although probably best

performed with a caregiver present. Some scoring strat-

egies are unclear. Under the item “2 ” (thought disor-

ders), the following are the anchor points: 0 ⫽ none; 1 ⫽

vivid dreaming; 2 ⫽ benign hallucinations with insight

retained; 3 ⫽ occasional to frequent hallucinations or

delusions, without insight, could interfere with daily

activities; 4 ⫽ persistent hallucinations, delusions, or

ﬂorid psychosis; not able to care for self. Thus, item 2 is

the most relevant item for psychosis, although cognitive

impairment, depression and reduced motivation fre-

quently co-occur with psychosis. The item combines

dreaming phenomena, hallucinations, and delusions in

one item. The UPDRS is currently under revision and

this subscale is subject to major changes.

Strengths: The thought disorder item can be adminis-

tered and scored very rapidly. Its inclusion in the most

widely used PD scale has made it widely used instrument

for assessment of psychiatric symptoms in epidemiolog-

ical studies, although psychometric studies are wanting.

It is brief and easily administered. The anchor points are

clear and clinically relevant.

Weaknesses: A single item is obviously not sufﬁcient

to explore the variety of psychotic phenomena in PD.

The item implicitly assumes that there is a continuum

from vivid dreams to formed hallucinations, a point that

is controversial based on current literature. The lumping

of dream phenomena, hallucinations and delusions into

one item does not always ﬁt with the clinical presenta-

tion. Although less common, it is possible for some

patients to experience delusions but not hallucinations.

Many symptoms that are not uncommon in PD patients

with psychosis, such as auditory or other hallucinations,

are not included. Finally, as a screening instrument it

may not be particularly sensitive to change, and to our

knowledge, studies addressing this issue do not exist.

Psychometric properties: A major problem is that a

“ﬂoor effect ” may be present in many patients (23%).

Interrater reliability is questionable. Internal consistency

is good (Cronbach ␣ 0.79). Item-to-scale mentation total

correlations were moderate (0.57– 0.66). Factor structure

conﬁrmed; the four items loaded on their own factor

(loadings 0.54 – 0.72). A self-rating version exists, which

has demonstrated good reliability with a clinical inter-

view.

Final Assessment: The UPDRS Part I and item 2

within Part I fulﬁll criteria as a listed scale for rating PD

psychosis.

Challenges Posed by Psychosis Rating Scales.

Several factors are implicit concerns in the assessment

of psychosis scales. Validity of any scale is necessarily

affected by the rater who conducts the interview. Patients

or informants may be more likely to endorse symptoms

when asked, for example, by a physician versus a re-

search assistant. Further, knowledge of a patient’s his-

tory and familiarity with previously reported hallucina-

tions or delusions provides a basis for the physician’s

inquiry about present symptoms and their impact. Clin-

ical judgment and experience of the interviewer will

inﬂuence the consistency of the informant/patient an-

swers and the extent to which the rater probes or notices

subtle features of psychosis. As an example, an untrained

rater with limited clinical experience may base ratings on

the patient’s initial answer to a direct question and miss

hallucinatory or delusional phenomena that only become

apparent during an open-ended interview or through dis-

closure that is based on trust in the physician. Some

scales attempt to limit this source of variation in symp-

tom ratings by providing scripted screening probes or

detailed anchors for scoring the ratings, but these restric-

tions may in fact underrate behaviors like psychosis that

are culturally sensitive.

The source of information also inﬂuence ratings, and

the various scales use a number of approaches, e.g.,

self-report, informant-derived information, medical

records. Some patients may not reveal their psychotic or

affective phenomena to others, diminishing the validity

of the instrument when it is administered to an informant

only. Conversely, cognitive impairment may limit the

information a patient can provide.

Some scales focus only on psychosis, whereas others

include a range of psychiatric symptoms in addition to

psychosis. Psychosis scales usually provide more de-

tailed description of psychosis, whereas the latter group

provide a broader psychopathological description.

A ﬁnal concern is that most psychosis scales were

developed in non-PD samples and have not been psy-

chometrically tested in PD.

