Differential impact of isolated psychotic symptoms on treatment outcome of major depressive disorder in the STAR*D cohort of Whites, Blacks and Latinos
OBJECTIVE: To determine whether isolated psychotic symptoms are more likely to be endorsed by depressed Latinos as opposed to other ethnic-racial groups; whether these symptoms affect Latinos similarly to other ethnic-racial groups in terms of treatment response; and whether they are more likely to be associated with anxiety disorders in depressed Latinos. METHODS: We analyzed data from STAR⁎D subjects who self identified as White, Black, or Latino. Rates of isolated psychotic symptoms were assessed by the self-rated Psychiatric Diagnostic Screening Questionnaire (PDSQ) and compared between ethnic-racial groups. Depressive remission outcomes were compared within each ethnic-racial group between subjects who endorsed psychotic symptoms versus no psychotic symptoms. Associations between isolated psychotic symptoms and anxiety disorders were also examined. RESULTS: Among 2597 eligible subjects with at least one post-baseline assessment and available PDSQ data excluding first-rank symptoms, the prevalence of auditory-visual hallucination was 2.5% in Whites (n=49/1928), 11.3% in Blacks (n=45/398) 6.3% in Latinos (n=17/270) (χ(2)=64.9; df=2; p<0.001). Prevalence of paranoid ideation was 15.5% in Whites (n=299/1927), 31.5% in Blacks (n=126/400), and 21.1% in Latinos (n=57/270) (χ(2)=57.3; df=2; p<0.001). Among Whites and Blacks but not Latinos, depressive remission rates were worse in subjects with auditory-visual hallucinations compared to those without them. Paranoid ideation had a significant negative impact on remission in Whites only. In all ethnic-racial groups, a significant association was found between auditory-visual hallucinations and PTSD and panic disorder. LIMITATIONS: The STAR*D study did not include any structured clinician-based assessment of psychotic symptoms. CONCLUSION: Latinos do not appear to have worse outcomes when treated for MDD with auditory-visual hallucinations, differently from Whites and Blacks.
Differential impact of isolated psychotic symptoms on treatment
outcome of major depressive disorder in the STAR*D cohort of Whites,
Blacks and Latinos
, Trina Chang
, Nhi-Ha Trinh
, Lee Baer
, Maurizio Fava
Depression Clinical and Research Program, Department of Psychiatry, Massachusetts General Hospital, Boston, MA 02114, United States
North Suffolk Mental Health Association, Chelsea Counseling Center, 301 Broadway, Chelsea, MA 02128, United States
Received 29 December 2012
Accepted 1 February 2013
Available online 13 March 2013
Major depressive disorder
Objective: To determine whether isolated psychotic symptoms are more likely to be endorsed by
depressed Latinos as opposed to other ethnic–racial groups; whether these symptoms affect Latinos
similarly to other et hnic–racial groups in terms of treatment response; and whether they are more
likely to be associated with anxiety disorders in depressed Latinos.
Methods: We analyzed data from STAR*D subjects who self identiﬁed as White, Black, or Latino. Rates
of isolated psychotic symptoms were assessed by the self-rated Psychiatric Diagnostic Screening
Questionnaire (PDSQ) and compared between ethnic–racial groups. Depressive remission outcomes
were compared within each ethnic–racial group between subjects who endorsed psychotic symptoms
versus no psychotic symptoms. Associations between isolated psychotic symptoms and anxiety
disorders were also examined.
Results: Among 2597 eligible subjects with at least one post-baseline assessment and available PDSQ
data excluding ﬁrst-rank symptoms, the prevalence of auditory–visual hallucination was 2.5% in
Whites (n¼ 49/1928), 11.3% in Blacks (n¼ 45/398) 6.3% in Latinos (n¼ 17/270) (
¼ 64.9; df¼ 2;
po 0.001). Prevalence of paranoid ideation was 15.5% in Whites (n¼ 299/1927), 31.5% in Blacks
(n¼ 126/400), and 21.1% in Latinos (n¼ 57/270) (
¼ 57.3; df¼ 2; po 0.001). Among Whites and Blacks
but not Latinos, depressive remission rates were worse in subjects with auditory–visual hallucinations
compared to those without them. Paranoid ideation had a signiﬁcant negative impact on remission in
Whites only. In all ethnic–racial groups, a sig niﬁcant association was found between auditory–visual
hallucinations and PTSD and panic disorder.
Limitations: The STAR*D stud y did not include any structured clinician-based assessment of psychotic
Conclusion: Latin os do not appear to have worse outcomes when treated for MDD with auditory–visual
hallucinations, differently from Whites and Blacks.
& 2013 Elsevier B.V. All rights reserved.
Previous research has shown that endorsement of psychotic
symptoms in subjects diagnosed with non-psychotic major
depressive disorder (MDD) is associated with poor outcome with
antidepressant treatment (Perlis et al., 2010). This ﬁnding
resulted from an analysis conducted on a diverse pool of patients,
which included a majority of Whites as well as Blacks, Latinos and
other ethnic–racial groups. The clinical relevance of the ﬁnding is
clear: for instance, depressed patients who might not appear
frankly psychotic but endorse features on the psychotic spectrum,
such as suspiciousness, might be less likely to respond to treat-
ment (Van Os, 2003; Sbrana et al., 2005; Shevlin et al., 2007).
There is a growing evidence that Lati nos disproporti onately
endorse isolated psychotic symptoms, and the actua l psychotic nature
of their experience may be questioned (Lewis-Fernandez et al., 2009;
Vega et al., 2006; Olfson et al., 2002; Minsky et al., 2003; Geltman and
Chang, 2004). In fact, putative psychotic symptoms in Latinos could
turn out to be psychosis-like manifestations, with noticeably different
phenomenological presentations (Cassano et al., 2012). Clinicians
who lack cultural training or experience with this population might
Contents lists available at ScienceDirect
journal hom epage: www.elsevier.com/locate/jad
Journal of Affective Disorders
0165-0327/$ - see front matter & 2013 Elsevier B.V. All rights reserved.
