- Email: [email protected]
Metacognitive beliefs in the at-risk mental state: A systematic review and meta-analysis
Behaviour Research and Therapy 90 (2017) 25e31
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Metacognitive beliefs in the at-risk mental state: A systematic review and meta-analysis Jack Cotter a, *, Alison R. Yung a, b, Rebekah Carney a, Richard J. Drake a a b
Division of Psychology and Mental Health, University of Manchester, UK Greater Manchester West Mental Health NHS Foundation Trust, Manchester, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 August 2016 Received in revised form 28 October 2016 Accepted 5 December 2016 Available online 8 December 2016
Dysfunctional metacognitive beliefs are common among people with psychosis. In this meta-analysis we examined whether these are also present in people meeting at-risk mental state (ARMS) criteria. We also explored the relationship between metacognitive beliefs and symptoms in the ARMS group. An electronic database search of Ovid MEDLINE, PsycINFO and Embase from inception until August 2016 was conducted using keyword search terms synonymous with ARMS and metacognition. Eligible studies were original research articles that examined metacognitive beliefs using the Metacognitions Questionnaire (MCQ) among people meeting ARMS criteria. Studies included in the meta-analyses also reported comparison MCQ data acquired from healthy controls, help-seeking individuals, or people with psychotic disorders. Eleven eligible studies were identified, reporting data from six unique ARMS samples. People with ARMS did not differ from those with established psychotic disorders on any MCQ subscale, but they reported significantly more dysfunctional metacognitive beliefs than healthy or helpseeking controls. Maladaptive metacognitive beliefs were associated with a range of symptoms in ARMS individuals, but evidence for associations with specific subthreshold psychotic phenomena was inconsistent. This evidence indicates how valuable assessment and treatment of dysfunctional metacognitive beliefs may be but suggests that specific aspects of methodology should be addressed. © 2016 Elsevier Ltd. All rights reserved.
Keywords: At-risk mental state Ultra-high risk Clinical high-risk Psychosis Metacognition Meta-analysis
1. Introduction Metacognition has been broadly defined as ‘thinking about thinking’ (Flavell, 1979), and includes the processes involved in the control, modification and interpretation of thought (Wells & Cartwright-Hatton, 2004). Certain metacognitive beliefs have been proposed to contribute to the development and maintenance of a range of mental health problems, including anxiety disorders (Ellis & Hudson, 2010; Hezel & McNally, (2015); Wells, 1995), alcohol abuse (Spada, Zandvoort, & Wells, 2007), eating disorders (Olstad, Solem, Hjemdal, & Hagen, 2015) and depression (Papageorgiou & Wells, 2003). Much of the research into the relationship between maladaptive metacognitive beliefs and psychopathology has been based on the self-regulatory executive function (S-REF) model proposed by Wells and Matthews (1996). This was originally developed to account for
* Corresponding author. Division of Psychology and Mental Health, University of Manchester, Jean McFarlane Building, Oxford Road, Manchester, M13 9PL, UK. E-mail address: [email protected] (J. Cotter). http://dx.doi.org/10.1016/j.brat.2016.12.004 0005-7967/© 2016 Elsevier Ltd. All rights reserved.
processes underlying affective disorders and refers to a cognitiveattentional syndrome in which heightened self-focused attention, reduced efficiency of cognitive functioning and repetitive rumination drive psychological dysfunction. Preoccupation with thoughts results in the depletion of resources needed to process information incompatible with dysfunctional beliefs. It also primes similar dysfunctional beliefs and makes the individual more sensitive to internal and external belief-congruent information. The most commonly used tools for assessing metacognitive beliefs are the Metacognitions Questionnaire (MCQ) (CartwrightHatton & Wells, 1997) and Metacognitions Questionnaire - short form (MCQ-30) (Wells & Cartwright-Hatton, 2004). Based on the SREF model, these self-report scales assess five dimensions of dysfunctional metacognitive beliefs originally derived using factor analyses; (1) ‘positive beliefs about worry’, which includes items suggesting worrying is beneficial for avoiding problems, remaining organised and helping one to cope; (2) ‘negative beliefs about uncontrollability of thoughts and corresponding danger’, which includes items emphasising the importance of controlling one's thoughts and potential mental and physical dangers associated
26
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
with not doing so; (3) ‘cognitive confidence’, which includes items concerned with perceived lack of self-confidence in one's memory and attention; (4) ‘negative beliefs about thoughts in general’, which is based around themes of superstition and punishment and includes items relating to the potential outcome of thoughts and feelings of responsibility for preventing those outcomes; (5) ‘cognitive self-consciousness’, which includes items reflecting one's tendency to be aware of and monitor one's thinking. Participants score individual items on a 4-point Likert scale based on the strength of their agreement with each statement. Relevant items are then summed to provide subscale scores for each of the five factors, with higher scores indicating more dysfunctional beliefs. High levels of dysfunctional metacognitive beliefs are reported among people with psychotic disorders (Sellers et al., 2016). These have been proposed to play a potential role in the onset and persistence of psychotic symptoms such as hallucinations and delusions (Morrison, 2001; Morrison, Wells, & Nothard, 2000; 2011). Positive beliefs about psychotic