Task Force Recommendations

Similar to the Quality Standards Subcommittee of the

American Academy of Neurology that concluded in their

496 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

review that there currently are “no validated tools for

psychosis screening in PD,”

it is evident from this

review that none of the current scales used in PD ade-

quately captures the entire phenomenology of PD psy-

chosis. It is important that the current scales do not have

the ability to assesses the incidence, prevalence and

severity (or impact) of all forms of discrete psychotic

phenomena. For treatment studies, it is essential that the

scale be useful for assessing individual responses as well

as group scores and responses. The BPRS and the SAPS

are scales that appear to be sensitive to change. The

SAPS is relatively complete but like most psychosis

scales derived from psychiatric research (BPRS, SAPS,

PANSS, etc), several items relevant to schizophrenia are

less useful to PD psychosis. The NPI is a good scale in

cataloging the presence or absence of psychotic phenom-

ena. It is a scale with a scripted interview that relies less

on the clinician’s judgment. Scales with open-ended

interviews rely more on the judgment of the rater. As

opposed to tremor, which is observable and fairly objec-

tive, clinical experience and judgment are needed to

classify in a reliable and valid way whether a perception

is indeed hallucinatory. Yet, despite the script in the NPI,

there are still concerns about how the constructs are

interpreted when the scale is used by clinicians who are

not psychiatrically trained. Trained raters are better at

picking up on cues that the patient does have paranoid

ideas or is hallucinating, and notice inconsistencies in

responses. The PPRS, PPQ and the UPDRS Part I were

designed speciﬁcally for PD but are still inadequate

scales in exploring and tracking the entire PD psychosis

phenomenology. The Rush Hallucination Inventory and

the Baylor Hallucination Questionnaire do not explore

delusions. The NOSIE-30 cannot be applied to the ma-

jority of PD patients who are community dwelling. See

Tables 3 and 4 for the summary of the general properties

of each scale reviewed. While the Task Force has labeled

some scales as “recommended ” or “suggested ” based

on the fulﬁlling deﬁned criteria, it does not mean that

these scales are “ideal.”

What Then Are the Properties of the “Ideal” Scale

for PD Psychosis?

Contentwise, the Task Force recommends that the

ideal scale have two parts. The ﬁrst should be a diagnos-

tic guide to determine whether the patient is, in fact,

psychotic. The scale should then be applied only to those

who meet this ﬁrst criterion, or at least, it should be

validated only for this objective. While the DSM criteria

are regarded as the “gold standard ” for establishing

psychiatric diagnoses, it does not include a speciﬁc di-

agnosis for PD-related psychosis. The Diagnostic Crite-

ria for Psychosis in Parkinson’s disease: Report of an

NINDS/NIMH Work Group

is therefore a good model

with which to start.

The second part should be the scale itself and should

rate the various psychotic symptoms by their presence

and by impact on the patient and the family. To do this,

frequency is involved, but should not be used as a mul-

tiplicative score, as in the NPI. Products of scores are

conceptually difﬁcult to interpret (e.g., what construct

does the product of symptom severity and distress rep-

resent?), yield noncontinuous scores at the extremes, and

the nature of the relationship between the constituent

scores (and of the scores with external constructs) may

be heterogeneous and difﬁcult to establish. There needs

to be a scoring system that allows the patient to rate the

presence of a (mild) “concern” versus a convincing be-

lief that a delusion is true. Determining the impact of a

symptom is critical, since delusions, for example, usually

have greater impact, even when minor, compared to

frequent hallucinations of a nonemotional nature. The

entire range of PD psychosis phenomenology needs to be

probed, including less common hallucinations (olfactory,

tactile, etc) “minor hallucinations ” (passage hallucina-

tions, sense of presence hallucinations), illusions, and

delusions. The delusion section should include delusions

sensu stricto, such as theft or jealousy delusions, and

items pertaining to misidentiﬁcation (such as Capgras

syndrome, “mirror sign,” reduplication, etc). It should be

equally applicable to cognitively intact and cognitively

impaired patients and should identify whether the patient

is demented and whether insight is preserved. The scale

should have the ability to distinguish hallucinations from

delirium; psychosis from RBD and vivid dreams. Asso-

ciated features such as mood disorders and anxiety

should be considered for inclusion, as they affect and are

affected by the hallucinations and delusions. Finally, a

global psychiatric measure should be embedded in the

scale, as in the CGIS, but should be restricted to the

psychiatric aspects alone and completely independent of

motor function.

