Psychiatry Research 272 (2019) 275–283

Contents lists available at ScienceDirect

Psychiatry Research
journal homepage: www.elsevier.com/locate/psychres

Targeting relapse prevention and positive symptom in first-episode T

schizophrenia using brief cognitive behavioral therapy: A pilot randomized
controlled study
⁎
Yan Liua,b, Xiaojie Yanga,b, Amy Gillespiec, Zhihua Guoa,b, Yun Maa,b, Runsen Chena,b,c, ,
⁎
Zhanjiang Lia,b,
a
The National Clinical Research Center for Mental Disorders & Beijing Key Laboratory of Mental Disorders, Beijing Anding Hospital, Capital Medical University, Beijing,
China
b
Advanced Innovation Center for Human Brain Protection, Capital Medical University, Beijing, China
c
Department of Psychiatry, University of Oxford, UK

A R T I C LE I N FO A B S T R A C T

Keywords: The present study aimed to provide preliminary evaluation of the effectiveness of a brief CBT intervention
First-episode focusing on relapse prevention and positive symptom in a Chinese first episode schizophrenia (FES) population.
CBT This randomized controlled trial recruited eighty outpatients with FES (as determined using the DSM-IV), aged
Chinese population 16–45 years, and on a current atypical antipsychotic. Patients were randomized to either 10 sessions of in-
One year follow up
dividual CBT (intervention group) adjunctive to treatment as usual (TAU) or TAU alone (control group).
Outcome assessment of symptoms, relapse, hospitalization, insight and social functioning were administered at
baseline and then post treatment (10 weeks), and at 6-month and 12-month follow ups. At 12 months, patients in
the intervention group had significantly greater improvements in positive symptoms, general psychopathology
and social functioning, as well as significantly lower rates of relapse, compared to the control group. Although
patients in both groups demonstrated significantly improved negative symptom and insight scores from baseline,
no group differences were found.
This RCT demonstrates that FES patient can greatly benefit from CBT designed to target relapse prevention
and positive symptom, with improvements sustained for 1 year following treatment.

1. Introduction treatment, relapse rates are 70–80% over the 5 years following first
onset and 33% of patients experience relapse at one year follow up
Schizophrenia has become the eighth leading cause of disability (Robinson et al., 2005; Üçok et al., 2006). Relapse prevention is im-
world-wide for people aged between 15–44 years (World Health portant in FES, as each relapse increases the risk of positive symptoms
Organization, 2001). It is of vital important that individuals having becoming persistent (Wiersma et al., 1998). Thus, previous studies have
their first episode of schizophrenia receive adequate treatment suggested combining antipsychotic treatment with psychosocial inter-
(Zhu et al., 2017), as meta-analyses suggest that a long duration of ventions such as CBT for people with early psychosis and schizophrenia
untreated psychosis is associated with worse clinical treatment out- (Alvarez-Jiménez et al., 2009; Penn et al., 2005).
comes, such as more severe symptoms, worse social functioning, and Several meta-analyses and reviews have concluded that CBT is an
failure to achieve remission (Perkins et al., 2005; Perkins et al., 2004). effective intervention for positive, negative and general symptoms, in-
Although patients with first episode schizophrenia (FES) usually re- sight, relapse and social functioning in chronic psychosis patients
spond well to antipsychotic treatment – potentially due to lower fre- (Wykes et al., 2008; Zimmermann et al., 2005); however, the picture is
quency of negative symptoms and better cognition function compared less clear for first-episode schizophrenia patients. Of the controlled
to patients with chronic schizophrenia - they may still experience dif- studies that exist, many report inconsistent findings (Drury et al., 1996,
ficulties with social interaction and stigma, may discontinue treatment 2000; Garety et al., 2008; Grawe et al., 2006; Haddock et al., 1999b;
(Ohlsen et al., 2004), and may relapse. Even with antipsychotic Jackson et al., 1998, 2008; Jolley et al., 2003; Lewis et al., 2002), with

⁎
Corresponding authors.
E-mail addresses: runsen.chen@psych.ox.ac.uk (R. Chen), lizhj8@ccmu.edu.cn (Z. Li).

https://doi.org/10.1016/j.psychres.2018.12.130
Received 9 July 2018; Received in revised form 15 November 2018; Accepted 23 December 2018
Available online 25 December 2018
0165-1781/ © 2018 Elsevier B.V. All rights reserved.
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

