Metacognition in first-episode psychosis and its association with positive and negative symptom profiles

Psychiatry Research 238 (2016) 14–23

Contents lists available at ScienceDirect

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

Metacognition in first-episode psychosis and its association with positive and negative symptom profiles$ Anne Marie Trauelsen a,b,c,n, Andrew Gumley d, Jens Einar Jansen b,c, Marlene Buch Pedersen c, Hanne-Grethe Lyse Nielsen c, Christopher Høier Trier e, Ulrik H. Haahr c,f, Erik Simonsen b,f a

Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark Psychiatric Research Unit, Region Zealand Psychiatry Roskilde, Roskilde, Denmark c Early Psychosis Intervention Center, Region Zealand Psychiatry Roskilde, Roskilde, Denmark d Institute of Health and Wellbeing, University of Glasgow, Glasgow, UK e Department of Psychology, University of Copenhagen, Copenhagen, Denmark f Institute of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark b

art ic l e i nf o

a b s t r a c t

Article history: Received 13 August 2015 Received in revised form 27 January 2016 Accepted 3 February 2016 Available online 15 February 2016

There is growing evidence that metacognitive abilities which include the ability to synthesize knowledge regarding mental states in self and others and use this ability to solve problems are impaired in nonaffective psychosis and associated with positive and negative symptom severity. We sought to (a) investigate the severity of metacognitive impairments in first-episode psychosis (FEP) compared to non-clinical controls and (b) explore associations with positive and negative symptom profiles. Ninetyseven people with FEP were compared to 101 control persons. Metacognition was assessed with interviews and the Metacognitive assessment scale-abbreviated. Four groups based on positive and negative symptoms were identified by cluster analysis and compared on metacognition, childhood adversities, duration of untreated psychosis and premorbid social and academic adjustment. Those with high levels of negative symptoms had poorer metacognitive abilities. Those with high positive and low negative symptoms did not have poorer metacognitive abilities than those with low positive and negative symptoms. None of the other predictors differed between the groups. The FEP group had poorer metacognitive abilities than the control group. Inclusion of metacognition in psychosis models may improve our understanding of negative symptoms, while previous findings of a relation with positive symptoms may have been confounded. Implications for current interventions are discussed. & 2016 Elsevier Ireland Ltd. All rights reserved.

Keywords: Case-control Risk positive symptoms Negative symptoms Self-reflectivity Mentalizing

1. Introduction There is accumulating evidence that metacognition is associated with positive and negative symptoms in people with non☆ This paper is part of the OPUS Zealand research project: Family Intervention in First-episode Psychosis. An investigation of metacognition, expressed emotion and caregiver experiences. From January 1. 2012 the project has the following research group: Erik Simonsen, MD, PhD (PI). Ulrik Haahr, MD (PI). Christopher Høier Trier, MSc in Psychology. Marlene Buch Pedersen, MSc in Psychology. Hanne-Grethe Lyse Nielsen, RN. Mette Sjoestroem Petersen, MA (ed) in Educational Psychology, Jens Einar Jansen, PhD. Ulf Soegaard, MD. Anne Marie Trauelsen, MD. From September 1. 2014 Signe Dunker Svendsen, RN. From Psychiatric Research Unit, Psychiatry Region Zealand, Denmark. Early Psychosis Intervention Center, Psychiatry East Region Zealand, Denmark. n Corresponding author at: Psychiatric Research Unit, Region Zealand Psychiatry Roskilde, Roskilde 4000, Denmark. E-mail address: [email protected] (A.M. Trauelsen).

http://dx.doi.org/10.1016/j.psychres.2016.02.003 0165-1781/& 2016 Elsevier Ireland Ltd. All rights reserved.

affective psychosis (Hamm et al., 2012; McLeod et al., 2014). Metacognition is concerned with how individuals make sense of their own and others' behaviour in terms of mental states and the utilisation of this capacity to solve problems and cope with mental states that cause distress (Lysaker et al., 2013, 2005; Semerari et al., 2003). Metacognition also includes the ability to synthesize basic elements of experience into complex wholes and acknowledge that different perspectives are possible. It is related to theory of mind (ToM), mentalisation and many aspects of social cognition (Penn et al., 2008). ToM is the ability to recognise mental states in oneself and others, and to understand that other people may view situations differently (Brune, 2005; Sprong et al., 2007). The concept of mentalisation is derived from ToM but more closely resembles metacognition. It refers to the ability to interpret the behaviour of oneself and others in terms of different mental intentions (Bateman and Fonagy, 2004; Brune, 2005; Choi-Kain and

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

Gunderson, 2008; Sprong et al., 2007). Social cognition is a broad construct related to processes of social interpretation. In psychosis research, social cognition especially implicates ToM, emotion recognition and attributional style, i.e. the attribution of responsibility for an event to oneself or others (Penn et al., 2008). ToM is most often studied using laboratory experiments, whereas mentalisation and metacognition are investigated using personal narratives and social cognition via many different methods. ToM can be regarded as the source of the other three concepts, which are broader, where mentalisation and metacognition are the most strongly rooted in everyday life experiences. In relation to executive functioning, metacognition represents a more personal and complex understanding and use of knowledge regarding oneself and others than what is represented by executive functioning. There is also evidence that impairments in metacognition are greater in those diagnosed with non-affective psychosis compared to controls with significant medical adversity (HIV) (Lysaker et al., 2011b)) and that metacognition is a predictor of non-affective psychosis (Lysaker et al., 2014). With the exception of three studies (Macbeth et al., 2014; McLeod et al., 2014; Vohs et al., 2014), studies exploring metacognition have largely concerned participants with prolonged non-affective psychosis. Vohs et al. (2014) found that people with a first-episode psychosis (FEP) had lower metacognitive abilities compared to those with prolonged psychosis and psychiatric controls. The authors noted that sample sizes were small and, thus, generalisability to the larger population was limited. The other studies did not include control groups. Such a comparison is important to establish the nature and extent of metacognitive difficulties in people with FEP. Positive and negative symptoms are defining features of psychosis; to date, negative symptoms have remained largely unaffected by both pharmacological and psychosocial interventions (Blanchard et al., 2011, 2005; Jordan et al., 2014; Kirkpatrick et al., 2006; Wykes et al., 2008). Several studies have found that metacognition is associated with negative symptoms in people with prolonged psychosis (Hamm et al., 2012; Lysaker et al., 2005; Nicolò et al., 2012; Rabin et al., 2014; Vohs et al., 2014), and a few have shown it in people with FEP (Macbeth et al., 2014; McLeod et al., 2014; Vohs et al., 2014). Seven studies show correlations between negative symptoms and metacognition (Hamm et al., 2012; Lysaker et al., 2005; Macbeth et al., 2014; McLeod et al., 2014; Nicolò et al., 2012; Rabin et al., 2014; Vohs et al., 2014). In addition, Hamm et al. (2012) show that metacognition at baseline predicts the level of negative symptoms six months later, and McLeod et al. (2014) shows that metacognition increases the prediction of recovery of negative symptoms after six and 12 months controlling for baseline symptom severity, the duration of untreated psychosis (DUP) and premorbid adjustment. Concerning associations between metacognition and positive symptoms, the results are less consistent. Three studies show no correlations with metacognition (Macbeth et al., 2014; Nicolò et al., 2012; Vohs et al., 2014), while two studies illustrate fewer correlations with positive (rather than negative) symptoms (Lysaker et al., 2005; McLeod et al., 2014). Hamm et al. (2012) confirm that metacognition correlates with positive symptoms, but that metacognition does not predict positive symptoms at six months follow-up. McLeod et al. (2014) show that metacognition at baseline also predicts positive symptoms at six and 12 months follow-up. Importantly, many of these studies have methodological issues, which may contribute to the lack of findings. For example, Macbeth et al. (2014) have a small sample size (N ¼34), and their FEP sample has low levels of positive symptoms with an average PANSS score for positive symptoms of 10.11 (SD 5.6), which may make it difficult to discover any associations. Nicolò et al. (2012) have a small sample size (N ¼ 45) that includes only stable (no recent hospitalisations) outpatients without substance abuse,