Mechanistically, the patient and at least one observer

need to be interviewed for the scale, if possible. When

there is only the patient available, then a notation should

be made that the scoring is based on the subject alone.

The test must be easy to administer and should take up to

15 to 20 minutes to administer, but not longer. A screen-

ing version that takes 5 to 10 minutes, allowing admin-

istration of whole instrument only if patient scores in a

certain range on the screen should be considered. The

scale should have a simple scoring metric. It should

contain clear instructions on how the scale is adminis-

tered and it should be capable of being accurately ad-

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 497

Movement Disorders, Vol. 23, No. 4, 2008

ministered by technicians. Detailed deﬁnitions or criteria

should be provided for all rating points, preferably with

a set of training tapes that can be used to evaluate

reliability.

Psychometrically, the ideal scale should, in a PD co-

hort, at least demonstrate concurrent validity with an-

other “gold” standard, and discriminant validity (e.g.,

from dementia severity, motor symptom severity). Items

from subscales (e.g., pertaining to visual hallucinations)

should have stronger relationship among themselves than

with other dimensions of psychosis or total score). Inter-

rater reliability is more important than test–retest reli-

ability. The scale should demonstrate sensitivity to

change and predict functional change with treatment. It

should be evaluated cross-culturally, and should thus

consider content that is relevant to different cultures.

On the basis of these considerations, the Task Force

therefore recommends the development of a new scale of

PD psychosis as none of the current scales contain all the

basic content, mechanistic and psychometric properties

outlined above. In the meantime, selection of the current

scales should be based on the goals of the assessment.

Different scales may be better for some settings versus

others. Any chosen scale should be adapted to use both

informant and patient information, and the rater will need

to make a judgment as to which information to use when

making a ﬁnal rating. Since one scale may not be able to

serve all needs, a scale used to measure clinical response

and change over time (such as the CGIS) may need to be

combined with another scale better at cataloging speciﬁc

features (such as the NPI or SAPS). At the present time,

for clinical trials on PD psychosis assessing new treat-

ments, the following are recommended primary outcome

scales: NPI (for the cognitively impaired PD population

or when a caregiver is required), SAPS, PANSS, or

BPRS (for the cognitively intact PD population or when

the patient is the sole informant). The CGIS is suggested

as a secondary outcome scale to measure change and

response to treatment over time.

Acknowledgments: The Writing and Steering Committee

thanks the MDS secretariat staff (Caley Kleczka, Director) for

their valuable assistance in this project.

REFERENCES

1. Fenelon G, Mahieux F, Huon R, Ziegler M. Hallucinations in

Parkinson’s disease: prevalence, phenomenology and risk factors.

Brain 2000;123:733–745.

2. Hely MA, Morris JG, Reid WG, Trafﬁcante R. Sydney multicenter

study of Parkinson’s disease: non-

L-dopa-responsive problems

dominate at 15 years. Mov Disord 2005;20:190 –199.

3. Sanchez-Ramos JR, Ortoll R, Paulson GW. Visual hallucinations

associated with Parkinson disease. Arch Neurol 1996;53:1265–

1268.

4. Inzelberg R, Kipervasser S, Korczyn AD. Auditory hallucinations

in Parkinson’s disease J Neurol Neurosurg Psychiatry 1998;64:

533–535.

5. Fenelon G, Mahieux F, Huon R, Ziegler M. Hallucinations in

Parkinson’s disease: prevalence, phenomenology and risk factors.

Brain 2000;123:733–745.

6. Papapetropoulos S, Mash DC. Psychotic symptoms in Parkinson’s

disease. From description to etiology. J Neurol 2005;252:753–764.

