Psychiatry Research 215 (2014) 46–51

Contents lists available at ScienceDirect

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

EQ-5D as a measure of programme outcome: Results from the Singapore early psychosis intervention programme Mythily Subramaniam a,n, Edimansyah Abdin a, Lye Yin Poon b, Janhavi Ajit Vaingankar a, Helen Lee b, Siow Ann Chong a, Swapna Verma b a b

Research Division, Institute of Mental Health, Singapore Department of Early Psychosis Intervention, Institute of Mental Health, Singapore

art ic l e i nf o

a b s t r a c t

Article history: Received 22 January 2013 Received in revised form 7 September 2013 Accepted 9 October 2013 Available online 21 October 2013

The current study aimed to establish the Health-Related Quality of Life (HRQoL) among participants with First Episode Psychosis (FEP) in Singapore, to elucidate the sociodemographic and clinical correlates of HRQoL, and ascertain the change after 1-year of treatment. Two hundred and forty one patients accepted into an Early Psychosis Intervention Programme (EPIP) from April 2009 to June 2011 and who had completed baseline EuroQol-5D (EQ-5D) assessments were included in this analysis. The mean (S.D.) EQ-5D index at baseline was 0.788 (0.258). One hundred thirty five (56.0%) patients who completed the EQ-5D assessment at the 12-month follow-up had a significantly higher EQ-5D index as compared to baseline. EQ-5D index was significantly higher among those patients who met criteria for remission at the 12-month interval than those who were not in remission. Our results suggest that the EQ-5D is responsive to improvement as it corresponded well to objective ratings of remission in our patients with FEP. & 2013 Elsevier Ireland Ltd. All rights reserved.

Keywords: Early psychosis EQ-5D Duration of untreated Psychosis Remission Singapore

1. Introduction Psychotic disorders are severe mental health disorders that have impaired reality testing as their core feature. This debilitating illness often emerges in late adolescence or early adulthood, and remains undetected and untreated for a long period of time. Psychosis results in enormous burden to the individual and society, both, in terms of human suffering and economic cost. The Australian Low Prevalence Disorders Study Group reported that – based on the 1-month prevalence of 4.7 per thousand in the 18–64-year old age group – psychosis costs the Australian Government at least $1.45 billion per annum, while societal costs were at least $2.25 billion per annum. This corresponded to 0.23% and 0.36% of GDP from the goverment and societal perspective respectively in 1999–2000. Adopting a conservative estimate of 12-month prevalence, societal costs which were estimated to be as high as $2.54 billion, increased to $2.62 billion when ‘time-loss’ costs due to mortality were included (Carr et al., 2002). Quality of Life (QoL) has come to be regarded as an important dimension of outcome in schizophrenia and other serious mental disorders. The World Health Organization (WHO) has defined QoL as individuals' perceptions of their position in life in the context of the culture and value systems in which they live and in relation to n Corresponding author. Research Division, Institute of Mental Health, Buangkok Green Medical Park, 10 Buangkok View, Singapore 539747, Singapore. Tel.: þ 6563 893 633; fax: þ 6563 437 962. E-mail address: [email protected] (M. Subramaniam).

0165-1781/$ - see front matter & 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.psychres.2013.10.002

their goals, expectations, standards and concerns (WHO QoL Group, 1994). Health-Related Quality of Life (HRQoL) narrows the concept to aspects of QoL that are affected positively or negatively by health and medical health care interventions. In a world of limited resources and continuous efforts to contain costs, HRQoL is often used as an outcome measure for comparison of therapies and programmes as well as for resource allocation (Awad and Voruganti, 2000). There are some inherent difficulties in measuring HRQoL generally and specifically in those with mental illnesses. Studies have demonstrated discrepancies between QoL that is self-rated (subjective QoL) and that assessed by trained raters (objective QoL) in schizophrenia (Fitzgerald et al., 2001; Adewuya and Makanjuola, 2010). People with schizophrenia often indicate better subjective QoL than would be expected from their living circumstances or objective QoL. It has been argued that patients' psychopathology or lack of insight may limit their ability to give valid self-reports (Doyle et al., 1999; Savilla et al., 2008). However, measures of subjective and objective QoL assess different constructs, and good test–retest reliability for psychotic patients' self-reported QoL has been demonstrated (Voruganti et al., 1998). In clinically stable patients with psychotic disorders, the selfreport method has been shown to be a valid and reliable way of evaluating their perceived wellness (Herrman et al., 2002; Angermeyer et al., 2001). HRQoL can be measured by instruments that are generic or disease-specific. Disease-specific questionnaires that assess HRQoL in schizophrenia include the Quality of Life Questionnaire in Schizophrenia (S-QoL) (Auquier et al., 2003)

M. Subramaniam et al. / Psychiatry Research 215 (2014) 46–51

and the Lancashire Quality of Life Profile (Lqo3LP) (Oliver et al., 1997). Generic instruments, on the other hand, summarise a spectrum of core concepts of HRQoL that may apply to different diseases and populations. Thus, generic instruments can be used to compare HRQoL of patient groups across different diseases, and can provide information to support health policy decisions and cost-effectiveness analyses (Patrick and Erickson., 1993). Studies have reported lower HRQoL among those with First Episode Psychosis (FEP) (Addington et al., 2003; Bechdolf et al., 2005). However, few have compared HRQoL of FEP with population norms or evaluated generic, self-reported HRQoL measures from a programme evaluation perspective. The aims of the current study were to (i) establish the HRQoL in terms of EuroQol-5D (EQ5D) indices among patients with FEP in Singapore and compare it to population norms and (ii) establish sociodemographic and clinical correlates of HRQoL in this sample at baseline and ascertain the change if any after 1-year of treatment provided by the Early Psychosis Intervention Programme (EPIP) in Singapore.