15

which may also render associations hard to find. In addition to these limitations, only one study (McLeod et al., 2014) includes several predictors of positive and negative symptoms. These include DUP (Boonstra et al., 2012; Crumlish et al., 2009; White et al., 2009), premorbid social and academic adjustment (PAS) (Chang et al., 2013; MacBeth and Gumley, 2008; Simonsen et al., 2010) and, less certainly, childhood adversity (Bentall et al., 2012; Sitko et al., 2014; Trotta et al., 2015; van Nierop et al., 2014). Childhood adversities are shown to predict the presence of positive symptoms (Bentall et al., 2012; Duhig et al., 2015; Sitko et al., 2014; van Nierop et al., 2014), and a recent metaanalysis concludes that childhood adversities increase the risk of persistence of psychotic experiences in both clinical and nonclinical populations (Trotta et al., 2015). Thus, prevailing studies have not been representative of non-affective FEP populations. In addition to these studies, a review showed that ToM more often associates with negative than positive symptoms (Harrington et al., 2005), while a meta-analysis shows that facial emotion perception is related to both negative and positive symptoms (Kohler et al., 2010). The mixed findings of the associations between metacognition and positive symptoms emphasise the continuing need for exploratory research. One possible explanation for the mixed findings is that found relations between metacognition and positive symptoms are confounded by the presence of negative symptoms. It has not been sufficiently researched how metacognitive abilities differ in relation to both positive and negative symptoms independently and together. Studies on metacognition explore associations with positive and negative symptoms separately (Hamm et al., 2012; Macbeth et al., 2014; McLeod et al., 2014; Nicolò et al., 2012) and do not take into account the possibility of their co-occurrence. In addition, only one of the above-mentioned studies and no other FEP studies compare the levels of metacognition in their samples to non-clinical control groups (Rabin et al., 2014), albeit it is the foundation of this research. 1.1. Aims and hypotheses Therefore, we aim to conduct a study exploring the association between metacognition and positive and negative symptoms in non-affective FEP. We wish to improve upon previous studies by controlling for several predictors of positive and negative symptoms. Building upon the methodology of Buck and colleagues (2014), we wish to identify positive and negative symptom profiles in people with FEP (Buck et al., 2014). We expect that the FEP group comprises people with all four-symptom profiles, as diagnoses are based on the presence of positive and negative symptoms, both in conjunction and separately. As conflicting studies show both positive and negative symptoms to be associated with poorer metacognitive abilities, we expect them to be associated both independently and together. In addition, we wish to compare the levels of metacognition in people with non-affective FEP to a well-matched, non-clinical control group to establish whether metacognition, as measured with the Metacognitive Assessment Scale-Abbreviated, is lower for people with FEP. We hypothesise that: i. Metacognitive abilities will be lower for people with non-affective FEP than for the non-clinical control group. ii. We will identify four clusters of positive and negative symptoms (High Negative/Low Positive; High Negative/High Positive; Low Negative/High Positive and Low Positive/Low Negative). iii. Poorer metacognitive abilities will be associated with high levels of positive and negative symptoms independently and together.

16

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

2. Methods 2.1. Participants 2.1.1. FEP group The sample has previously been described by Trauelsen et al. (2015). Denmark has a nationwide out-patient, early intervention program (OPUS) for people with FEP (Petersen et al., 2005). Inclusion criteria for OPUS during the study were an ICD 10 diagnosis F20  29, except F21, and an age of 18  35 years. Exclusion criteria were a previous diagnosis of non-affective psychosis. The catchment area was Region Zealand (population ¼ 816,359) and everyone included in treatment between April 1, 2011 and April 1, 2013 was approached for participation. Additional criteria for the study were sufficient Danish skills to complete the interviews. The FEP participants were assessed as soon as possible, but the median time from treatment start to assessment was 94 days (range 456 days). Forty-one percent were psychotic in the week prior to either the first or second interview, when the PANSS was performed. 2.1.2. Control group Inclusion criteria were living in the catchment area and being 17–34 years of age. Exclusion criteria were previous or current psychiatric disorders and insufficient Danish skills to complete the interview. Control persons were matched 1:1 by gender, age (7 1 year), and parental education (7 1 one on a 5-point scale). Control persons were recruited through advertisements in newspapers, at educational institutions, libraries, sport clubs and by word of mouth. They were included from October 1, 2013 to May 22, 2014. 2.2. Measures 2.2.1. Indiana psychiatric illness interview (IPII) The IPII is a semi-structured interview where the interviewees are asked to talk about their life. This includes their life story; the psychiatric problems that they may or may not think they have, and how they have been affected by them; how they control and are controlled by these problems; how they are affected by and affect other people; and how their future will be (Lysaker et al., 2002). The control participants go through the same line of questioning, but instead of psychiatric problems, they are asked about psychological difficulties that have occurred within the last two years. The interview is conversational with minimal amount of direction, though participants are encouraged to answer all questions in the interview and are given plenty of time. This allows for a spontaneous narrative. The interview lasts about 30 min, and the responses are audio taped and subsequently transcribed. To minimize the interviewer's influence on the interviewee, the IPII was the first interview to be performed for both participant groups. 2.2.2. Metacognition assessment scale-abbreviated (MAS-A) MAS-A (Lysaker et al., 2005) is an elaboration of the metacognition assessment scale (MAS) which was developed for metacognitive assessment in psychotherapy (Semerari et al., 2003). The MAS-A brings a dimensional structure to the original MAS and demands a presence of lower order metacognitive skills before the highest scores are awarded. The ratings were based on the IPII transcripts and thus measured the presence of observations and elaborations conveyed by the interviewees. The MAS-A consists of the four subscales Self-reflectivity, Awareness of the Mind of the Other (AoM), Decentration and Mastery. Self-reflectivity consists of nine levels, which range from identifying to integrating knowledge about oneself with the identification of affective and cognitive operations as a benchmark for achieving higher ratings. AoM consists of seven levels, which

range from identifying to integrating knowledge about others, in a progressively more complex way. Decentration consists of three levels. It reflects a person's ability to understand that other people lead their lives and interact with each other independent of the person. Mastery consists of nine levels and concerns the ability to use understanding of self and others to solve or overcome psychological difficulties. The MAS-A total score is the sum of the four subscales. The MAS-A has been found to have good inter-rater reliability with intra-class coefficients ranging from 0.71 to 0.91 (Lysaker et al., 2005). The current study calculated reliability for 20% randomly drawn IPII transcripts rated by the first and third author. Cronbach's alphas for the control and FEP group were 0.89 for the total MAS score, 0.84 for Self-reflectivity, 0.60 for AoM, 0.73 for Decentration and 0.90 for Mastery. 2.2.3. Childhood adversities The Childhood Trauma Questionnaire (CTQ) was used for trauma assessment (Bernstein et al., 2003). We used the Danish validated version (Bernstein and Fink, 2011). The CTQ consists of five subcategories, each represented by five questions. The CTQ subcategories were dichotomized using the cut-off scores from moderate to severe as suggested by Bernstein et al. (Bernstein and Fink, 2011). Separation and institutionalization were assessed with the Childhood Experience of Care and Abuse Questionnaire (CECA.Q) (Bifulco et al., 2005; Smith et al., 2002). The number of adversities ranged from zero to seven and was calculated by combining the five CTQ subcategories with separation and institutionalization. 2.2.4. Psychopathology The OPCRIT diagnostic system was used to obtain ICD  10 diagnoses, based on patient records and a Positive and Negative Symptom Scales (PANSS) interview (Kay et al., 1987). The latter was extended to include life-long symptoms (McGuffin et al., 1991). For analyses, we used the five factor categorization suggested and validated by van der Gaag (2006). A psychologist or medical doctor, i.e. AMT, JEJ, CHT or MBP, administered the instruments and had all received training by a senior psychiatrist. Control participants were screened with the Mini International Neuropsychiatric Interview (MINI) 6.0 for any prior and present psychiatric diagnoses (Sheehan et al., 2008, p. 0). MINI 6.0 was administered by the first author. 2.2.5. Duration of untreated psychosis and premorbid adjustment DUP was defined as having a psychotic symptom score equal to or above 4 on the PANSS for the whole day a week in a row or several times a week for several weeks, as described by (Melle et al., 2006). Any periods without symptoms were subtracted from the DUP. The DUP was considered over when antipsychotic medication, OPUS or inpatient treatment started. Any uncertainties regarding DUP ratings were resolved between the authors and information drawn from electronic patient files and staff. Test– retest reliability for this method of determining DUP from is reported as good (intraclass coefficient r¼.96, p b 0.01) (Larsen et al., 1998). The Premorbid Adjustment Scale (Cannon-Spoor et al., 1982) was used for assessment of premorbid social and academic adjustment. The PAS has shown good validity and reliability (van Mastrigt and Addington, 2002). The median score for all eligible periods were presented separately for social and academic adjustment, as suggested by MacBeth and Gumley (2008). Seventeen PANSS interview videos and vignettes were randomly drawn and rated by the four raters. For diagnoses, there was an overall agreement of 82% and a median kappa of 0.52. The ICC (2, k) coefficients for DUP was 0.94. For the PANSS van der Gaag components, the ICC (2, k) coefficients were: Positive component score