7. Graham JM, Grunewald RA, Sagar HJ. Hallucinosis in idiopathic

Parkinson’s disease. J Neurol Neurosurg Psychiatry 1997;63:434 –

440.

8. Inzelberg R, Kipervasser S, Korczyn AD. Auditory hallucinations

in Parkinson’s disease. J Neurol Neurosurg Psychiatry 1998;64:

533–535.

9. de Maindreville AD, Fenelon G, Mahieux F. Hallucinations in

Parkinson’s disease: a follow-up study. Mov Disord 2005;20:212–

217.

10. Marsh L. Psychosis in Parkinson’s disease. Curr Treat Options

Neurol 2004;6:181–189.

11. Fernandez HH, Trieschmann ME, Okun MS. Rebound psychosis:

effect of discontinuation of antipsychotics in Parkinson’s disease.

Mov Disord 2005;20:104 –105.

12. Pollak P, Tison F, Rascol O, et al. Clozapine in drug induced

psychosis in Parkinson’s disease: a randomised, placebo controlled

study with open follow up. J Neurol Neurosurg Psychiatry 2004;

75:689 – 695.

13. Factor SA, Feustel PJ, Friedman JH, et al. Longitudinal outcome of

Parkinson’s disease patients with psychosis. Neurology 2003;60:

1756 –1761.

14. Aarsland D, Larsen JP, Karlsen K, Lim NG, Tandberg E. Mental

symptoms in Parkinson’s disease are important contributors to

caregiver distress. Int J Geriatr Psychiatry 1999;14:866 –874.

15. Aarsland D, Larsen JP, Tandberg E, Laake K. Predictors of nursing

home placement in Parkinson’s disease: a population-based, pro-

spective study. J Am Geriatr Soc 2000;48:938 –942.

16. Goetz CG, Stebbins GT. Mortality and hallucinations in nursing

home patients with advanced Parkinson’s disease. Neurology

1995;45:669 – 671.

17. Fernandez HH, Donnelly EM, Friedman JH. Long-term outcome

of clozapine use for psychosis in parkinsonian patients. Mov

Disord 2004;19:831– 833.

18. American Psychiatric Association. Diagnostic and statistical man-

ual of mental disorders, 4th ed. Washington, DC: American Psy-

chiatric Association; 1994.

19. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clin-

ical interview for DSM-IV axis I disorders-patient edition. Wash-

ington, DC: American Psychiatric Press; 1996.

20. Ravina B, Marder K, Fernandez HH, et al. Diagnostic criteria for

psychosis in Parkinson’s disease: report of an NINDS/NIMH work

group. Mov Disord 2007;22:1061–1068.

21. Nunnally JC, Psychometric theory, 2nd ed. New York: McGraw-

Hill; 1978.

22. Sattler JM, Assessment of children: cognitive applications, 4th ed.

San Diego: Jerome M. Sattler; 2001.

23. Streiner DL, Norman GR. Health measurement scales: a practical

guide to their development and use, 2nd ed. Oxford: Oxford

University Press; 1995.

24. Goetz CG, Fahn S, Martinez-Martin P, et al. Movement Disorder

Society-sponsored revision of the Uniﬁed Parkinson’s Disease

Rating Scale (MDS-UPDRS). Mov Disord 2007;22:41– 47.

25. Friedberg G, Zoldan J, Melamed E. Parkinson psychosis rating

scale: a practical instrument for grading psychosis in Parkinson’s

disease. Clin Neuropharmacol 1998;21:280 –284.

26. Brandstaedter D, Spiker S, Ulm G, et al. Development and eval-

uation of the Parkinson psychosis questionnaire: a screening in-

strument for the early diagnosis of drug-induced psychosis in

Parkinson’s disease. J Neurol 2005;252:1060 –1066.

498 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008

27. Goetz CG, Leurgans S, Pappert EJ, Raman R, Stemer AB. Pro-

spective longitudinal assessment of hallucinations in Parkinson’s

disease. Neurology 2001;57:2078 –2082.

