www.elsevier.com/locate/schres

Schizophrenia Research

The prodromal questionnaire (PQ): Preliminary validation

of a self-report screening measure for prodromal

and psychotic syndromes

Rachel L. Loewya,T, Carrie E. Beardenb, Jennifer K. Johnsonb,

Adrian Rainec, Tyrone D. Cannona,b

aDepartment of Psychology, University of California, Los Angeles, 1285 Franz Hall, Box 951563, Los Angeles, CA 90095-1563, United StatesbDepartment of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, Box 956968, Rm 2265, 300 Medical Plaza,

Los Angeles, CA 90095-6968, United StatescDepartment of Psychology, University of Southern California, S.G.M. Building, Los Angeles, CA 90089-1061, United States

Received 30 September 2004; received in revised form 2 March 2005; accepted 14 March 2005

Available online 17 May 2005

Abstract

This study aims to establish preliminary validity of the Prodromal Questionnaire (PQ), a 92-item self-report screening

measure for prodromal and psychotic symptoms. Adolescents and young adults (N =113) referred to a prodromal psychosis

research clinic completed the PQ and the Structured Interview for Prodromal Syndromes (SIPS [Miller, T.J., McGlashan, T.H.,

Woods, S.W., Stein, K., Driesen, N., Corcoran, C.M., Hoffman, R., Davidson, L., 1999. Symptom assessment in schizophrenic

prodromal states. Psychiatric Quarterly 70(4), 273–287]), an interview with established validity for predicting future psychosis.

When maximizing selection of true positive cases, scores on a subset of PQ items that assess positive symptoms predicted a

concurrent prodromal or psychotic SIPS diagnosis with 90% sensitivity and 49% specificity. The PQ shows good preliminary

validity in detecting individuals with an interview-diagnosed prodromal or psychotic syndrome, but it is less sensitive to the

threshold between prodromal and full-blown psychosis.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Assessment; Prodrome; Psychosis; Schizophrenia; Screening

0920-9964/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.schres.2005.03.007

T Corresponding author. Tel.: +1 310 794 9673; fax: +1 310 794

9740.

E-mail addresses: rloewy@ucla.edu (R.L. Loewy),

cbearden@mednet.ucla.edu (C.E. Bearden),

jjohnson@psych.ucla.edu (J.K. Johnson), raine@usc.edu

(A. Raine), cannon@psych.ucla.edu (T.D. Cannon).

1. Introduction

During the past decade, over a dozen research

clinics have been established to investigate the onset,

pathophysiology, and treatment of the prodromal

phase of schizophrenia by assessing symptomatic

79 (2005) 117–125

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125118

individuals who may be at risk for psychosis.

However, the difficulty of participant ascertainment

has limited the ability of these clinics to produce

empirical results in a timely manner (Heinssen et al.,

2003). Where such data have been reported, approx-

imately one-third of referrals meet criteria for inter-

view-based diagnoses of high-risk status (Yung et al.,

2003).

While clinicians and researchers have long noted

the signs that appear in patients just prior to a

psychotic break, the first set of prodromal criteria

with empirical validity in predicting future psychosis

were recently developed by Yung and McGorry

(1996). Based on descriptions of the prodrome

compiled from retrospective accounts of first-episode

or remitted schizophrenia patients and their relatives,

case studies, high-risk studies, and studies of the

prodrome to psychotic relapse, researchers developed

a set of prodromal criteria that define three bat-riskQsyndromes, which have been incorporated into a semi-

structured interview, the Comprehensive Assessment

of at Risk Mental States (CAARMS; Yung et al.,

2003). These criteria were modified slightly to form

the basis of a similar interview, the Structured

Interview for Prodromal Syndromes (SIPS), along

with the Scale of Prodromal Symptoms (SOPS),

which were used in the current study (Miller et al.

2002, 1999). The syndromes refer specifically to the

blate prodrome,Q a period of illness just prior to full

psychosis (Addington, 2004; Yung et al., 1998).