2. Methods Singapore is an island state in South East Asia; the population in 2011 was just about 5.2 million of which 3.8 million were Singapore residents (citizens and permanent residents). Of its residents, 74.1% are of Chinese descent, 13.4% are Malays, and 9.2% are of Indian descent (Singapore Department of Statistics, 2012). The EPIP is a nationwide programme, launched in 2001 at the Institute of Mental Health and Woodbridge Hospital, to address the needs of those experiencing FEP and to prevent adverse consequences (McGorry, 2002; McGorry and Killackey, 2002). One of the stated goals of the Singapore EPIP was to improve clinical outcomes and QoL of those with psychosis (Chong et al., 2004). Patients in this programme fulfil the following inclusion criteria: (i) age between 15 and 40 years, (ii) first episode psychotic disorder with no prior or minimal treatment, defined as, o 12 weeks of antipsychotic medications, (iii) no current history of substance abuse, and (iv) no history of major medical or neurological illness.

2.1. Sample This study assessed patients with FEP presenting to EPIP from April 2009 to June 2011. A total of 573 patients were screened for eligibility to be accepted in this programme between April 2009 and June 2011. Two hundred and forty one patients (42.1%) who completed baseline EQ-5D assessments were included in this analysis, 180 (31.4%) patients subsequently refused to do the EQ-5D rating and 78

47

(13.6%) patients did not complete all the EQ-5D fields. Three (0.5%) patients were excluded because they were assessed subsequently to be non-psychotic cases and 6 (1%) died. Eleven (1.9%) defaulted on their treatment and were lost to follow-up; 54 patients (9.4%) had been discharged from the service before the end of the followup period – they had either moved out of the country or wanted to continue their care with other psychiatric services. Finally, 135 patients completed EQ-5D ratings at both baseline and 12-months of follow-up (Fig. 1).

2.2. Assessments Trained case managers collected sociodemographic data on age, gender, ethnicity, educational level, marital status, occupation and living situation from all patients using a semi-structured questionnaire. Diagnosis of participants was established using the Structured Clinical Interviews for DSM-IV (SCID-clinical version) (First et al., 1996) at the first contact (baseline). Duration of Untreated Psychosis (DUP) was operationalised as the time (in months) between onset of psychotic symptoms (delusions, hallucinations, disorganised behaviour) and the time when a definitive diagnosis and treatment were established. Patients and their primary caregivers were interviewed by the clinical team and asked to date the onset of psychotic symptoms and the DUP was estimated after combining information from the interviews and case records. Severity of psychopathology was assessed by Positive and Negative Scale for Schizophrenia (PANSS) (Kay et al., 1987). The PANSS assesses the levels of positive, negative and general psychopathology symptoms that are associated with psychosis. It consists of 30 items scored on a 1 (absent) to 7 (extreme) scale and a higher score reflects a greater psychopathology. The Global Assessment of Functioning (GAF) was used to assess level of functioning (Goldman et al., 1992). The GAF assesses symptom severity and levels of psychological, social and occupational functioning, on a 1–100 rating scale that is divided into 10 deciles, each of which provides a description of functioning and symptom severity. A lower score on the GAF denotes a worse response. These ratings at baseline, 6, and 12 months were conducted by experienced clinical psychiatrists who were trained in the use of these instruments. All raters participated in periodic inter-rater reliability sessions to avoid rater drift. Sociodemographic data was collected using a semi-structured questionnaire by trained case managers. We used the criteria for symptomatic remission as proposed by the Schizophrenia Working Group (Andreasen et al., 2005), that is, achieving and maintaining a PANSS rating of three or less for a duration of at least 6 months on the following items: delusions (P1), unusual thought contents (G9), hallucinatory behaviour (P3), conceptual disorganisation (P2), mannerisms (G5), blunted affect (N1), social withdrawal (N4) and lack of spontaneity (N6) (7). Functional remission was defined a priori as having a GAF disability score of Z61 at 1 year with engagement in ageappropriate vocation when assessed 12 months after the baseline assessment (Verma et al., 2012). Patients who fulfilled the criteria for both symptomatic and functional remission were classified as being in remission for the purpose of this study.

573 (Eligible patients)

180(31.4%) Refusal

. 78(136%) Did not complete EQ-5D 3 (0.5%) Non-psychotic cases 6 (1.0%) Died 11 (1.9%)

Defaulted cases 54 (9.4%) Discharged from the service

241 (42.1%) Completed EQ-5D at Baseline Fig. 1. Flow chart of patients.