28. Ondo WG, Tinter R, Voung MA, Lai D, Ringholz G. Double-

blind, placebo-controlled, unforced titration, parallel trial of

quetiapine for dopaminergic-induced hallucinations in Parkinson’s

disease. Mov Disord 2005;20:958 –963.

29. Cummings JL, Mega M, Gray K, Rosenberg-Thompson S, Carusi

DA, Gornbein J. The neuropsychiatric inventory: comprehensive

assessment of psychopathology in dementia. Neurology 1994;44:

2308 –2314.

30. Juncos JL, Roberts VJ, Evatt ML, et al. Quetiapine improves

psychotic symptoms and cognition in Parkinson’s disease. Mov

Disord 2004;19:29 –35.

31. Breier A, Sutton VK, Feldman PD, et al. Olanzapine in the treat-

ment of dopamimetic-induced psychosis in patients with Parkin-

son’s disease. Biol Psychiatry 2002;52:438 – 444.

32. Marsh L, Lyketsos C, Reich S. Olanzapine for the treatment of

psychosis in patients with Parkinson’s disease and dementia. Psy-

chosomatics 2001;42:477– 481.

33. Reisberg B, Borenstein J, Salob SP, et al. Behavioral symptoms in

Alzheimer’s disease: phenomenology and treatment. J Clin Psy-

chiatry 1987;48 (Suppl 5):9 –15.

34. Auer SR, Monteiro IM, Reisberg B. The empirical behavioral

pathology in Alzheimer’s disease (E-BEHAVE-AD) rating scale.

Int Psychogeriatr 1996;8:247–266.

35. De Deyn PP, Rabheru K, Rasmussen A, et al. A randomized trial

of risperidone, placebo, and haloperidol for behavioral symptoms

of dementia. Neurology 1999;53:946 –955.

36. Reisberg et al. Behavior Pathology in Alzheimer’s Disease Rating

Scale (BEHAVE-AD). In: Task Force for the Handbook of Psy-

chiatric Measures, editor. Handbook of psychiatric measures.

Washington, DC: American Psychiatric Association; 2000. p 401–

404.

37. Harwood DG, Ownby RL, Barker WW, Duara R. The behavioral

pathology in Alzheimer’s Disease Scale (BEHAVE-AD): factor

structure among community-dwelling Alzheimer’s disease pa-

tients. Int J Geriatr Psychiatry 1998;13:793– 800.

38. Chacko RC, Hurley RA, Harper RG, Jankovic J, Cardoso F.

Clozapine for acute and maintenance treatment of psychosis in

Parkinson’s disease. J Neuropsychiatry Clin Neurosci 1995;7:471–

475.

39. Overall JE, Gorham DR, The brief psychiatric rating scale. Psychol

Rep 1962;10:799 – 812.

40. Scholtz E, Dichgans J. Treatment of drug-induced exogenous

psychosis in parkinsonism with clozapine and ﬂuperlapine. Eur

Arch Psychiatry Clin Neurosci 1985;235:60 – 64.

41. Ostergaard K, Dupont E. Clozapine treatment of drug-induced

psychotic symptoms in late stages of Parkinson’s disease. Acta

Neurol Scand 1988;78:349 –350.

42. Friedman J. Lannon M. Clozapine in the treatment of psychosis in

Parkinson’s disease. Neurology 1990;39:1219 –1221.

43. Kahn N, Freeman A, Juncos J, et al. Clozapine is beneﬁcial for

psychosis in Parkinson’s disease. Neurology 1991;41:1699 –1700.

44. Wolk S, Douglas C. Clozapine treatment of psychosis in Parkin-

son’s disease: a report of ﬁve consecutive cases. J Clin Psychiatry

1992;53:373–376.

45. Factor S, Brown D. Clozapine prevents recurrence of psychosis in

Parkinson’s disease. Mov Disord 1992;7:125–131.

46. Lew M, Waters C. Clozapine treatment of Parkinsonism with

psychosis. J Am Geriatr Soc 1993;41:669 – 671.