Therefore, patients diagnosed with these syndromes

whose transition to psychosis would be expected to do

so quickly (i.e. within 1–2 years); initial validity data

for the two structured clinical interviews show that

approximately 40–54% of diagnosed patients’ tran-

sition to full psychosis within a year (Miller et al.,

2002; Yung et al., 2003).

While promising, the administration of these

interviews requires specific training and several

hours of clinicians’ time. In order to increase the

yield of participants within a clinic-referred sample,

we developed the Prodromal Questionnaire, a self-

report prodromal psychosis survey. The goal of this

study was to examine the preliminary psychometric

properties of the screening measure, primarily

through assessing concurrent validity with suspected

prodromal and psychotic syndrome diagnoses based

on structured clinical interview.

2. Methods

2.1. Materials

The PQ is a 92-item self-report questionnaire that

takes approximately 20 min to complete. Most items

were adapted from the Schizotypal Personality Ques-

tionnaire (Raine, 1991) and from probe questions in

the SIPS (Miller et al., 2002); some original items

were also added. The items are answered true/false

and sum to form four major subscales: 1) Positive

symptoms (e.g. unusual thinking and perceptual

abnormalities), 2) Negative symptoms (e.g. flat affect

and social isolation), 3) Disorganized symptoms (e.g.

odd behavior) and 4) General symptoms (e.g. depres-

sion and role functioning). Sample items include

bSometimes I think that people can read my mind,Qand bI tend to avoid social activities with other

people.QPresence of a bprodromal syndromeQ was diag-

nosed using the Criteria for Prodromal Syndromes

(COPS), as assessed by the SIPS. The COPS

identifies three bprodromal syndromesQ defined by:

1) Attenuated positive symptoms (APPS), 2) Brief

intermittent psychotic symptoms (BIPS) and 3)

Decline in role functioning and either a diagnosis of

schizotypal personality disorder or a first-degree

relative with psychotic disorder (Genetic Risk and

Deterioration Prodromal Syndrome, GRDPS). The

SIPS assesses for a psychotic syndrome using the

criteria defining the Presence of Psychotic Syndromes

(POPS). Diagnosis of a bpsychotic syndromeQ may

not correspond precisely to a DSM-IV psychotic

disorder diagnosis but, rather, refers to psychotic

symptoms of particular duration and intensity (Miller

et al., 2002). Interviewers also rate the same four

symptom domains assessed by the PQ on the Scale of

Prodromal Symptoms (SOPS).

2.2. Participants and assessments

Study participants were 158 individuals, aged 12–

35, who were referred to the Staglin Music Festival

Center for Assessment and Prevention of Prodromal

States (CAPPS) at UCLA because of suspected

prodromal psychotic symptoms. Participants were

referred to CAPPS from the following sources: the

UCLA NPI’s Division of Child and Adolescent

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125 119

Psychiatry (approximately 36% of referrals), school

counselors (28%), outpatient mental health clinicians

(27%), self-referral in response to advertisements

placed in local newspapers and newsletters at local

universities, community colleges and local chapters of

the National Alliance for the Mentally Ill (6%), or

another source, such as a current patient or person

familiar with CAPPS (3%). The majority of referrals

were made by mental health clinicians who had

previously seen a presentation from CAPPS staff that

described the Center and inclusion/exclusion criteria

for our study.

Participants referred to the CAPPS program con-

tacted the intake coordinator, who conducted a short

phone screen. The coordinator briefly described the

program and the types of qualifying symptoms, as

well as exclusion criteria (diagnoses of schizophrenia

or schizoaffective disorder, neurological disorder, IQ

less than 70, substance dependence or abuse related to

onset of symptoms). Callers were encouraged to

consider the relevance of these criteria to their case

and discuss any questions with the coordinator but

were not required to provide any information over the

phone. Callers were then offered the opportunity to

schedule an assessment. Once they arrived at CAPPS,

93% of patients completed the study procedures. For

the 7% (N =9) who were unable to complete study

procedures, this was most often due to the patient

being considered an acute danger to self or others,

and/or too disorganized to complete the interview,

leading to an inpatient hospitalization. A few cases

were deemed incapable of giving informed consent

due to the severity of symptoms at the time of

interview.