issues regarding small sample size (Haddock et al., 1999b; Jackson 2. Methods
et al., 1998; Jolley et al., 2003), variable additional treatment, and
highly heterogeneous samples (Drury et al., 1996, 2000; Lewis et al., 2.1. Sampling and study design
2002). Importantly, two systematic reviews (Morrison, 2009; Penn
et al., 2005) of CBT for first episode psychosis (FEP) concluded that Participants were recruited from Beijing Anding Hospital affiliated
there was limited evidence of significant improvements in symptoms Capital Medical University between June 2012 and March 2014.
(auditory hallucinations and hopelessness), recovery and functioning, Outpatients were referred to this program by psychiatrists. Inclusion
as well as reducing rates of relapse or hospital readmission for first- criteria were: (1) A diagnosis of schizophrenia as assessed by experi-
episode patients given CBT intervention. enced research psychiatrists using the Structured Clinical Interview for
Few previous FEP studies have specifically looked at using CBT to DSM-IV Axis I Disorders-Clinician Version; (2) Age between 16–45
target relapse prevention, while studies in chronic patients have de- years; (3) The current illness episode was their first episode of schizo-
monstrated that CBT focused on relapse prevention is effective in re- phrenia, and any continuous treatment had been for less than 1 month
ducing relapse rates. For example, Gumley et al. (2003) found that or was inadequate treatment of less than 3 months verified through
relapse prevention focused CBT greatly reduced hospital admission and clinical interview and medical chart review; (4) Use of a single atypical
relapse rates in patients with chronic schizophrenia, as well as sig- antipsychotic drug for at least two weeks; (5) A score of 4 or more on
nificantly improving symptoms, global psychopathology and social the delusion (P1) or hallucination (P3) symptom scales of the Positive
functioning. This study emphasized the importance of early relapse and Negative Syndrome Scale (PANSS); (6) Their first psychotic
indicators, which can trigger negative beliefs about relapse and hospi- symptoms had occured less than 3 years ago. Exclusion criteria were:
talization. However, a study of relapse prevention therapy (combing (1) Co-morbid diagnosis of mental disability or primary substance de-
individual and family CBT therapy) in early psychosis patients pendence; (2) Inability to communicate, or lack of spontaneity and flow
(Gleeson et al., 2009) found that while this was indeed effective in of conversation (5 or above on PANSS in conceptual disorganization);
reducing relapse rates over 7 months follow up, this did not generalize (3) Electroconvulsive therapy within the 1 month prior to entry into the
to other outcomes, with no improvement in medication adherence, study; (4) Serious or unstable physical health condition; (5) Currently
psychosocial functioning or quality of life. Thus, the authors suggested receiving any other form of systematic psychotherapy.
that focusing on relapse prevention alone was not enough, and further Eighty patients met the criteria and were included in the trial. Fig. 1
consideration of the relevant targets for early CBT interventions is re- provides a CONSORT diagram illustrating the flow of participants
quired. through the study.
Positive symptoms, such as paranoia, hallucinations and delusions, This study was approved by the Human Research and Ethics
are stronger predictors of relapse than negative symptom (Lavretsky, Committee of Beijing Anding Hospital, Capital Medicine University.
2008; Patel et al., 2014). A recent study in which CBT primarily tar- Informed consent was obtained from the patients and guardians. This
geted negative symptoms in schizophrenia patients found that this ap- clinical trial is registered with the Chinese Clinical Trial Registry
proach effectively reduced negative symptoms, but nearly 30% patients (ChiCTR-TRC-13003929).
in the CBT group and 35.1% patients in the cognitive remediation
group were readmitted to hospital within 1 year following treatment 2.2. Assessment
(Klingberg et al., 2011). Similarly, Jackson et al. (1998) found that
while CBT significantly improved negative symptoms in FEP, no group Face-to-face assessments of all participants were performed by three
differences were found for rates of relapse and hospitalization. There- clinicians who were blind to the treatment condition, during treatment
fore, the CBT model may need to be modified to provide effective as usual sessions. The clinicians were trained once every two months
treatment for reducing both positive symptoms and preventing relapse and reviewed the standards of the scales, to ensure assessment accuracy
in first episode psychosis. and consistency. At baseline and weeks 10, 36 and 62, severity of
There is a great need for research investigating CBT for FES across psychopathology, relapse, number of hospital admissions, insight, and
diverse ethnic backgrounds, especially in developing countries social functioning, were assessed, based on the following guidelines:
(Pontes et al., 2013), with very few studies investigating the effects of Severity of psychopathology was assessed according to the Chinese
CBT in FES in China (Wang et al., 2003). One reason for this is the version of the PANSS (Kay et al., 1987; Tianmei et al., 2004). The
limited mental health services in China which create a barrier for Chinese version of this scale has shown good reliability (Cronbach's
conducting psychological interventions of long duration for large a = 0.871). It also has a good level of construct validity, with total
groups of patients (Chen et al., 2018a,b). Due to this limitation and variance explaining 59% of the variance in symptoms in Chinese pa-
considering the heightened need for cost-effective treatments in most tients with schizophrenia (Tianmei et al., 2004). This scale includes 30
developing countries (Saraceno et al., 2007), the current study ex- items, each of which is scored on a seven-point Likert scale (1 = ab-
amines a brief CBT intervention rather than costly multimodal treat- sence of psychopathology; 7 = very severe symptom). PANSS scores
ment. were calculated using three dimensions: positive symptoms, negative
The aim of this randomized controlled trial was to investigate the symptoms, and general psychopathology.
effectiveness of a brief CBT intervention designed specifically to target The Psychotic Symptoms Rating Scale (PSYRATS) is an 11-item
relapse prevention and positive symptom in a Chinese first episode rating scale with good validity in schizophrenia patients which evalu-
schizophrenia population. The primary outcomes were relapse and ates the severity of different dimensions of the psychotic symptoms. The
hospitalization rates over 12 months, and change in psychotic symp- Chinese version of PSYRAT has good reliability (Cronbach's a = 0.943)
toms from baseline to month 12. The secondary outcomes were changes and validity. The correlation of auditory hallucination, delusion and
in insight, social functioning, and occupational functioning from base- total scores of the PSYRAT with hallucination, delusion and positive
line to month 12. We hypothesized that brief CBT intervention would subscale of PANSS is reported to be 0.909, 0.833 and 0.737 respectively
have better positive symptom reduction and lower relapse rates com- (Xu et al., 2012). The items include the frequency, duration, controll-
pared to the treatment as usual condition. ability, location, loudness; severity and intensity of stress; amount and
degree of negative content; and beliefs about the origin of voices and
disruption (Haddock et al., 1999a).
The criteria for relapse was deterioration in regards to worsened
psychotic symptoms, as identified by either a rating of 6 or 7 on PANSS
thought-disorder items (P2, P3, P5 and G9) OR two or more of these

276
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

Fig. 1. . CONSORT (Consolidated Standards of Reporting Trials) diagram

CBT, Cognitive-behavioural therapy; TAU, Treatment as usual.

items rating as 5 or above. combination of participant interview and review of medical records.
Insight was assessed using the Schedule for Assessing Insight
(David, 1990). The SAI comprises of questions to assess three dimen- 2.3. Procedure
sions of insight: awareness, relabeling of symptoms, and attitudes to
treatment. The SAI includes seven items, each of which is scored on a Participants were randomly allocated to the CBT + TAU group or
three-point Likert scale from 0 (no insight) to 2 (good insight). The the TAU group (1:1 randomization) after providing informed consent
range of total score is from 0 to 14. The Chinese version of SAI has been and completing baseline assessments. Randomization was conducted by
demonstrated with good psychometric properties (Xu et al., 2013): the computer-generated (SPSS Version 20.0) blocks of four random num-
internal consistency of the total scale was high, with Cronbach's α of bers using 1:1 assignment ratio to allocate eligible participants to either
0.890 and correlation of the SAI with insight of the PANSS was −0.635. of the two groups. Randomization was conducted by a researcher who
Social functioning was rated using the Chinese version of the was not involved in the study. The outcome measures were evaluated
Personal and Social Performance Scale (PSP).The Chinese version of by clinicians blind to the study protocol and treatment assignment.
PSP shows good internal consistency (Cronbach's a = 0.840) and con-
struct validity, with statistically significant correlations with the Global
2.4. Treatment group
Assessment of Functioning scale (Tianmei et al., 2011). The PSP is
comprised of four areas of functioning for patients with schizophrenia:
The brief CBT programme was a manualised individual treatment
(1) participation in socially useful activities; (2) personal and social
comprising ten sessions over 10 weeks, with each session typically
relationships; (3) self-care; (4) interruptive or aggressive behaviour.
lasting 45 min. In weeks 1–2, there were 2 sessions per week; in weeks
Each area of functioning is rated on a six-point Likert scale based on the
3–6, there was 1 session per week; in weeks 7–10, there was 1 session
degree of difficulties ranging from 0 (absence of difficulty) to 6 (severe
per fortnight. The manual was developed based on
difficulty).
Turkington et al. (2002) and Wright (2009) and tailored to Chinese
At week 62 follow-up assessment, participants were evaluated on
patients with first episode schizophrenia. The main goals of this inter-
their employment status and any relapse or hospitalization by a
vention were to apply the general principles of CBT, such as case

