Académique Documents
Professionnel Documents
Culture Documents
Schizophrenia Research
journal homepage: www.elsevier.com/locate/schres
National Institute for Health and Welfare, Department of Mental Health and Substance Abuse Services, P.O. Box 30, FI-00271 Helsinki, Finland
University of California, San Francisco, Department of Psychiatry, Box 0984 PAR, 401 Parnassus Ave., LangPorter, San Francisco, CA 94143, United States
Yale University, Department of Psychology, Box 208205, 2 Hillhouse Ave., New Haven, CT 06520-8205, United States
a r t i c l e
i n f o
Article history:
Received 27 February 2014
Received in revised form 18 June 2014
Accepted 24 June 2014
Available online 22 July 2014
Keywords:
Adolescent
Psychosis prodrome
Follow-up
Self-report instrument
Screening
a b s t r a c t
The Prodromal Questionnaire (PQ) identies psychiatric help-seekers in need of clinical interviews to diagnose
psychosis risk. However, some providers use the PQ alone to identify risk. Therefore, we tested its predictive
utility among 731 adolescent psychiatric help-seekers, with a 39-year register-based follow-up. Nine latent
factors corresponded well with postulated subscales. Depersonalization predicted later hospitalization with a
psychosis diagnosis (HR 1.6 per SD increase), and Role Functioning predicted any psychiatric hospitalization
(HR 1.3). Published cut-off scores were poor predictors of psychosis; endorsement rates were very high for
most symptoms. Therefore, we do not recommend using the PQ without second-stage clinical interviews.
2014 Elsevier B.V. All rights reserved.
1. Introduction
Psychotic disorders often develop gradually, and most patients have a
health care contact due to psychiatric symptoms before the onset of psychosis (Anderson et al., 2013). Initial psychiatric care should thus be an
excellent opportunity for detecting such a disease course. Though structured interviews focused on the early detection of psychoses can be
employed, they are time consuming and require special training. Consequently, several specic questionnaires have been constructed for the
screening of psychotic symptoms (e.g. Heinimaa et al., 2003; Ord et al.,
2004; Liu et al., 2013). These questionnaires can select patients for
targeted interviews. Some of the new instruments have been validated
against gold-standard interview methods for detecting incipient psychosis (Loewy et al., 2005), but research on prospective predictive value is
still scant. Despite this, some providers use the screeners without referring for clinical interview. Therefore, we sought to test the accuracy of
the Prodromal Questionnaire (PQ) in predicting psychosis on its own.
The rst visit to a non-specic psychiatric outpatient clinic is an
optimal situation for testing the utility of such inventories, balancing
base risk and volume in a situation acceptable to the patients. Accordingly,
http://dx.doi.org/10.1016/j.schres.2014.06.031
0920-9964/ 2014 Elsevier B.V. All rights reserved.
As models of up to nine Oblimin-rotated factors were readily interpretable, this number of factors was retained (RMSEA 0.014, CFI .99). Factor
loadings, thresholds, and factor correlations are shown in Supplementary
Table 1.
KaplanMeier curves of hospitalization with a psychosis diagnosis
and any psychiatric hospitalization are depicted in Fig. 1, by gender.
After three years, which was the shortest follow-up time, 3.6% of
females and 7.2% of the males had been hospitalized for psychosis.
Cox regression results are reported in Table 1. Four predictors of hospitalization with a psychosis diagnosis were statistically signicant
individually (in order of effect size): Depersonalization, PQ Total sum
score, Role Functioning, and Dysphoria. When the strongest predictor
Depersonalization (HR 1.6, 95% CI 1.22.2, p = 0.005) was accounted
for, none of the other variables offered signicant improvements to
the model (p = .94, .14 and .35, respectively). Five predictors of any
psychiatric hospitalization were statistically signicant individually:
a) Psychosis
4. Analyses
No PQ item had more than 0.6% missing responses, and the overall
missingness rate was 0.2%. The means (SDs) for the PQ Positive subscale
sum score and PQ Total sum score were 10.6 (7.8) and 31.3 (18.7),
respectively.
A conrmatory factor analysis of the PQ indicated that the a
priori four-dimensional model showed only a moderately good t
(RMSEA = 0.04, CFI b .90). Therefore, to determine the empiric factorial structure of the PQ, exploratory factor analysis of the 731 response
sets was conducted with the WLSMV algorithm, Oblimin rotation, and
default parameters in Mplus version 7.11 (Muthn and Muthn,
2012). In this model, a response threshold parameter is calculated for
each item, and one factor loading parameter for each item-factor combination is estimated from the tetrachoric item correlations. The model
was computed with an increasing number of dimensions, until there
was no improvement in RMSEA.