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

0.81, negative component score 0.90, depressive component score 0.86, excitement component score 0.64 and cognitive component score 0.81. These numbers represent excellent 4 .75 and fair-togood reliability 0.40–0.75 (Fleiss, 1999). 2.2.6. Demographic data Parental education was defined as the highest educational degree achieved by either parent with whom the patient had lived. We wished to match on this parameter to avoid confounding, as parental education is associated with increased risk of childhood adversity and increased risk of psychosis (Croudace et al., 2000; Sidebotham and Heron, 2006). Lower skilled referred to education that required special training, usually with participation in longterm (42 years) courses. Higher skilled referred to theoreticalpractical education of about 4 years with an emphasis on the practical. Longer theoretical referred to education of about 4 years with an emphasis on the theoretical. Academic referred to having a University Degree. 2.3. Ethics All participants received oral and written study information. For persons with FEP, it was clearly stated that they could withdraw their consent at any time and that participation had no impact on their treatment. Control persons received DKK 400 equal to EURO 54 as compensation for their time and contribution. Patients did not receive any compensation as the Danish National Ethical Committee does not allow payment of research participation if people are receiving psychiatric treatment. The protocol was submitted to the Regional Ethics Committee and pre-approval was found unnecessary. The Data Protection Council, Region Zealand, approved data management (Journal no. 12  660). 2.4. Statistical analysis The IBM SPSS Statistics for Windows, Version 22.0 was used for all analysis. Frequency graphs of continuous measures were inspected visually to determine approximation of the normal distribution. Ttests were performed to compare variables between the FEP and the control group. PAS social and DUP were transformed by logarithmic functions due to skewed distributions securing good enough normality distribution within groups for all variables. If the variable was not continuous or normally distributed, as for age, a non-parametric test was performed. The K-Means cluster analysis was applied for the identification of four groups based on positive and negative symptom z-scores. Two outliers were excluded from the cluster analysis, as they were too influential; they were allocated to their subgroups subsequently (Steinley, 2006). The groups were derived based on distribution which secures a minimal within-groups difference relative to the between-groups difference (Dillon and Goldstein, 1984). The groups were labelled in correspondence with their values as 1) Low Positive/Low Negative; 2) High Positive/Low Negative; 3) Low Positive/High Negative and 4) High Positive/High Negative. Before performing a one-way multivariate analysis of variance (MANOVA) to test for differences between the groups, correlation analyses were performed. The MANOVA was performed to compare the groups on metacognition, number of adversities, DUP and premorbid social and academic adjustment. Multicollinearity was not an issue with the generally low correlations. The Hosmer-Lemeshow goodness-of-fit tests were nonsignificant and variances were therefore assumed equal. Based on the Box's M value of 29.14 and associated value of 0.99 the covariance matrices between the groups were also assumed to be equal for the purposes of the MANOVA (Tinsley and Brown, 2000).

17

We chose Pillai's trace as it is the least sensitive to violations of assumptions. Subsequently, the Fisher's Least Squared Difference method was applied for a series of post-hoc analyses for individual mean difference comparisons across all four groups and the five dependent variables.

3. Results 3.1. Descriptive data There were 194 eligible persons with FEP, of which 101 (52%) fully consented to participate. Exclusions were due to lack of a wish to participate (51); inability to obtain contact (18); withdrawal of consent (14); inclusion before the CTQ was included in the study protocol (11) and insufficient Danish skills (3). This paper contained additional missing data for the 101 included FEP participants. These were due to a lack of PAS data because of very early onset of psychosis (N ¼2); lack of DUP due to schizophrenia simplex (F20.6) (N ¼ 3) and partially missing data due to incomplete data assessment (N ¼13). Thus, 83 persons had full datasets and were included in the symptom profile comparisons. Four of the 101 participants with FEP had missing metacognitive data. Therefore, 97 persons were compared to the control group on metacognition. They did not differ from the excluded group in relation to age (Mann–Whitney U 4160, p ¼0.065) but held fewer females (43% compared to 26%, χ2 ¼5.42, p ¼0.02) and more people with schizophrenia (68% compared to 92%, χ2 ¼24.6, p o0.000). 3.2. Case control comparisons All the metacognitive subscales were significantly lower in the FEP group with the largest difference for Self-reflection (zscore ¼5.29, p o0.001) and the lowest for AoM (z-score¼ 4.37, p o0.001). The number of adversities was significantly higher in the FEP group with an average of 2.87 adversities compared to 0.65 adversities in the control group (p o0.001). The details of the adversities are presented in Trauelsen et al. (2015). Demographic and clinical characteristics are presented in Table 1. 3.3. Cluster analysis The K-means cluster analysis identified four symptom profiles. Group 1 had Low Positive/Low Negative symptoms; positive mean score 8.82 (SD 1.88) and negative mean score 12.44 (SD 3.07). Group 2 had High Positive/Low Negative); positive mean 16.21 (SD 2.72) and negative mean 10.71 (SD 2.09). Group 3 had Low Positive/High Negative); positive mean 8.17 (SD 2.17) and negative mean 24.75 (SD 3.14). Group 4 had High Positive/High Negative); positive mean 16.30 (SD 3.60) and negative mean 19.48 (SD 3.57). These are presented in Fig. 1. 3.4. Symptom profile comparisons There was a significant difference between the symptom groups for the MANOVA including premorbid adjustment, DUP, metacognition and number of adversities between (Pillais' Trace¼0.42, F(15, 231) ¼2.51, p¼ 0.002). The effect size, Partial Eta Squared, was 0.14, which implies that 14% of the variance in the shared constructed dependent variable (canonically derived) was accounted for by positive and negative symptom severity. To determine which dependent variables had the greatest MANOVA effect the standardized discriminant function coefficients were calculated. They suggested that the symptom groups were maximally differentiated by a canonical variable with greatest weight

18

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

Table 1 Demographic and clinical data for persons with first-episode psychosis (n¼ 101) and matched control persons (n¼ 101). Demographic variable

FEP, n (%)

Controls, n (%)

χ2

p

Gender Male Female Age, median (range) Ethnicity

75 (74) 26 (26) 22.5 (18  34)

75 (74) 26 (26) 22 (18  33)

U ¼  5060a

0.931

White, Danish Black African Asian Education

93(92) 1 (1) 1 (1)

98 (97) 1 (1) 1 (1)

Not finished public school Public school (9  10 years) Gymnasium (12  13 years) University/completed vocational education Parental education b

5 (5) 68 (68) 13 (13) 14 (14)

0 19 (19) 57 (56) 25 (25)

64.4

0.000

No education Lower skilled Higher skilled Longer theoretical Academic Number of adversities c

14 (14) 14 (14) 38 (38) 28 (28) 7 (7) 2.87 (1.98)

10 (10) 10 (10) 40 (40) 28 (28) 13 (13) 0.65 (1.13)

3.19

0.527

Z ¼  8.69

Mean (SD)

Median (range)

Mean (SD)

Median (range)

Self-reflectivity (N¼ 97) AoM (N¼97) Decentration (N¼ 97) Mastery (N ¼97) Premorbid adjustment

5.24 (1.50) 3.42 (0.95) 1.20 (0.62) 4.74 (1.47)

5.00 (5.50) 3.00 (5.00) 1.00 (3.00) 5.00 (7.00)