277
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

formulation, goal setting, homework and the cognitive process and 2.7. Statistical analysis
behavior experiment, with a focus on treatment of relapse prevention
and primary positive symptoms. All data were analyzed using the Statistical Package for Social
The intervention can be divided into three stages. In the first stage, Sciences (SPSS) Version 20.0. The demographic and baseline char-
therapists built a good rapport with patients, assessed patients’ life acteristics of the two groups were compared using independent-sample
experiences, and developed patients’ problem list. In the intermediate t-test, Mann-Whitney U test or Chi-square test depending on the nature
stage, patients primarily learnt how to manage positive symptoms (such (continuous/ dichotomous, normal distribution/skewed distribution) of
as delusions, hallucinations, thought disorder) via behavioural experi- those variables. Due to the small sample size, analysis of the outcome
ments, graded exposure, reattribution of symptoms and coping strate- measures followed an intention-to-treat (ITT) framework. Regardless of
gies. Emotional problems were also discussed. In the last stage, thera- study completing, all cases were included in the statistical analysis of
pists focused on stress management, problem solving and medication the curative effect, and the last observation carried forward (LOCF)
adherence, as well as relapse prevention. Patients in stress management method was used to handle the missing data. If the outcome data in-
sessions discussed predisposing factors, stressors (e.g. sleep deprivation, dicated a skewed distribution, reciprocal transformation was used. The
trauma and posttraumatic stress symptom), protective factors (e.g. so- four time periods were treated as a four-level repeated measure in
cial skills, social support and help seeking) and risk factors (e.g. social ANOVA analysis comparing both groups. Main effects of intervention
isolation, substance abuse). In relapse prevention sessions, patients group, time and the interaction between treatment group and time were
learned: 1) 6 stages of relapse (quiet period, stress period, hopeless also estimated (post-hoc comparison). All tests were two-tailed with α
period, loss of control period, mental confusion and psychotic symptom set at 0.05. Zero-inflated passion models were used to compare count
occurrence); 2) the early signs of relapse; 3) how to make their own list data on hospitalization and relapse between groups. Kaplan-Meier
of signs of relapse with unique features; 4) how to monitor their survival analyses were used to calculate the estimated time from
symptoms in daily life; 5) how to make emergency plans for early baseline to response and relapse. The analyses included patients who
warning signs including grading signs, prevention strategies and plans met the criteria for response or relapse and those who were lost to
for seeking help. At the beginning of each session, therapists reviewed follow-up without a documented response or relapse, as well as those
patients’ homework to ensure patients had mastered the skills covered who did not meet response or relapse criteria at their last assessment.
in previous sessions. The number of sessions attended by patients was also compared. Nested
In addition, all participants received treatment as usual; patients ANOVA was conducted on the primary outcomes to explore the nesting
were seen monthly by a psychiatrist across the duration of the trial. effect of different therapists. We also used stepwise multivariate re-
gression model to estimate which factors affected the PANSS Total score
2.4.1. Therapists and fidelity and the Positive score.
Four therapists provided CBT for the treatment group. All four
clinical psychologists held master's degrees in Clinical Psychology with 3. Results
5–20 years’ experience using psychotherapy in hospitals on psychiatric
patients with mental disorder. They had been trained in the application 3.1. Participant flow and characteristics
of cognitive therapy for psychosis by experienced cognitive–behavioral
therapists and received over 200 h of supervision each. Fig. 1 shows the CONSORT diagram. After exclusion of patients who
During the study, therapists were supervised in two ways: peer su- did not meet the inclusion criteria, eighty patients were randomly as-
pervision and expert supervision. During peer supervision, therapists signed to CBT + TAU (treatment group) or TAU group. During the
presented the case formulation, treatment plan and therapy progress for study, a total of 13 patients withdrew. Patients assigned to CBT re-
every case during the first six sessions, and peer therapists provided ceived a mean of 7.4 sessions (Range 4–10; SD = 2.1). 31 (77.5%) of
feedback and suggestions. Therapists then submitted written case re- the 40 patients had at least six sessions, suggesting good adherence to
ports, case formulations, treatment plans, therapy processes and team treatment group. The 62-week follow-up was completed by 35 parti-
members’ questions arising from the peer supervision sessions to two cipants (87.5%) in the treatment group and 29 participants (72.5%) in
experts for supervision. Expert supervision was delivered once every 2 the TAU group. There was no significant difference between two groups
weeks face-to-face. Another expert therapist in CBT for psychosis from in terms of the proportion or the demographic and clinical character-
the UK was invited to provide supervision via phone, Skype or email on istics of patients who failed to complete the assessment at any time
a monthly basis. During supervision, compliance with the manual was point (See Supplemental Table 1). There was no significant difference in
also checked by experts to ensure the fidelity of therapy. terms of the demographic and clinical characteristics of patients be-
tween completers and non-completers at week 62 (See Supplemental
2.5. Treatment as usual (TAU) Tables 2–4).
The mean age was 26.35 (SD = 7.41) for treatment group, 28.6
TAU included medication and case management (including some (SD = 5.86) for TAU, and approximately 35% were woman in treat-
psychological health education and social support). Psychiatrists eval- ment group and 50% in the control group. There were no statistically
uated all patients and formulated appropriate drug treatment, and saw significant differences at baseline between the two groups in demo-
patients and their family members in the outpatient unit each month, in graphic characteristics (age, gender, education background, marital and
order to supervise and encourage medication adherence and provide occupation status) or in clinical characteristics (duration of illness,
guidance to family members on managing adverse drug reactions. PANSS, SAI, PSYRATS and PSP scores) (Table 1). The treatment group
When patients presented with impulsiveness or serious adverse drug took a 391–389 mg equivalent of chlorpromazine, and the TAU group
reactions, psychiatrists initiated an emergency response mechanism. took a 356–339 mg equivalent of chlorpromazine from baseline to 62
weeks. There were also no significant difference in antipsychotic
2.6. Medication treatment medication use at baseline to week 62 in chlorpromazine equivalents
(Table 2). Overall, participants were clinically stable at baseline – while
Medication prescription was not affected by the trial protocol. The they had scores of above 4 on P1 or P3 on the PANSS (as stated in the
patients in both groups remained under their usual psychiatric care. inclusion criteria), no participants scored above 10 when totaling the
Medication decisions were made by their primary treating team based sub-set of items measuring thought-disorder (conceptual disorganisa-
on clinical needs. The doses of antipsychotic medication were recorded tion (P2), hallucinatory behaviour (P3), grandiosity (P5) and unusual
and converted into equivalent doses of chlorpromazine (Table 3). thought content (G9)) and none of these items scored 6 or above.