Correspondence to: Depression Clinical and Research Program,
Department of Psychiatry, Massachusetts General Hospital, 1 Bowdoin Square, 6th
Floor, Boston, MA 02114, United States. Tel.: þ 1 617 643 9622;
fax: þ 1 617 724 3028.
E-mail addresses: email@example.com,
firstname.lastname@example.org (P. Cassano).
Journal of Affective Disorders 150 (2013) 578–584
misinterpret these manifestations as genuine psychotic symptoms.
The potential for misinterpretati on is even greater when the patient
endorses these symptoms on a self-report rating scale, because this
format does not allow exploratory follow-up questions aimed at
characterizing the psychotic nature of the experience. Examples of
psychosis-like symptoms include hearing ‘‘someone calling’’, ‘‘knock-
ing at the door’’ or the ‘‘phone ringing in the house’’ and then realizing
it was a false impression, after checking or after asking others.
In this report we investigated whether Latinos in the
Sequenced Treatment Alternatives to Relieve Depression cohort
(STAR*D cohort) were more likely to endorse isolated psychotic
symptoms (presumably psychosis-like symptoms) compared to
other ethnic–racial groups. We also examined the impact of
isolated psychotic symptoms on treatment outcome of depression
in Whites, Blacks and Latinos. In addition, we studied the
association of isolated psychotic symptoms with PTSD and other
comorbid anxiety disorders in the same three ethnic–racial
groups. We hypothesized that isolated psychotic (presumably
psychosis-like) symptoms would be more prevalent in Latinos
(Olfson et al., 2002; Minsky et al., 2003) and that their impact on
treatment outcome of depression would be less prominent in
Latinos. Finally, we predicted that isolated psychotic symptoms
would be associated with trauma and PTSD (Lewis-Fernandez
et al., 2010), and possibly with other anxiety disorders, at higher
rates in Latinos than in other ethnic–racial groups (Olfson et al.,
2002; Vega et al., 2006; Lewis-Fernandez et al., 2009).
Study aims and analyses for this new examination of the
STAR*D cohort database were derived from an initial consensus
among the authors (PC, DM, TC, NHT, LB). The NIMH reviewed our
proposal and approved it prior to study initiation.
2.1. STAR*D cohort
STAR*D is an NIMH-sponsored multicenter trial on non-psychotic
MDD that involved 14 regional centers and 41 clinical sites across the
United States. The goal of the parent trial was to compare different
treatment strategies such as antidepressant augmentation and anti-
depressant switch, pharmacological augmentation and cognitiv e-
behavioral therapy for the treat ment of resistant depression. The
treatment algorithm did not include any antipsychotic medication.
Subjects were diagnosed with MDD by clinical interview and could
enter the study if their depressive severity was at least moderate
(as indicated by a Hamilton Depression Scale-17 (HAM-D
14 or higher). Lifetime psychosis, mania or hypomania was exclu-
sionary, as was current obsessive–compulsive disorder or any eating
disorder. Sub stance or alcohol use disorders were allowed unless
acute treatment interventions were requir ed. Intolerance or resis-
tance to any of the medications used in the level 1 or level 2 of the
trial was exclusionary; therefore subjects who had been already
unsuccessfully medicated for their current depressive episode were
likely to be excluded. Subjects who were pregnant or intended to
conceive were excluded as well. For more details on the STAR*D
original trial, please refer to Rush et al. (2004).
2.2. Isolated psychotic symptoms
We used the Psychiatric Diagnostic Screening Questionnaire
(PDSQ) to deﬁne the concept of isolated psychotic symptoms,
such as auditory–visual hallucinations or paranoid ideation
(Zimmerman and Chelminski, 2006). The PDSQ is a self-report
scale of psychiatric symptoms used as a screening tool for multi-
ple DSM-IV axis-I disorders; the scale includes a section for
psychosis as well. For every disorder explored by the PDSQ, there
are published sensitivity/speciﬁcity cut-off scores. The scale was
completed by each subject at study entry.
For the present study, we selected three of the six items
explored by this scale for psychosis. To assess auditory–visual
hallucinations, we selected subjects who endorsed: ‘‘Did you hear
voices that other people did not hear, or see things that other
people did not see?’’ To assess paranoid ideation, we selected
subjects who endorsed at least one of the following items: ‘‘Were
you convinced that other people were watching you, talking about
you, or spying on you?’’or ‘‘Did you think that you were in danger
because someone was plotting to hurt you?’’ The scale explored the
presence of the aforementioned symptoms in the prior 2 weeks.
Subjects who endorsed symptoms of the psychotic domain of
the PDSQ that suggested ﬁrst-rank Schneiderian symptoms were
excluded from our analyses (namely: ‘‘Did you think that some
force or power from the outside was controlling your body or
mind?’’ and ‘‘Did you think that you had special powers other
people did not have?’’). In fact our goal was to select subjects who
speciﬁcally endorsed auditory or visual hallucinations or paranoid
ideation. These manifestations are in fact more likely (as opposed
to ﬁrst-rank symptoms) to be endorsed by subjects with psychosis-
like manifestations but not frank psychosis (Cassano et al., 2012).
The only additional item of the PDSQ domain that we did not
include was: ‘‘Did things happen that you knew were true, but
other people told you that were your imagination?’’ We antici-
pated that some psychosis-like symptoms might be at times
normalized by family and friends, who might offer culturally-
bound explanations to the subject who experiences them. For
instance, in the case of a deceased family member calling, some-
one might validate the experience by saying that the deceased
family member is trying to communicate with the living. Due to
the potential risk of false negatives, and consequent underestima-
tion of the psychotic or psychosis-like symptoms, this item was
not included in our deﬁnition of isolated psychotic symptoms.