47. Factor S, Brown D, Molho ES, Podskalny GD. Clozapine: a 2-year

trial open trial in Parkinson’s disease patients with psychosis.

Neurology 1994;44:544 –546.

48. Ford B, Cote L, Fahn S. Risperidone in Parkinson’s disease (letter).

Lancet 1994;344:681.

49. Wagner M, Deﬁlippi J, Menza M, et al. Clozapine for the treatment

of psychosis in Parkinson’s disease: chart review of 49 patients.

J Neuropsychiatr 1996;3:276 –280.

50. The Parkinson Study Group. Low-dose clozapine for the treatment

of drug-induced psychosis in Parkinson’s disease. N Engl J Med

1999;340:757–763.

51. Fernandez HH, Friedman JH, Jacques C, Rosenfeld M. Quetiapine

for the treatment of drug-induced psychosis in Parkinson’s disease.

Mov Disord 1999;14:484 – 487.

52. Ellis T, Cudkowicz ME, Sexton PM, Growdon JH. Clozapine and

risperidone treatment of psychosis in Parkinson’s disease. J Neu-

ropsychiatry Clin Neurosci 2000;12:364 –369.

53. Fernandez HH, Trieschmann ME, Okun MS. Rebound psychosis:

effect of discontinuation of antipsychotics in Parkinson’s disease.

Mov Disord 2005;20:104 –105.

54. Hedlund JL, Vieweg BW. The Brief Psychiatric Rating Scale

(BPRS): a comprehensive review. J Oper Psychiatr 1980;11:48–

65.

55. Overall JE, Gorham DR. Brief psychiatric rating scale in task force

for the handbook of psychiatric measures, handbook of psychiatric

measures. Washington, DC: American Psychiatric Association;

1962. p 490 – 494.

56. Nicholson IR, Chapman JE, Neufeld RW. Variability in BPRS

deﬁnitions of positive and negative symptoms. Schizophr Res

1995;17:177–185.

57. Shafer A. Meta-analysis of the brief psychiatric rating scale factor

structure. Psychol Assess 2005;17:324 –335.

58. Overall JE, Gorham DR. The Brief Psychiatric Rating Scale

(BPRS): recent developments in ascertainment and scaling. Psy-

chopharmacol Bull 1988;24:97–99.

59. Kay SR, Opler LA, Lindermayer JP. The Positive and Negative

Syndrome Scale (PANSS): rationale and standardisation. Br J

Psychiatry 1989;155 (Suppl 7):59 – 65.

60. Chacon JR, Duran E, Duran JA, Alvarez M. Usefulness of olan-

zapine in the levodopa-induced psychosis in patients with Parkin-

son’s disease. Neurologia 2002;17:7–11.

61. French Clozapine Study Group. Clozapine in drug-induced psy-

chosis in Parkinson’s disease. Lancet 1999;353:2041–2042.

62. Kohmoto J, Kihira T, Miwa H, Kondo T. Effect of quetiapine

fumarate on drug-induced psychosis in patients with Parkinson’s

disease. No To Shinkei 2002;54:489 – 492.

63. Mohr E, Mendis T, Hildebrand K, De Deyn PP. Risperidone in the

treatment of dopamine-induced psychosis in Parkinson’s disease:

an open pilot trial. Mov Disord 2000;15:1230 –1237.

64. Bell et al., 1992.

65. Norman RM, Malla AK, Cortese L, Diaz F. A study of the

interrelationship between and comparative interrater reliability of

the SAPS, SANS and PANSS. Schizophr Res 1996;19:73–85.

66. Muller MJ, Rossbach W, Davids E, Wetzel H, Benkert O. [Eval-

uation of standardized training for the “Positive and Negative

Syndrome Scale” (PANSS)]. Nervenarzt 2000;71:195–204.

67. Kay SR, Fiszbein A, Opler LA. The positive and negative syn-

drome scale (PANSS) for schizophrenia. Schizophr Bull 1987;13:

261–276.

68. Van den Oord EJ, Rujescu D, Robles JR, et al. Factor structure and

external validity of the PANSS revisited. Schizophr Res 2006;82:

213–223.

