Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Rate and predictors of disengagement from a 2-year early intervention program for psychosis in Hong Kong

Schizophrenia Research



This study aims to examine the prevalence and predictors of disengagement in a longitudinal cohort of first-episode psychosis (FEP) patients.


Seven hundred FEP patients aged 15 to 25 enrolled into the Early Assessment Service for Young People with Psychosis (EASY) from 2001 to 2003 were recruited into the study. Data on sociodemographics, clinical characteristics, baseline symptoms and functioning and medication adherence were collected. Rate and predictors of service disengagement were the outcomes of interest. Predictors were examined using Cox proportional hazards model.


Ninety four patients (13%) were disengaged from the EASY program. Fewer negative symptoms at initial presentation, a diagnosis other than schizophrenia-spectrum disorder and poorer medication compliance in the first month of treatment were significant predictors of disengagement from service.


Early intervention teams should pay attention to factors associated with disengagement, and monitor at risk patients closely to detect signs of non-adherence.

Keywords: First-episode psychosis, Disengagement, Treatment adherence, Schizophrenia.

1. Introduction

Over the past 20 years, different early intervention programs for psychotic disorders have been implemented in places over the world. Numerous studies demonstrated that patients who received early intervention had better clinical and functional outcomes (Jackson and Birchwood, 1996, Craig et al, 2004, and Chen et al, 2011). However, the benefits of these programs may be undermined by the high rate of service disengagement found in early psychosis patients. Studies found that the disengagement rate in early intervention programs ranged from 14% to 33% in 2 years (Turner et al, 2009 and Zheng et al, 2013). Other studies found that poor adherence and treatment discontinuation are key factors associated with relapse and poor prognosis (Lieberman et al, 1996 and Wiersma et al, 1998).

Despite the relatively high rate of service disengagement found in first episode psychosis (FEP) patients and its importance to illness outcomes, there is a paucity of data with regard to disengagement and its predictors in early intervention programs, especially for those implemented in Chinese population. Previous studies demonstrated that socio-demographic factors, including not living with family and no family involvement in the program predicted disengagement from early intervention service (Schimmelmann et al, 2006, Conus et al, 2010, and Stowkowy et al, 2012). For clinical characteristics, substance abuse and duration of untreated psychosis (DUP) were found to be associated with disengagement (Turner et al, 2007, Conus et al, 2010, and Stowkowy et al, 2012). Some studies demonstrated that longer DUP was associated with disengagement ( Turner et al., 2007 ) while others found shorter DUP instead ( Stowkowy et al., 2012 ). Positive and negative symptom severities as well as level of functioning were also found to be associated with disengagement (Schimmelmann et al, 2006, Turner et al, 2007, Turner et al, 2009, Conus et al, 2010, and Stowkowy et al, 2012). However, there was a discrepancy in findings across different studies regarding the direction of association between these factors with disengagement. In Asian countries, a study in Singapore found that 14% of patients disengaged from early psychosis program at the second year ( Zheng et al., 2013 ). Disengaged patients were more likely to have Malay ethnicity and lower educational attainment. Another study evaluating the same program revealed non-schizophrenia diagnosis and better illness improvements to be possible predictors ( Verma et al., 2012 ).

To date, studies generated inconsistent findings concerning the predictors of disengagement. In general, studies seemed to demonstrate an association between substance use and lower symptom severity with disengagement. The relationships of other factors such as DUP, functioning, family involvement and insight with disengagement were relatively less consistent. It is worth noting that there is a substantial variation across different regions in sociocultural context, health care system and early intervention services regarding the content and intensity of service provided, which thus limits generalizability of study results to other populations. In this study, we sought to examine the prevalence and predictors of early disengagement using a large representative cohort of Chinese young people presenting with first-episode psychosis who were enrolled into the Early Assessment Service for Young People with Psychosis (EASY program). The aims of the present study were to 1) examine the rate of disengagement and to 2) investigate the predictors of disengagement from the EASY program. Based on previous literature, it was hypothesized that lower symptom severity, diagnosis of non-schizophrenia-spectrum psychotic disorder, and substance abuse would predict disengagement.