2.3. Study sample
From the original cohort of 4041 STAR*D subjects, we selected
those who deﬁned themselves by only one racial–ethnic back-
ground, speciﬁcally white non-Hispanic (Whites), black non-
Hispanic (Blacks) and white and Hispanic (Latinos). Subjects also
needed to have completed the items of the psychotic domain of
the PDSQ considered in our study (items 77–81 of the STAR*D
version of the PDSQ). Given that we also sought to evaluate the
impact of isolated psychotic symptoms on treatment outcomes,
we included only subjects who had at least one follow-up
assessment of their depression (HAM-D
) after the baseline visit.
The assessments of depressive symptoms used for our analyses
and for the selection of the study sample were carried out by
phone interviews by research outcome assessors (ROA) who were
blinded to the treatment assignment. Subjects were excluded
from this analysis if they endorsed more than one race, did not
complete the PDSQ scale, could not be reached by the ROA, or
were lost to follow-up after study screen.
Our analyses addressed three scientiﬁc questions:
(1) Are isolated psychotic symptoms more likely to be endorsed by
depressed Latinos as opposed to other ethnic–racial groups?
(2) Do isolated psychotic symptoms affect depressed Latinos in
the same way as other ethnic–racial groups in terms of
clinical outcome and treatment response?
P. Cassano et al. / Journal of Affective Disorders 150 (2013) 578–584 579
(3) Are isolated psychotic symptoms more likely to be associated
with PTSD and/or other anxiety disorders in depressed Latinos
than in other ethnic–racial groups?
We carried out descriptive analyses to characterize the sample.
The same comparisons were performed separately in each ethnic–
racial group. Baseline variables, including demographics, index
episode duration, and course of illness, were compared in subjects
who did or did not endorse isolated psychotic symptoms. Endor-
sement of auditory–visual hallucinations and endorsement of
paranoid ideation were tested separately for their correlations
with other variables of interest. Continuous variables were
compared by the independent samples t-test for equality of
means (p-value at 0.05 for 2-tailed signiﬁcance with equal
variance assumed); categorical (dichotomous) variables were
compared by chi-square (2 2 tables with p-value set at 0.05
for exact signiﬁcance, 2-sided). Several demographic variables
were derived from the STAR*D database and dichotomized
for presenting the baseline characteristics of the three com-
parison pairs. Current marital status was deﬁned as ‘‘married or
cohabiting with partner’’ versus ‘‘single, separated, divorced or
widowed’’. Current student status included both part-time and
full-time studies. Similarly, current employment status included
part-time, full-time as well as self-employment versus ‘‘unem-
ployed or retired not-working’’. Language preference, English or
Spanish, was deﬁned at screening based on clinic and phone
interviews and based on self-rating scales. For each subject,
information on language had to be available on at least half of
the screening assessments and language preference was coded
only when all available assessments were concordant. For exten-
sive description and direct comparison of the baseline character-
istics of the three ethnic–racial groups, as a whole, refer to Lesser
et al. (2007).
To compare prevalence of isolated psychotic symptoms across
the three aforementioned ethnic–racial groups, a chi-square
analysis was conducted ﬁrst for endorsement of auditory–visual
hallucinations and then for paranoid ideation (2 3 table with p-
value set at 0.05 for exact signiﬁcance, 2-sided). Secondary chi-
square analyses were performed to determine signiﬁcant differ-
ences between individual groups (2 2 tables with p-value set at
0.05 for exact signiﬁcance, 2-sided).
The impact of isolated psychotic symptoms on depressive
outcome was assessed separately for auditory–visual hallucina-
tions and for paranoid ideation. Three sets of Cox regression
models were constructed separately for each of the three groups
of interest: Whites, Blacks and Latinos. Presence of the isolated
psychotic symptom was deﬁned as the independent variable
while depressive remission was set as the dependent variable
(p-value at 0.05 for signiﬁcance). Tables were plotted to show
separate curves indicating the likelihood of persistent depressive
symptoms overtime in subjects who did or did not endorse the
speciﬁc psychotic symptom of interest. Depressive remission was
deﬁned as a HAM-D
of 7 or less at the ROA assessment. The
timeframe to remission could span 48 weeks or more, since
4 sequential levels of treatment for depression were offered in
STAR*D, each of which could last up to 12–16 weeks.
In subsequent exploratory analyses, the associations between
auditory–visual hallucinations and anxiety disorders (PTSD, panic
disorder, agoraphobia), trauma, and derealization were tested by
binary logistic regression (with p-value set at 0.05 for exact
signiﬁcance) and chi-square (2 2 tables with p-value set at
0.05 for exact signiﬁcance, 2-sided). No corrections for multiple
comparisons were used, given the exploratory nature of the
analyses. PTSD, panic disorder, and agoraphobia were deﬁned
according to the relative domains of the PDSQ scale. Published
cut-offs with high speciﬁcity (89–91%) were used for case
deﬁnition (Zimmerman and Chelminski, 2006).We chose to test
for PTSD because there is a high index of suspicion for a
preferential association of auditory–visual hallucinations and
PTSD in Latinos (Lewis-Fernandez et al., 2010). We included other
common anxiety disorders such as panic disorder and agorapho-
bia because they are frequently associated with psychosis-like
symptoms in Latinos (Olfson et al., 2002 ; Vega et al., 2006; Lewis-
Fernandez et al., 2009). We intentionally excluded generalized
anxiety disorder due to its high overlap with MDD. Presence of
trauma was assessed by the endorsement of the following PDSQ
items 29 or 30: ‘‘have you ever experienced (or witnessed) a
traumatic event such as combat, rape, assault, sexual abuse, or
any other extremely upsetting event?’’ Similarly, PDSQ item 39
was used as a proxy of derealization: ‘‘did you feel distant
and cutoff from other people because of having experienced a
3.1. Prevalence of isolated psychotic symptoms
Of the 4041 STAR*D subjects 3749 identiﬁed themselves
exclusively as Whites (n¼ 2614), Blacks (n ¼ 694) or Latinos
(n¼ 441). Among them, 3707 had completed the PDSQ items of
the psychotic domain for the assessment of the prevalence of
isolated psychotic symptoms, and 3538 were included after ﬁrst-
rank psychotic symptoms were ruled out. Finally, only 2597 of
these subjects had at least one follow-up assessment of their
depressive symptoms (HAM-D
) by an independent evaluator.