6.33 (1.21) 4.02 (0.98) 1.65 (0.59) 5.68 (1.09)

6.50 (5.00) 4.00 (5.00) 1.50 (4.50) 6.00 (5.00)

Social Academic DUP in weeksd PANSS 5 factorse

1.91 (1.4) 2.48 (1.3) 168 (225)

1.8 (6.0) 2.3 (5.5) 68.0 (1072)

11.6 (4.73) 16.3 (6.21) 5.03 (1.56) 8.47 (2.81) 8.71 (3.40)

16.0 (24.0) 4.00 (6.00) 8.00 (15.0) 8.00 (14.0)

o 0.001 p

Z-score

Metacognition (MAS)

Positive Negative Excitatory Disorganized Emotional

 5.29  4.37  5.13  4.64

o 0.001 o 0.001 o 0.001 o 0.001

Notes: AoM, awareness of the mind of the other. a

Mann–Whitney test. The highest parental vocational education for either parent. Including CTQ subcategories with moderate-severe cut-off scores, institutionalization and separation from primary caregivers. d Three missing due to schizophrenia simplex diagnosis. e Van der Gaag's 5 factor model (van der Gaag, 2006). b c

Table 2 Pearson correlation analysis of baseline data in persons with FEP with full datasets (N ¼ 83).

1. MAS sum score 2. DUP 3.Number of adversities 4. PAS academic 5. PAS social mean 6. Positive symptoms† 7. Negative symptoms† 8. Self-reflectivity 9. AoM 10. Decentration 11. Mastery

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

1  0015 .0139  0.138 0.092  0.113  0.417nn 0.902nn 0.757nn 0.776nn 0.828nn

1 0.240n  0.014 0.176 0.352nn  0.176  0.031 0.033 0.111  0.074

1 0.000  0.060 0.120  0.065 0.145 0.130 0.162 0.056

1 0.200 0.126  0.014  0.95  0.220  0.210  0.028

1 .0228n 0.050 0.124  0.069 0.063 0.122

1  0.047  0.094  0.065  0.025  0.140

1  0.415nn  0.327nn  0.355nn  0.286nn

1 0.564nn 0.660nn 0.650nn

1 0.701nn 0.424nn

1 0.444nn

1

Notes: AoM, awareness of the mind of the other. n

po 0.05. po 0.01. Positive and negative symtoms as defined by van der Gaag et al. (2006).

nn

†

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

19

Table 3 Comparison of baseline characteristics between groups based on positive and negative symptoms (N ¼83).

ANOVA a MAS-A total DUPb Number of adversitiesc Premorbid academic Premorbid social b

η2

Standardized discriminant function coefficients

1 Low negative/ low positive (n ¼ 34) Mean (SD)

2 Low negative/ high positive (n ¼14) Mean (SD)

3 High negative/ low positive (n¼ 12) Mean (SD)

4 High negative/ high positive (n ¼23) Mean (SD)

F(15, 231)

P¼

0.96 0.54  0.40

16.03 (3.19) 119.09 (177.06) 2.88 (1.97)

14.57 (3.33) 179.21(214.17) 2.50 (1.61)

11.71 (3.80) 129.92 (248.68) 2.92 (1.98)

13.35 (2.99) 192.26 (225.54) 3.04 (1.97)

6.40 2.57 0.24

o 0.001 0.20 143, 4; 243 0.06 0.09 3o 2, 4 0.87 0.01

0.31

2.36 (1.15)

1.61(1.37)

2.25 (0.98)

2.93 (1.53)

1.38

0.26

0.05

 0.44

1.90 (1.28)

2.83(1.23)

1.78 (1.62)

2.38 (1.45)

1.15

0.35

0.04

Post hoc comparisons p o0.05

a MANOVA comparing metacognitive total score (MAS-A), DUP, number of adversities, premorbid academic and social functioning. Pillais' Trace¼ 0.42, F(15, 231) ¼2.51, p ¼0.002, Partial Eta Squared¼ 0.14. b DUP and premorbid social adjustment were entered as logarithmically transformed variables. c Including CTQ subcategories with moderate-severe cut-off scores, institutionalization and separation from primary caregivers.

from the MAS-A total score (0.96), followed by DUP (0.54) and premorbid social adjustment ( 0.44). This corresponded with the correlations between the dependent variables and the canonically derived scores which were 0.80 for the MAS-A total score and 0.27 for DUP, and smaller than 7 0.10 for the last three variables. The post-hoc analyses showed that the MAS-A total score differed significantly between the groups (p o0.001). The groups with high negative symptom scores had lower MAS-A total scores than the groups without, showing an independent relation between metacognition and negative symptoms. On the contrary, the group with high positive symptoms and low negative symptoms did not have significantly lower MAS-A total than the group with low positive and negative symptom scores. The correlations between MAS-A subscales and positive symptoms were all insignificant (r ¼  0.14 to  0.03), whereas those with negative symptoms were all significant (r ¼  0.29 to  0.415). The DUP trended towards significance (p ¼0.06) reflecting that the groups with high positive symptom scores had longer DUPs than the groups without, independent of negative symptom scores. The correlation analyses, MANOVA and post-hoc analyses are presented in Tables 2–4. To address the concern of the importance of the timing of assessment, we performed correlation analyses between time from treatment start to assessment and the MAS-A subscales. There were no correlations between the time and Self-reflectivity (rs ¼0.03, p ¼0.80), AoM (rs ¼0.13, p ¼0.20), Decentration (rs ¼ 0.04, p ¼0.71) or Mastery (rs ¼  0.03, p ¼0.75). We also performed t-test comparisons of those with and without positive symptom remission at time of interview. This also showed that there were no differences on any MAS-A subscale: Self-reflectivity (t ¼  0.66, p ¼0.51), AoM (t¼  0.74, p ¼0.46), Decentration (t ¼  1.25, p ¼0.21) or Mastery (t ¼ 0.44, p ¼0.67).

3.5. Post-hoc symptom group comparisons of metacognitive subscales The MAS-A total score was the only variable that differed significantly between the groups. Therefore, and in addition to our original hypotheses, we explored whether these differences were due to any particular MAS-A subscale. A MANOVA analysis could not be conducted due to unequal variances between the groups (Levene's Test of Equality of Error Variances for AoM was F(3, 90) ¼ 3.71, p¼ 0.01 and F(3, 90) ¼ 4.79, p ¼0.004 for Mastery). The nonparametric Kruskal–Wallis test was performed instead and showed significant differences between the groups on Self-reflectivity (χ2 ¼16.73, p ¼0.001); Decentration (χ2 ¼10.75, p ¼0.01); and Mastery (χ2 ¼ 17.80, p ¼0.01). These results were followed by non-parametric Mann–Whitney U tests between the individual groups. Bonferroni corrections were applied for each scale, which consisted of six analyses. Overall, Self-reflectivity (0.001), Decentration (0.006) and Mastery (0.002) scores were all higher in the Low Positive/Low Negative group compared to the Low Positive/ High Negative group. The Low Positive/Low Negative group was higher than the High Positive/High Negative group on Self-reflectivity (0.001) and Mastery (o0.001). Of note, the subscale comparisons were based on larger groups than for the MANOVA since the variables with missing data (DUP, PAS and PANSS) were not included in these post-hoc analyses.

4. Discussion The aims of this study were to compare metacognitive abilities in representative samples of people with FEP and non-clinical

Table 4 Comparisons of metacognitive subscales between symptom groups (N ¼ 94).

Self-reflectivity AoM Decentration Mastery

1 Low positive/low negative (n ¼39) Mean (SD) Mean rank

2 High positive/low negative (n ¼ 15) Mean (SD) Mean rank

3 Low positive/high negative (n¼ 16) Mean (SD) Mean rank

4 High positive/high negative (n ¼ 24) Mean (SD) Mean rank

χ2(3)

6.00 (1.45) 3.68 (1.14) 1.44 (0.65) 5.47 (1.15)

5.20 (1.31) 3.53 (0.83) 1.20 (0.53) 4.30 (1.78)

4.50 3.06 0.88 3.88

4.69 (1.38) 3.21 (0.66) 1.04 (0.51) 4.40 (1.08)

16.73 o 0.01 1o 3, 4 6.19 0.10 10.75 0.01 1o3 17.80 o 0.001 1o 3, 4

60.19 53.21 56.97 60.99

46.80 53.17 48.03 42.60

(1.28) (0.81) (0.59) (1.75)

33.31 37.06 34.00 35.00

Notes: Kruskal–Wallis analysis of variance and post hoc Mann–Whitney. AoM, awareness of the mind of the other. A total of 94 persons with FEP was divided into groups based on positive and negative symptoms.

a

b c

Mann–Whitney post hoc tests between all groups. Bonferroni correction applied for six tests per subscale, p ¼ 0.008.