2. Methods

2.1. Participants and setting

This study was part of a larger study adopting a historical case–control design which aimed at evaluating the effectiveness of the EASY program ( Chen et al., 2011 ). In Hong Kong (HK), the EASY program was launched in 2001. The EASY program was a government-funded territory-wide early intervention program with 3 main components: public awareness and education about psychosis, easily accessible referral channel, and 2-year intensive phase-specific intervention for patients aged 15 to 25 years presenting with first-episode psychosis (FEP) ( Tang et al., 2010 ). The program adopted a case-management approach in which each patient was assigned with a case manager who provided protocol-based psychosocial interventions taking reference to international clinical guidelines for early psychosis ( International Early Psychosis Association Writing Group, 2005 ) with local cultural adaptations ( So, 2013 ). These interventions comprised three main modules including i) enhancement of psychological adjustment to early psychosis (via in-depth engagement, comprehensive psychoeducation, stress and coping management, etc.); ii) psychotherapy for secondary psychiatric morbidity; and iii) cognitive therapy for treatment-refractory symptoms ( So, 2013 ). In this program, standardized clinical assessments and multi-disciplinary case conferences were regularly conducted to monitor patient's clinical progress and treatment outcomes.

Seven hundred FEP patients who attended the EASY program from 2001 to 2003 were identified from the hospital database (Psychiatric Case Register). Patients were included in the study if they had diagnoses of psychotic disorder according to the International Classification of Diseases, Tenth Revision (ICD-10) ( World Health Organization, 1992 ). Patients were excluded if they had drug-induced psychosis, significant organic condition or moderate to severe mental retardation. The study was approved by relevant local institutional review board and ethics committees. Previous findings on gender differences, prediction of remission and recovery, and relationship between DUP and treatment outcomes in the same cohort from the EASY program have been reported elsewhere (Chang et al, 2011, Chang et al, 2012a, and Chang et al, 2012b).

2.2. Assessment

Data from clinical records of the patients were systematically retrieved according to standardized procedures. For each patient, data were acquired by trained research assistant from outpatient and inpatient clinical records as well as computerized clinical information based on operationalized definitions. Only data that could be reliably retrieved were used for analysis.

Data were collected on socio-demographics, clinical characteristics including DUP, diagnosis, as well as symptom severity and functioning. Diagnostic assignment of each patient was based on longitudinal approach considering the possibility of a diagnostic change over time (Chang et al, 2009a and Chang et al, 2009b). In the EASY program, diagnosis of each case was derived from consensus results of inter-disciplinary conference utilizing multiple sources of information and regular diagnostic reviews incorporating additional clinical information emerged during the course of illness. The final 2-year clinical diagnosis in this study was based on ICD-10 criteria ( World Health Organization, 1992 ). Following the method used by previous research (Conus et al, 2010 and Haro et al, 2011), the Clinical Global Impressions — Severity Scale (CGI-S; Guy, 1976 ) was applied to medical records to determine the positive and negative symptom severity. Functional performance was measured using Social and Occupational Functioning Assessment Scale (SOFAS; Goldman et al., 1992 ). Cross-sectional symptomatic remission at index month of disengagement was defined as having CGI-S scores of < 3 for both positive and negative symptoms. Service adherence and utilization variables included drug compliance, default status, and length of stay in service. Rating on medication compliance was derived from multiple sources of information including case management notes, medical records documented by doctors, as well as collateral information provided by caregivers. Disengagement was defined as continuous default of the EASY out-patient appointments up till the end of the 2-year service despite therapeutic need and active tracing from staff for psychiatric follow-up. All data were obtained via clinical records on a monthly basis, from baseline (month 1) till the end of the 24-month follow-up.

To ensure data quality, weekly consensus meetings were held during data collection period. Validity and inter-rater reliability were measured for variables such as DUP and functioning. Twelve cases were used for this practice, intra-class correlation coefficient (ICC) was calculated. To measure validity, ratings between clinician and research staff were compared (ICC ranged from 0.78 to 0.83). For inter-rater reliability, ratings between 2 research staff were compared (ICC ranged from 0.70 to 0.98). All scores reflected satisfactory validity and reliability in ratings.

2.3. Data analysis

Kaplan–Meier survival curve was generated to estimate the likelihood of disengagement as a function of time since the start of the EASY program. Cox-proportional hazards regression model was used to investigate the predictive capacity of baseline socio-demographics, clinical characteristics, symptoms, functioning, and other variables on disengagement. Variables whose hazard ratios had p values less than 0.1 were included in the multivariate analysis. Significance level was set at 0.05. All statistical analyses were carried out using the Statistical Package for the Social Sciences version 20.