135(56.0%) Completed EQ-5D at Baseline and 12-months

48

M. Subramaniam et al. / Psychiatry Research 215 (2014) 46–51

2.2.1. EQ-5D EQ-5D is a generic instrument for subjectively describing and valuing HRQoL that has been developed by the EuroQol Group (1990). It generates a health profile as well as preference-based index scores for HRQoL that may be used as weights for calculating Quality-Adjusted Life Years (QALYs) in the economic evaluation of health care (Drummond et al., 1997). A number of studies have evaluated the validity of EQ-5D in patients with psychosis and found it to be a valid and useful scale for evaluation of HRQoL (Prieto et al., 2004; Stochl et al., 2013; Bobes et al., 2005). EQ-5D comprises a descriptive system and a Visual Analogue Scale (VAS) (EuroQol Group, 1990). The descriptive system assesses five domains (i.e., mobility, self-care, usual activities, pain/discomfort, anxiety/depression) and participants are asked to rate their health on that day of assessment on a three-point scale (no problem/moderate problem/extreme problem). Responses to these five dimensions are converted into one of 243 different EQ-5D health state descriptions, which range between no problems on all five dimensions (11111) and severe/extreme problems on all five dimensions (33333). The utility of EQ-5D health states was elicited using the time trade-off method from a representative sample of the UK general population to value a number of potential EQ-5D states (the time trade-off seeks to establish by how much would one be willing to reduce one's life expectancy in order to obtain full health) (Dolan, 1997). This utility-based EQ-5D index score ranges from  0.59 to 1.00, with negative values representing health states worse than being dead, 0 representing being dead, and 1.00 representing the state of full health. The EQ VAS records the participant's self-rated health on a vertical, VAS where the endpoints are labelled ‘Best imaginable health state’ (100) and ‘Worst imaginable health state’ (0). EQ-5D was self-administered by patients at baseline and 12 months into the programme. In this analysis, we used the EQ-5D index scores for HRQoL and QALY estimations as utility values using time trade-off are available for various populations (Dolan, 1997). EQ-5D was first proposed as an outcome indicator when the EPIP was included in the National Mental Health Blueprint of Singapore (Chong, 2007) and incorporated as a measure in 2009 for this programme.

2.3. Statistical analyses All statistical analyses were performed using STATA and SAS version 9.2 for Windows. Descriptive statistics were computed for the basic demographic and clinical variables. Mean and standard deviations (S.D.s) were calculated for continuous variables, and frequencies and percentages for categorical variables. The EQ-5D score was compared with Singaporean population norms using one-sample t-test. Normality of quantitative data was checked using the Kolmogorv–Smirnov 1-sample and Shapiro–Wilk's tests. Spearman correlation and multiple linear regression analyses (Ordinary Least Square (OLS) Approach) were used to determine the relationship between EQ-5D index score and sociodemographic variables (i.e., age, sex, race, marital status) and clinical scales at baseline. The sociodemographic variables were selected based on research in Asian population that suggests that age, gender, ethnicity and marital status are important determinants of HRQoL (Abdin et al., 2013). In order to establish the best model that can explain the relationship between EQ-5D utility scores and sociodemographic and clinical correlates with the presence of a ceiling effect, we decided to include another two models in our analyses namely Tobit regression (Tobin, 1958) and Censored Lease Absolute Deviations (CLAD) (Powell, 1984). The selection of these models was based on their frequency of use and applicability to estimate health utility scores (Payakachat et al., 2009). It has been reported that Tobit and CLAD models are preferable to OLS in the presence of a censored dependent variables or ceiling effect (Huang et al., 2008). We have compared the performance of these three models using Mean Absolute Prediction Error (MAPE), Mean Error (ME) and Root Mean Square Errors (RMSE) indices. These indices provided better picture of the model predictive accuracy than using R2 goodness of fit measures (Austin, 2002; Sullivan and Ghushchyan, 2006; Forster, 2002). Because there is no cut-off point for selecting the best model, a small value for MAPE, ME, and RMSE was considered a desirable predictive model. MAPE was the major criteria to select the best predictive models in this study (Payakachat et al., 2009). EQ-5D scores measured over time were analysed using repeated measures Analysis of Variance (ANOVA). Differences between clinical groups were tested by independent t-test and Mann–Whitney U test for normal and non-normal continuous variables respectively whenever appropriate. Level of significance was set at p value o 0.05. We also estimated the QALYs participants had experienced over 1 year in EPIP using “areas under (health utility) curves” method (Manca et al., 2005).