symptoms (for example that suspiciousness is good and keeps an individual safe) are argued to contribute to more frequent and severe symptoms, whereas negative beliefs about these thoughts (such as that they are uncontrollable or dangerous) are posited to lead to distress (Morrison, 2001; Morrison et al., 2015). Over the past two decades, criteria have been developed to identify individuals vulnerable to developing a psychotic disorder (Miller et al., 2002; Yung et al., 1996; 2003). These have been referred to as the prodromal, ultra-high risk (UHR), clinical highrisk (CHR) and at-risk mental state (ARMS) criteria (Fusar-Poli et al., 2013). Recent estimates suggest approximately 36% of this group will go on to develop a psychotic disorder over the following 3 years (Fusar-Poli et al., 2012a), though people continue to be at risk of transition upwards of ten years after initially presenting to clinical services (Nelson et al., 2013). In addition, young people with ARMS frequently present with, or go on to develop, high rates of mood and anxiety disorders (Addington et al., 2011; Fusar-Poli, Nelson, Valmaggia, Yung, & McGuire, 2014; Lin et al., 2015). Maladaptive metacognitive beliefs are therefore a potentially relevant target for clinical intervention for a range of mental health problems in this population. However, no reviews to our knowledge have examined metacognitive dysfunction in the ARMS group. Reducing both psychiatric symptom severity and associated distress may ultimately lead to reduced vulnerability to both psychotic and non-psychotic clinical outcomes. The aim of this review was to examine whether young people with ARMS report more maladaptive metacognitive beliefs compared with healthy controls, help-seeking individuals who do not meet ARMS criteria, and people diagnosed with psychotic disorders. We also sought to examine the relationship between metacognitive beliefs and clinical symptoms in the ARMS group. 2. Method This review was conducted in line with the PRISMA guidelines for reporting systematic reviews and meta-analyses (Moher, Liberati, Tetzlaff, Altman, & PRISMAGroup, 2009). 2.1. Eligibility criteria Studies eligible for inclusion in the narrative synthesis were original research articles that examined metacognitive beliefs using the MCQ or MCQ-30 among people meeting ARMS criteria (FusarPoli et al., 2013). In order to be included in the meta-analyses, studies also needed to report additional MCQ comparison data acquired from healthy controls, help-seeking individuals who did not meet ARMS criteria or people diagnosed with a full-threshold
psychotic disorder. Studies that included only subjects at genetic risk who had not met formal ARMS criteria, case studies and review articles were ineligible. No language restrictions were placed on articles for inclusion. 2.2. Search strategy On 1st August 2016 we conducted an electronic database search of Ovid MEDLINE, PsycINFO and Embase (from inception) using the following keyword search terms: “metacogniti*” and “at risk mental state” or “ultra high risk” or “UHR” or “clinical high risk” or “CHR” or “prodrom*” and “psychosis” or “psychotic” or “schizophrenia”. In addition, a basic search of Google Scholar was conducted, recent conference abstracts were screened, authors were contacted for unpublished data and the reference lists of retrieved articles were also reviewed to identify any additional eligible studies. 2.3. Study selection and data extraction Two reviewers (J.C. and R.C.) independently screened articles for eligibility. For all eligible studies, a data extraction spreadsheet was used to record: (1) study characteristics (year of publication, country where the work was performed); (2) sample demographics (sample size, gender composition, mean age); (3) the screening instrument used to assess ARMS status; (4) metacognitive data (MCQ measure used, mean sample scores for each subscale); (5) summary of study findings. 2.4. Statistical analyses Meta-analyses were performed to examine group differences in metacognitive beliefs based on the five factors derived from the MCQ. Separate analyses were performed to examine differences between (1) ARMS and healthy controls; (2) ARMS and people with psychotic disorders; (3) ARMS and help-seeking individuals who did not meet ARMS criteria. Data analyses were performed using Comprehensive Meta-Analysis version 3.0 (Borenstein, Hedges, Higgins, & Rothstein, 2007). Standardised mean differences (effect sizes) were calculated for each of the five MCQ subscales using Hedges’ g. A random-effects model was applied throughout, better accounting for observed heterogeneity. Estimates are more conservative but such models perform better than fixed-effect approaches (Brockwell & Gordon, 2001). Heterogeneity was quantified using the Q-test and I2 statistic. 3. Results 3.1. Eligible studies The study selection process is summarised in Fig. 1. We identified eleven papers eligible for inclusion in the narrative synthesis, collectively reporting data obtained from six unique ARMS samples (due to multiple publication) (Table 1). The meta-analyses included data from each of these six samples: metacognitive beliefs in five ARMS samples were compared to healthy controls (Brett, Johns, Peters, & McGuire, 2009; Leicester, 2013; Morrison, French, & Wells, 2007, 2006; Taylor, 2010; Welsh, Cartwright-Hatton, Wells, Snow, & Tiffin, 2014), in three samples were compared to people diagnosed with a psychotic disorder (Brett et al., 2009; Morrison et al., 2007, 2006; Taylor, 2010), and two to help-seeking individuals who did not meet ARMS criteria (Barbato et al., 2014; Taylor, 2010). Metacognitive data acquired at baseline from the Early Detection and Intervention Evaluation (EDIE) trial was used in a series of publications. In the meta-analyses we included data from
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
27
Fig. 1. PRISMA flow diagram.