278
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

Table 1
Comparison of baseline demographic and clinical characteristics between CBT + TAU and TAU group.
Variables CBT + TAU group (n = 40) TAU group (n = 40) t/χ2/z P

Age, years: mean (s.d.) 26.35 (7.41) 28.60 (5.71) t = −1.485 0.132
Gender χ2 = 1.841 0.175
Male 26 20
Female 14 20
Education, years: mean (s.d.) 13.30 (2.95) 13.00 (2.29) t = 0.496 0.613
Marital status (n) χ2 = 2.581 0.108
Married 28 21
Single 12 19
Employment (n) χ2 = 0.487 0.485
Yes 16 13
No 24 27
PANSS total: mean (s.d.) 68.48 (15.67) 69.08 (15.91) t = −0.170 0.866
Positive Scale: mean (s.d.) 17.88 (5.40) 18.68 (5.76) t = −0.641 0.523
Negative Scalea: media (range) 15 (7–24) 15 (7–24) z = 0.275 0.783
General Scalea: media (range) 33.5 (20–52) 36.5 (19–55) z = 0.419 0.675
PSYRATSa: media (range) 27.5 (10–55) 32.5 (11–57) z = −0.905 0.365
SAI: mean (s.d.) 7.10 (3.55) 6.80 (3.63) t = 0.374 0.709
PSP: mean (s.d.) 52.88 (13.72) 53.18 (14.83) t = 0.094 0.925

CBT, Cognitive-behavioural therapy; TAU, Treatment as usual; PANSS, Positive and Negative Syndrome Scale; SAI, Schedule for Assessing Insight; PSYRATS,
Psychotic Symptoms Rating Scale; PSP, Personal and Social Performance.
a
For skewed distributions, Mann–Whitney U tests were used.

3.2. Outcomes (see Figs. 2 and 3).

Table 3 shows the main effect of time, group and time × group
3.2.1. Primary outcomes interaction on each symptom outcome using ITT analysis. A main effect
By the end of the trial, four participants (10%, 4/40) in the CBT of time was found for PANSS-Total, PANSS–Positive, PANSS-Negative
group had relapsed compared to thirteen participants (32.5%, 13/40) and PANSS-General subscales, as well as PSYRATS, indicating a sig-
in the TAU group. Similarly, three participants (7.5%, 3/40) in the CBT nificant decrease over time across groups for all symptom assessments.
group had been admitted to hospital compared to eight participants However, for total symptoms, there was a significant time × group
(20%, 8/40) in the TAU group, though this was not statistically sig- interaction, indicating that the CBT group showed greater decreases in
nificant (χ2 = 3.529, P = 0.06). To overcome overdispersion caused by total symptoms than the TAU group at follow-up (At baseline, F
zero-inflation of the count data, zero-inflated Poisson models were used (1,78) = 0.029, P = 0.866; At 10 weeks, F(1,78) = 3.136, P = 0.080;
to compare counts of hospitalization and relapse between groups. At 36 weeks, F(1,78) = 5.987, P = 0.017; At 62 weeks, F
Between the two group there was no significant difference in hospita- (1,78) = 14.135, P < 0.001). Similarly, for general symptoms, there
lization (Z = −1.397, P = 0.162), but there was significant difference was a significant time × group interaction, indicating that the CBT
in relapse (Z = −2.019, P = 0.044). group showed greater decreases in general symptoms over time (At
Using the criterion of a 25% or greater reduction on the PANSS total baseline, F(1,78) = 0.066, P = 0.798; At 10 weeks, F(1,78) = 2.319,
score, 47.5% (19/40) of those in the CBT group experienced a clinically P = 0.132; At 36 weeks, F(1,78) = 6.146, P = 0.015; At 62 weeks, F
significant improvement compared with only 25% (10/40) of those in (1,78) = 13.890, P = 0.000) (See Supplemental Table 5).
TAU alone group. This difference was statistically significant For positive symptoms, there was also a main effect of group but no
(χ2 = 4.381, P = 0.036) and indicates that the CBT intervention had a significant time × group interaction (At baseline, F(1,78) = 0.411,
more clinically meaningful effect on overall symptoms. The number P = 0.523; At 10 weeks, F(1,78) = 2.329, P = 0.131; At 36 weeks, F
needed to treat for improvement in overall symptoms was 4, which (1,78) = 1.809, P = 0.182; At 62 weeks, F(1,78) = 11.866, P = 0.001).
means that for every four patients treated with CBT there was one Similarly, for PSYRATS, there was a main effect of group, but there was
additional patient who demonstrated a clinically significant improve- no significant time × group interaction, indicating that the CBT group
ment over TAU. Similarly, using the criterion of a 25% or greater re- had significantly lower scores than the TAU group on delusions and
duction in PANSS positive score, 85% (34/40) of those in the CBT group auditory hallucinations (At baseline, F(1,78) = 0.791, P = 0.376; At 10
experienced a significant improvement compared with 65% (26/40) of weeks, F(1,78) = 10.270, P = 0.002; At 36 weeks, F(1,78) = 9.022,
those in TAU alone group. This difference was statistically significant P = 0.004; At 62 weeks, F(1,78) = 14.234, P < 0.001). For negative
(χ2 = 14.267, P = 0.039). The number needed to treat for improve- symptoms, there was no significant time × group interaction or main
ment in positive symptoms was 5. The difference between the survival effect of group, indicating that the two conditions resulted in significant
curves was significant for both response in positive symptoms (log reductions of negative symptoms, but with no significant difference
rank = 5.413, P = 0.020) and relapse (log rank = 6.073, P = 0.014) across conditions (At baseline, F(1,78) = 0.044, P = 0.834; At 10

Table 2
Changes in prescribed antipsychotic drugs.
Assessment CBT + TAU group (n = 40) TAU group (n = 40) t P

Chlorpromazine equivalents, mg
Baseline 391.20(199.68) 356.00(201.06) t = 0.776 0.440
10 weeks 383.70(185.19) 354.68(205.25) t = 0.656 0.514
36 weeks 389.33(185.74) 349.42(201.32) t = 0.910 0.365
62 weeks 389.33(185.74) 339.95(199.54) t = 1.125 0.264

CBT, Cognitive-behavioural therapy; TAU, Treatment as usual.

279
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

Fig. 2. . Kaplan-Meier estimates of times to response; CBT, Cognitive-behavioural therapy; TAU, Treatment as usual.