3. Results A total of 241 patients who completed EQ-5D at baseline were included in the analysis. The mean (S.D.) age of the participants was 26.1 (6.7) years and ranged from 16 to 40 years. The sample was predominantly Chinese (73.9%), single (84.5%) and majority were diagnosed with schizophrenia spectrum disorders including

schizophrenia, schizophreniform and schizoaffective disorder (68.1%). The mean (S.D.) of the DUP was 10.6 (16.5) months and median was 4 months. The mean (S.D.) PANSS total score and the GAF total score at baseline were 75.5 (19.1) and 40.0 (12.5) respectively. The mean (S.D.) EQ-5D index of the 241 patients who completed EQ-5D at baseline was 0.788 (0.258). Eighty eight patients (36.5%) reported “full health” with the EQ-5D and had a health index score of 1.0. We found the EQ-5D utility scores were not distributed normally and skewed significantly to the left due to ceiling effect. The mean EQ-5D index of patients with FEP was significantly lower as compared to the general population mean of Singapore (Abdin et al., 2013) (0.788 vs. 0.951, Mean Difference ¼  0.163, p value o0.001). In order to take into account age and gender distribution differences between the FEP sample and that of the general population, age and gender distribution of the general population were extracted from the previous study (Abdin et al., 2013) and stratified according to the FEP sample's distribution. While the proportion of females and males was quite similar, the mean age of respondents was slightly higher in the population sample than in the FEP sample. This was due to the slightly narrow age range of the previous study. After standardizing the age and gender distribution, we found that the mean EQ5D in the FEP sample was still significantly lower as compared to that of the general population (0.788 vs. 0.963, p value o0.001). There were significant differences in baseline scores by sociodemographic characteristics and clinical diagnosis among the patients. Females scored higher than males on the EQ-5D index and those who were married had a higher score than those who were single and separated or divorced. Those with a diagnosis of brief psychotic disorder had the highest EQ-5D score while those with a diagnosis of depression with psychotic features had the lowest (Table 1). Moderate or extreme problem was most frequently reported in the domain ‘anxious/depressed’ (49.0%), followed by ‘pain/discomfort’ and ‘usual activities’ (32.8%), whilst ‘self-care’ (7.9%) was the least frequently reported problem in this sample. The EQ-5D index was negatively correlated with the DUP, PANSS general psychopathology and total scores at baseline (Table 2). Table 3 shows the relationship between EQ-5D utility scores and sociodemographic and clinical correlates based on three models (OLS, Tobit and CLAD). We found the OLS model produced the lowest MAPE value and was considered the best model to explain the relationship. In multiple linear regression analysis, DUP (Beta ¼  0.002, p value ¼0.025) and baseline PANSS general psychopathology score (Beta ¼  0.008, p value o0.001) were negatively and significantly associated with EQ-5D index (Table 3). While EQ-5D indices were higher among those with less severe clinical symptoms at baseline (PANSS r76; 76 being the median PANSS score of the cohort), there were no significant differences between the two groups (Table 4). However, only four patients had baseline GAF scores higher than 60, hence it was not possible to comment on the differences in EQ-5D indices between those with higher functioning vs. those with lower functioning using a GAF cut-off of 60 (Table 4). One hundred thirty five (56.0%) patients completed the EQ-5D assessment at the 12-month follow-up. There were no significant differences in terms of sociodemographic characteristics between those who completed 12-month follow-up and those who did not complete follow-up (N ¼106 for those without follow up). At the 12-month time point, the mean (S.D.) EQ-5D index was 0.891 (0.197). Repeated measures ANOVA indicated significant improvement in mean EQ-5D index over 1-year follow-up (p value o 0.001). It revealed that the mean EQ-5D index at 12month follow-up was significantly higher as compared to baseline (0.891 vs. 0.799, mean change of 0.092, p value o0.001). Assuming

M. Subramaniam et al. / Psychiatry Research 215 (2014) 46–51

49

Table 1 Demographic and clinical characteristics of the sample (n¼ 241). Variable Sex Female Male Race Chinese Malay Indian Others (specify) Marital Single/unmarried Married Separated and divorced Missing SCID diagnosis Schizophrenia spectrum (Schizophrenia, Schizophreniform and Schizoaffective) Bipolar (with or without psychotic features) Delusional disorder Brief psychotic disorder Psychosis not otherwise specified Depression (with psychotic features) Age DUP since onset of symptoms (in months) Baseline PANSS positive Baseline PANSS negative Baseline PANSS general Psychopathology Baseline PANSS total Baseline GAF total Baseline GAF symptoms Baseline GAF disability EQ-5D Index n

Percent

Mean EQ-5D

p valuen

125 116

51.9 48.1

0.799 0.776

0.044

178 33 20 10

73.9 13.7 8.3 4.2

0.801 0.708 0.810 0.773

0.476

202 29 4 4

84.5 12.1 1.7 1.7

0.774 0.903 0.711 0.931

0.038

164 21 12 18 10 16

68.1 8.7 5.0 7.5 4.2 6.6

0.804 0.741 0.778 0.885 0.720 0.623

0.006

Mean 26.1 10.6 21.2 15.5 38.9 75.5 40.0 40.8 45.2 0.788

S.D. 6.7 16.5 6.1 7.5 10.4 19.1 12.5 12.6 11.7 0.258

Significant differences between subgroups were set at p value o 0.05.