the largest ARMS sample and largest healthy control and psychosis comparator samples reported across these publications (Morrison et al., 2006; 2007) in order to maximise power to detect group differences. One study reported data for ARMS individuals with and without hallucinatory experiences separately (Leicester, 2013). As both groups had distinct clinical profiles, they were both included and treated as separate data points in the meta-analysis.
subscale scores in people with ARMS relative to people with psychotic disorders are reported in Table 3. There were no significant differences between people with ARMS and those with psychotic disorders on any of the five MCQ subscales (all p 0.198) (Supplementary Figures 6-10). Moderate-to-high levels of statistical heterogeneity were observed between studies included in each of these estimates.
3.1.1. Metacognitive beliefs among people with ARMS relative to healthy controls Effect sizes, sample size and heterogeneity statistics for MCQ subscale scores in people with ARMS relative to healthy controls are reported in Table 2. ARMS individuals scored significantly higher on all measures of dysfunctional metacognitive beliefs compared to healthy controls (all p < 0.001), except for the ‘positive beliefs about worry’ subscale which met only trend-level significance (p ¼ 0.053) (Supplementary Figures 1-5). There was evidence of low-tomoderate statistical heterogeneity for the ‘cognitive confidence’, ‘negative beliefs about thoughts in general’ and ‘cognitive selfconsciousness’ subscale estimates.
3.1.3. Metacognitive beliefs among people with ARMS relative to help-seeking controls Effect sizes, sample size and heterogeneity statistics for MCQ subscale scores in people with ARMS relative to help-seeking controls are reported in Table 4. People with ARMS scored significantly higher than help-seeking controls on the ‘negative beliefs about uncontrollability’, ‘cognitive confidence’ and ‘negative beliefs about thoughts in general’ subscales of the MCQ (all p 0.004). There were no significant differences between the groups for the ‘positive beliefs about worry’ (p ¼ 0.147) and ‘cognitive self-consciousness’ subscales (p ¼ 0.291) (Supplementary Figures 11-15). There was no evidence of statistical heterogeneity for any of these subscale scores.
3.1.2. Metacognitive beliefs among people with ARMS relative to people with psychotic disorders Effect sizes, sample size and heterogeneity statistics for MCQ
3.1.4. Metacognitive beliefs and symptoms in the ARMS group Due to variation in both the use and reporting of symptom
28
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
Table 1 Sample characteristics. Study þ country
Barbato et al. (2014) Canada, USA Brett et al. (2009) - UK Leicester (2013) Australia (HG) Leicester (2013) Australia (NHG) Welsh et al. (2014) - UK EDIE trial sample Barkus et al. (2010) - UK Morrison et al. (2002) - UK Morrison et al. (2006) - UK Morrison et al. (2007) - UK EDIE-2 trial sample Morrison et al. (2015) - UK Palmier-Claus et al. (2013) - UKa Taylor (2010) - UK
Metacognitive measure
ARMS group
Healthy control group
Psychotic disorder group
Help-seeking group
Screening instrument
N (M/F)
Age mean (SD)
N (M/F)
Age mean (SD)
N (M/F)
Age mean (SD)
N (M/F)
Age mean (SD)
MCQ
SIPS
153 (88/65)
19.7 (4.2)
e
e
e
e
68 (40/28)
19.7 (4.2)
MCQ MCQ-30
CAARMS CAARMS
32 (21/11) 38 (15/23)