Symptom factor scores and the a priori PQ Total and Positive Symptom subscale sum scores were the hypothesized predictors in Cox proportional hazards models of a) any psychiatric hospitalization (n =
120) and b) hospitalization with a psychosis diagnosis (n = 41) during
the individual's full follow-up time. Predictors were rst used singly,
and those that were signicant at the p = .01 level individually were included in a forward-stepping Cox model with the same p = .01 criterion
for entry. For comparability of coefcient estimates, all factor scores and
PQ sum scores were normalized before survival analyses. Age was used
as a covariate in all analyses. Due to a larger baseline psychosis and hospitalization risk among boys in our sample, all survival analyses were
conducted with gender as a stratum.
In addition, for facilitating comparison with previously published
results, we assessed the one-year predictive values for psychosis of
the previously proposed (Loewy et al, 2007) cut-offs for the Total and
Positive Symptom subscale sum scores.
b) Psychiatric hospitalization
5. Results
The multidimensional WLSMV latent factor models of the PQ provided
improving t using RMSEA, CFI, WRMR indices with up to 10 dimensions.
Fig. 1. Survival curves for psychosis and psychiatric hospitalization outcomes in register
follow-up, by gender. a) Psychosis and b) psychiatric hospitalization.
95% CI for HR
Lower
Upper
1.45
1.28
1.49
1.23
0.97
1.26
1.24
1.04
1.41
1.60
1.37
1.04
0.93
1.08
0.89
0.70
0.92
0.91
0.75
1.00
1.16
0.99
2.02
1.75
2.07
1.70
1.34
1.73
1.68
1.43
1.99
2.23
1.89
0.027
0.130
0.017
0.206
0.864
0.156
0.174
0.821
0.047
0.005
0.056
1.28
1.20
1.32
1.16
1.18
1.08
1.24
1.17
1.18
1.23
1.17
1.07
1.01
1.11
0.97
0.99
0.91
1.05
1.00
0.98
1.03
0.98
1.53
1.43
1.58
1.37
1.39
1.29
1.47
1.38
1.43
1.47
1.39
0.007
0.035
0.002
0.097
0.059
0.355
0.013
0.057
0.080
0.024
0.085
a) Psychosis, n = 731.
PQ Total Score
PQ Positive Symptoms
F1 Role Functioning
F2 Delusional Ideation
F3 Hallucinations
F4 Odd
F5 Social Avoidance
F6 Magical Thinking
F7 Dysphoria
F8 Depersonalization
F9 Anhedonia
b) Hospitalization, n = 654
PQ Total Score
PQ Positive Symptoms
F1 Role Functioning
F2 Delusional Ideation
F3 Hallucinations
F4 Odd
F5 Social Avoidance
F6 Magical Thinking
F7 Dysphoria
F8 Depersonalization
F9 Anhedonia
a
HR is per standard deviation increase of the factor. Predictors with p b0.05 are marked
in bold.
Role Functioning, PQ Total sum score, Positive Subscale sum score, Social Avoidance, and Depersonalization. When Role Functioning (HR
1.3, 95% CI 1.11.6, p = 0.002) was accounted for, none of the other
variables improved the model (p = .38, .30, .18, and .22, respectively).
At the one-year follow-up, the Positive Symptom subscale score
criterion of 14 or more provided a sensitivity of 48% and a specicity of
68% for predicting psychosis, with a 5% PPV and 97% NPV. The Total
Score criterion of 36 or more had a 64% sensitivity and a 57% specicity,
with a 5% PPV and 98% NPV. Adjusting cutoffs or including only the
items of the shortened PQ-B (Loewy et al., 2011) did not signicantly improve results.
6. Discussion
In the exploratory factor analysis, we found interpretable latent
factors, demonstrating structural validity of the PQ. Of the measured
constructs, Depersonalization, Role Functioning, and Dysphoria predicted later hospitalization for psychosis, though none of the others improved prediction after Depersonalization was taken into account. That
Depersonalization was the strongest predictor in questionnaire-based
psychosis screening is to our knowledge a novel nding, which has
been anticipated in theoretical work on the hallucination-predisposing
effects of loss of atmospheric cues (Graux et al., 2012). Nevertheless,
the Hazard Ratio of 1.6 per SD increase in Depersonalization is insufcient for effective screening, at least when used without a second
stage interview.
Also the predictive value of previously suggested sum score cut-offs
intended for two-stage screening was poor; the rate of false positives to
true positives was approximately twenty-fold. In sum, our results indicate that the instrument should only be used as originally intended,
i.e. for selecting respondents to secondary screening with a structured
interview.