Table 2 Correlation between EQ-5D index and clinical scales (n ¼241). Variable

Spearman correlation coefficient

DUP since onset of symptoms (in months) Baseline PANSS positive Baseline PANSS negative Baseline PANSS general psychopathology Baseline PANSS total Baseline GAF total Baseline GAF symptoms Baseline GAF disability

 0.134n  0.001  0.103  0.179nn  0.128n 0.008 0.038 0.001

n

N

Correlation is significant at the 5% level. Correlation is significant at the 1% level.

nn

linear changes in utility values over time, the QALY gain in this cohort using the AUC approach was 0.092 (95% CI, 0.045–0.140). The 12-month EQ-5D index was negatively and significantly associated with DUP after controlling for sociodemographic variables and baseline EQ-5D index in multiple linear regression analysis (Beta¼  0.304, p value o0.001). Out of the 135 patients (54%) who had completed EQ-5D at 12-months follow up, 73 patients (54.1%) had remitted. We found that EQ-5D index was significantly higher among those who were in remission at 12-months than those who were not in remission (0.926 vs. 0.850, p value¼0.039) (Table 4).

4. Discussion The EQ-5D index among the cohort of patients with FEP at baseline was significantly lower as compared to the general population mean of Singapore. There was a significant change of 0.092 (95% CI 0.045–

0.140) in the mean EQ-5D index from baseline [0.799 (95% CI, 0.757– 0.840)] to 1-year follow-up [0.891 (95% CI of 0.858–0.925)], yet the mean index after 1 year of treatment was still lower compared to general population. This is consistent with studies that have shown that a psychotic disorder like schizophrenia tends to be chronic and with significant negative impact on emotional, physical and functional domains of health and that despite successful treatment of psychotic symptoms, patients continue to experience poorer QoL than the general population (Harvey et al., 2004; Ritsner, 2007). Our values are slightly higher than those reported by Barton et al. (2009) of 0.676 (95% CI 0.604–0.748) at baseline compared with 0.743 (95% CI 0.671–0.816) at 9 months, giving a mean change of 0.043 (95% CI 0.034–0.122) from an early psychosis trial in Norfolk, UK. We found that EQ-5D index at baseline was negatively correlated with baseline PANSS, general psychopathology and total scores. EQ-5D indices were also higher among those with less severe PANSS scores [using a cut off of 76 (the median value in our sample)]. Our findings are similar to that of the study by Barton et al. (2009) among people with psychosis that found differences in EQ-5D values between groups with mild and more severe illness. Unfortunately, we were unable to establish whether there were differences in EQ-5D indices according to differences in baseline functioning as only four patients were assessed to be “high functioning” using a GAF cut-off score of 60. However at 1-year follow-up, EQ-5D index was significantly higher among those who were in remission than those who were not. Remission is an important outcome indicator and hence in our study we had used criteria for both symptomatic and functional improvement to define remission, and we were able to demonstrate that EQ5D indices corresponded well with both symptomatic and functional improvement. Prieto et al. (2004) investigated the construct validity of EQ-5D in 2657 patients with schizophrenia. The patients completed EQ-5D at baseline and after 3–6 months of therapy with different

50

M. Subramaniam et al. / Psychiatry Research 215 (2014) 46–51

Table 3 Factors associated with EQ-5D index at baseline (n¼ 236). Variable

OLS

Age Sex Male Female Marital status Single Married Separated/divorced Ethnicity Chinese Malay Indian Others Diagnosis Schizophrenia spectrum (Schizophrenia, Schizophreniform and Schizoaffective) Bipolar (with or without psychotic features) Delusional disorder Brief psychotic disorder Psychosis not otherwise specified Depression (with psychotic features) DUP symptoms PANSS positive PANSS negative PANSS general psychopathology GAF symptoms GAF disability Goodness of fit indices MAPE ME RMSEA