69. Andreasen NC. Scale for the assessment of positive symptoms

(SAPS). Iowa City, Iowa: University of Iowa; 1984.

70. Malla AK, Norman RM, Williamson P. Stability of positive and

negative symptoms in schizophrenia. Can J Psychiatry 1993;38:

617– 621.

71. Malla AK, Norman RM, Williamson P, Cortese L, Diaz F. Three

syndrome concept of schizophrenia. A factor analytic study.

Schizophr Res 1993;10:143–150.

72. Andreasen NC, Grove WM. Evaluation of positive and negative

symptoms in schizophrenia. Psychiatr Psychobiol 1986;1:108–

121.

73. Peralta V, Cuesta MJ. Dimensional structure of psychotic symp-

toms: an item-level analysis of SAPS and SANS symptoms in

psychotic disorders. Schizophr Res 1999;38:13–26.

SCALES TO ASSESS PSYCHOSIS IN PARKINSON⬘S DISEASE 499

Movement Disorders, Vol. 23, No. 4, 2008

74. Honigfeld G, Klett CJGillis RD. NOSIE-30: a treatment sensitive

ward behaviour scale. Psychol Rep 1966;19:180 –182.

75. Wolters EC, Hurwitz TA, Mak E, et al. Clozapine in the treatment

of parkinsonian patients with dopaminomimetic psychosis. Neu-

rology 1990;40:832– 834.

76. Zoldan J, Friedberg G, Livneh M, Melamed E. Psychosis in ad-

vanced Parkinson’s disease: treatment with ondansetron, a 5-HT3

receptor antagonist. Neurology 1995;45:1305–1308.

77. Dingemans PM. The Brief Psychiatric Rating Scale (BPRS) and

the Nurses’ Observation Scale for Inpatient Evaluation (NOSIE) in

the evaluation of positive and negative symptoms. J Clin Psychol

1990;46:168 –174.

78. Hafkenscheid A. Psychometric evaluation of the Nurses Observa-

tion Scale for Inpatient Evaluation in The Netherlands. Acta Psy-

chiatr Scand 1991;83:46 –52.

79. Guy W, ed. ECDEU assessment manual for psychopharmacology.

Rockville, MD: Alcohol, Drug Abuse, and Mental Health Admin-

istration, 1976 (DHEW publication no. ADM 76-338). p 217–222.

80. Morgante L, Epifanion A, Spina E, et al. Quetiapine and clozapine

in parkinsonian patients with dopaminergic psicosis. Clin Neuro-

pharmacol 2004;27:153–156.

81. Guy W. Clinical Global Impressions (CGI) scale. In: Task Force

for the Handbook of Psychiatric Measures, editor. Handbook of

Psychiatric Measures. Washington, DC: American Psychiatric As-

sociation; 2000. p 100 –102.

82. Fahn S, Elton RL. UPDRS program members. Uniﬁed Parkinson’s

disease rating scale. In Fahn S, Marsden CD, Goldstein M, Calne

D, editors. Recent developments in Parkinson’s disease, Vol. 2.

Florham Park, NJ: Macmillan Healthcare Information; 1987. p

153–163.

83. Martinez-Martin P, Forjaz MJ. Metric attributes of the uniﬁed

Parkinson disease rating scale 3.0 battery: part I, feasibility, scaling

assumptions, reliability and precision. Mov Disord 2006;29:1182–

1188.

84. Louis ED, Lynch T, Marder K, Fahn S. Reliability of patient

completion of the historical portion of the uniﬁed Parkinson dis-

ease rating scale. Mov Disord 1996;11:185–192.

85. Miyasaki JM, Shannon K, Voon V, et al. Practice parameter:

evaluation and treatment of depression, psychosis and dementia in

Parkinson disease (an evidence based review). Report of the Qual-

ity Standards Subcommittee of the American Academy of Neurol-

ogy. Neurology 2006;66:996 –1002.

500 H.H. FERNANDEZ ET AL.

Movement Disorders, Vol. 23, No. 4, 2008