36.77 41.65 40.77 36.98

p¼

Ub p o 0.008c

20

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

control persons, and to examine whether poorer metacognitive abilities were associated with high levels of positive and negative symptoms independently and together. We found significantly lower levels of metacognition in participants with FEP compared with age, gender and parental education matched non-clinical controls. As predicted, we identified four clusters of positive and negative symptoms (High Negative/Low Positive; High Negative/ High Positive; Low Negative/High Positive and Low Positive/Low Negative). Based on this cluster analysis we found that impairments in metacognition were only noted in subgroups with higher negative symptoms. Post-hoc analyses revealed that these findings were due to differences in the metacognitive subscales Self-Reflectivity, Decentration and Mastery. There were no differences between the symptom profiles on DUP, premorbid adjustment or number of childhood adversities. Therefore, the findings of this study suggested that impairments in metacognitive abilities were associated with negative symptoms independent of the level of positive symptoms, while there was no consistent association between metacognition and positive symptoms. This corresponds with the existing literature that showed mixed findings of the relation between metacognition and positive symptoms and consistent findings for negative symptoms (Hamm et al., 2012; Lysaker et al., 2005; Macbeth et al., 2014; McLeod et al., 2014; Nicolò et al., 2012). An important difference in our study was that the groups with either High Positive or High Negative symptoms were protected from confounding by the other symptom. For example, McLeod et al. (2014) found no correlation between metacognition and positive symptoms, but found that metacognition increased the prediction of positive symptoms at follow-up. The model was not adjusted for negative symptoms and, thus, the symptoms may have confounded the model. Others found a correlation between metacognition and positive symptoms, but no association in subsequent regression models, again suggesting that findings may have been confounded (Hamm et al., 2012). Our findings suggested that metacognition did not affect the level of positive symptoms, and future studies should replicate our finding or adjust for negative symptoms when exploring this relation. The strong relation between metacognition and negative symptom severity also corresponded well with the literature (Hamm et al., 2012; Lysaker et al., 2011a; Macbeth et al., 2014; Nicolò et al., 2012; Rabin et al., 2014; Vohs et al., 2014). The only unexpected finding was that AoM was not poorer for the High Negative groups (p¼ 0.10). The significant negative correlation between AoM and negative symptoms in the whole group (r ¼ 0.327) could have been due to a low power to detect the relationship, as the group with the highest negative symptom score (High Neg/Low Pos) was small (n ¼12). The finding that metacognitive abilities were only dependent on negative and not positive symptom severity suggested that theories should prioritise the explanation of how low metacognitive abilities may affect negative symptoms. For example, it was suggested that negative symptoms arose from problems with understanding the self and others during social interactions (Salvatore et al., 2008, 2007). Instead of swiftly reaching a conclusion about the other's mental state, several hypotheses remained plausible and eventually caused confusion and arousal. To avoid such uncomfortable internal states, the person would begin to avoid and withdraw from social situations (Salvatore et al., 2008, 2007). In more general terms, metacognition may affect responses to experiences by affecting how they are perceived and interpreted (Lysaker and Hasson-Ohayon, 2014). The levels of metacognitive abilities were in line with the largest FEP sample (McLeod et al., 2014), and the non-clinical control group was similar to the non-clinical control group in a case– control study of people with depression (Ladegaard et al., 2014).

However, two studies with smaller FEP sample sizes found lower metacognitive levels on some of the MAS-A subscales (Macbeth et al., 2014; Vohs et al., 2014). Macbeth et al. (2014) found that AoM, and not Self-Reflectivity, was impaired, and suggested this might apply to all FEP groups. Our analyses suggested otherwise, as we found almost equal differences for all metacognitive subscales (z ¼  5.29 to 4.37). The two studies had small sample sizes, which might explain the different results (N ¼34 and 26). We found a trend towards significance for between-group differences for DUP, but none for the number of adversities or premorbid adjustment. The lack of associations may have been due to the three-month time-period from service entry to assessment where symptom severity should have improved due to the instatement of antipsychotic treatment (Lewis et al., 2002). This may have lowered symptom severity and, thereby, made it more difficult to detect differences. It was unexpected that DUP was not longer for the Low Positive/High Negative group, as metaanalytic evidence associated DUP with negative symptom severity (Boonstra et al., 2012; Marshall et al., 2005; Perkins et al., 2005). Our results suggested that metacognition was a stronger predictor of negative symptoms than DUP and PAS, and that DUP was a stronger predictor of positive symptoms. This was indicative of the independence between DUP, PAS and metacognition, and suggested that they contribute to different aspects of psychosis. Therefore, we recommend that metacognition is included in aetiopathological and recovery models of psychosis, and be considered as a target for interventions, as suggested by Lysaker and Dimaggio (2014). The lack of association between childhood adversities and symptom profiles implied that childhood adversity was unimportant for the level of symptoms in people with FEP. The findings from other FEP studies were mixed, but mostly they found no or weak associations. One study found no correlation between total adversity scores and positive or negative symptoms (Ramsay et al., 2011), while another found a weak correlation with positive and none with negative symptoms (Duhig et al., 2015); a third found correlations between total adversity scores and positive but not negative symptoms (Ucok and Bikmaz, 2007). All studies used the CTQ. No reviews examined the associations in FEP samples, but Trotta et al. (2015), in a meta-analysis of psychotic experiences, found that childhood adversities increased the risk of their persistence in both clinical and non-clinical populations. Thus, the associations may change in the different phases of psychosis. It was possible that the association with childhood adversities was overshadowed by other factors in the acute first phase of psychosis. These questions require further examination. 4.1. Strengths and limitations The cross-sectional design did not allow for causal inferences. It was, therefore, possible that interview performance was affected by negative symptom severity, resulting in lower metacognitive scores and not only vice versa. The timing of assessment was another weakness, as especially positive symptom severity was possibly affected by early treatment response (Lewis et al., 2002). Blinding was not applied to the ratings of metacognition, which possibly biased the results, most likely in the direction of lowered scores for persons with psychosis and higher for control persons. We did not assess intelligence or executive function, which was considered a minor weakness, as metacognition was largely found to be an independent quality (Brüne et al., 2011; Koelkebeck et al., 2010; Sprong et al., 2007). Recall biases are inevitable in retrospective assessment, but studies demonstrated that the recall of adversities was reasonably stable in people with FEP (Fisher et al., 2011) and that they were comparable to people without psychopathology (Fergusson et al., 2000). However, 20% of the sample was randomly picked after rating completion to ensure reliability

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

21

Financial support The study was supported by The Regional Health and Research Foundation, Zealand (A.M.T. Grant number 12  95); and The Lundbeck Foundation (A.M.T. Grant number R93-A8804).

Conflict of interest None.

Acknowledgements We are grateful for the time and effort that the interviewees and clinicians have put into the study and for the supervision and collaboration provided by Professor Paul Lysaker and his research group.