24.3 (3.6) 19.24 (2.84)
32 (18/14) 35 (19/16)
27.7 (7.5) 19.20 (2.98)
27 (12/15) e
32.4 (11.2) e
e e
e e
MCQ-30
CAARMS
32 (18/14)
19.03 (2.40)
35 (19/16)
19.20 (2.98)
e
e
e
e
MCQ-30
CAARMS
31 (15/16)
15.8 (1.4)
76 (29/47)
14.82 (1.1)
e
e
e
e
MCQ MCQ
PANSS PANSS
58 (40/18) 31 (22/9)
22.1 (4.4) 23.2 (4.78)
95 (37/58) 50 (8/42)
22.79 (6.93) 21.7 (7.71)
e e
e e
e e
e e
MCQ
PANSS
58 (40/18)
22.1 (4.4)
56 (19/37)
22.9 (5.4)
e
e
e
e
MCQ
PANSS
43 (31/12)
22.6 (4.7)
188 (144/44)
27.6 (11.1)
73 (53/20)
41.2 (10.3)
e
e
MCQ-30
CAARMS
117 (71/46)
20.3
e
e
e
e
e
e
MCQ-30
CAARMS
27 (14/13)
22.6 (4.4)
e
e
e
e
e
e
MCQ-30
CAARMS
113 (67/46)
20.4 (4.3)
30 (8/22)
22.8 (3.7)
20 (14/5)
22.4 (5.4)
28 (23/5)
21.3 (3.4)
Abbreviations: ARMS: At-risk mental state; CAARMS: Comprehensive Assessment of At-Risk Mental States; EDIE: Early Detection and Intervention Evaluation; MCQ: Metacognitions Questionnaire; MCQ-30: Metacognitions Questionnaire short form; PANSS: Positive and Negative Syndrome Scale; SIPS: Structured Interview for Prodromal Syndromes; HG: ARMS hallucinating group; NHG: ARMS non-hallucinating group. a Sample partially overlaps with the EDIE-2 trial sample.
measures across studies, we were unable to analyse statistically the relationship between metacognitive beliefs and symptom severity in the ARMS group. Instead we have provided a detailed narrative synthesis of the literature to date. Higher scores on ‘cognitive self-consciousness’, the ‘negative beliefs about uncontrollability’ and ‘negative beliefs about thoughts
in general’ subscales were significantly associated with more severe affective symptoms (Morrison et al., 2015, 2006; Palmier-Claus, Dunn, Taylor, Morrison, & Lewis, 2013; Taylor, 2010), more severe manic symptoms (Welsh et al., 2014), elevated levels of perceived stress (Morrison et al., 2006), and poor psychosocial functioning (Welsh et al., 2014). Maladaptive metacognitive beliefs have also
Table 2 Mean weighted effect sizes, sample sizes and heterogeneity statistics for MCQ subscales among ARMS individuals relative to healthy controls. MCQ subscale
k
N (ARMS)
N (Control)
Hedges' g
95% CI
Z
p
Q
I2
Positive beliefs about worry Negative beliefs about uncontrollability Cognitive confidence Negative beliefs about thoughts Cognitive self-consciousness
6 6 6 6 6
286 287 287 288 288
395 393 395 394 395
0.16 1.50 0.92 1.09 0.57
0.01 to 0.33 1.31 to 1.68 0.71 to 1.13 0.85 to 1.33 0.34 to 0.80
1.94 15.96 8.60 8.93 4.87
0.053 <0.001 <0.001 <0.001 <0.001
3.486 1.794 7.050 8.869 8.811
0% 0% 29% 44% 43%
Abbreviations: CI: confidence interval; k: number of comparisons; MCQ: Metacognitions Questionnaire; Q: homogeneity analysis.
Table 3 Mean weighted effect sizes, sample sizes and heterogeneity statistics for MCQ subscales among ARMS individuals relative to people with psychotic disorders. MCQ subscale
k
N (ARMS)
N (Control)
Hedges' g
95% CI
Positive beliefs about worry Negative beliefs about uncontrollability Cognitive confidence Negative beliefs about thoughts Cognitive self-consciousness
3 3 3 3 3
185 186 186 187 187
115 115 115 114 115
0.24 0.12 0.16 0.17 0.14
0.61 0.28 0.67 0.51 0.24
to to to to to
0.13 0.52 0.36 0.18 0.52
Z
p
Q
I2
1.29 0.58 0.60 0.95 0.73
0.198 0.563 0.548 0.343 0.463
3.812 4.563 7.399 3.402 4.049
48% 56% 73% 41% 51%
Abbreviations: CI: confidence interval; k: number of comparisons; MCQ: Metacognitions Questionnaire; Q: homogeneity analysis.