All clinical evaluations were completed by MA,

PhD, or MD level mental health specialists or by

other trained staff members with commensurate

extensive prior training in clinical interviewing. The

original CAPPS clinical interviewer team was trained

on the SIPS by Dr. Tandy Miller (Miller et al., 2003,

1999), one of the developers of the instrument.

During this training, raters completed a didactic

component followed by independently rating six

videotaped SIPS interviews. Certification requires

that interviewers achieve good to excellent reliability

(ICCs N .75 for all SOPS symptom ratings, and

kappas z .80 for prodromal syndrome diagnoses) on

SIPS training videotapes as compared with bgold

standardQ ratings developed by Dr. Miller and her

clinical team at Yale, as well as achieve this reliability

standard on three live interview cases, as compared to

ratings by experienced assessment supervisors. For

further quality assurance, members of a clinical

consensus team, which includes the Director of

Clinical Assessment (CEB) and other authors (RLL,

JKJ), independently completed SIPS ratings for the

first 12 CAPPS cases. Reliability analyses indicated

good to excellent inter-rater reliability regarding

individual symptoms (ICCsN .75), prodromal syn-

dromes (kappasz .90), and GAF scores (ICCsz .80).

.80). Any cases with diagnostic discrepancies were

brought to regular meetings of the clinical consensus

team or the Director of Clinical Assessment for

resolution. The study protocol and informed consent

procedures were approved by the UCLA Institutional

Review Board.

Participants (N =149) completed the PQ and at

least the positive symptom scale of the SIPS. Because

positive symptoms provide the basis for inclusion

diagnoses, the additional SIPS sections were not

administered to individuals who clearly did not meet

criteria for a prodromal syndrome when time was

limited. Interviews were conducted at the intake

assessment with the participant and a parent or other

informant, when available. Written informed consent

and/or child assent with parent consent was obtained

after the procedures had been fully explained.

Participants with an estimated IQ less than 70, a

neurological disorder, or serious substance use that

was temporally related to the onset of prodromal

symptoms were not included in the current analyses

(N =12). Of the remaining 113 subjects with complete

data, 37 met criteria for a present psychotic syndrome,

39 participants met criteria for a prodromal syndrome

(5 BIPS, 31 APPS, 1 GRDPS, 2 APPS and GRDPS),

and 37 did not meet criteria for either a prodromal or

psychotic syndrome as assessed by the SIPS.

Demographic characteristics of the sample are

presented in Table 1. As can be seen in the table,

the overall sample is predominantly (65%) male and

ethnically diverse. Age, gender, ethnicity, and parental

education did not differ significantly between diag-

nostic groups, although there was a trend toward

fewer males in the prodromal group, as compared to

both the psychotic group and cases who were neither

prodromal nor psychotic ( p =.06). PQ scores were not

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

False Positive Rate (1- Specificity)

Fig. 1. ROC curve: PQ positive symptoms predicting no SIPS

diagnosis vs. prodromal/psychotic diagnoses.

Table 1

Demographic characteristics

SIPS diagnosis F or v2 p

Psychotic

(N =37)

Prodromal

(N =39)

Neither

(N =37)

Age (years) 19.3F5.3 17.3F3.8 17.6F5.0 0.63 0.19

Male 76% 51% 72% 5.6 0.06

Caucasian 41% 46% 33% 1.3 0.52

Parental

education

(years)

15.4F2.6 15.9F3.9 14.3F3.7 1.83 0.17

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

False Positive Rate (1 - Specificity)

Fig. 2. ROC curve: PQ total symptoms predicting no SIPS diagnosis

vs. prodromal/psychotic diagnoses.

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125120

statistically related to age (r =.06, p =.53) or parental

education (r=� .10, p= .30).