3. Results

3.1. Characteristics of the sample

Seven hundred patients participated in the study. In our sample, the mean age was 20.65 years (SD = 3.43). There were 360 male participants (51.43%). The average years of education were 10.9 years (SD = 2.33). In terms of clinical characteristics, the mean age of onset was 20.46 years (SD = 3.38). The median DUP was 91 days (mean = 239.80, SD = 373.37); most patients had a gradual onset of psychotic disorder (69.14%). A majority of the patients had schizophrenia spectrum disorder (schizophrenia or schizoaffective disorders; 69.14%).

3.2. Estimated rate of disengagement

Results of the Kaplan–Meier plot showed that the estimated rate of disengagement from the EASY program was 0.05 at 6 months (36 cases between months 0 and 6), 0.09 at the end of first year (24 cases between month 6 and month 12) and 0.13 at the end of second year (34 cases between month 12 and month 24). A total of 94 patients disengaged within the two years. Fig. 1 displays the likelihood of staying in the EASY program over time. Over the 2 years, the mean estimated time of patients staying in service was 671.8 days (95% CI 659.512–684.022).


Fig. 1 The likelihood of remaining in the EASY program over time.

3.3. Predictors of service disengagement

Univariate analysis of potential predictors is presented in Table 1 . Variables with hazard ratios that had p values of 0.1 or below were identified. It was found that patients who were more likely to disengage had psychotic disorders other than schizophrenia-spectrum disorders (acute and transient psychosis, unspecified non-organic psychosis, affective disorders with psychotic features), lower baseline positive and negative symptom scores, higher baseline SOFAS score, fair or poor medication compliance at the first month of treatment, substance abuse history within the initial 6 months of the program, and were not taking second generation antipsychotics at baseline. Further analysis was conducted to examine the role of symptomatic remission on service disengagement. It was found that 63% of disengaged patients achieved symptomatic remission at the month of disengagement.

Table 1 Sociodemographic and clinical predictors of service disengagement.

  Univariate analyses
Mean (SD) or no. (%) Hazard ratio 95% CI χ 2 p
Age 20.65 (3.43) 1.01 0.95–1.07 0.16 0.69
 Male 360 (51.43) 1.14 0.76–1.71 0.41 0.52
 Female 340 (48.57) (Reference category)      
Years of education 10.90 (2.33) 0.99 0.90–1.08 0.11 0.74
Clinical characteristics
Age at onset 20.46 (3.38) 1.01 0.95–1.07 0.1 0.76
Onset mode          
 < 1 month 216 (30.86) 1.08 0.70–1.66 0.11 0.74
 > 1 month 484 (69.14) (Reference category)      
DUP median split          
 < 91 days 354 (50.57) 0.91 0.61–1.36 0.22 0.64
 > 91 days 346 (49.43) (Reference category)      
 Schizophrenia spectrum 484 (69.14) 0.41 0.28–0.62 17.93 < 0.001
 Other psychosis 216 (30.86) (Reference category)      
Baseline characteristics
CGI-S positive symptom scores 4.18 (0.90) 0.75 0.60–0.92 6.99 0.01
CGI-S negative symptom scores 2.62 (1.29) 0.71 0.60–0.83 17.61 < 0.001
SOFAS score 46.2 (12.58) 1.03 1.01–1.04 9.64 0.002
Baseline medication          
 Taking SGAs 517 (73.96) 0.62 0.40–0.94 4.69 0.03
 Not taking SGAs 182 (26.04) (Reference category)      
Baseline medication compliance          
 Good 580 (85.04) 0.39 0.25–0.61 14.29 < 0.001
 Fair/poor 102 (14.96) (Reference category)      
Substance abuse history within first 6 months          
 Yes 13 (1.86) 4.25 1.86–9.71 7.94 0.001
 No 687 (98.14) (Reference category)      

CGI-S = the Clinical Global Impressions — Severity Scale; DUP = duration of untreated psychosis; other psychosis = bipolar affective disorder, severe depression with psychotic symptoms, acute and transient psychosis, and unspecified non-organic psychosis; schizophrenia spectrum disorders = schizophrenia and schizoaffective disorders based on the International Classification of Diseases, Tenth Revision; SGAs = second generation antipsychotics; SOFAS = Social and Occupational Functioning Assessment Scale.