Table 4 Mean weighted effect sizes, sample sizes and heterogeneity statistics for MCQ subscales among ARMS individuals relative to help-seeking controls. MCQ subscale
k
N (ARMS)
N (Control)
Hedges' g
95% CI
Z
p
Q
I2
Positive beliefs about worry Negative beliefs about uncontrollability Cognitive confidence Negative beliefs about thoughts Cognitive self-consciousness
2 2 2 2 2
220 221 221 222 222
89 88 89 88 88
0.18 0.37 0.39 0.49 0.13
0.06 to 0.43 0.12 to 0.62 0.14 to 0.63 0.24 to 0.74 0.12 to 0.38
1.45 2.89 3.03 3.82 1.06
0.147 0.004 0.002 <0.001 0.291
0.252 0.660 0.010 0.728 0.277
0% 0% 0% 0% 0%
Abbreviations: CI: confidence interval; k: number of comparisons; MCQ: Metacognitions Questionnaire; Q: homogeneity analysis.
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
been significantly associated with sub-threshold positive psychotic symptoms, but there was no consistent link across studies between scores on particular metacognitive subscales and specific subthreshold psychotic symptoms (Brett et al., 2009; Leicester, 2013; Morrison et al., 2015, 2006; Palmier-Claus et al., 2013; Taylor, 2010; Welsh et al., 2014). Leicester (2013) subdivided 70 young people with ARMS into those who had and had not reported auditory hallucinatory experiences. Both groups reported significantly more maladaptive metacognitive beliefs than healthy controls but there were no significant differences between the two ARMS groups on any of the MCQ-30 subscales. In contrast, Welsh et al. (2014) reported significant positive correlations between ‘negative beliefs about the uncontrollability of thoughts’ and both more intense and more distressing perceptual abnormalities among 31 adolescents with ARMS. They also reported a weak correlation between ‘positive beliefs about worry’ and higher frequency of unusual thought content. Morrison et al. (2015) examined the cross-sectional association between persecutory ideation and scores on the ‘cognitive selfconsciousness’, ‘negative beliefs about uncontrollability’ and ‘negative beliefs about thoughts in general’ subscales of the MCQ30. They reported that these were positively correlated with both perceived persecution and associated deservedness. In multivariate analyses controlling for a range of mood, anxiety and demographic variables, ‘negative beliefs about thoughts in general’ remained significantly associated with perceived deservedness of persecution, but not with self-reported persecution itself. In the same sample, Taylor (2010) reported that ‘negative beliefs about uncontrollability’ and ‘negative beliefs about thoughts in general’ also appeared to mediate many of the associations between both the frequency and severity of persecutory ideation and unwanted thoughts, and their associations with depression, social anxiety and symptom-related distress. Palmier-Claus et al. (2013) examined the temporal relationship between metacognitive beliefs and both psychotic and mood symptoms over 6 days in 27 ARMS individuals. They used an experience sampling approach in which participants completed a diary when prompted by an electronic wristwatch. They reported that heightened ‘cognitive self-consciousness’ preceded hallucinatory experiences but only among individuals who scored highly on the ‘negative beliefs about thoughts’ subscale (Palmier-Claus et al., 2013). Delusional thoughts were unrelated to metacognitive beliefs. They also reported that total scores on the MCQ-30 moderated the association between stressors and affective symptoms, but not psychotic symptoms. In the only long-term follow-up study conducted in this area to date, Barbato et al. (2014) reported that MCQ subscale scores and both positive and negative psychotic symptoms significantly improved at 6-month follow-up among a large ARMS cohort. They also reported preliminary evidence for a link between metacognitive dysfunction and conversion to psychotic disorder. Significantly higher scores at baseline on the ‘positive beliefs about worry’ and the two ‘negative belief’ subscales of the MCQ were reported among ARMS and help-seeking controls who subsequently developed a psychotic disorder compared to those who did not over the following two-year period. 4. Discussion 4.1. Summary of results These results suggest that young people with ARMS were more likely to view their thoughts as dangerous and uncontrollable and had greater concerns over their memory and
29
attention compared to both healthy and help-seeking controls. They were also more likely to monitor and be aware of their thoughts compared to healthy controls. However, people with ARMS did not significantly differ from people with full-threshold psychotic disorders on any of the MCQ subscales. Overall, these findings were consistent with the S-REF model which suggests that dysfunctional metacognitive beliefs may contribute to psychological dysfunction (Wells & Matthews, 1996), though the relationship between metacognitive beliefs and the development and persistence of specific sub-threshold psychotic symptoms remains unclear. 