2.3. Statistical analyses

Concurrent validity of the PQ against the SIPS

was examined in several ways. First, stepwise

logistic regression analyses were performed with

PQ scales as the independent variables and SIPS

diagnosis as the dependent variable, dichotomized

first as no SIPS diagnosis versus prodromal/psychotic

SIPS diagnoses, and then classified as no SIPS

diagnosis/prodromal versus psychotic syndrome SIPS

diagnosis. Next, receiver operating characteristic

(ROC) analyses were used to assess the accuracy of

the PQ against the SIPS. ROC curves plot the true

positive rate against the false positive rate for the

different possible cutpoints of a test, showing the

tradeoff between sensitivity and specificity (any

increase in sensitivity will be accompanied by a

decrease in specificity). The closer the curve follows

the left-hand border and then the top border of the

ROC space, the more accurate the test, while the

closer the curve comes to the 458 diagonal of the

ROC space, the less accurate the test. Test accuracy

is considered statistically significant when the con-

fidence interval for the area under the curve (AUC) is

greater than .50, indicating that the test predicts

diagnosis better than chance (Zweig and Campbell,

1993). Correlations of PQ scale scores with SOPS

scale scores also examined agreement between the

measures. Internal consistency of the PQ subscales

was measured using Cronbach’s alpha statistic.

Statistical analyses were performed using SAS

version 8.2 (SAS, 1999) and MedCalc for Windows,

version 7.4.1.0 (Medcalc, 2005).

3. Results

Stepwise logistic regression analysis with PQ

subscale scores as independent variables and SIPS

diagnosis as the dependent variable revealed a

significant effect of total PQ score on diagnosis for

prodromal/psychotic syndrome versus no SIPS diag-

nosis (v2=25.6, df =1, p b .0001). Positive symptom

score also significantly predicted SIPS diagnostic

status by logistic regression (v2 = 27.4, df = 1,

p b .0001; positive symptoms b =.14, p =.002). How-

ever, the other individual PQ subscale scores did not

Table 2

SIPS diagnostic classification accuracy by PQ scores

PQ score cutoff Sensitivity Specificity PPVa (%) NPVa (%) Positive LRa

Value (%) 95% CI Value (%) 95% CI

Outcome classification: SIPS diagnosis of prodromal/psychotic versus neither

PQ positive symptomsb z8 90 80–95 49 32–66 78 69 1.74

PQ positive symptomsc z14 71 60–81 81 65–92 89 58 3.76

PQ total symptomsb z19 91 82–96 38 23–55 75 67 1.5

PQ total symptomsc z36 70 58–80 84 68–94 90 57 4.3

Outcome classification: SIPS diagnosis of psychotic versus prodromal/neither

PQ positive symptomsc z17 68 50–82 63 51–74 47 80 1.83

PQ positive symptomsb z26 35 20–52 90 80–95 62 74 3.34

PQ total symptomsc z36 76 59–88 60 47–70 48 83 1.86

PQ total symptomsb z59 27 14–44 90 82–96 50 70 2.05

a PPV=positive predicative value, NPV=negative predictive value, LR=likelihood ratio.b PQ cutoff values chosen to result in 90% correct classification of SIPS prodromal cases.c PQ cutoff values at point that maximizes overall accuracy.

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125 121

significantly predict SIPS-based prodromal status

(negative symptoms b = .09, p = .21; disorganized

symptoms b =� .11, p =.36; general/affective symp-

toms b =� .02, p =.77).

Receiver operating characteristic (ROC) curves

were plotted for PQ positive symptom items to predict

no SIPS diagnosis versus SIPS prodromal/psychotic

syndrome diagnoses. The area under the curve was

significant for positive symptom score (See Fig. 1,

AUC=.79, SE= .042, 95% CI= .699–.857) and total

score (See Fig. 2, AUC= .77, SE= .044, 95%

CI= .685–.847). For use as a screening measure, PQ

prediction thresholds (8 or more positive symptom

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

False Positive Rate (1 - Specificity)

Fig. 3. ROC curve: PQ positive symptoms predicting no SIPS

diagnosis/prodromal vs. psychotic diagnosis.

items) were set to achieve 90% sensitivity, resulting in

49% specificity and good overall classification accu-

racy. Alternatively, a cutoff score (14 or more positive

symptom items) can be selected to balance sensitivity

(71%) and specificity (81%). Table 2 shows further

details of these analyses. The PQ was less sensitive to

the distinction between prodromal and psychotic

syndrome diagnoses. When assessing PQ classifica-

tion accuracy of SIPS psychotic syndrome diagnoses

versus prodromal/no SIPS diagnosis, the area under

the ROC curve (see Fig. 3) was significant for positive

symptom score (AUC=.67, SE= .056, 95% CI= .573–

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1False positive rate (1 - Specificity)

Fig. 4. ROC curve: PQ total symptoms predicting no SIPS

diagnosis/prodromal vs. psychotic diagnosis.