The prevalence of auditory–visual hallucinations was 2.5% in
Whites (n¼ 49/1928), 11.3% in Blacks (n¼45/398), and 6.3% in
Latinos (n¼ 17/270) (
¼ 64.9; df¼ 2; po 0.001). Signiﬁcantly
greater prevalence was found in Blacks compared to Latinos
¼ 4.8; df¼ 1; p¼0.030) and in Latinos compared to Whites
¼ 11.5; df¼ 1; p¼ 0.002). The prevalence of paranoid ideation
was 15.5% in Whites ( n¼ 299/1927), 31.5% in Blacks (n¼ 126/400),
and 21.1% in Latinos (n¼57/270) (
¼ 57.3; df¼ 2; po 0.001).
As with our ﬁndings for hallucinations, the prevalence of paranoid
ideation was signiﬁcantly higher in Blacks than Latinos (
df¼ 1; p¼0.003) and in Latinos versus Whites (
¼ 5.5; df¼ 1;
p¼ 0.022). When the prevalence of the same psychotic symptoms
was tested in the larger sample of 3538 STAR*D subjects,
independent of outcome data, ﬁndings were similar (data not
shown).Tables 1 and 2 show the demographic characteristics of
subjects who did versus did not endorse isolated psychotic
symptoms, broken down by race/ethnicity.
3.2. Isolated psychotic symptoms and treatment outcome
Rates of depressive remission were worse in White and Black
subjects with auditory–visual hallucinations compared to those
without them (rates in Whites: 39% versus 56%; n¼ 1928;
O.R.¼ 1.70; CI
: 1.08–2.68; p¼ 0.022; rates in Blacks: 22% versus
42%; n¼ 398; O.R.¼ 2.11; CI
: 1.11–4.02; p¼ 0.023) (Figs. 1 and 2).
Conversely, Latinos had nearly identical rates of depressive
remission whether or not they endorsed auditory–visual halluci-
nations (rates: 47% versus 51%; n¼ 270; O.R.¼0.98; CI
2.00; p ¼ n.s.) (Fig. 3).
Paranoid ideation also had a negative impact on the likelihood
of remission in Whites (rates: 45% versus 57%; n¼ 1927;
O.R.¼ 1.44; CI
: 1.21–1.73; po 0.001); however, the comparison
did not reach signiﬁcance in Blacks (rates: 36% versus 42%;
n¼ 400; O.R.¼ 1.34; CI
: 0.95–1.89; p¼n.s.) or Latinos (rates:
47% versus 51%; n¼ 270; O.R.¼ 1.17; CI
: 0.77–1.78; p¼ n.s.)
P. Cassano et al. / Journal of Affective Disorders 150 (2013) 578–584580
(Figs. 4–6). Demographic characteristics for each group compar-
ison pair are reported in Tables 1 and 2.
Both the analyses for auditory–visual hallucinations and for
paranoid ideation were repeated after reintroducing the subjects
who had been excluded due to endorsement of ﬁrst-rank symptoms.
Findings were essentially equivalent, except that differences in
remission rates with versus without paranoid ideation approached
signiﬁcance in both Latinos and Blacks (data not shown).
Demographic characteristics of subjects who endorsed versus did not endorse paranoid ideation across racial and ethnic groups.
Baseline variable Whites Blacks Latinos
Endorsement of paranoid ideation Yes n¼ 299 No n¼ 1628 p-values Yes n¼ 126 No n ¼ 274 p-values Yes n¼ 57 No n ¼ 213 p-values
Mean age (years 7 SD) 39.2712.7 42.7713.4 o 0.001 41.07 12.2 43.77 12.3 o 0.05 38.37 10.5 41.07 12.3 n.s.
Gender (women %) 57 60 n.s. 60 70 n.s. 74 75 n.s.
(married or cohabiting with partner) (%)
44 44 n.s. 33 35 n.s. 49 49 n.s.
Education (mean years of schooling7 SD) 13.272.5 14.373.0 o 0.001 12.87 2.5 13.372.7 n.s. 12.17 3.6 12.07 4.3 n.s.
Currently a student (%) 12 15 n.s. 18 14 n.s. 14 11 n.s.
Language preference (Spanish) n/a n/a n/a n/a 12% 19% n.s.
Employment (%) 58 62 n.s. 54 47 n.s. 53 60 n.s.
Monthly household income (mean $) 221572193 303773702 o .001 1377 7 1,343 19907 2,673 o .05 16567 1794 169171653 n.s.
Care setting specialty (%) 68 66 n.s. 60 54 n.s. 44 47 n.s.
Severity of depression (mean HRDS total
24.175.1 21.774.6 o 0.001 25.575.2 23.77 5.6 o
0.005 23.37 4.8 22.2 7 4.9 n.s.
Duration of episode (mean days7 SD) 63671391 6657 1422 n.s. 11137 2148 865 7 1424 n.s. 92471468 10147 1863 n.s.
Age of onset for MDD (mean years7 SD) 23.0712.9 25.67 14.6 o 0.005 25.5 7 14.4 26.67 13.9 n.s. 25.57 12.7 26.27 14.3 n.s.
n.s.: non-signiﬁcant p-value.
n/a: not applicable.
Demographic characteristics of subjects who endorsed versus did not endorse auditory–visual hallucinations across racial and ethnic groups.
Baseline variable Whites Blacks Latinos
Endorsement of auditory–visual
Yes n¼ 49 No n¼ 1879 p-values Yes n¼ 45 No n¼ 353 p-values Yes n¼ 17 No n ¼ 253 p-values
Age (mean years7SD) 42.3714.3 42.17 13.3 n.s. 42.0711.0 42.97 12.5 n.s. 41.6713.7 40.47 11.9 n.s.