Fig. 3. . Kaplan-Meier estimates of times to relapse; CBT, Cognitive-behavioural therapy; TAU, Treatment as usual.

weeks, F(1,78) = 1.064, P = 0.306; At 36 weeks, F(1,78) = 1.732, week 62 (P < 0.001). Similarly, lower baseline PANSS total score
P = 0.193; At 62 weeks, F(1,78) = 4.443, P = 0.038) (See (P < 0.001) and more treatment sessions (P = 0.014) predicted lower
Supplemental Table 5). PANSS positive score at week 62 (See Supplemental Table 7 and 8).
Our nested ANOVA analysis indicated that there was a significant
difference between therapists regarding improvement in PANSS
Positive scale and PSYRATS (F = 2.902, P = 0.040; F = 4.735, 3.2.2. Secondary outcomes
P = 0.004); there were no significant differences for PANSS Total, Analyses of the SAI total score and PSP both revealed main effects of
Negative Scale, General Scale, SAI, PSP (F = 2.603, P = 0.058; time, suggesting an increase in insight and functioning over time for
F = 0.511, P = 0.676; F = 2.244, P = 0.090; F = 0.733, P = 0.536; both groups. However, there was a main effect of time x group inter-
F = 0.853, P = 0.470). Sub analyses indicated that the difference was action for PSP scores indicating that the CBT group had significantly
driven by therapist 2 and 4 (See Supplemental Table 6). higher levels of functioning compared with the TAU group at follow up
The stepwise multivariate regression analyses indicated that lower (At baseline, F(1,78) = 0.001, P = 0.980; At 10 weeks, F
baseline negative PANSS score predicted lower PANSS total score at (1,78) = 5.329, P = 0.024; At 36 weeks, F(1,78) = 4.523, P = 0.037;
At 62 weeks, F(1,78) = 6.616, P = 0.012). In contrast, there was no

280
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

Table 3
Repeated-measures analysis of variance (ANOVA) in clinical symptoms, insight and social functioning by group at the four time points.
Outcomes, group Baseline (n = 40) 10 weeks (n = 40) 36 weeks (n = 40) 62 weeks (n = 40) Value (ITT)
Time Group time × group

PANSS total F = 102.757 P = 0.000 F = 5.475 P = 0.022 F = 3.376 P = 0.042

CBT + TAU 68.48(15.67) 51.65(11.22) 44.88(8.06) 43.20(7.89)
TAU 69.07(16.33) 55.05(9.56) 49.60(10.47) 52.00(13.92)
Positive scale F = 102.122 P = 0.000 F = 4.603 P = 0.035 F = 1.342 P = 0.265
CBT + TAU 17.87(5.40) 11.92(4.10) 9.97(2.48) 9.42 (2.62)
TAU 18.68(5.76) 13.17(3.17) 10.85(3.29) 12.03(3.99)
Negative scalea F = 17.513a P = 0.000a F = 1.753a P = 0.189a F = 1.443a P = 0.237a
CBT + TAU 15 (7–24) 13(7–20) 11 (7–19) 11 (1–19)
TAU 15(7–24) 14(7–20) 13(7–22) 11.5(7–26)
General scalea F = 82.534a P = 0.000 F = 5.143a P = 0.026a F = 3.884 a P = 0.027a
a

CBT + TAU 33.5(20–52) 27(16–39) 23 (16–34) 23(16–30)

TAU 36.5(19–55) 27.5(18–42) 24.5(16–43) 26 (17–46)
PSYRATSa F = 97.319a P = 0.000a F = 12.215a F = 1.945a P = 0.024a
P = 0.001a
CBT + TAU 27.5(10–56) 9.5(1–40) 3(1–23) 2(1–26)
TAU 32.5(11–57) 18.5(4–42) 7(1–47) 7(1–50)
SAI F = 37.760 P = 0.000 F = 2.428 P = 0.123 F = 0.814 P = 0.458
CBT + TAU 7.10(3.55) 10.00(3.15) 11.08(3.27) 11.02(3.03)
TAU 6.80(3.63) 9.10(3.12) 9.98(3.48) 9.58(3.85)
PSP F = 78.444 P = 0.000 F = 4.475 P = 0.055 F = 3.119 P = 0.039
CBT + TAU 52.88(13.72) 66.42(11.86) 74.75(10.14) 75.73(11.88)
TAU 52.79(14.83) 59.00(16.42) 68.74(14.62) 67.85(15.18)

CBT, Cognitive-behavioural therapy; TAU, Treatment as usual; PANSS, Positive and Negative Syndrome Scale; PSYRATS, Psychotic Symptoms Rating Scale; PSP,
Personal and Social Performance.
a
For skewed distributions, the original data were turned into reciprocal firstly, then performed repeated-measures analysis of variance.