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125122

.753) and total score (See Fig. 4, AUC= .64,

SE= .057, 95% CI= .546–.729). Although statistically

significant, the classification accuracy was substan-

tially lower than that for PQ score prediction of no

SIPS diagnosis versus prodromal/psychotic syndrome

diagnoses and would be less than adequate in practice,

as outlined in the bottom portion of Table 2.

PQ subscales were moderately correlated with

corresponding SIPS scales for positive symptoms

(r(102)= .60, p b .0001), negative symptoms (r(57)=

.53, p b .0001), general symptoms (r(54) = .49,

p =.0001) and for the total scales (r(53)= .49, p=

.0002), with a low correlation for disorganized

symptoms (r(55)= .27, p =.039). Cronbach’s alpha

was .92 for positive symptoms (45 items), .88 for

negative symptoms, (19 items), .79 for disorganized

symptoms (13 items), .85 for general symptoms (15

items), and .96 for the total scale (92 items).

4. Discussion

The concurrent validity data indicate that the PQ

can distinguish participants with prodromal/psy-

chotic-syndrome SIPS diagnoses from those with no

SIPS diagnosis. As expected, given that SIPS

diagnoses are primarily defined by positive symp-

toms, the PQ positive symptom subscale showed the

greatest association with SIPS diagnosis. For use as a

screening instrument, we chose to maximize sensi-

tivity when differentiating prodromal/psychotic cases

from those with no SIPS diagnosis, thereby lowering

the rate of rejected true prodromal/psychotic cases.

Thus, the PQ would correctly reject approximately 5

out of 10 non-prodromal/non-psychotic participants

(i.e., a 50% reduction in time and resources spent

interviewing SIPS-negative referrals), at the cost of

excluding approximately one out of ten actual

prodromal/psychotic cases. Of note, approximately

one-third of patients referred to our clinic for

screening were found to be experiencing fully

psychotic syndromes, suggesting that referring clini-

cians may be unaware of the presence or severity of

psychotic symptoms in some patients they are

treating. Therefore, the PQ may be useful in

identifying young people who may otherwise go

undetected and, therefore, untreated for their psy-

chotic illness.

Differentiating psychotic participants from prodro-

mal subjects using the PQ resulted in less than

adequate classification accuracy. These results are

consistent with our expectation that a simple true–

false self-report screening measure would not be

sensitive to the threshold between the late prodromal

stage and full-blown psychosis. The detailed diag-

nostic definitions of psychosis provided by the SIPS

are based on the severity, frequency and duration of

symptoms, dimensions that cannot be efficiently

assessed with a brief screen and, thus, are not assessed

by the PQ.

Most PQ scales showed moderate correlations with

the corresponding SIPS scales, also supporting con-

current validity. The disorganized scale showed only a

modest correlation, likely because many disorganized

symptoms on the SIPS (e.g., odd behavior) are rated

by observation, while the PQ is based exclusively on

self-report. Internal consistency of the PQ scales was

high.

Currently, we propose that the PQ be used only for

the purpose of screening in clinical high-risk and early

psychosis research clinics and not outside of mental

health settings for the following reasons. First, the

concurrent validity statistics presented in this paper

were based on an benrichedQ sample with a very high

base rate of the target syndrome. Sensitivity would be

much weaker in samples with low base rates of true

prodromal/psychotic cases (O’Toole, 2000). Second,

the PQ does not specifically assess distress or treat-

ment-seeking behavior related to prodromal symp-

toms, although these factors are integral to the

prodrome construct. Rather, we expect to use the

screen in settings where individuals present for

treatment due to distressing symptoms that they or

their mental health clinician suspect may represent the

psychosis prodrome. All participants in this study

were treatment-seeking to some degree, whether they

wished to obtain assessment, psychotherapy, psychi-

atric services, or bsecond opinionsQ regarding psychi-

atric diagnoses, and most were already receiving

mental health services.