Gender (women) (%) 55 59 n.s. 53 69 o 0.05 71 75 n.s.
(married or cohabiting with partner) (%)
51 44 n.s. 36 35 n.s. 35 50 n.s.
Education (mean years of schooling7SD) 12.0 7 2.6 14.27 2.9 o 0.001 12.372.0 13.272.7 o 0.05 11.87 3.0 12.17 4.2 n.s.
Currently a student (%) 2 15 o 0.01 7 15 n.s. 12 11 n.s.
Language preference (Spanish) n/a n/a n/a n/a 22% 17% n.s.
Employment (%) 39 62 o 0.005 40 50 n.s. 41 59 n.s.
Monthly household income (mean $) 17107 2104 2938 7 3543 o 0.05 12327 1243 18667 2,447 n.s. 15277 1780 16957 1677 n.s.
Care setting specialty (%) 71 66 n.s. 62 55 n.s. 47 46 n.s.
Severity of depression (mean HRDS total
24.874.3 22.07 4.8 o
0.001 26.875.4 23.97 5.4 ¼ 0.001 23.275.3 22.4 7 4.9 n.s.
Duration of episode (mean days7SD) 84471929 6567 1401 n.s. 102871487 9367 1719 n.s. 1146 7 2342 98571746 n.s.
Age of onset for MDD (mean years 7 SD) 24.77 13.4 25.2 7 14.4 n.s. 23.37 11.8 26.67 14.2 n.s. 24.97 14.0 26.27 14.0 n.s.
n.s.: non-signiﬁcant p-value.
n/a: not applicable.
Fig. 1. Remission of MDD in subjects who did and did not endorse ‘‘seeing or
hearing things that other people did not’’ among Whites.
Fig. 2. Remission of MDD in subjects who did and did not endorse ‘‘seeing or
hearing things that other people did not’’ among Blacks.
P. Cassano et al. / Journal of Affective Disorders 150 (2013) 578–584 581
3.3. Isolated psychotic symptoms, PTSD and other anxiety disorders
In all ethnic–racial groups, a signiﬁcant association was found
between auditory–visual hallucinations and PTSD and panic
disorder. The association between auditory–visual hallucinations
and derealization, and history of trauma, reached signiﬁcance
in Whites and in Blacks, but not in Latinos (data not shown).
Finally, the presence of agoraphobia was similar in Latinos with
and without endorsement of auditory–visual hallucinations
(18% versus 20%), but not in Whites (43% versus 11%) and Blacks
(42% versus 17%) (statistics not shown).
The major ﬁnding of our study is that the endorsement of
auditory–visual hallucinations by Latinos did not signiﬁcantly
impact their response to treatment of MDD. This contrasted with
what both Whites and Blacks experienced in the same STAR*D
This ﬁnding suggests that the perceptions that led to the
endorsement of auditory–visual hallucinations in Latinos might
have been different in nature from those of Whites and Blacks
(Cassano et al., 2012). Abnormal auditory and visual perceptions
are often reported by Latinos with MDD as well as by Latinos with
other mood and anxiety disorders (Posternak and Zimmerman,
2005; David et al., 1999; Mischoulon et al., 2005). In a separate
paper, we attempted to systematically differentiate psychosis-like
manifestations from genuine psychotic presentations in Latinos
(Cassano et al., 2012). Unfortunately in the STAR*D cohort these
differential features were not explored.
Our unproven working hypothesis is that genuine isolated
auditory–visual hallucinations in Latinos with MDD might well
have the same prognostic validity as in other ethnic–racial
groups; but psychosis-like perceptions are likely to mask this
effect in our study results. This ﬁnding implies that at least with
depressed Latinos, taking the report of isolated auditory–visual
hallucinations and other psychotic symptoms at face value might
be misleading and could lead to unnecessary treatment.
A second important ﬁnding of our study was that endorsement
of paranoid ideation was signiﬁcantly associated with treatment
outcome in depressed Whites, but not in Blacks or Latinos. It is
possible that the sample size and distribution were not enough to
yield a signiﬁcant effect. Yet the sample of subjects who endorsed
paranoid ideation was larger than the sample of subjects who
endorsed auditory–visual hallucinations, especially for Blacks.
Alternatively, the endorsement of paranoid ideation might
reﬂect both cases of excessive and maladaptive suspiciousness
as well as cases of justiﬁed distrust based on living circumstances.
In our sample, both Blacks and Latinos appear to have a low
average income level and might be living in more disadvantaged
and dangerous neighborhoods (Combs et al., 2002; Shoff and
Yang, 2012). In these settings a greater sense of distrust might be
not only justiﬁed but also adaptive. As with the endorsement of
auditory–visual hallucinations, clinicians should not take the
endorsement of paranoid ideation at face value but assess its
Blacks were the group that most often reported isolated
psychotic symptoms. This was in contrast with our hypothesis
that Latinos would endorse these symptoms the most (David
et al., 1999; Olfson et al., 2002; Posternak and Zimmerman, 2005 ).
Recently, a greater likelihood of psychosis in Blacks has been
: 1.21 – 1.73
Fig. 4. Remission of MDD in subjects who did and did not report being ‘‘convinced
that others are spying’’ and/or convinced they were ‘‘in danger because someone is
plotting against [me]’’ in Whites.
Fig. 5. Remission of MDD in subjects who did and did not report being ‘‘convinced
that others are spying’’ and/or convinced they were ‘‘in danger because someone is
plotting against [me]’’ in Blacks.
Fig. 6. Remission of MDD in subjects who did and did not report being ‘‘convinced
that others are spying’’ and/or convinced they were ‘‘in danger because someone is
plotting against [me]’’ in Latinos.
Fig. 3. Remission of MDD in subjects who did and did not endorse ‘‘seeing or
hearing things that other people did not’’ among Latinos.