Fig. 1. The distribution of positive and negative symptoms in symptom groups (N¼ 94). Notes: Cluster groups based on positive and negative symptom z-scores.

and validity. The MAS-A raters calibrated their rating skills with the Lysaker research group. In addition, their internal reliability was overall satisfactory for the applied statistical analyses. The control (but not the FEP) group was paid for their participation in the study. This may have caused control participants to be more informative in the IPII interview and, thus, perhaps caused higher MAS-A ratings. The Ethical National Ethical Committee does not permit patient compensation during periods of treatment. The organisation of the Danish Health Service ensures that almost all cases of FEP are referred to the early intervention service. Therefore, the FEP group was representative of the catchment area, even though half of the eligible group was not included. The current study was one of the largest to examine metacognition using the MAS-A in an FEP group and compare it with a well-matched control group. 4.2. Future research and clinical implications The strong association between the high negative symptom groups and metacognition conveyed both clinical and research implications. It signified that, especially those with high levels of negative symptoms, had difficulties recognising and formulating their inner states. This could have impeded communication of mental states with the service, such as asking for help when dealing with disturbing inner states. At the same time, the clinicians may find it difficult to understand and attune to the service user if this information is inaccessible. Consequently, it may be harder to tailor and adjust interventions for this group. The explicit assessment of metacognitive difficulties for those with high negative symptom levels might improve both the service users and clinicians’ recognition of this possible barrier to treatment and serve as a starting point for metacognitive improvements. We suggest future studies examine whether metacognitive improvements decrease the level of negative symptoms, both to establish the association that we have found and to develop interventions that can diminish the negative symptoms. This is supported by the findings that metacognitive development is dependent upon social interactions (Colvert et al., 2008; Ontai and Thompson, 2008; Yagmurlu et al., 2005) and that negative symptoms have been found to improve with improved social environments (Wing and Brown, 1970). Our study focused on positive and negative symptoms. We did not investigate metacognition in relation to patterns of disorganisation symptoms; further research could address this in the future.

References Bateman, A., Fonagy, P., 2004. Psychotherapy for Borderline Personality Disorder: Mentalization-based Treatment, first edition. OUP Oxford, Oxford ; New York. Bentall, R.P., Wickham, S., Shevlin, M., Varese, F., 2012. Do specific early-life adversities lead to specific symptoms of psychosis? A study from the 2007 the adult psychiatric morbidity survey. Schizophr. Bull. 38, 734–740. http://dx.doi. org/10.1093/schbul/sbs049. Bernstein, D.P., Fink, L., 2011. CTQ Vejledning, Dansk Version, 2011th ed. Pearson Assessment, Stockholm. Bernstein, D.P., Stein, J.A., Newcomb, M.D., Walker, E., Pogge, D., Ahluvalia, T., Stokes, J., Handelsman, L., Medrano, M., Desmond, D., Zule, W., 2003. Development and validation of a brief screening version of the childhood trauma questionnaire. Child. Abus. Negl. 27, 169–190. http://dx.doi.org/10.1016/ S0145 2134(02)00541-0. Bifulco, A., Bernazzani, O., Moran, P.M., Jacobs, C., 2005. The childhood experience of care and abuse questionnaire (CECA.Q): validation in a community series. Br. J. Clin. Psychol. 44, 563–581. http://dx.doi.org/10.1348/014466505  35344 [doi]. Blanchard, J.J., Horan, W.P., Collins, L.M., 2005. Examining the latent structure of negative symptoms: is there a distinct subtype of negative symptom schizophrenia? Schizophr. Res. 77, 151–165. http://dx.doi.org/10.1016/j. schres.2005.03.022. Blanchard, J.J., Kring, A.M., Horan, W.P., Gur, R., 2011. Toward the next generation of negative symptom assessments: the collaboration to advance negative symptom assessment in schizophrenia. Schizophr. Bull. 37, 291–299. http://dx.doi. org/10.1093/schbul/sbq104. Boonstra, N., Klaassen, R., Sytema, S., Marshall, M., De Haan, L., Wunderink, L., Wiersma, D., 2012. Duration of untreated psychosis and negative symptoms — a systematic review and meta-analysis of individual patient data. Schizophr. Res. 142, 12–19. http://dx.doi.org/10.1016/j.schres.2012.08.017. Brune, M., 2005. “Theory of mind” in schizophrenia: a review of the literature. Schizophr. Bull. 31, 21–42, doi:sbi002 [pii];10.1093/schbul/sbi002 [doi]. Brüne, M., Schaub, D., Juckel, G., Langdon, R., 2011. Social skills and behavioral problems in schizophrenia: the role of mental state attribution, neurocognition and clinical symptomatology. Psychiatry Res. 190, 9–17. http://dx.doi.org/ 10.1016/j.psychres.2010.03.015. Buck, K.D., McLeod, H.J., Gumley, A., Dimaggio, G., Buck, B.E., Minor, K.S., James, A.V., Lysaker, P.H., 2014. Anhedonia in prolonged schizophrenia spectrum patients with relatively lower vs. Higher levels of depression disorders: associations with deficits in social cognition and metacognition. Conscious. Cogn. 29, 68–75. http://dx.doi.org/10.1016/j.concog.2014.07.005. Cannon-Spoor, H.E., Potkin, S.G., Wyatt, R.J., 1982. Measurement of premorbid adjustment in chronic schizophrenia. Schizophr. Bull. 8, 470–484. Chang, W.C., Tang, J.Y.M., Hui, C.L.M., Wong, G.H.Y., Chan, S.K.W., Lee, E.H.M., Chen, E.Y.H., 2013. The relationship of early premorbid adjustment with negative symptoms and cognitive functions in first-episode schizophrenia: a prospective three-year follow-up study. Psychiatry Res. 209, 353–360. http://dx.doi.org/ 10.1016/j.psychres.2013.02.014. Choi-Kain, L.W., Gunderson, J.G., 2008. Mentalization: ontogeny, assessment, and application in the treatment of borderline personality disorder. Am. J. Psychiatry 165, 1127–1135. http://dx.doi.org/10.1176/appi.ajp.2008.07081360. Colvert, E., Rutter, M., Kreppner, J., Beckett, C., Castle, J., Groothues, C., Hawkins, A., Stevens, S., Sonuga-Barke, E.J., 2008. Do theory of mind and executive function deficits underlie the adverse outcomes associated with profound early deprivation?: findings from the English and Romanian adoptees study. J. Abnorm. Child. Psychol. 36, 1057–1068, doi:10.1007/s10802  008  9232-x [doi]. Croudace, T.J., Kayne, R., Jones, P.B., Harrison, G.L., 2000. Non-linear relationship between an index of social deprivation, psychiatric admission prevalence and the incidence of psychosis. Psychol. MEd. 30, 177–185.