Variables with p values of 0.1 or below were included in the multivariate analyses. Four significant predictors were identified: fewer negative symptoms at baseline, a diagnosis other than schizophrenia spectrum disorder, poorer medication compliance at baseline and substance abuse within first 6 months of program all contributed independently to predicting disengagement from the EASY program. In an attempt to avoid bias introduced by low rate of substance abuse on the final prediction model of disengagement, multivariate regression analysis was repeated with substance abuse being excluded from the model. Consistent results were generated with fewer baseline negative symptoms, a diagnosis other than schizophrenia spectrum disorder and poor medication compliance in the first month of treatment being identified as independent predictors of disengagement. Table 2 shows the hazard ratios for the final model.

Table 2 Final prediction model for disengagement from the EASY program.

Variables df Parameter estimate Standard error Hazard ratio 95% CI p
CGI-S positive symptom scores 1 − 0.14 0.13 0.87 0.67–1.13 0.29
CGI-S negative symptom scores 1 − 0.28 0.09 0.76 0.64–0.90 0.002
SOFAS score 1 0.004 0.01 1.004 0.99–1.02 0.69
Schizophrenia spectrum disorders 1 − 0.83 0.21 0.44 0.29–0.66 < 0.001
Good medication compliance 1 − 1.00 0.23 0.37 0.23–0.58 < 0.001
Taking SGAs 1 − 0.13 0.26 0.88 0.53–1.45 0.61

CGI-S = the Clinical Global Impressions — Severity Scale; SGAs = second generation antipsychotics; schizophrenia spectrum disorders = schizophrenia and schizoaffective disorders based on the International Classification of Diseases, Tenth Revision; SOFAS = Social and Occupational Functioning Assessment Scale.

4. Discussion

To date, few studies have explored the rate and predictors of disengagement from early intervention program for psychosis, especially in Chinese population. In this study, it was found that 13% of first-episode psychosis patients disengaged from the 2-year EASY program. A similar study conducted in Singapore revealed a 14% disengagement rate ( Zheng et al., 2013 ). Overall, disengagement rate of our first-episode cohort was in the lower end (i.e., 14% to 33% within 2-year period) of that reported in the literature for early psychosis programs. The disengagement rate found in our study was slightly lower than that in Western studies. It might be possible that the close collaborations between the EASY case managers and family caregivers who are critically involved in patient management in HK via treatment adherence monitoring and detection of early relapse signs etc. acted as a protective factor in minimizing risk of service disengagement (Schimmelmann et al, 2006, Conus et al, 2010, and Stowkowy et al, 2012). In fact, most patients enrolled in the EASY program live with their families.

Our results revealed three key predictors for disengagement in our first-episode psychosis sample. Patients who had fewer baseline negative symptoms, a diagnosis other than schizophrenia-spectrum disorder and poor medication compliance in the first month of treatment were more likely to disengage. These findings were partially in keeping with those of previous studies. Across different studies, lower symptom severity was frequently revealed as predictor for disengagement from early psychosis program (Schimmelmann et al, 2006, Turner et al, 2007, Conus et al, 2010, and Stowkowy et al, 2012). In this study, we found that a decrease of 1 unit of CGI-S negative symptom score led to a 1.32 times increase in the hazard rate. When compared to first-episode psychosis patients with lower symptom severity, those with severe symptoms may be more motivated to receive treatment, their family and intervention teams may also put greater effort in ensuring treatment compliance.

It was also found that patients who had psychotic disorder other than schizophrenia-spectrum disorders were more likely to disengage. These diagnoses included affective disorders with psychotic features, acute and transient psychosis, and unspecified non-organic psychosis, which were associated with a two-fold risk of disengagement. This finding was similar to that of a Singapore study, in which disengaged patients were more likely to have non-schizophrenia psychotic disorders ( Verma et al., 2012 ). These patients may, in general, have an apparently more benign course at an initial stage of illness or better outcome in the early course of the illness when compared with patients with schizophrenia. They may feel that they are well enough to terminate treatment at an early stage.