main effect of group or significant time × group interaction for SAI 2008; Patel et al., 2014). Third, the relapse and hospitalization rate in
total score (At baseline, F(1,78) = 0.140, P = 0.709; At 10 weeks, F this study was low compared to previous studies with inpatient sample
(1,78) = 1.650, P = 0.203; At 36 weeks, F(1,78) = 2.126, P = 0.149; population. For instance, Jolley et al. (2003) report a hospitalization
At 62 weeks, F(1,78) = 3.501, P = 0.065) (See Table 3 and Supple- rate of 25% for the cognitive therapy group and 22.2% in TAU group in
mental Table 5). early psychosis patients, and Tarrier et al. (2004) found the relapse
The CBT group had significantly better employment outcomes than rates/hospital readmission in CBT, supportive counselling, TAU group
TAU Group (χ2 = 4.053, P = 0.044); 24 participants in the CBT group were 54.6%/33%, 52.1%/29% and 51.1%/36%, respectively. However,
had a current employment position at the end of 62 weeks compared to the present study targeted an outpatient sample with less severe
15 participants in the TAU group. symptoms. Future studies should evaluate the effectiveness of this CBT
intervention for hospitalized FES patients with more severe psychotic
symptoms.
4. Discussion Part of patients’ acceptance of their illness and self-knowledge is
recognising the importance of treatment, which is associated with
This study demonstrates that a brief CBT targeting relapse preven- subsequent improved adherence to medication (Bedford and David,
tion and positive symptom was effective in first episode schizophrenia 2014; Wang et al., 2003) and reduced relapse rates and hospital read-
patients. Specifically, compared with treatment-as-usual, it was asso- missions (Kemp et al., 1996). A number of previous studies have de-
ciated with significantly greater reductions in relapse, as well as sig- monstrated that CBT can improve insight in first episode psychosis
nificantly greater improvements in positive, general and overall psy- patients (Jackson et al., 1998; Wang et al., 2003), potentially via psy-
chotic symptoms and social functioning (including employment status). choeducation or behavioural experiments (Rathod et al., 2005). Inter-
This study also demonstrates that this adapted brief CBT model is fea- estingly, in the present study we found that both groups had high SAI
sible and acceptable in a Chinese FES population, as indicated by a low scores at 1 year with significant improvements in insight over time.
drop-out rate. Although patients in the CBT group showed higher insight scores than
Previous CBT interventions focusing on general psychotic symptoms TAU, the failure to find a significant group difference in this study may
have shown inconsistent results for reducing positive symptoms in first in part be due to patients in the TAU group experiencing regular contact
episode psychosis (Haddock et al., 1999b; Jackson et al., 2008; Lewis with their psychiatrist and clinicians assessing them for research, who
et al., 2002). The present study suggests that the positive symptom may have offered informal supportive consulting.
management aspect of the intervention in our study may have played Impairment in social and occupational functioning is common fol-
an important role in overall symptom reduction. lowing first episode psychosis (Penn and Mueser, 1996). Our study
A previous systematic review concluded that CBT intervention were reports that the CBT group had significant improvements in their gen-
not especially effective in reducing rates of relapse or hospital admis- eral psychosocial functioning, which is in line with previous studies
sion in FEP (Morrison, 2009). However, our study reports significantly (Jackson et al., 1998; Penn et al., 2005; Power et al., 2003). This may
greater reductions in relapse for participants who received CBT com- have been driven by standard CBT intervention techniques, such as
pared to TAU alone, and a trend for reduced hospitalization. There are problem-solving skills, stress management and behavior activation, or
number of explanations for this. First, the targeted relapse prevention the targeted focused on reducing relapse rates. We also found that CBT
focus of our CBT intervention may have had a direct, unique and spe- group had better occupational functioning (reflected in their current
cific effect on reducing relapse risk. Second, the CBT intervention was employment status) than TAU. Previous evidence indicates that CBT is a
associated with greater positive symptom compared with those in TAU particularly promising intervention to improve work-related beliefs and
group, which may be critical in preventing relapse in FES (Lavretsky,

281
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

behaviors, resulting in enhanced employment outcome (Kukla et al., assignment, with significant differences between therapist 2 and
2016; Lysaker et al., 2009). Importantly, a recent study conducted by therapist 4 on improvement measured by both PANSS Positive sub-scale
Kukla et al. (2018), found that participants with schizophrenia spec- and PSYRATS. It is unclear whether this may have been due to non-
trum disorders who were allocated to receive CBT with cognitive re- random assignment of therapists. Finally, this study did not use active
mediation (CR) had significantly more positive improvements in global treatment as a control group, meaning that non-specific therapeutic
work performance and work quality compared with being allocated to factors may have driven group differences – for example, the format of
receive the vocational support group and CBT alone. The authors sug- more frequent regular sessions and subsequent increased focus and
gested that improving memory and capacity for learning and problem attention on the participant may drive the effect in the treatment group.
solving through CR intervention could amplify the effects of CBT for Therefore, we cannot conclusively say which element of the brief CBT
better employment outcomes (Kukla et al., 2018). It would be worth- intervention drives it efficacy or whether a non-specific intervention
while for future research to examine CBT with CR in FES populations. would find similar effects.
It is worth noting that definitions and measurements of relapse, In summary, this is the first RCT study indicating effectiveness of a
when relying on hospitalization, may vary between countries. In many brief CBT intervention focused on relapse prevention in patients with
developed countries, relapse can be treated at outpatient and commu- first-episode schizophrenia. This positive finding suggests that more
nity clinic services without being admitted to hospital due to the es- research is warranted exploring the effects of CBT for first episode