CLAD

95% CI

Tobit

Beta

95% CI

Beta

95% CI

0

 0.004

0.007

0.002

 0.003

0.007

0.003

 0.005

0.012

Ref. 0.034

 0.034

0.103

Ref. 0.086

0.01

0.162

Ref. 0.073

 0.026

0.172

Ref. 0.108  0.005

 0.005  0.189

0.22 0.179

Ref. 0.054  0.032

 0.264  0.356

0.371 0.293

Ref. 0.097  0.057

 0.204  0.341

0.399 0.227

Ref.  0.08 0.028  0.048

 0.174  0.147  0.205

0.015 0.091 0.109

Ref.  0.068  0.076  0.036

 0.214  0.362  0.173

0.079 0.211 0.101

Ref.  0.119  0.055  0.047

 0.255  0.228  0.276

0.017 0.118 0.183

Ref.  0.048  0.047 0.026  0.08  0.078  0.002  0.001  0.001  0.008  0.002 0

 0.175  0.196  0.097  0.245  0.218  0.005  0.007  0.004  0.013  0.007  0.005

0.078 0.102 0.15 0.085 0.062  0.001 0.006 0.007  0.004 0.003 0.003

Ref. 0.046  0.062 0.082  0.109  0.071  0.001  0.005 0  0.006  0.002  0.001

 0.01  0.379  0.13  0.325  0.207  0.005  0.01  0.008  0.013  0.007  0.008

0.197 0.255 0.293 0.107 0.065 0.002 0.003 0.008 0.001 0.004 0.005

Ref.  0.027  0.1 0.056  0.114  0.112  0.004  0.002 0.004  0.012  0.003  0.004

 0.21  0.311  0.135  0.343  0.305  0.007  0.012  0.004  0.018  0.01  0.008

0.156 0.111 0.247 0.116 0.082  0.001 0.008 0.012  0.006 0.004 0.007

0.219 0.011 0.151

Beta

0.225  0.061 0.183

0.236  0.081 0.198

MAPE, Mean Absolute Prediction Error; ME, Mean Error; RMSE, Root Mean Square Errors. Table 4 EQ-5D index among various clinical groups (N ¼ 135). Variable Baseline GAF total score Baseline PANSS total score Remission at 12-monthsb

a b

n Mild ( 4 61) More severe ( o 60) More severe ( 4 77) Mild ( o 76) Yes No

4 131 62 73 73 62

EQ-5D score

Difference, mean (95% CI)

p valuea

0.667 0.803 0.759 0.832 0.926 0.850

0.135 (  0.111, 0.382)

–

0.073 (  0.010, 0.157)

0.129

0.076 (0.009, 0.142)

0.039

Mann–Whitney U Test. Remission was estimated based on PANSS and GAF score.

antipsychotics. They found a positive association between the GAF and CGI scores (clinician rated) and the scores on the EQ-5D. In addition, the EQ-5D was able to identify differences on QoL among patients with different degrees of severity which led the authors to conclude that the EQ-5D had acceptable construct validity in patients with schizophrenia. Our study is among the first to identify a significant negative correlation between DUP and EQ-5D both at baseline and at end of 1-year follow-up. Long DUP has been shown to be associated with negative outcomes in a number of studies (Addington et al., 2004; Barnes et al., 2008), and its negative association with HRQoL as well emphasises the need for aggressive outreach and early intervention to reduce DUP. HRQoL has become an important outcome in clinical practice and research. However, the validity of self-report instruments has been questioned among patients with schizophrenia and one study had concluded that in general, quality-of-life measures did not correlate well with symptoms or clinician-assessed outcomes in mental illnesses except in the case of depression (Saarni et al., 2010). On the

other hand, Stochl et al. (2013) studied the validity of individual questionnaire items using an Item Response/Latent Trait Theory modelling approach and concluded that EQ-5D instrument items provide a valid and useful scale for evaluation of HRQoL in a diverse grouped sample of young adults with FEP. Similarly in our study, EQ5D significantly correlated with both disease severity and symptomatic and functional improvement in patients with FEP, lending support to the use of EQ-5D as a valid outcome indicator in an early psychosis programme. There are several limitations that need to be considered when interpreting the results of this study. First, the study does not take into account comorbid medical conditions; and while the patients in our study were young (16–40 years), prevalence of comorbid medical conditions is known to be high among those with chronic mental illnesses and it is likely that the health state utilities of patients with psychosis and medical disorders would be significantly reduced. Second, the EQ-5D tariffs most commonly used for the translation of health states into utilities are derived using the time trade-off method (TTO). The first TTO-based EQ-5D tariff used