4.2. The relationship between metacognitive beliefs, symptoms and cognition Metacognitive beliefs were associated with a range of symptoms. However, small sample sizes, cross-sectional study designs and heterogeneity in both measures and reporting makes confident inferences about the significance and nature of these associations difficult. Though significant associations between metacognitive subscales and both perceptual abnormalities and persecutory ideation were reported, there was no consistent pattern between the presence of maladaptive metacognitive beliefs and the intensity or frequency of specific sub-threshold psychotic phenomena. Instead dysfunctional metacognitive beliefs appear to be more strongly associated with mood and anxiety symptoms (Brett et al., 2009; Palmier-Claus et al., 2013; Welsh et al., 2014). Other reviews have also reported limited evidence that maladaptive metacognitive beliefs play a causal role in the development and persistence of psychotic symptoms (Varese & Bentall, 2011). It is unclear why metacognitive beliefs’ associations with psychotic phenomena are inconsistent when existing theories predict they should be specific and consistent (Morrison, 2001; Morrison et al., 2000). Notably, the MCQ was originally developed to assess metacognitive beliefs associated with worry. In future, modified versions of the MCQ or symptom-specific measures enquiring directly about metacognitive appraisals of hallucinations or delusions may be more suited to examining the development and course of psychotic symptoms. Young people with ARMS also exhibit significant deficits in cognition, particularly in memory and attention (Fusar-Poli et al., 2012b). This is likely to be reflected in the poorer perceived ‘cognitive confidence’ among the ARMS group relative to both the healthy and help-seeking controls. However, it is unclear whether these beliefs accurately reflect cognitive performance or whether they overestimate deficits in these domains. Further research is also needed to establish whether maladaptive metacognitive beliefs are predictors of social and occupational functioning in this group (Cotter et al., 2014; 2015). 4.3. Strengths and limitations To our knowledge, this is the first meta-analytic review to examine metacognitive dysfunction in the ARMS group. Studies in this area have typically been small and potentially underpowered to detect between-group differences. As a result of the crosssectional design of most of these studies, it is also difficult to establish whether maladaptive metacognitive beliefs occur as a cause or consequence of psychiatric symptoms. Similarly, the stability of metacognitive beliefs and their relationship with transition to psychotic disorder in the ARMS group remains largely unexplored and poorly understood. Young people with ARMS are helpseeking and often in crisis at presentation. However, sub-threshold psychotic symptoms and other mental health issues may only be transitory for some people. Metacognitive beliefs may reflect state
30
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
rather than trait factors and arise as a result of concurrent clinical symptoms and distress rather than vulnerability to psychosis. There is evidence that both positive and negative psychotic symptoms and dysfunctional metacognitive beliefs improve over time in this group (Barbato et al., 2014). Further prospective, longitudinal research is needed to establish the stability of maladaptive metacognitive beliefs, as well as the direction of causality and the clinical implications of the relationship between metacognitive dysfunction and psychiatric symptoms. There is also currently a lack of studies examining differences between ARMS individuals and non-ARMS help-seeking controls. Metacognitive dysfunction may be primarily associated with mood and anxiety symptoms, providing an alternative explanation for the significant differences observed between the ARMS and healthy control groups. Only two studies to date have examined metacognitive beliefs among both help-seeking controls and ARMS samples. Both of these studies comprised relatively small helpseeking control groups. Though we found significant differences on three of the MCQ subscales between these groups, further research is needed comparing people with ARMS to those with only affective symptoms before more definitive conclusions can be drawn. Moderate-to-high levels of heterogeneity were observed in the ARMS compared to psychosis group MCQ subscale estimates. This may have been driven by variation in the age, gender composition and diagnoses of people in the psychosis groups across studies. Studies assessing both group differences in metacognitive beliefs and the association between metacognition and clinical symptoms have often failed to use multivariate analyses to adjust for potentially confounding factors. There is evidence that group differences in metacognitive beliefs as well as their association with positive psychotic symptoms became non-significant after controlling for anxiety and depressive symptoms (Brett et al., 2009; Morrison et al., 2015). This should be an important consideration for future research.
4.4. Clinical implications Assessing psychological factors in this help-seeking group which may contribute towards and exacerbate mental health issues is important. Cognitive behavioural assessments should therefore include examination of metacognitive beliefs. Strategies to help modify maladaptive metacognitive beliefs may also be a target for clinical intervention in this group. Metacognitive therapy involves identifying and challenging metacognitive beliefs (Wells, 2009), and has been shown to be effective for the treatment of both anxiety and depression (Normann, van Emmerik, & Morina, 2014). Early evidence suggests it may also be beneficial for improving psychotic symptoms in people with a diagnosis of schizophrenia (Hutton, Morrison, Wardle, & Wells, 2014; Morrison et al., 2014). It would therefore be useful to examine whether it has a similar impact on clinical symptoms in the ARMS group, given the range of co-morbid mental health problems frequently present among ARMS individuals (Addington et al., 2011; Fusar-Poli et al., 2014; Lin et al., 2015). Current predictors of transition to psychotic disorder in the ARMS population have been limited largely to the same range of clinical, demographic and neurocognitive variables. Evaluation of metacognition may potentially provide additional prognostic value over and above these factors and warrants further investigation. Barbato et al. (2014) have provided preliminary evidence that metacognitive dysfunction is a potential marker for conversion to psychotic disorder; however, additional longitudinal replication studies are required.