The version of the instrument used in this study was unfortunately
limited to a Yes/No response format, where response bias and item
comprehension play a large part. The revisions coding distress and frequency (Loewy et al., 2007) may thus be preferable. Notably, a version
of the PRIME Screen (Kobayashi et al., 2008), which used a sevenpoint agreement scale and symptom duration, achieved fairly good predictive power in a similar study with a 6-month follow-up.
Our results do not support the use of the PQ for strict screening
of psychosis risk, though it provides an easy way to discuss and uncover psychotic-like experiences in a young patient; these experiences
deserve attention in their own right, especially when associated with
distress (Yung et al., 2006). Our ndings show that psychotic-like experiences also predict a general functional outcome as assessed by
psychiatric hospitalization; however, the added value was small after
general Role Functioning was accounted for.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.schres.2014.06.031.
Role of funding sources
Funding for this study was provided by the Academy of Finland Research Programme
Strategy Grant 115479 and a personal grant to S.T. by the Finnish Graduate School in
Psychiatry. Neither of these institutions had any further role in the study design, in the
collection, analysis, or interpretation of data, in the writing of the report, or in the decision
to submit the paper for publication.
Contributors
T.D.C. designed the study and wrote the protocol. All authors collaborated in designing
the data collection or analyses. S.T. undertook the statistical analyses and wrote the rst
manuscript draft. All authors contributed to and have approved the nal manuscript.
Conict of interest
All authors declare that they have no conicts of interest.
Acknowledgment
We thank all of our respondents and our collaborators at the Hospital District of
Helsinki and Uusimaa for making this study possible.
10
20. Things that I see appear different from the way they usually do
(brighter, duller, larger, smaller, or changed in some other way).
21. I tend to be very quiet and keep in the background on social
occasions.
22. People sometimes stare at me because of my odd appearance.
23. I wander off the topic or ramble on too much when I am speaking.
24. I believe in telepathy, psychic forces, or fortune-telling.
25. I often feel that others have it in for me.
26. My sense of smell sometimes becomes unusually strong.
27. Sometimes I have felt that I'm not in control of my own ideas or
thoughts.
28. I have been feeling unhappy or depressed lately.
29. Everyday things affect me more than they used to.
30. I believe that I am especially important or have abilities that are out
of the ordinary.
31. Other people think that I am a little strange.
32. Sometimes my thoughts seem to be broadcast out loud so that other
people know what I am thinking.
33. I often feel that I have nothing to say or very little to say.
34. I am unusually sensitive to noise.
35. I am superstitious.
36. I have heard my own thoughts as if they were outside of my head.
37. I have trouble focusing on one thought at a time.
38. I often feel that other people are watching me or talking about
me.
39. I get very nervous when I have to make polite conversation.
40. People comment on my unusual mannerisms and habits.
41. I am less interested in school or work lately.
42. I nd it hard to be emotionally close to other people.
43. I tend to avoid social activities with other people.
44. I feel very guilty.
45. I am an odd, unusual person.
46. I sometimes feel that things I see on television or read in the newspaper have a special meaning for me.
47. My moods are highly changeable and unstable.
48. I have been unable to enjoy things that I used to enjoy.
49. My thinking feels confused, muddled, or disturbed in some way.
50. Sometimes I feel suddenly distracted by distant sounds that I am not
normally aware of.
51. Recently, I have begun talking to myself.
52. I have had the sense that some person or force is around me, even
though I could not see anyone.
53. I am in danger of failing out of school, or have been red from my job.
54. I have some eccentric (odd) habits.
55. At times I worry that something may be wrong with my mind.
56. I have felt that I don't exist, the world does not exist, or that I am dead.
57. I have been confused at times whether something I experienced
was real or imaginary.
58. People nd me aloof and distant.
59. I tend to keep my feelings to myself.
60. I have experienced unusual bodily sensations (tingling, pulling,
pressure, aches, burning, cold, numbness, shooting pains, vibrations
or electricity).
61. I hold beliefs that other people would nd unusual or bizarre.
62. People say that my ideas are strange or illogical.
63. I feel worthless.
64. I feel that parts of my body have changed in some way, or that parts
of my body are working differently than before.
65. My thoughts are sometimes so strong that I can almost hear them.
66. I am not very good at returning social courtesies and gestures.
67. I sometimes see special meanings in advertisements, shop
windows, or in the way things are arranged around me.
68. I often pick up hidden threats or put-downs directed at me in what
people say or do.
69. I sometimes use words in unusual ways.
70. I am often angry, easily irritated or offended.