M. Subramaniam et al. / Psychiatry Research 215 (2014) 46–51

valuations of 42 health states from a British general population sample to model utilities for the 243 states (Dolan,1997). These values have been used extensively and we used the same for the purposes of our study. However, applying the UK tariff to nonBritish study populations may not be appropriate because health preferences may vary between countries and cultures (Greiner et al., 2005; Badia et al., 2001). Lastly, as EQ-5D was collected as part of the programme evaluation, we were unable to follow up on patients who either defaulted or were transferred from the programme. These limitations notwithstanding, our results suggest that the EQ5D is responsive to clinical improvements as it corresponded well to objective ratings of remission in our patients with FEP. Our results also suggest that EQ-5D can be used for the purpose for evaluating effectiveness of mental health programmes and interventions. Utility scores are needed for calculating QALYs, which are used as indicators of effectiveness or outcome in economic evaluations (Gold et al., 1996; Pyne et al., 1997). Cost-effectiveness studies are required to help clinicians and health care decision-makers in determining the impact of new programmes in terms of the costs. Gain in QALYs and costutility analyses allow comparisons across diseases and help determine the allocation of scarce resources in healthcare. There is a need for further research in both determining the appropriateness of HRQoL measures for mental health conditions and their use in measuring the cost utility of mental health programmes. References Abdin, E., Subramaniam, M., Vaingankar, J.A., Nan, L., Chong, S.A., 2013. Measuring health-related quality of life among adults in Singapore: population norms for the EQ-5D. Qual. Life Res. 2013 (April), 3. (Epub ahead of print). Addington, J., Young, J., Addington, D., 2003. Social outcome in early psychosis. Psychol. Med. 33, 1119–1124. Addington, J., Van Mastrigt, S., Addington, D., 2004. Duration of untreated psychosis: impact on 2-year outcome. Psychol. Med. 34, 277–284. Adewuya, A.O., Makanjuola, R.O., 2010. Subjective life satisfaction and objective living conditions of patients with schizophrenia in Nigeria. Psychiatry Serv. 61, 314–316. Andreasen, N.C., Carpenter Jr., W.T., Kane, J.M., Lasser, R.A., Marder, S.R., Weinberger, D.R., 2005. Remission in schizophrenia: proposed criteria and rationale for consensus. Am. J. Psychiatry 162, 441–449. Angermeyer, M.C., Holzinger, A., Kilian, R., Matschinger, H., 2001. Quality of life-as defined by schizophrenic patients and psychiatrists. Int. J. Soc. Psychiatry 47, 34–42. Auquier, P., Simeoni, M.C., Sapin, C., Reine, G., Aghababian, V., Cramer, J., Lancon, C., 2003. Development and validation of a patient-based health related quality of life questionnaire in schizophrenia: the S-QoL. Schizophr. Res. 63, 137–149. Austin, P.C., 2002. A comparison of methods for analyzing health-related quality-oflife measures. Value Health 5, 329–337. Awad, A.G., Voruganti, L.N., 2000. Intervention research in psychosis: issues related to the assessment of quality of life. Schizophr. Bull. 26, 557–564. Badia, X., Roset, M., Herdman, M., Kind, P., 2001. A comparison of United Kingdom and Spanish general population time trade-off values for EQ-5D health states. Med. Decis. Making 21, 7–16. Barnes, T.R., Leeson, V.C., Mutsatsa, S.H., Watt, H.C., Hutton, S.B., Joyce, E.M., 2008. Duration of untreated psychosis and social function: 1-year follow-up study of first-episode schizophrenia. Br. J. Psychiatry 193, 203–209. Barton, G.R., Hodgekins, J., Mugford, M., Jones, P.B., Croudace, T., Fowler, D., 2009. Measuring the benefits of treatment for psychosis: validity and responsiveness of the EQ-5D. Br. J. Psychiatry 195, 170–177. Bechdolf, A., Pukrop, R., Köhn, D., Tschinkel, S., Veith, V., Schultze-Lutter, F., Ruhrmann, S., Geyer, C., Pohlmann, B., Klosterkötter, J., 2005. Subjective quality of life in subjects at risk for a first episode of psychosis: a comparison with first episode schizophrenia patients and healthy controls. Schizophr. Res. 79, 137–143. Bobes, J., García-Portilla, P., Sáiz, P.A., Bascarán, T., Bousoño, M., 2005. Quality of life measures in schizophrenia. Eur. Psychiatry 20, S313–S317. Carr, V., Neil, A., Halpin, S., Holmes, S., 2002. Costs of Psychosis in Urban Australia. A Bulletin of the Low Prevalence Disorders Study, National Survey of Mental Health and Wellbeing. Available online at 〈http://www.health.gov.au/internet/ main/publishing.nsf/Content/3C7E39B92E056E60CA257229001704E6/$File/ costcov.pdf〉. Chong, S.A., Lee, C., Bird, L., Verma, S., 2004. A risk reduction approach for schizophrenia: the early psychosis intervention programme. Ann. Acad. Med. Singap. 33, 630–635. Chong, S.A., 2007. Mental health in Singapore: a quiet revolution? Ann. Acad. Med. Singap. 36, 795–796.