Poor medication compliance in the initial treatment stage was also a significant predictor in this study. This was not surprising as it served as an early sign reflecting a patient's treatment adherence in general. Consistent with most previous studies, substance abuse history within the first 6 months of the EASY program was found to predict service disengagement. Owing to a relatively low rate of substance abuse in our cohort when compared to that of first-episode populations in western countries, caution should therefore be exercised in interpreting our findings on the impact of substance abuse on disengagement from early intervention service. It is noteworthy that the current study did not reveal significant predictive effect of DUP on service disengagement, which was contrary to findings of several studies (Turner et al, 2007 and Stowkowy et al, 2012) but in line with others (Turner et al, 2009 and Conus et al, 2010). Inconsistency in terms of the direction of association between DUP and disengagement were found even for studies revealing significant predicting effect of DUP. Some studies demonstrated an association between shorter DUP and disengagement ( Stowkowy et al., 2012 ) while others found longer DUP instead ( Turner et al., 2007 ). Variations in terms of sample characteristics and DUP measurement may provide explanation for the discrepancies across study findings. Future studies may further delineate the relationship between DUP and service disengagement using standardized instrument for DUP assessment.

The current findings have important clinical implications; patients with low severity of negative symptoms and non-schizophrenia psychotic disorders may have better treatment response in the initial stage of illness. However the current study revealed that these patients with an apparently less severe illness may be at significantly higher risk of disengagement from early intervention service. Early disengagement from service may be detrimental to their longer-term illness outcome. First, owing to fluidity of symptom manifestations in the early course of psychotic illness, literature has repeatedly shown that there is a high rate of diagnostic change from those less frequent but apparently “good-prognostic” (with initial good treatment response) diagnostic entities such as acute and transient psychotic disorders toward schizophrenia-spectrum disorder in the first few years after treatment initiation for first-episode psychosis (Chang et al, 2009a and Chang et al, 2009b) Second, relapse rate of psychosis is high in the early illness stage, which is associated with worse long-term outcomes (Lieberman, 1996 and Wiersma et al, 1998). One recent study even revealed that 79% of first-episode psychosis patients, who were symptomatically remitted with antipsychotic treatment, experienced relapse in one year after planned medication discontinuation ( Chen et al., 2010 ). Third, it is hypothesized that the first few years of psychotic illness represents a critical window of opportunity for intensive intervention to determine long-term prognosis ( Birchwood and Fiorillo, 2000 ). The early intervention team should therefore pay special effort in engaging these patients and avoid any early disengagement.

There were several limitations in this study. One limitation was that all information on socio-demographics, clinical characteristics, symptoms and functioning were retrieved from clinical records, which may have been biased by varying degrees of documentation quality. In particular, owing to the limitations posed by the retrospective design of the current study, positive and negative symptom severity was determined by CGI-S on the basis of case notes review, which is a less specific symptom measure than those more refined psychopathological rating instruments such as Positive and Negative Syndrome Scale (PANSS). Nevertheless it should be noted that all of the data was retrieved systematically by trained research assistants from various sources. There was a weekly consensus meeting held to ensure data quality, and the ICC scores calculated on clinician and research staff's ratings demonstrated satisfactory validity and reliability. The use of CGI-S in rating symptom severity was adopted from similar studies relying on case note review and was shown to be a reliable method of symptom assessment (Conus et al, 2010 and Haro et al, 2011). Another limitation was that some variables that may predict disengagement from early intervention service could not be addressed in our study due to the retrospective study design. Previous studies found that insight and lack of family involvement in the program predicted disengagement (Schimmelmann et al, 2006, Turner et al, 2007, Conus et al, 2010, and Stowkowy et al, 2012). Further prospective studies should be conducted to examine the impact of these variables on disengagement from early intervention.

In conclusion, the current study showed that, in a large and clinically homogeneous cohort of Chinese young people presenting with FEP to specialized early intervention program in HK, 13% were disengaged from the 2-year service. We have confirmed reports in previous studies, which were mostly conducted in western populations, that diagnosis of a non-schizophrenia-spectrum psychotic disorder, lower baseline negative symptom severity, poor medication compliance and substance abuse history predicted service disengagement. More research is required to clarify the inconsistency concerning relationship between DUP and disengagement. As well, future studies focusing on other important variables such as family involvement and insight may help identify specific interventions for minimizing risk of service disengagement in the early course of the illness.

Role of the funding source

The study was supported by Health and Health Service Research Fund of the Food and Health Bureau, the Government of Hong Kong Special Administrative Region. The grant number was 03040141. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.


Study rationale (EYHC); search of the literature, data entry and statistical analyses (TCWC, WCC, CLMH); writing of the article (TCWC); proof reading (WCC, SKWC, EHML).

Conflict of interest

EYHC has participated in paid advisory board for Otsuka, has received educational grant support from Janssen-Cilag, and has received research funding from AstraZeneca, Janssen-Cilag, Pfizer, Eli Lilly, Sanofi-Aventis, and Otsuka.