We emphasize distress and treatment-seeking in

defining psychotic and prodromal syndromes because

several recent studies have shown that psychotic

experiences commonly occur in the general popula-

tion outside of a diagnosable psychotic disorder. In an

unselected general population sample, 18.1% of adults

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125 123

reported some type of psychotic experience in their

lifetime that was not characterized by distress or

treatment-seeking, while only 1.5% of the sample was

diagnosed with psychotic disorder (Hanssen et al.,

2003). In another investigation, high school students

reported unusual psychotic-like experiences at rates

far beyond the prevalence of psychotic-spectrum

disorders (McGorry et al., 1995). Recently, we found

that when a large sample of undergraduate college

students completed the PQ, 43% of the sample scored

above the cutoff suggested in the present paper (8 or

more positive symptom items), although only 2%

rated at least 8 items as distressing, and a negligible

proportion of the sample reported previously seeking

mental health services (Loewy et al., in preparation).

These studies underscore the importance of distin-

guishing between unusual experiences and true

symptoms or disorder, which must be complemented

by distress and/or an effect on role functioning. A

version of the PQ modified to assess level of distress

and treatment seeking is currently under development

to investigate the possibility of wider screening in

non-clinical settings, such as schools and general

medical facilities.

The PQ is not diagnostic but, rather, is intended

for use as a tool to pre-select patients for more

intensive interviewing. We suggest a careful

approach when screening young people for prodro-

mal psychosis to minimize unnecessary stigma,

confusion, anxiety, and potential treatment with

antipsychotic medications (Peterson, 2000; Post,

2001; Rosen, 2000; Yung and McGorry, 1997).

Informing patients, their families and their clinicians

of the nature of their risk status in a clear and

understandable manner ensures that they receive

accurate information without undue alarm. All

parties should be informed that this screening

measure is not diagnostic and that it has a margin

of error. They can be encouraged to return for further

assessment if there is any future change in symptoms

and/or functioning, and they can be provided with a

description of psychotic symptoms to help track

potential worsening of symptoms. It is possible that

some apparently false negative cases on the PQ are

defined as bprodromalQ on the SIPS but will not

actually become fully psychotic. For included

patients (i.e., SIPS positive cases), the SIPS and

PQ are administered at multiple follow-up time

points, which will allow us to examine this question

in future analyses.

One limitation of the PQ is that all the items are

keyed in the btrueQ direction, thereby subjecting scoresto the effects of response bias, or the tendency to

respond in one direction. Response bias can be

influenced by personality factors such as agreeable-

ness, social desirability, demand characteristics of the

research environment, or cognitive processes engaged

when completing the questionnaire (Krosnick, 1999;

Peltier and Walsh, 1990). However, positive symptom

items on the PQ were constructed to describe specific

experiences of prodromal symptoms that are not

easily phrased in the negative direction, such as bIsometimes feel that things I see on television or read

in the newspaper have a special meaning for me.Q Inthat case, positively keyed items increase validity

because respondents are more likely to understand the

item accurately than when it is phrased in a negative

direction (Schrieshem and Hill, 1981). Therefore, a

positively keyed questionnaire seemed the best

approach to facilitate accurate responses.

Participants who were classified as neither pro-

dromal nor psychotic often reported symptoms con-

sistent with the negative and affective symptoms that

characterize the prodromal phase of psychosis, with-

out the positive symptoms sufficient to warrant a

prodromal or psychotic diagnosis according to the

COPS/POPS. However, the only diagnostic instru-

ment administered at intake was the SIPS. Because

the SIPS is a highly detailed, time-intensive interview

measure, staff costs and patient burden precluded the

completion of an additional clinical interview (e.g.,

SCID) with subjects already known to be ineligible

for the larger CAPPS study. Therefore, we are unable

to provide DSM-IV diagnoses for these excluded

subjects. Of the 39 patients diagnosed as putatively

prodromal at intake, 33 have agreed to follow-up

assessment within our clinic and 6 of these patients

have been followed for a full 2 years. Consequently,

we are not yet able to provide an accurate estimate of

the psychosis transition rate within our sample.