51

Dolan, P., 1997. Modelling valuations for EuroQol health states. Med. Care 35, 1095–1108. Doyle, M., Flanagan, S., Browne, S., Clarke, M., Lydon, D., Larkin, C., O'Callaghan, E., 1999. Subjective and external assessments of quality of life in schizophrenia: relationship to insight. Acta Psychiatr. Scand. 99, 466–472. Drummond, M.F., O'Brien, B., Stoddart, G.L., Torrance, G.W., 1997. Methods for the Economic Evaluation of Health Care Programmes. Oxford University Press, Oxford, UK. First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 1996. Structured Clinical Interview for DSM-IV Axis I Disorders, Clinician Version (SCID-CV). American Psychiatric Press, Inc, Washington, D.C. Fitzgerald, P.B., Williams, C.L., Corteling, N., Filia, S.L., Brewer, K., Adams, A., de Castella, A.R., Rolfe, T., Davey, P., Kulkarni, J., 2001. Subject and observerrated quality of life in schizophrenia. Acta Psychiatr. Scand. 103, 387–392. Forster, M., 2002. Predictive accuracy as an achievable goal of science. Philos. Sci. 69, S124–S134. Gold, M.R., Siegal, J.E., Russell, L.B., Weinstein, M.C. (Eds.), 1996. Cost-effectiveness in Health and Medicine. Oxford University Press, New York. Goldman, H.H., Skodol, A.E., Lave, T.R., 1992. Revising axis V for DSM–IV: a review of measures of social functioning. Am. J. Psychiatry 149, 1148–1156. Greiner, W., Claes, C., Busschbach, J.J., von der Schulenburg, J.M., 2005. Validating the EQ5D with time trade-off for the German population. Eur. J. Health Econ. 6, 124–130. Harvey, P.D., Green, M.F., Keefe, R.S., Velligan, D.I., 2004. Cognitive functioning in schizophrenia: a consensus statement on its role in the definition and evaluation of effective treatments for the illness. J. Clin. Psychiatry 65, 361–372. Herrman, H., Hawthorne, G., Thomas, R., 2002. Quality of life assessment in people living with psychosis. Soc. Psychiatry Psychiatr. Epidemiol. 37, 510–518. Huang, I.C., Frangakis, C., Atkinson, M.J., Willke, R.J., Leite, W.L., Vogel, W.B., Wu, A. W., 2008. Addressing ceiling effects in health status measures: a comparison of techniques applied to measures for people with HIV disease. Health Serv. Res. 43, 327–339. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. Manca, A., Hawkins, N., Sculpher, M.J., 2005. Estimating mean QALYs in trial-based cost-effectiveness analysis: the importance of controlling for baseline utility. Health Econ. 14, 487. (486). McGorry, P.D., 2002. The recognition and optimal management of early psychosis: an evidence-based reform. World Psychiatry 1, 76–83. McGorry, P.D., Killackey, E.J., 2002. Early intervention in psychosis: a new evidence based paradigm. Epidemiol. Psichiatr. Soc. 11, 237–247. Oliver, J.P.J., Huxley, P.J., Priebe, S., Kaiser, W., 1997. Measuring the quality of life of severely mentally ill people using the Lancashire quality of life profile. Soc. Psychiatry Psychiatr. Epidemiol. 32, 76–83. Patrick, D.L., Erickson., P., 1993. Health Status and Health Policy. Oxford University Press, New York. Payakachat, N., Summers, K.H., Pleil, A.M., Murawski, M.M., Thomas, J., Jennings, K., Anderson, J.G., 2009. Predicting EQ-5D utility scores from the 25-item National Eye Institute Vision Function Questionnaire (NEI-VFQ 25) in patients with agerelated macular degeneration. Qual. Life Res. 18, 801–813. Powell, J.L., 1984. Least absolute deviations estimation for the censored regression model. J. Econom. 25, 303. (235). Prieto, L., Sacristan, J.A., Hormaechea, J.A., Casado, A., Badia, X., Gomez, J.C., 2004. Psychometric validation of a generic health-related quality of life measure (EQ-5D) in a sample of schizophrenic patients. Curr. Med. Res. Opin. 20, 827–835. Pyne, J.M., Patterson, T.L., Kaplan, R.M., Gillin, J.C., Koch, W.L., Grant, I., 1997. Assessment of quality of life of patients with major depression. Psychiatr. Serv. 48, 224–230. Ritsner, M.S., 2007. The distress/protection vulnerability model of quality of life impairment syndrome. In: Ritsner, MS, Awad, AG (Eds.), Quality of Life Impairment in Schizophrenia, Mood and Anxiety Disorders. New Perspectives on Research and Treatment. Springer, London, pp. 3–19. Saarni, S.I., Viertiö, S., Perälä, J., Koskinen, S., Lönnqvist, J., Suvisaari, J., 2010. Quality of life of people with schizophrenia, bipolar disorder and other psychotic disorders. Br. J. Psychiatry 197, 386–394. Savilla, K., Kettler, L., Galletly, C., 2008. Relationships between cognitive deficits, symptoms and quality of life in schizophrenia. Aust. N. Z. J. Psychiatry 42, 496–504. Singapore Department of Statistics. Singapore in Figures, 2012. Available online at 〈http://www.singstat.gov.sg/pubn/reference/sif2012.pdf〉 (last assessed on 11 October 2012). Stochl, J., Croudace, T., Perez, J., Birchwood, M., Lester, H., Marshall, M., Amos, T., Sharma, V., Fowler, D, Jones, P.B., 2013. The National Eden Study Team, 2013. Usefulness of EQ-5D for evaluation of health-related quality of life in young adults with first-episode psychosis. Qual. Life Res., 1055–1063. (Epub ahead of print). Sullivan, P.W., Ghushchyan, V., 2006. Mapping the EQ-5D Index from the SF-12: US general population preferences in a nationally representative sample. Med. Decis. Making, 26; , pp. 401–409. The EuroQol Group, 1990. EuroQol: a new facility for the measurement of healthrelated quality of life. Health Policy 16, 199–208. Tobin, J., 1958. Estimation of relationships for limited dependent variables. Econometrica 26, 24–26. Verma, S., Subramaniam, M., Abdin, E., Poon, L.Y., Chong, S.A., 2012. Symptomatic and functional remission in patients with first-episode psychosis. Acta Psychiatr. Scand. 126, 282–289. Voruganti, L., Heslegrave, R., Awad, A.G., Seeman, M.V., 1998. Quality of life measurement in schizophrenia: reconciling the quest for subjectivity with the question of reliability. Psychol. Med. 28, 165–172. WHOQOL Group, 1994. Development of the WHOQOL: rationale and current status. Int. J. Ment. Health 23, 24–56.