The authors would like to thank the hospital staff, the psychiatric and medical records units for their assistance.


  • Birchwood and Fiorillo, 2000 M. Birchwood, A. Fiorillo. The critical period for early intervention. Psychiatr. Rehabil. Skills. 2000;4(2):182-198 Crossref
  • Chang et al., 2009a W.C. Chang, S.S.M. Chan, D.W.S. Chung. Diagnostic stability of functional psychosis: a systematic review. Hong Kong J. Psychiatry. 2009;19:30-41
  • Chang et al., 2009b W.C. Chang, S.L.K. Pang, D.W.S. Chung, S.S.M. Chan. Five-year stability of ICD-10 diagnoses among Chinese patients presented with first-episode psychosis in Hong Kong. Schizophr. Res.. 2009;115(2–3):351-357 Crossref
  • Chang et al., 2011 W.C. Chang, J.Y.M. Tang, C.L.M. Hui, C.P.Y. Chiu, M.M.L. Lam, G.H.Y. Wong, D.W.S. Chung, C.W. Law, S. Tso, K.P.M. Chan, S.F. Hung, E.Y.H. Chen. Gender differences in patients presenting with first-episode psychosis in Hong Kong: a three-year follow up study. Aust. N. Z. J. Psychiatry. 2011;45(3):199-205 Crossref
  • Chang et al., 2012a W.C. Chang, J.Y. Tang, C.L. Hui, M.M. Lam, S.K. Chan, G.H. Wong, C.P. Chiu, E.Y. Chen. Prediction of remission and recovery in young people presenting with first-episode psychosis in Hong Kong: a 3-year follow-up study. Aust. N. Z. J. Psychiatry. 2012;46(2):100-108 Crossref
  • Chang et al., 2012b W.C. Chang, J.Y.M. Tang, C.L.M. Hui, M.M.L. Lam, G.H.Y. Wong, S.K.W. Chan, C.P.Y. Chiu, D.W.S. Chung, C.W. Law, S. Tso, K. Chan, S.F. Hung, E.Y.H. Chen. Duration of untreated psychosis: relationship with baseline characteristics and three-year outcome in first-episode psychosis. Psychiatry Res.. 2012;198(3):360-365 Crossref
  • Chen et al., 2010 E.Y.H. Chen, C.L.M. Hui, M.M.L. Lam, C.P.Y. Chiu, C.W. Law, D.W.S. Chung, S. Tso, E.P.F. Pang, K.T. Chan, Y.C. Wong, F.Y.M. Mo, K.P.M. Chan, T.J. Yao, S.F. Hung, W.G. Honer. Maintenance treatment with quetiapine versus discontinuation after one year of treatment in patients with remitted first episode psychosis: randomised controlled trial. Br. Med. J.. 2010;341:c4024 Crossref
  • Chen et al., 2011 E.Y. Chen, J.Y. Tang, C.L. Hui, C.P. Chiu, M.M. Lam, C.W. Law, C.W. Yew, G.H. Wong, D.W. Chung, S. Tso, K.P. Chan, K.C. Yip, S.F. Hung, W.G. Honer. Three-year outcome of phase-specific early intervention for first-episode psychosis: a cohort study in Hong Kong. Early Interv. Psychiatry. 2011;5(4):315-323 Crossref
  • Conus et al., 2010 P. Conus, M. Lambert, S. Cotton, C. Bonsack, P.D. McGorry, B.G. Schimmelmann. Rate and predictors of service disengagement in an epidemiological first-episode psychosis cohort. Schizophr. Res.. 2010;118(1–3):256-263 Crossref
  • Craig et al., 2004 T.K. Craig, P. Garety, P. Power, N. Rahaman, S. Colbert, M. Fornells-Ambrojo, G. Dunn. The Lambeth Early Onset (LEO) Team: randomised controlled trial of the effectiveness of specialised care for early psychosis. Br. Med. J.. 2004;329(7474):1067-1070
  • Goldman et al., 1992 H.H. Goldman, A.E. Skodol, T.R. Lave. Revising Axis-V for DSM-IV: a review of measures of social functioning. Am. J. Psychiatry. 1992;149:1148-1156
  • Guy, 1976 W. Guy. Clinical global impressions. ECDEU Assessment Manual for Psychopharmacology (revised) (National Institute of Mental Health, Rockville, MD, 1976) 217-221
  • Haro et al., 2011 J.M. Haro, D. Novick, J. Bertsch, J. Karagianis, M. Dossenbach, P.B. Jones. Cross-national clinical and functional remission rates: Worldwide Schizophrenia Outpatient Health Outcomes (W-SOHO). BMJ. 2011;203(6):194-201 Crossref