However, we expect to examine the predictive

validity of the PQ once a greater number of patients

in our sample have been followed for a sufficient

period of time.

Evaluation of sample characteristics is crucial in

determining the utility of a clinical measure such as

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125124

the PQ. Patients with attenuated prodromal syn-

dromes represent the largest portion of the validation

samples for the SIPS (Miller et al., 2003), the

CAARMS (Yung et al., 2003), and the present study

(85%), and may represent the prodromal patients

most likely to seek treatment (Davidson and Weiser,

2004). Therefore, the validity of the PQ as assessed

in this study refers largely to accurate identification

of attenuated prodromal syndromes. Definitions of

the prodrome are likely to change with time as

researchers refine the symptom criteria and identify

additional factors that predict psychosis (Knowles

and Sharma, 2004; Lencz et al., 2004; Yung et al.,

2004). As such, future versions of the PQ may

reflect those changes once their predictive power is

established. For now, the definition of the schizo-

phrenia prodrome used in the present study repre-

sents the best current attempts in the field to define

at-risk syndromes.

According to these preliminary results, the PQ

shows promise as a simple and cost-effective way to

screen for prodromal and previously undiagnosed

psychotic symptoms in clinic-referred populations.

While the PQ shows good agreement with concurrent

diagnoses of putatively bprodromalQ or psychotic

syndromes, its predictive validity has not yet been

investigated. The present findings indicate that further

validation studies of the PQ are warranted, and

therefore, data collection for the current study is

ongoing and expanding. These additional data will

allow us to accrue a larger validation sample size

adequate for factor analyses and item analyses. We

intend to use the results of these analyses to reduce the

number of items on the PQ in order to improve

efficiency of administration.

References

Addington, J., 2004. The diagnosis and assessment of individuals

prodromal for schizophrenic psychosis. CNS Spectr. 9 (8),

588–594.

Davidson, M., Weiser, M., 2004. Prodromal schizophrenia: the

dilemma of prediction and early intervention. CNS Spectr. 9 (8),

578.

Hanssen, M., Peeters, F., Krabbendam, L., Radstake, S., Verdoux,

H., van Os, J., 2003. How psychotic are individuals with non-

psychotic disorders? Soc. Psychiatry Psychiatr. Epidemiol. 38

(3), 149–154.

Heinssen, R.K., Cuthbert, B.N., Breiling, J., Colpe, L.J., Dolan-

Sewell, R., 2003. Overcoming barriers to research in early

serious mental illness: issues for future collaboration. Schizophr.

Bull. 29 (4), 737–745.

Knowles, L., Sharma, T., 2004. Identifying vulnerability markers in

prodromal patients: a step in the right direction for schizophre-

nia prevention. CNS Spectr. 9 (8), 595–602.

Krosnick, J.A., 1999. Survey research. Annu. Rev. Psychol. 50,

537–567.

Lencz, T., Smith, C., Auther, A., Correll, C., Cornblatt, B., 2004.

Nonspecific and attenuated negative symptoms in patients at

clinical high-risk for schizophrenia. Schizophr. Res. 68 (1),

37–48.

Loewy, R.L., Riese, S.R., Johnson, J.K., and Cannon, T.D., in

preparation. How unusual are unusual experiences? Self-report

of attenuated psychotic symptoms in a college population.

Manuscript submitted for publication.

McGorry, P.D., McFarlane, C., Patton, G.C., Bell, R., Hibbert,

M.E., Jackson, H.J., Bowes, G., 1995. The prevalence of

prodromal features of schizophrenia in adolescence: a prelimi-

nary survey. Acta Psychiatr. Scand. 92 (4), 241–249.