tablishment of community mental health teams, but that is not true for psychosis, specifically targeting relapse prevention and positive
China. Although China has established a three-tier network (from town symptom.
to district and city levels) based on community health services for the
treatment and prevention of psychosis, most of the community health Declaration of conflicting interests
service remain limited, suffering from a severe lack of mental health
professional workforce, service capabilities and inadequate regulatory The authors declare that there is no conflict of interest.
resources (Tse et al., 2013). There is also a significant urban-rural
disparity in community health service development (Liu et al., 2011). Funding
The key responsibility for community health services is to reduce the
possibility of criminal offending among people with severe psychosis This study was supported by Capital Clinical Characteristic
(Xiang et al., 2012). Furthermore, specialist psychiatrists are con- Application Research (Z111107058811078), the funding body had no
centrated in large general hospitals or specialized psychiatry hospitals role in the design of the study and collection, analysis, and inter-
and practitioners at community health clinics are not similarly pretation of data and in writing the manuscript.
equipped with the sufficient and experienced qualifications. Therefore,
we would have expected a higher rate of hospitalisation for relapse in Acknowledgements
China than other countries and it is surprising that we found lower
rates. We thank all the patients who participated in this study.
In addition, while few studies have investigated the association
between stigma and psychiatric re-hospitalization, some studies - for References
instance, Alex (2012) – have found that stigma about mental disorder
within a family can lead to higher rates of re-hospitalization for re- Alvarez-Jiménez, M., Parker, A.G., Hetrick, S.E., McGorry, P.D., Gleeson, J.F., 2009.
latives with psychosis; perhaps due to neglecting the need for treatment Preventing the second episode: a systematic review and meta-analysis of psychosocial
and pharmacological trials in first-episode psychosis. Schizophr. Bull. 37 (3),
leading to worsened symptoms. In Chinese culture, the role of 619–630.
Confucianism is significant and impacts upon understanding of mental Bedford, N.J., David, A.S., 2014. Denial of illness in schizophrenia as a disturbance of self-
health disorders and stigma. The collectivistic tradition encourages reflection, self-perception and insight. Schizophr. Res. 152 (1), 89–96.
Chen, R., Xi, Y., Wang, X., Li, Y., He, Y., Luo, J., 2018a. Perception of inpatients following
people to maintain social and familial harmony, and the notion of remission of a manic episode in bipolar I disorder on a group-based Psychoeducation
“saving face” is of crucial importance to social identity and hierarchy program: a qualitative study. BMC Psychiatry 18 (1), 26.
(Xi et al., 2017). Therefore, people suffering from mental health dis- Chen, R., Zhu, X., Capitao, L.P., Zhang, H., Luo, J., Wang, X., et al., 2018b.
Psychoeducation for psychiatric inpatients following remission of a manic episode in
orders are equated with extreme shame for the whole family. Although bipolar I disorder: a randomized controlled trial. Bipolar Disord.
numerous studies have investigated the stigma and self-stigma among David, A.S., 1990. Insight and psychosis. Br. J. Psychiatry 156 (6), 798–808.
schizophrenia patients in Chinese populations (Lv et al., 2013; Phillips Drury, V., Birchwood, M., Cochrane, R., 2000. Cognitive therapy and recovery from acute
psychosis: a controlled trial: 3. Five-year follow-up. Br. J. Psychiatry 177 (1), 8–14.
et al., 2002), none have examined the influence of stigma and re-hos-
Drury, V., Birchwood, M., Cochrane, R., MacMillan, F., 1996. Cognitive therapy and re-
pitalization. Again, while literature from other countries may have in- covery from acute psychosis: a controlled trial. I. Impact on psychotic symptoms. Br.
dicated increased hospitalization rates due to increased stigma in J. Psychiatry 169 (5), 593–601.
China, we found unusually low rates; this may reflect a different re- Garety, P.A., Fowler, D.G., Freeman, D., Bebbington, P., Dunn, G., Kuipers, E., 2008.
Cognitive–behavioural therapy and family intervention for relapse prevention and
lationship with stigma than expected (stigma may delay or discourage symptom reduction in psychosis: randomised controlled trial. Br. J. Psychiatry 192
seeking necessary hospitalisation) or may be reflective of the specific (6), 412–423.
hospital or region. Gleeson, J., Cotton, S.M., Alvarez-Jimenez, M., Wade, D., Gee, D., Crisp, K., et al., 2009. A
randomized controlled trial of relapse prevention therapy for first-episode psychosis
The present study has several limitations. Firstly, while we assessed patients. J. Clin. Psychiatry 70 (4), 477–486.
patients at numerous follow ups over one year, this follow up period is Grawe, R., Falloon, I., Widen, J., Skogvoll, E., 2006. Two years of continued early
still too brief to explore the full long-term effects of CBT on symptoms, treatment for recent‐onset schizophrenia: a randomised controlled study. Acta
Psychiatrica Scandinavica 114 (5), 328–336.
functioning, insight and relapse in FES patients. Secondly, the sample Gumley, A., O'GRADY, M., McNay, L., Reilly, J., Power, K., Norrie, J., 2003. Early in-
did not include patients with severe symptoms requiring hospitaliza- tervention for relapse in schizophrenia: results of a 12-month randomized controlled
tion, impacting on generalizability. Future study should examine trial of cognitive behavioural therapy. Psychol. Med. 33 (3), 419–431.
Haddock, G., McCarron, J., Tarrier, N., Faragher, E., 1999a. Scales to measure dimensions
whether a brief CBT intervention has similar outcomes for FES in- of hallucinations and delusions: the psychotic symptom rating scales (PSYRATS).
patients or patients with more severe symptoms. Thirdly, we did not use Psychol. Med. 29 (4), 879–889.
a standardized measure of treatment fidelity such as the Cognitive Haddock, G., Tarrier, N., Morrison, A., Hopkins, R., Drake, R., Lewis, S., 1999b. A pilot
study evaluating the effectiveness of individual inpatient cognitive-behavioural
Therapy Scale to check treatment fidelity. This is because such a scale
therapy in early psychosis. Soc. Psychiatry Psychiatr. Epidemiol. 34 (5), 254–258.
has not been developed in China; we would strongly encourage future Jackson, H., McGorry, P., Edwards, J., Hulbert, C., Henry, L., Francey, S., et al., 1998.
research to pursue this to help ensure high quality of clinical trials. Cognitively-oriented psychotherapy for early psychosis (COPE): preliminary results.
Fourthly, there was variability in outcomes related to therapist Br. J. Psychiatry.