  • International Early Psychosis Association Writing Group, 2005 International Early Psychosis Association Writing Group. International clinical practice guidelines for early psychosis. Br. J. Psychiatry. 2005;187(Suppl. 48):120-124
  • Jackson and Birchwood, 1996 C. Jackson, M. Birchwood. Early intervention in psychosis: opportunities for secondary prevention. Br. J. Clin. Psychol.. 1996;35(Pt 4):487-502 Crossref
  • Lieberman, 1996 J.A. Lieberman. Evidence for sensitization in the early stage of schizophrenia. Eur. Neuropsychopharm.. 1996;6:155
  • Lieberman et al., 1996 J.A. Lieberman, A.R. Koreen, M. Chakos, B. Sheitman, M. Woerner, J.M.J. Alvir, R. Bilder. Factors influencing treatment response and outcome of first-episode schizophrenia: implications for understanding the pathophysiology of schizophrenia. J. Clin. Psychiatry. 1996;57:5-9
  • Schimmelmann et al., 2006 B.G. Schimmelmann, P. Conus, M. Schacht, P. McGorry, M. Lambert. Predictors of service disengagement in first-admitted adolescents with psychosis. J. Am. Acad. Child Adolesc. Psychiatry. 2006;45(8):990-999 Crossref
  • So, 2013 S.H.W. So. Implementing Psychological Intervention Programme in Early Psychosis (PIPE). E.Y.H. Chen, H. Lee, G.H.K. Chan, G.H.Y. Wong (Eds.) Early Psychosis Intervention: A Culturally Adaptive Clinical Guide (Hong Kong University Press, Hong Kong, 2013) 137-157
  • Stowkowy et al., 2012 J. Stowkowy, D. Addington, L. Liu, B. Hollowell, J. Addington. Predictors of disengagement from treatment in an early psychosis program. Schizophr. Res.. 2012;136(1–3):7-12 Crossref
  • Tang et al., 2010 J.Y. Tang, G.H. Wong, C.L. Hui, M.M. Lam, C.P. Chiu, S.K. Chan, D.W. Chung, S. Tso, K.P. Chan, K.C. Yip, S.F. Hung, E.Y. Chen. Early intervention for psychosis in Hong Kong — the EASY programme. Early Interv. Psychiatry. 2010;4(3):214-219 Crossref
  • Turner et al., 2007 M. Turner, C. Smith-Hamel, R. Mulder. Prediction of twelve-month service disengagement from an early intervention in psychosis service. Early Interv. Psychiatry. 2007;1(3):276-281 Crossref
  • Turner et al., 2009 M.A. Turner, J.M. Boden, C. Smith-Hamel, R.T. Mulder. Outcomes for 236 patients from a 2-year early intervention in psychosis service. Acta Psychiatr. Scand.. 2009;120(2):129-137 Crossref
  • Verma et al., 2012 S. Verma, L.Y. Poon, M. Subramaniam, E. Abdin, S.A. Chong. The Singapore Early Psychosis Intervention Programme (EPIP): a programme evaluation. Asian J. Psychiatry. 2012;5(1):63-67 Crossref
  • Wiersma et al., 1998 D. Wiersma, F.J. Nienhuis, C.J. Slooff, R. Giel. Natural course of schizophrenic disorders: a 15-year followup of a Dutch incidence cohort. Schizophr. Bull.. 1998;24(1):75-85 Crossref
  • World Health Organization, 1992 World Health Organization. The ICD-10 Classification of Mental and Behavioral Disorders: Clinical Description and Diagnostic Guidelines. (WHO, Geneva, 1992)
  • Zheng et al., 2013 S. Zheng, L.Y. Poon, S. Verma. Rate and predictors of service disengagement among patients with first-episode psychosis. Psychiatr. Serv.. 2013;64(8):812-815 Crossref


a Department of Psychiatry, The University of Hong Kong, Queen Mary Hospital, Pokfulam Road, Hong Kong

b The State Key Laboratory of Brain and Cognitive Sciences, The University of Hong Kong, Hong Kong

lowast Corresponding author. Tel.: + 852 22554486; fax: + 852 28551345.