5. Conclusions Maladaptive metacognitive beliefs are common in young people meeting ARMS criteria. Further research is needed to clarify the relationship between metacognitive dysfunction and the development and persistence of both psychotic and non-psychotic clinical symptoms. Assessment and treatment aimed at alleviating metacognitive dysfunction may be useful for improving a range of clinical outcomes. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.brat.2016.12.004. References Addington, J., Cornblatt, B. A., Cadenhead, K. S., Cannon, T. D., McGlashan, T. H., Perkins, D. O., et al. (2011). At clinical high risk for psychosis: Outcome for nonconverters. American Journal of Psychiatry, 168(8), 800e805. Barbato, M., Penn, D. L., Perkins, D. O., Woods, S. W., Liu, L., & Addington, J. (2014). Metacognitive functioning in individuals at clinical high risk for psychosis. Behavioural and Cognitive Psychotherapy, 42(5), 526e534. Barkus, E., Stirling, J., French, P., Morrison, A., Bentall, R., & Lewis, S. (2010). Distress and metacognition in psychosis prone individuals: Comparing high schizotypy to the at-risk mental state. Journal of Nervous and Mental Disease, 198(2), 99e104. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2007). Comprehensive metaanalysis: A computer program for meta-analysis. Englewood, NJ: Biostat Inc. Brett, C. M., Johns, L. C., Peters, E. P., & McGuire, P. K. (2009). The role of metacognitive beliefs in determining the impact of anomalous experiences: A comparison of help-seeking and non-help-seeking groups of people experiencing psychotic-like anomalies. Psychological Medicine, 39(6), 939e950. Brockwell, S. E., & Gordon, I. R. (2001). A comparison of statistical methods for meta-analysis. Statistics in Medicine, 20(6), 825e840. Cartwright-Hatton, S., & Wells, A. (1997). Beliefs about worry and intrusions: The meta-cognitions questionnaire and its correlates. Journal of Anxiety Disorders, 11(3), 279e296. Cotter, J., Bartholomeusz, C., Papas, A., Allott, K., Nelson, B., Yung, A. R., et al. (2015). Examining the association between social cognition and functioning in individuals at ultra-high risk for psychosis. Australian and New Zealand Journal of Psychiatry (in press), http://dx.doi.org/10.1177/0004867415622691. Cotter, J., Drake, R. J., Bucci, S., Firth, J., Edge, D., & Yung, A. R. (2014). What drives poor functioning in the at-risk mental state? A systematic review. Schizophrenia Research, 159(2e3), 267e277. Ellis, D. M., & Hudson, J. L. (2010). The metacognitive model of generalized anxiety disorder in children and adolescents. Clinical Child and Family Psychology Review, 13(2), 151e163. Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive-developmental inquiry. American Psychologist, 34(10), 906e911. Fusar-Poli, P., Bonoldi, I., Yung, A. R., Borgwardt, S., Kempton, M. J., Valmaggia, L., et al. (2012a). Predicting psychosis: meta-analysis of transition outcomes in individuals at high clinical risk. Archives of General Psychiatry, 69(3), 220e229. € ssler, A., SchultzeFusar-Poli, P., Borgwardt, S., Bechdolf, A., Addington, J., Riecher-Ro Lutter, F., et al. (2013). The psychosis high-risk state: A comprehensive state-ofthe-art review. JAMA Psychiatry, 70(1), 107e120. Fusar-Poli, P., Deste, G., Smieskova, R., Barlati, S., Yung, A. R., Howes, O., et al. (2012b). Cognitive functioning in prodromal psychosis: A meta-analysis. Archives of General Psychiatry, 69(6), 562e571. Fusar-Poli, P., Nelson, B., Valmaggia, L., Yung, A. R., & McGuire, P. K. (2014). Comorbid depressive and anxiety disorders in 509 individuals with an at-risk mental state: Impact on psychopathology and transition to psychosis. Schizophrenia Bulletin, 40(1), 120e131. Hezel, D. M., & McNally, R. J. (2015). A theoretical review of cognitive biases and deficits in obsessiveecompulsive disorder. Biological Psychology, 121(B), 221e232. Hutton, P., Morrison, A. P., Wardle, M., & Wells, A. (2014). Metacognitive therapy in treatment-resistant psychosis: A multiple-baseline study. Behavioural and Cognitive Psychotherapy, 42(2), 166e185. Leicester, S. B. (2013). Cognitive processing associated with hallucinations for at risk mental state (Unpublished doctoral thesis). Australia: Melbourne School of Psychological Sciences, Faculty of Medicine, Dentistry & Health Sciences, The University of Melbourne. Lin, A., Wood, S. J., Nelson, B., Beavan, A., McGorry, P., & Yung, A. R. (2015). Outcomes of nontransitioned cases in a sample at ultra-high risk for psychosis. American Journal of Psychiatry, 172(3), 249e258. Miller, T. J., McGlashan, T. H., Rosen, J. L., Somjee, L., Markovich, P. J., Stein, K., et al. (2002). Prospective diagnosis of the initial prodrome for schizophrenia based on the structured Interview for prodromal syndromes: Preliminary evidence of interrater reliability and predictive validity. American Journal of Psychiatry,