22

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

Crumlish, N., Whitty, P., Clarke, M., Browne, S., Kamali, M., Gervin, M., McTigue, O., Kinsella, A., Waddington, J.L., Larkin, C., O'Callaghan, E., 2009. Beyond the critical period: longitudinal study of 8-year outcome in first-episode non-affective psychosis. Br. J. Psychiatry 194, 18–24. http://dx.doi.org/10.1192/bjp. bp.107.048942. Dillon, W.R., Goldstein, M., 1984. Multivariate Analysis: Methods and Applications, 1 edition. Wiley, New York. Duhig, M., Patterson, S., Connell, M., Foley, S., Capra, C., Dark, F., Gordon, A., Singh, S., Hides, L., McGrath, J.J., Scott, J., 2015. The prevalence and correlates of childhood trauma in patients with early psychosis. Aust. N. Z. J. Psychiatry 49, 651–659. http://dx.doi.org/10.1177/0004867415575379. Fergusson, D.M., Horwood, L.J., Woodward, L.J., 2000. The stability of child abuse reports: a longitudinal study of the reporting behaviour of young adults. Psychol. Med. 30, 529–544. Fisher, H.L., Craig, T.K., Fearon, P., Morgan, K., Dazzan, P., Lappin, J., Hutchinson, G., Doody, G.A., Jones, P.B., McGuffin, P., Murray, R.M., Leff, J., Morgan, C., 2011. Reliability and comparability of psychosis patients’ retrospective reports of childhood abuse. Schizophr. Bull. 37, 546–553, doi:sbp103 [pii];10.1093/schbul/ sbp103 [doi]. Fleiss, J. L., 1999. Reliability of Measurement, in: The Design and Analysis of Clinical Experiments, John Wiley & Sons, Inc., Hoboken, NJ, USA, http://dx.doi.org/10. 1002/9781118032923.ch1. Hamm, J.A., Renard, S.B., Fogley, R.L., Leonhardt, B.L., Dimaggio, G., Buck, K.D., Lysaker, P.H., 2012. Metacognition and social cognition in schizophrenia: stability and relationship to concurrent and prospective symptom assessments. J. Clin. Psychol. 68, 1303–1312. http://dx.doi.org/10.1002/jclp.21906. Harrington, L., Siegert, R.J., McClure, J., 2005. Theory of mind in schizophrenia: a critical review. Cognit. Neuropsychiatry 10, 249–286. http://dx.doi.org/10.1080/ 13546800444000056. Jordan, G., Lutgens, D., Joober, R., Lepage, M., Iyer, S.N., Malla, A., 2014. The relative contribution of cognition and symptomatic remission to functional outcome following treatment of a first episode of psychosis. J. Clin. Psychiatry 75, e566–e572. http://dx.doi.org/10.4088/JCP.13m08606. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–275. Kirkpatrick, B., Fenton, W.S., Carpenter, W.T., Marder, S.R., 2006. The NIMH-MATRICS consensus statement on negative symptoms. Schizophr. Bull. 32, 214–219. http://dx.doi.org/10.1093/schbul/sbj053. Koelkebeck, K., Pedersen, A., Suslow, T., Kueppers, K.A., Arolt, V., Ohrmann, P., 2010. Theory of mind in first-episode schizophrenia patients: correlations with cognition and personality traits. Schizophr. Res. 119, 115–123. http://dx.doi.org/ 10.1016/j.schres.2009.12.015. Kohler, C.G., Walker, J.B., Martin, E.A., Healey, K.M., Moberg, P.J., 2010. Facial emotion perception in schizophrenia: a meta-analytic review. Schizophr. Bull. 36, 1009–1019. http://dx.doi.org/10.1093/schbul/sbn192. Ladegaard, N., Larsen, E.R., Videbech, P., Lysaker, P.H., 2014. Higher-order social cognition in first-episode major depression. Psychiatry Res. 216, 37–43. http: //dx.doi.org/10.1016/j.psychres.2013.12.010. Larsen, T.K., Johannessen, J.O., Opjordsmoen, S., 1998. First-episode schizophrenia with long duration of untreated psychosis. pathways to care. Br. J. Psychiatry Suppl. 172, 45–52. Lewis, S., Tarrier, N., Haddock, G., Bentall, R., Kinderman, P., Kingdon, D., Siddle, R., Drake, R., Everitt, J., Leadley, K., Benn, A., Grazebrook, K., Haley, C., Akhtar, S., Davies, L., Palmer, S., Faragher, B., Dunn, G., 2002. Randomised controlled trial of cognitive-behavioural therapy in early schizophrenia: acute-phase outcomes. Br. J. Psychiatry Suppl. 43, s91–s97. Lysaker, P.H., Carcione, A., Dimaggio, G., Johannesen, J.K., Nicolò, G., Procacci, M., Semerari, A., 2005. Metacognition amidst narratives of self and illness in schizophrenia: associations with neurocognition, symptoms, insight and quality of life. Acta Psychiatr. Scand. 112, 64–71. http://dx.doi.org/10.1111/ j.1600 0447.2005.00514.x. Lysaker, P.H., Clements, C.A., Plascak-Hallberg, C.D., Knipscheer, S.J., Wright, D.E., 2002. Insight and personal narratives of illness in schizophrenia. Psychiatry Interpers. Biol. Process 65, 197–206. http://dx.doi.org/10.1521/ psyc.65.3.197.20174. Lysaker, P.H., Dimaggio, G., 2014. Metacognitive capacities for reflection in schizophrenia: implications for developing treatments. Schizophr. Bull. 40, 487–491. http://dx.doi.org/10.1093/schbul/sbu038. Lysaker, P.H., Gumley, A., Brüne, M., Vanheule, S., Buck, K.D., Dimaggio, G., 2011a. Deficits in the ability to recognize one's own affects and those of others: associations with neurocognition, symptoms and sexual trauma among persons with schizophrenia spectrum disorders. Conscious. Cogn. 20, 1183–1192. http: //dx.doi.org/10.1016/j.concog.2010.12.018. Lysaker, P.H., Gumley, A., Dimaggio, G., 2011b. Metacognitive disturbances in persons with severe mental illness: theory, correlates with psychopathology and models of psychotherapy. Psychol. Psychother. 84, 1–8. http://dx.doi.org/ 10.1111/j.2044  8341.2010.02007.x. Lysaker, P.H., Gumley, A., Luedtke, B., Buck, K.D., Ringer, J.M., Olesek, K., Kukla, M., Leonhardt, B.L., Popolo, R., Dimaggio, G., 2013. Social cognition and metacognition in schizophrenia: evidence of their independence and linkage with outcomes. Acta Psychiatr. Scand. 127, 239–247. http://dx.doi.org/10.1111/ acps.12012. Lysaker, P.H., Hasson-Ohayon, I., 2014. Metacognition in schizophrenia: introduction to the special issue. Isr. J. Psychiatry Relat. Sci. 51, 4–7. Lysaker, P.H., Vohs, J., Hamm, J.A., Kukla, M., Minor, K.S., de Jong, S., van Donkersgoed, R., Pijnenborg, M.H.M., Kent, J.S., Matthews, S.C., Ringer, J.M., Leonhardt,