P. Cassano et al. / Journal of Affective Disorders 150 (2013) 578–584582
reported with assessments blind to ethnicity and race (Gara et al.,
2012). In our study, presenting a parallel ﬁnding for isolated
psychotic symptoms in Blacks, the self-reported nature of the
collected data similarly averts a clinician bias towards psychosis.
Prior reports from the STAR*D cohort have shown that Blacks
have signiﬁcantly poorer outcome as compared to Whites and
Latinos when treated for MDD (Lesser et al., 2007). Our ﬁndings
extend the previous ones, since depressed Blacks were also more
affected by concomitant, self-report auditory and visual halluci-
nations (O.R. 2.11 for lack of remission) than Whites (O.R. 1.70) or
Latinos (O.R. 0.98).
The third and last aim of our study was to explore the nature
of the isolated psychotic symptoms expressed by Latinos and to
investigate a putative preferential association with PTSD, other
anxiety disorders, trauma and derealization (Lewis-Fernandez
et al., 2010; Freeman and Fowler, 2009). Our ﬁndings did not
support this hypothesis. The endorsement of auditory–visual
hallucinations was equally associated with several of the afore-
mentioned conditions in both Latinos and non-Latinos.
Our study was conducted on a large sample of subjects
assessed prospectively and by independent raters, blind to treat-
ment assignment. The cohort of patients is highly representative
of everyday depressed patients seen in primary and specialty
care, which makes our ﬁndings generalizable. Nonetheless, this
work has several limitations. The diagnosis of non-psychotic MDD
was clinical in nature, lacking a standardized assessment of
psychotic symptoms, and this might have led to erroneous
inclusion of some patients with pronounced psychotic features.
Likewise, a sizeable proportion of STAR*D patients were enrolled
from primary care settings, which suggests that the assessment of
the research coordinators was reviewed by a generalist instead of
a psychiatrist. Although all staff had been trained to diagnose and
assess non-psychotic MDD, it could be argued that in the primary
care setting there might have been a greater rate of misdiagnosis
due to lack of specialty training. This bias seems unlikely, since
the rates of endorsed isolated psychotic symptoms were similar
in both primary and specialty care settings. It is also noteworthy
that self-assessment of psychosis is far from being the gold
standard. Finally, while patients at times can endorse isolated
and subthreshold psychotic symptoms in the absence of full-
blown psychosis (Sbrana et al., 2005), no formal diagnosis of
personality disorders was conducted in STAR*D, so we could not
rule out that symptoms might have been present as traits of a
DSM-IV cluster A personality disorder. It is likely that in many
cases a subthreshold presentation of the psychotic spectrum was
present, such as vague and even excessive suspiciousness not
rising to delusional quality, or ﬂeeting and rare hallucinatory
experiences as opposed to pervasive and recurrent hallucinations
(American Psychiatric Association and American Psychiatric
Association Task Force on DSM-IV, 2000).
Besides lacking any speciﬁc assessment for psychosis-like
symptoms in Latinos, we were also unable to stratify patients
based on their origin, immigration status, or generation and years
lived in the United States. Furthermore, the gender distribution
highly favored women over men in the Latino group, which is
reﬂective of clinical practice. Yet no gender effect was seen in the
prevalence of isolated psychotic symptoms.
Finally, the sample of Latinos (n¼ 270) was much smaller than
Whites (n¼ 1928) and Blacks (n¼ 398). It could therefore be
argued that the lack of predictive effect of endorsed auditory–
visual hallucinations on remission in Latinos reﬂects a lack of
power. Yet the effect size for the aforementioned analyses on
remission is markedly different in the three ethnic–racial groups:
The difference in remission rates was 4% in Latinos and 17% and
20% in Whites and Blacks, respectively. Even if the study had been
powered to detect a 4% difference in rates, the same difference
would not be clinically informative.
The lack of signiﬁcant differences in outcome for Latinos, and
at times for Blacks, endorsing isolated psychotic symptoms
cannot be ascribed to a direct antipsychotic effect of treatment,
particularly since in STAR*D antipsychotic augmentation was not
used. In view of the above, we recommend avoiding preemptive
antipsychotic augmentation of the treatment of MDD with iso-
lated auditory–visual hallucinations or paranoid ideation. Instead,
a careful psychopathological assessment of the nature of the
psychotic symptoms is recommended prior to initiating an anti-
psychotic drug, especially in Latinos. If the endorsement of a
psychotic symptom by itself does not translate into poorer
outcome with traditional treatment strategies for MDD, the
iatrogenic risks of the antipsychotic augmentation for MDD might
not be justiﬁed (Henderson, 2008).
We are fully aware of this and other policies governing
publications in the Journal of Affective Disorders.
Role of funding source
The STAR*D project has been funded with Federal funds from the National
Institute of Mental Health, National Institutes of Health, Bethesda, MD, under
Contract N01MH90003 to UT Southwestern Medical Center at Dallas (P.I.: AJ Rush).
Conﬂict of interest
Dr. P. Cassano has received research support from Photothera Inc. in the past
year. Drug donation from TEVA. Travel reimbursement from Pharmacia-Upjohn.
Dr. T. Chang has received research support for serving as a study rater for the
following companies in the last three years: Alkermes, AstraZeneca, CeNeRx,
Euthymics, Forest, GlaxoSmithKline, J&J, NeuralStem, Pﬁzer. Travel reimbursement
from Bristol-Myers Squibb for presenting at their advisory board. Honorarium for
this presentation; paid to institution and not to individual.
Dr. N. Trinh and Dr. L. Baer report no competing interests.