282
Y. Liu et al. Psychiatry Research 272 (2019) 275–283

Jackson, H., McGorry, P., Killackey, E., Bendall, S., Allott, K., Dudgeon, P., et al., 2008. Power, P., Bell, R., Mills, R., Herrman‐Doig, T., Davern, M., Henry, L., et al., 2003. Suicide
Acute-phase and 1-year follow-up results of a randomized controlled trial of CBT prevention in first episode psychosis: the development of a randomised controlled
versus befriending for first-episode psychosis: the ACE project. Psychol. Med. 38 (5), trial of cognitive therapy for acutely suicidal patients with early psychosis. Austr. N.
725–735. Z. J. Psychiatry 37 (4), 414–420.
Jolley, S., Garety, P., Craig, T., Dunn, G., White, J., Aitken, M., 2003. Cognitive therapy in Rathod, S., Kingdon, D., Smith, P., Turkington, D., 2005. Insight into schizophrenia: the
early psychosis: a pilot randomized controlled trial. Behav. Cognit. Psychother. 31 effects of cognitive behavioural therapy on the components of insight and association
(4), 473–478. with sociodemographics—data on a previously published randomised controlled
Kay, S.R., Fiszbein, A., Opfer, L.A., 1987. The positive and negative syndrome scale trial. Schizophr. Res. 74 (2), 211–219.
(PANSS) for schizophrenia. Schizophr. Bull. 13 (2), 261. Robinson, D.G., Woerner, M.G., Delman, H.M., Kane, J.M., 2005. Pharmacological
Kemp, R., Hayward, P., Applewhaite, G., Everitt, B., David, A., 1996. Compliance therapy treatments for first-episode schizophrenia. Schizophr. Bull. 31 (3), 705–722.
in psychotic patients: randomised controlled trial. Bmj 312 (7027), 345–349. Saraceno, B., van Ommeren, M., Batniji, R., Cohen, A., Gureje, O., Mahoney, J., et al.,
Klingberg, S., Wölwer, W., Engel, C., Wittorf, A., Herrlich, J., Meisner, C., et al., 2011. 2007. Barriers to improvement of mental health services in low-income and middle-
Negative symptoms of schizophrenia as primary target of cognitive behavioral income countries. Lancet 370 (9593), 1164–1174.
therapy: results of the randomized clinical TONES study. Schizophr. Bull. 37 Tarrier, N., Lewis, S., Haddock, G., Bentall, R., Drake, R., Kinderman, P., et al., 2004.
(suppl_2), S98–S110. Cognitive-behavioural therapy in first-episode and early schizophrenia: 18-month
Kukla, M., Bell, M.D., Lysaker, P.H., 2018. A randomized controlled trial examining a follow-up of a randomised controlled trial. Br. J. Psychiatry 184 (3), 231–239.
cognitive behavioral therapy intervention enhanced with cognitive remediation to Tianmei, S., Jianzhong, Y., Liang, S., 2004. The reliability, validity of PANSS and its
improve work and neurocognition outcomes among persons with schizophrenia implications. Chin. Mental Health J. 1, 015.
spectrum disorders. Schizophr. Res. Tianmei, S., Liang, S., Yun'ai, S., Chenghua, T., Jun, Y., Jia, C., et al., 2011. The Chinese
Kukla, M., Strasburger, A.M., Lysaker, P.H., 2016. A CBT intervention targeting compe- version of the personal and social performance scale (PSP): validity and reliability.
titive work outcomes for persons with mental illness. Psychiatr. Serv. 67 (6) 697-697. Psychiatry Res. 185 (1), 275–279.
Lavretsky, H., 2008. History of schizophrenia as a psychiatric disorder. Clinical Handbook Tse, S., Ran, M.-S., Huang, Y., Zhu, S., 2013. Mental Health Care Reforms in Asia: the
of Schizophrenia. pp. 1. urgency of now: building a recovery-oriented, community mental health service in
Lewis, S., Tarrier, N., Haddock, G., Bentall, R., Kinderman, P., Kingdon, D., et al., 2002. China. Psychiatr. Serv. 64 (7), 613–616.
Randomised controlled trial of cognitive-behavioural therapy in early schizophrenia: Turkington, D., Kingdon, D., Turner, T., 2002. Effectiveness of a brief cognitive—beha-
acute-phase outcomes. Br. J. Psychiatry 181 (S43), s91–s97. vioural therapy intervention in the treatment of schizophrenia. Br. J. Psychiatry 180
Liu, J., Ma, H., He, Y.L., Xie, B., Xu, Y.F., Tang, H.Y., et al., 2011. Mental health system in (6), 523–527.
China: history, recent service reform and future challenges. World Psychiatry 10 (3), Üçok, A., Polat, A., Çakır, S., Genç, A., 2006. One year outcome in first episode schizo-
210–216. phrenia. Eur. Arch. Psychiatry Clin. Neurosci. 256 (1), 37–43.
Lv, Y., Wolf, A., Wang, X., 2013. Experienced stigma and self-stigma in Chinese patients Wang, C., Li, Y., Zhao, Z., Pan, M., Feng, Y., Sun, F., et al., 2003. Controlled study on long-
with schizophrenia. Gen. Hosp. Psychiatry 35 (1), 83–88. term effect of cognitive behavior intervention on first episode schizophrenia. Chin.
Lysaker, P.H., Davis, L.W., Bryson, G.J., Bell, M.D., 2009. Effects of cognitive behavioral Men. Health J. 17 (3), 200–202.
therapy on work outcomes in vocational rehabilitation for participants with schizo- Wiersma, D., Nienhuis, F.J., Slooff, C.J., Giel, R., 1998. Natural course of schizophrenic
phrenia spectrum disorders. Schizophr. Res. 107 (2-3), 186–191. disorders: a 15-year followup of a Dutch incidence cohort. Schizophr. Bull. 24 (1), 75.
Morrison, A.P., 2009. Cognitive behaviour therapy for first episode psychosis: good for World Health Organization, 2001. The World Health Report 2001. Mental Health: New
nothing or fit for purpose? Psychosis 1 (2), 103–112. Understanding, New Hope. WHO, Geneva.
Ohlsen, R.I., O'Toole, M.S., Purvis, R.G., Walters, J.T., Taylor, T.M., Jones, H.M., et al., Wright, J.H., 2009. Cognitive-behavior therapy for severe mental illness: an illustrated
2004. Clinical effectiveness in first-episode patients. Eur. Neuropsychopharmacol. 14, guide. Am. Psychiatr. Pub.
S445–S451. Wykes, T., Steel, C., Everitt, B., Tarrier, N., 2008. Cognitive behavior therapy for schi-
Patel, K.R., Cherian, J., Gohil, K., Atkinson, D., 2014. Schizophrenia: overview and zophrenia: effect sizes, clinical models, and methodological rigor. Schizophr. Bull. 34
treatment options. Pharm. Ther. 39 (9), 638. (3), 523–537.
Penn, D.L., Mueser, K.T., 1996. Research update on the psychosocial treatment of schi- Xi, Y., Chen, R., Yan, F., Ma, X., Rakofsky, J.J., Tang, L., et al., 2017. Low post-traumatic
zophrenia. The Am. J. Psychiatry 153 (5), 607. stress disorder rate in Chinese in Beijing, China. Asian J. Psychiatry 30, 79–83.
Penn, D.L., Waldheter, E.J., Perkins, D.O., Mueser, K.T., Lieberman, J.A., 2005. Xiang, Y.-T., Yu, X., Sartorius, N., Ungvari, G.S., Chiu, H.F., 2012. Mental health in China:
Psychosocial treatment for first-episode psychosis: a research update. Am. J. challenges and progress. Lancet 380 (9855), 1715–1716.
Psychiatry 162 (12), 2220. Xu, Z.-T., Guo, Z.H., Fu, Z.Y., Na, W., Zhang, Y., 2013. Reliability and validity of the
Perkins, D.O., Gu, H., Boteva, K., Lieberman, J.A., 2005. Relationship between duration of Chinese version of the schedule for assessment of insight. Chin. J. Behav. Med. Brain
untreated psychosis and outcome in first-episode schizophrenia: a critical review and Sci. 22 (8), 752–754.
meta-analysis. Am. J. Psychiatry 162 (10), 1785–1804. Xu, Z.J., Li, Z.J., Guo, Z.H., Chen, Q., Zhang, Y., 2012. Reliability and validity of the
Perkins, D.O., Lieberman, J.A., Gu, H., Tohen, M., McEvoy, J., Green, A.I., et al., 2004. Chinese version of the psychotic symptom rating scales. Chin. J. Clin. Psychol. 20 (4),
Predictors of antipsychotic treatment response in patients with first-episode schizo- 445–447.
phrenia, schizoaffective and schizophreniform disorders. Br. J. Psychiatry 185 (1), Zhu, Y., Krause, M., Huhn, M., Rothe, P., Schneider-Thoma, J., Chaimani, A., et al., 2017.
18–24. Antipsychotic drugs for the acute treatment of patients with a first episode of schi-
Phillips, M.R., Pearson, V., Li, F., Xu, M., Yang, L., 2002. Stigma and expressed emotion: a zophrenia: a systematic review with pairwise and network meta-analyses. Lancet
study of people with schizophrenia and their family members in China. Br. J. Psychiatry 4 (9), 694–705.
Psychiatry 181 (6), 488–493. Zimmermann, G., Favrod, J., Trieu, V., Pomini, V., 2005. The effect of cognitive beha-
Pontes, L.M., Martins, C.B., Napolitano, I.C., Fonseca, J.R., Oliveira, G.M., Iso, S.M., et al., vioral treatment on the positive symptoms of schizophrenia spectrum disorders: a
2013. Cognitive training for schizophrenia in developing countries: a pilot trial in meta-analysis. Schizophr. Res. 77 (1), 1–9.
Brazil. Schizophr. Res. Treat. 2013.

283