Miller, T.J., McGlashan, T.H., Woods, S.W., Stein, K., Driesen, N.,

Corcoran, C.M., Hoffman, R., Davidson, L., 1999. Symptom

assessment in schizophrenic prodromal states. Psychiatr. Q. 70

(4), 273–287.

Miller, T.J., McGlashan, T.H., Rosen, J.L., Somjee, L., Markovich,

P.J., Stein, K., Woods, S.W., 2002. Prospective diagnosis of the

initial prodrome for schizophrenia based on the Structured

Interview for Prodromal Syndromes: preliminary evidence of

interrater reliability and predictive validity. Am. J. Psychiatry

159 (5), 863–865.

Miller, T.J., McGlashan, T.H., Rosen, J.L., Cadenhead, K.,

Ventura, J., McFarlane, W., Perkins, D.O., Pearlson, G.D.,

Woods, S.W., 2003. Prodromal assessment with the

Structured Interview for Prodromal Syndromes and the

Scale of Prodromal Symptoms: predictive validity, interrater

reliability, and training to reliability. Schizophr. Bull. 29 (4),

703–715.

O’Toole, B.I., 2000. Screening for low prevalence disorders. Aust.

N. Z. J. Psychiatry 34, S39–S46 (Suppl.).

Peltier, B.D., Walsh, J.A., 1990. An investigation of response bias in

the Chapman Scales. Educational and Psychological Measure-

ment 50 (4), 803–815.

Peterson, D.R., 2000. The ethics of research into schizophrenia

prevention: a carer’s perspective. Aust. N. Z. J. Psychiatry 34,

S201–S203 (Suppl.).

Post, S.G., 2001. Preventing schizophrenia and Alzheimer disease:

comparative ethics. Schizophr. Res. 51 (1), 103–108.

Raine, A., 1991. The SPQ: a scale for the assessment of schizotypal

personality based on DSM-III-R criteria. Schizophr. Bull. 17 (4),

555–564.

Rosen, A., 2000. Ethics of early prevention in schizophrenia. Aust.

N. Z. J. Psychiatry 34, S208–S212 (Suppl.).

SAS Version 8.2 [Computer Software], 1999. SAS Institute Inc.,

Cary, NC.

Schoonjans, F., 2005. MedCalc (version 7.4.1.0) [Computer

Software]. MedCalc Software, Mariakerke, Belgium.

R.L. Loewy et al. / Schizophrenia Research 79 (2005) 117–125 125

Schrieshem, C.A., Hill, K.D., 1981. Controlling acquiescence

response bias by item reversals: the effect on questionnaire

validity. Educ. Psychol. Meas. 41, 1101–1114.

Yung, A.R., McGorry, P.D., 1996. The prodromal phase of first-

episode psychosis: past and current conceptualizations. Schiz-

ophr. Bull. 22 (2), 353–370.

Yung, A.R., McGorry, P.D., 1997. Is pre-psychotic intervention

realistic in schizophrenia and related disorders? Aust. N. Z. J.

Psychiatry 31 (6), 799–805.

Yung, A.R., Phillips, L.J., McGorry, P.D., McFarlane, C.A.,

Francey, S., Harrigan, S., Patton, G.C., Jackson, H.J., 1998.

Prediction of psychosis. A step towards indicated prevention of

schizophrenia. Br. J. Psychiatry 172 (33), 14–20 (Suppl.).

Yung, A.R., Phillips, L.J., Yuen, H.P., Francey, S.M., McFarlane,

C.A., Hallgren, M., McGorry, P.D., 2003. Psychosis prediction:

12-month follow up of a high-risk (bprodromalQ) group.

Schizophr. Res. 60 (1), 21–32.

Yung, A.R., Phillips, L.J., Yuen, H.P., McGorry, P.D., 2004. Risk

factors for psychosis in an ultra high-risk group: psychopathol-

ogy and clinical features. Schizophr. Res. 67 (2–3), 131–142.

Zweig, M.H., Campbell, G., 1993. Receiver-operating characteristic

(ROC) plots: a fundamental evaluation tool in clinical medicine.

Clin. Chem. 39 (4), 561–577.