Dr. M. Fava has received research support from Abbot Laboratories, Alkermes,
Inc., Aspect Medical Systems, AstraZeneca, BioResearch, BrainCells Inc., Bristol-
Myers Squibb, CeNeRx BioPharma, Cephalon, Clintara, LLC, Covance, Covidien, Eli
Lilly and Company, ElMindA, Ltd., EnVivo Pharmaceuticals, Inc., Euthymics
Bioscience, Inc., Forest Pharmaceuticals, Inc., Ganeden Biotech, Inc., GlaxoSmithK-
line, Icon Clinical Research, i3 Innovus/Ingenix, Johnson & Johnson Pharmaceutical
Research & Development, Lichtwer Pharma GmbH, Lorex Pharmaceuticals,
National Alliance for Research on Schizophrenia & Depression (NARSAD), National
Center for Complementary and Alternative Medicine (NCCAM), National Institute
of Drug Abuse (NIDA), National Institute of Mental Health (NIMH), Novartis AG;
Organon Pharmaceuticals, PamLab, LLC., Pﬁzer Inc., Pharmavite
Roche Pharmaceuticals, RCT Logic, LLC (formerly Clinical Trials Solutions, LLC),
Sanoﬁ-Aventis US LLC, Shire; Solvay Pharmaceuticals, Inc., Synthelabo, Wyeth-
He has been Advisor/Consultant for Abbott Laboratories, Affectis Pharmaceuticals
AG, Alkermes, Inc., Amarin Pharma Inc., Aspect Medical Systems, AstraZeneca, Auspex
Pharmaceuticals, Bayer AG, Best Practice Project Management, Inc., BioMarin Pharma-
ceuticals, Inc., Biovail Corporation, BrainCells Inc, Bristol-Myers Squibb, CeNeRx
BioPharma, Cephalon, Inc., CNS Response, Inc., Compellis Pharmaceuticals, Cypress
Pharmaceutical, Inc., DiagnoSearch Life Sciences (P) Ltd., Dinippon Sumitomo
Pharma Co. Inc., Dov Pharmaceuticals, Inc., Edgemont Pharmaceuticals, Inc., Eisai
Inc., Eli Lilly and Company, EnVivo Pharmaceuticals, Inc., ePharma Solutions, EPIX
Pharmaceuticals, Inc., Euthymics Bioscience, Inc., Fabre-Kramer Pharmaceuticals, Inc.,
Forest Pharmaceuticals, Inc., GenOmind, LLC, GlaxoSmithKline, Grunenthal GmbH, i3
Innovus/Ingenis, Janssen Pharmaceutica, Jazz Pharmaceuticals, Inc., Johnson & John-
son Pharmaceutical Research & Development, LLC, Knoll Pharmaceuticals Corp.,
Labopharm Inc., Lorex Pharmaceuticals, Lundbeck Inc., MedAvante, Inc., Merck &
Co., Inc., MSI Methylation Sciences, Inc., Naurex, Inc., Neuralstem, Inc., Neuronetics,
Inc., NextWave Pharmaceuticals, Novartis AG, Nutrition 21, Orexigen Therapeutics,
Inc., Organon Pharmaceuticals, Otsuka Pharmaceuticals, Pamlab, LLC., Pﬁzer Inc.,
P. Cassano et al. / Journal of Affective Disorders 150 (2013) 578–584 583
LLC., PharmoRx Therapeutics, Precision Human Biolabora-
tory, Prexa Pharmaceuticals, Inc., Puretech Ventures, PsychoGenics, Psylin Neuros-
ciences, Inc., Rexahn Pharmaceuticals, Inc., Ridge Diagnostics, Inc., Roche, Sanoﬁ-
Aventis US LLC., Sepracor Inc., Servier Laboratories, Schering-Plough Corporation,
Solvay Pharmaceuticals, Inc., Somaxon Pharmaceuticals, Inc., Somerset Pharmaceu-
ticals, Inc., Sunovion Pharmaceuticals, Supernus Pharmaceuticals, Inc., Synthelabo,
Takeda Pharmaceutical Company Limited, Tal Medical, Inc., Tetragenex Pharmaceu-
ticals, Inc., TransForm Pharmaceuticals, Inc., Transcept Pharmaceuticals, Inc., Vanda
He is on Speaking/Publishing board for Adamed, Co, Advanced Meeting
Partners, American Psychiatric Association, American Society of Clinical Psycho-
pharmacology, AstraZeneca, Belvoir Media Group, Boehringer Ingelheim GmbH;
Bristol-Myers Squibb; Cephalon, Inc.; CME Institute/Physicians Postgraduate Press,
Inc.; Eli Lilly and Company; Forest Pharmaceuticals, Inc.; GlaxoSmithKline;
Imedex, LLC; MGH Psychiatry Academy/Primedia, MGH Psychiatry Academy/Reed
Elsevier, Novartis AG, Organon Pharmaceuticals, Pﬁzer Inc., PharmaStar, United
BioSource,Corp., Wyeth-Ayerst Laboratories.
Equity Holdings: Compellis.
Royalty/patent, other income: patent for Sequential Parallel Comparison
Design (SPCD), which are licensed by MGH to RCT Logic, LLC, and patent
application for a combination of azapirones and bupropion in Major Depressive
Copyright for the MGH Cognitive & Physical Functioning Questionnaire (CPFQ),
Sexual Functioning Inventory (SFI), Antidepressant Treatment Response Ques-
tionnaire (ATRQ), Discontinuation-Emergent Signs & Symptoms (DESS), and
SAFER, Lippincott, Williams & Wilkins, Wolkers Kluwer, World Scientiﬁc Publish-
ing Co. Pte. Ltd.
Dr. Mischoulon has received research support from the Bowman Family
Foundation, Bristol-Myers Squibb Co., Cederroth, FisherWallace, Ganeden,
Lichtwer Pharma, Nordic Naturals, Laxdale (Amarin), and SwissMedica. He has
received honoraria for consulting, speaking, and writing from Pamlab, Bristol-
Myers Squibb Co., Nordic Naturals, Virbac, Pﬁzer, and Reed Medical Education. He
has received royalties from Back Bay Scientiﬁc for PMS Escape, and from
Lippincott Williams & Wilkins for published book ‘‘Natural Medications for
Psychiatric Disorders: Considering the Alternatives’’. No payment has exceeded
American Psychiatric Association, American Psychiatric Association Task Force on
DSM-IV, 2000. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV.
American Psychiatric Association, Washington, DC.