B.L., Francis, M.M., Buck, K.D., Dimaggio, G., 2014. Deficits in metacognitive capacity distinguish patients with schizophrenia from those with prolonged medical adversity. J. Psychiatr. Res. 55, 126–132. http://dx.doi.org/10.1016/j. jpsychires.2014.04.011. MacBeth, A., Gumley, A., 2008. Premorbid adjustment, symptom development and quality of life in first episode psychosis: a systematic review and critical reappraisal. Acta Psychiatr. Scand. 117, 85–99, doi:ACP1134 [pii];10.1111/ j.1600  0447.2007.01134.x [doi]. Macbeth, A., Gumley, A., Schwannauer, M., Carcione, A., Fisher, R., McLeod, H.J., Dimaggio, G., 2014. Metacognition, symptoms and premorbid functioning in a first episode psychosis sample. Compr. Psychiatry 55, 268–273. http://dx.doi. org/10.1016/j.comppsych.2013.08.027. Marshall, M., Lewis, S., Lockwood, A., Drake, R., Jones, P., Croudace, T., 2005. Association between duration of untreated psychosis and outcome in cohorts of first-episode patients: a systematic review. Arch. Gen. Psychiatry 62, 975–983. http://dx.doi.org/10.1001/archpsyc.62.9.975. McGuffin, P., Farmer, A., Harvey, I., 1991. A polydiagnostic application of operational criteria in studies of psychotic illness. development and reliability of the OPCRIT system. Arch. Gen. Psychiatry 48, 764–770. McLeod, H.J., Gumley, A.I., Macbeth, A., Schwannauer, M., Lysaker, P.H., 2014. Metacognitive functioning predicts positive and negative symptoms over 12 months in first episode psychosis. J. Psychiatr. Res. 54, 109–115. http://dx.doi. org/10.1016/j.jpsychires.2014.03.018. Melle, I., Johannesen, J.O., Friis, S., Haahr, U., Joa, I., Larsen, T.K., Opjordsmoen, S., Rund, B.R., Simonsen, E., Vaglum, P., McGlashan, T., 2006. Early detection of the first episode of schizophrenia and suicidal behavior. Am. J. Psychiatry 163, 800–804. http://dx.doi.org/10.1176/appi.ajp.163.5.800. Nicolò, G., Dimaggio, G., Popolo, R., Carcione, A., Procacci, M., Hamm, J., Buck, K.D., Pompili, E., Buccione, I., Lagrotteria, B., Lysaker, P.H., 2012. Associations of metacognition with symptoms, insight, and Neurocognition in clinically stable outpatients with schizophrenia. J. Nerv. Ment. Dis. 200, 644–647. http://dx.doi. org/10.1097/NMD.0b013e31825bfb10. Ontai, L.L., Thompson, R.A., 2008. Attachment, parent–child discourse and theoryof-mind development. Soc. Dev. 17, 47–60. http://dx.doi.org/10.1111/ j.1467 9507.2007.00414.x. Penn, D.L., Sanna, L.J., Roberts, D.L., 2008. Social cognition in schizophrenia: an overview. Schizophr. Bull. 34, 408–411. http://dx.doi.org/10.1093/schbul/sbn014. Perkins, D.O., Gu, H., Boteva, K., Lieberman, J.A., 2005. Relationship between duration of untreated psychosis and outcome in first-episode schizophrenia: a critical review and meta-analysis. Am. J. Psychiatry 162, , 1785–1804, doi:162/ 10/1785 [pii];10.1176/appi.ajp.162.10.1785 [doi]. Petersen, L., Jeppesen, P., Thorup, A., Abel, M.-B., Øhlenschlæger, J., Christensen, T.Ø., Krarup, G., Jørgensen, P., Nordentoft, M., 2005. A randomised multicentre trial of integrated versus standard treatment for patients with a first episode of psychotic illness. Br. Med. J. 331, 602. http://dx.doi.org/10.1136/ bmj.38565.415000.E01. Rabin, S.J., Hasson-Ohayon, I., Avidan, M., Rozencwaig, S., Shalev, H., Kravetz, S., 2014. Metacognition in schizophrenia and schizotypy: relation to symptoms of schizophrenia, traits of schizotypy and social quality of life. Isr. J. Psychiatry Relat. Sci. 51, 44–53. Ramsay, C.E., Flanagan, P., Gantt, S., Broussard, B., Compton, M.T., 2011. Clinical correlates of maltreatment and traumatic experiences in childhood and adolescence among predominantly African American, socially disadvantaged, hospitalized, first-episode psychosis patients. Psychiatry Res. 188, 343–349, doi:S0165  1781(11)00385  4 [pii];10.1016/j.psychres.2011.05.019 [doi]. Salvatore, G., Dimaggio, G., Lysaker, P.H., 2007. An intersubjective perspective on negative symptoms of schizophrenia: implications of simulation theory. Cognit. Neuropsychiatry 12, 144–164. http://dx.doi.org/10.1080/13546800600819921. Salvatore, G., Dimaggio, G., Popolo, R., Lysaker, P.H., 2008. Deficits in mindreading in stressful contexts and their relationships to social withdrawal in schizophrenia. Bull. Menn. Clin. 72, 191–209. http://dx.doi.org/10.1521/bumc.2008.72.3.191. Semerari, A., Carcione, A., Dimaggio, G., Falcone, M., Nicolo, G., Procacci, M., Alleva, G., 2003. How to evaluate metacognitive functioning in psychotherapy? the Metacognition assessment scale and its applications. Clin. Psychol. Psychother. 10, 238–261. Sheehan, D., Janavs, J., Harnett-Sheehan, K., Sheehan, M., Gray, C., 2008. Mini International Neuropsychiatric Interview (MINI). Danish Version 6.0.0. Sidebotham, P., Heron, J., 2006. Child maltreatment in the “children of the nineties”: a cohort study of risk factors. Child. Abus. Negl. 30, 497–522, doi: S0145  2134(06)00103  7 [pii];10.1016/j.chiabu.2005.11.005 [doi]. Simonsen, E., Friis, S., Opjordsmoen, S., Mortensen, E.L., Haahr, U., Melle, I., Joa, I., Johannessen, J.O., Larsen, T.K., Rossberg, J.I., Rund, B.R., Vaglum, P., McGlashan, T.H., 2010. Early identification of non-remission in first-episode psychosis in a two-year outcome study [references]. Acta Psychiatr. Scand., 375–383. http: //dx.doi.org/10.1111/j.1600  0447.2010.01598.x. Sitko, K., Bentall, R.P., Shevlin, M., O'Sullivan, N., Sellwood, W., 2014. Associations between specific psychotic symptoms and specific childhood adversities are mediated by attachment styles: an analysis of the national Comorbidity survey. Psychiatry Res. 217, 202–209. http://dx.doi.org/10.1016/j.psychres.2014.03.019. Smith, N., Lam, D., Bifulco, A., Checkley, S., 2002. Childhood experience of care and abuse questionnaire (CECA.Q). validation of a screening instrument for childhood adversity in clinical populations. Soc. Psychiatry Psychiatr. Epidemiol. 37, 572–579. http://dx.doi.org/10.1007/s00127  002 0589  9 [doi]. Sprong, M., Schothorst, P., Vos, E., Hox, J., van, E.H., 2007. Theory of mind in schizophrenia: meta-analysis. Br. J. Psychiatry 191, 5–13, doi:191/1/5 [pii];10.1192/ bjp.bp.107.035899 [doi].

A.M. Trauelsen et al. / Psychiatry Research 238 (2016) 14–23

Steinley, D., 2006. K-means clustering: a half-century synthesis. Br. J. Math. Stat. Psychol. 59, 1–34. http://dx.doi.org/10.1348/000711005  48266. Tinsley, H.E.A., Brown, S.D., 2000. Handbook of Applied Multivariate Statistics and Mathematical Modeling, 1 edition. Academic Press, San Diego. Trauelsen, A.M., Bendall, S., Jansen, J.E., Nielsen, H.-G.L., Pedersen, M.B., Trier, C.H., Haahr, U.H., Simonsen, E., 2015. Childhood adversity specificity and dose–response effect in non-affective first-episode psychosis. Schizophr. Res. . http: //dx.doi.org/10.1016/j.schres.2015.03.014 Trotta, A., Murray, R.M., Fisher, H.L., 2015. The impact of childhood adversity on the persistence of psychotic symptoms: a systematic review and meta-analysis. Psychol. MEd. FirstView, 1–18. http://dx.doi.org/10.1017/S0033291715000574. Ucok, A., Bikmaz, S., 2007. The effects of childhood trauma in patients with firstepisode schizophrenia. Acta Psychiatr. Scand. 116, 371–377, doi:ACP1079 [pii];10.1111/j.1600  0447.2007.01079.x [doi]. van der Gaag, M., Hoffman, T., Remijsen, M., Hijman, R., de Haan, L., van Meijel, B., van Harten, P.N., Valmaggia, L., de Hert, M., Cuijpers, A., Wiersma, D., 2006. The five-factor model of the positive and negative syndrome scale II: a ten-fold cross-validation of a revised model. Schizophr. Res. 85, 280–287. http://dx.doi. org/10.1016/j.schres.2006.03.021. van Mastrigt, S., Addington, J., 2002. Assessment of premorbid function in firstepisode schizophrenia: modifications to the Premorbid adjustment scale. J. Psychiatry Neurosci. JPN 27, 92–101.

23

van Nierop, M., Lataster, T., Smeets, F., Gunther, N., van Zelst, C., de Graaf, R., ten Have, M., van Dorsselaer, S., Bak, M., Myin-Germeys, I., Viechtbauer, W., van Os, J., van Winkel, R., 2014. Psychopathological mechanisms linking childhood traumatic experiences to risk of psychotic symptoms: analysis of a large. Represent. Popul.-based Sample Schizophr. Bull. 40 (Suppl 2), S123–S130. http: //dx.doi.org/10.1093/schbul/sbt150. Vohs, J.L., Lysaker, P.H., Francis, M.M., Hamm, J., Buck, K.D., Olesek, K., Outcalt, J., Dimaggio, G., Leonhardt, B., Liffick, E., Mehdiyoun, N., Breier, A., 2014. Metacognition, social cognition, and symptoms in patients with first episode and prolonged psychoses. Schizophr. Res. 153, 54–59. http://dx.doi.org/10.1016/j. schres.2014.01.012. White, C., Stirling, J., Hopkins, R., Morris, J., Montague, L., Tantam, D., Lewis, S., 2009. Predictors of 10-year outcome of first-episode psychosis. Psychol. Med. 39, 1447–1456. http://dx.doi.org/10.1017/S003329170800514X. Wing, J.K., Brown, G.W., 1970. Institutionalism and Schizophrenia: A Comparative Study of Three Mental Hospitals. Cambridge University Press, Cambridge, England, pp. 1960–1968. Wykes, T., Steel, C., Everitt, B., Tarrier, N., 2008. Cognitive behavior therapy for schizophrenia: effect sizes, clinical models, and methodological rigor. Schizophr. Bull. 34, 523–537. http://dx.doi.org/10.1093/schbul/sbm114. Yagmurlu, B., Berument, S.K., Celimli, S., 2005. The role of institution and home contexts in theory of mind development. J. Appl. Dev. Psychol. 26, 521–537.