Cassano, P., Fava, M., Mischoulon, D., 2012. Major depressive disorder with
psychosis-like symptoms among Latinos. Psychiatric Services 63, 482–487.
Combs, D.R., Penn, D.L., Fenigstein, A., 2002. Ethnic differences in subclinical
paranoia: an expansion of norms of the paranoia scale. Cultural Diversity &
Ethnic Minority Psychology 8, 248–256.
David, D., Kutcher, G.S., Jackson, E.I., Mellman, T.A., 1999. Psychotic symptoms in
combat-related posttraumatic stress disorder. Journal of Clinical Psychiatry 60,
Freeman, D., Fowler, D., 2009. Routes to psychotic symptoms: trauma, anxiety and
psychosis-like experiences. Psychiatry Research 169, 107–112.
Gara, M.A., Vega, W.A., Arndt, S., Escamilla, M., Fleck, D.E., Lawson, W.B., Lesser, I.,
Neighbors, H.W., Wilson, D.R., Arnold, L.M., Stratowski, S.M., 2012. Inﬂuence of
patient race and ethnicity on clinical assessment in patients with affective
disorders. Archives of General Psychiatry 69, 593–600.
Geltman, D., Chang, G., 2004. Hallucinations in Latino psychiatric outpatients:
a preliminary investigation. General Hospital Psychiatry 26, 153–157.
Henderson, D.C., 2008. Managing weight gain and metabolic issues in patients
treated with atypical antipsychotics. Journal of Clinical Psychiatry 69, e04.
Lesser, I.M., Castro, D.B., Gaynes, B.N., Gonzalez, J., Rush, A.J., Alpert, J.E., Trivedi,
M., Luther, J.F., Wisniewski, S.R., 2007. Ethnicity/race and outcome in the
treatment of depression: results from STAR
D. Medical Care 45, 1043–1051.
Lewis-Fernandez, R., Gorritz, M., Raggio, G.A., Pelaez, C., Chen, H., Guarnaccia, P.J.,
2010. Association of trauma-related disorders and dissociation with four
idioms of distress among Latino psychiatric outpatients. Culture, Medicine
and Psychiatry 34, 219–243.
Lewis-Fernandez, R., Horvitz-Lennon, M., Blanco, C., Guarnaccia, P.J., Cao, Z.,
Alegria, M., 2009. Signiﬁcance of endorsement of psychotic symptoms by US
Latinos. Journal of Nervous and Mental Disease 197, 337–347.
Minsky, S., Vega, W., Miskimen, T., Gara, M., Escobar, J., 2003. Diagnostic patterns
in Latino, African American, and European American psychiatric patients.
Archives of General Psychiatry 60, 637–644.
Mischoulon, D., Lagomasino, I.T., Harmon, C., 2005. Atypical psychotic symptoms
in a Hispanic population: diagnostic dilemmas and implications for treatment.
Psychiatry 2, 38–46.
Olfson, M., Lewis-Fernandez, R., Weissman, M.M., Feder, A., Gameroff, M.J.,
Pilowsky, D., Fuentes, M., 2002. Psychotic symptoms in an urban general
medicine practice. American Journal of Psychiatry 159, 1412–1419.
Perlis, R.H., Uher, R., Ostacher, M., Goldberg, J.F., Trivedi, M.H., Rush, A.J., Fava, M.,
2010. Association between bipolar spectrum features and treatment outcomes
in outpatients with major depressive disorder. Archives of General Psychiatry.
Posternak, M.A., Zimmerman, M., 2005. Elevated rates of psychosis among
treatment-seeking Hispanic patients with major depression. Journal of Ner-
vous and Mental Disease 193, 66–69.
Rush, A.J., Fava, M., Wisniewski, S.R., Lavori, P.W., Trivedi, M.H., Sackeim, H.A.,
Thase, M.E., Nierenberg, A.A., Quitkin, F.M., Kashner, T.M., Kupfer, D.J.,
Rosenbaum, J.F., Alpert, J., Stewart, J.W., Mcgrath, P.J., Biggs, M.M., Shores-
Wilson, K., Lebowitz, B.D., Ritz, L., Niederehe, G., 2004. Sequenced treatment
alternatives to relieve depression (STAR
D): rationale and design. Controlled
Clinical Trials 25, 119–142.
Sbrana, A., Dell
osso, L., Benvenuti, A., Rucci, P., Cassano, P., Banti, S., Gonnelli, C.,
Doria, M.R., Ravani, L., Spagnolli, S., Rossi, L., Raimondi, F., Catena, M., Endicott, J.,
Frank, E., Kupfer, D.J., Cassano, G.B., 2005. The psychotic spectrum: validity and
reliability of the Structured Clinical Interview for the Psychotic Spectrum.
Schizophrenia Research 75, 375–387.
Shevlin, M., Murphy, J., Dorahy, M.J., Adamson, G., 2007. The distribution of
positive psychosis-like symptoms in the population: a latent class analysis of
the National Comorbidity Survey. Schizophrenia Research 89, 101–109.
Shoff, C., Yang, T.C., 2012. Untangling the associations among distrust, race, and
neighborhood social environment: a social disorganization perspective. Social
Science & Medicine 74, 1342–1352.
Van Os, J., 2003. Is there a continuum of psychotic experiences in the general
population? Epidemiologia e Psichiatria Sociale 12, 242–252.
Vega, W.A., Sribney, W.M., Miskimen, T.M., Escobar, J.I., Aguilar-Gaxiola, S., 2006.
Putative psychotic symptoms in the Mexican American population: preva-
lence and co-occurrence with psychiatric disorders. Journal of Nervous and
Mental Disease 194, 471–477.
Zimmerman, M., Chelminski, I., 2006. A scale to screen for DSM-IV Axis I disorders
in psychiatric out-patients: performance of the Psychiatric Diagnostic Screening
Questionnaire. Psychological Medicine 36, 1601–1611.
P. Cassano et al. / Journal of Affective Disorders 150 (2013) 578–584584