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31 159(5), 863e865. Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & PRISMA Group. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ, 339, b2535. Morrison, A. P. (2001). The interpretation of intrusions in psychosis: An integrative cognitive approach to hallucinations and delusions. Behavioural and Cognitive Psychotherapy, 29(3), 257e276. Morrison, A. P., Bentall, R. P., French, P., Walford, L., Kilcommons, A., Knight, A., et al. (2002). Randomised controlled trial of early detection and cognitive therapy for preventing transition to psychosis in high-risk individuals. Study design and interim analysis of transition rate and psychological risk factors. British Journal of Psychiatry, 43, s78es84. Morrison, A. P., French, P., Lewis, S. W., Roberts, M., Raja, S., Neil, S. T., et al. (2006). Psychological factors in people at ultra-high risk of psychosis: Comparisons with non-patients and associations with symptoms. Psychological Medicine, 36(10), 1395e1404. Morrison, A. P., French, P., & Wells, A. (2007). Metacognitive beliefs across the continuum of psychosis: Comparisons between patients with psychotic disorders, patients at ultra-high risk and non-patients. Behaviour Research and Therapy, 45(9), 2241e2246. Morrison, A. P., Gumley, A. I., Ashcroft, K., Manousos, I. R., White, R., Gillan, K., et al. (2011). Metacognition and persecutory delusions: Tests of a metacognitive model in a clinical population and comparisons with non-patients. British Journal of Clinical Psychology, 50(3), 223e233. Morrison, A. P., Pyle, M., Chapman, N., French, P., Parker, S. K., & Wells, A. (2014). Metacognitive therapy in people with a schizophrenia spectrum diagnosis and medication resistant symptoms: A feasibility study. Journal of Behavior Therapy and Experimental Psychiatry, 45(2), 280e284. Morrison, A. P., Shryane, N., Fowler, D., Birchwood, M., Gumley, A. I., Taylor, H. E., et al. (2015). Negative cognition, affect, metacognition and dimensions of paranoia in people at ultra-high risk of psychosis: A multi-level modelling analysis. Psychological Medicine, 45(12), 2675e2684. Morrison, A. P., Wells, A., & Nothard, S. (2000). Cognitive factors in predisposition to auditory and visual hallucinations. British Journal of Clinical Psychology, 39(1), 67e78. Nelson, B., Yuen, H. P., Wood, S. J., Lin, A., Spiliotacopoulos, D., Bruxner, A., et al. (2013). Long-term follow-up of a group at ultra high risk (“prodromal”) for psychosis: The PACE 400 study. JAMA Psychiatry, 70(8), 793e802. Normann, N., van Emmerik, A. A., & Morina, N. (2014). The efficacy of metacognitive therapy for anxiety and depression: A meta-analytic review. Depression and Anxiety, 31(5), 402e411.
31
Olstad, S., Solem, S., Hjemdal, O., & Hagen, R. (2015). Metacognition in eating disorders: Comparison of women with eating disorders, self-reported history of eating disorders or psychiatric problems, and healthy controls. Eating Behaviors, 16, 17e22. Palmier-Claus, J. E., Dunn, G., Taylor, H., Morrison, A. P., & Lewis, S. W. (2013). Cognitive-self consciousness and metacognitive beliefs: Stress sensitization in individuals at ultra-high risk of developing psychosis. British Journal of Clinical Psychology, 52(1), 26e41. Papageorgiou, C., & Wells, A. (2003). An empirical test of a clinical metacognitive model of rumination and depression. Cognitive Therapy and Research, 27(3), 261e273. Sellers, R., Varese, F., Wells, A., & Morrison, A. P. (2016). A meta-analysis of metacognitive beliefs as implicated in the self-regulatory executive function model in clinical psychosis. Schizophrenia Research, 179, 75e84. Spada, M. M., Zandvoort, M., & Wells, A. (2007). Metacognitions in problem drinkers. Cognitive Therapy and Research, 31(5), 709e716. Taylor, H. E. (2010). Cognitive processes across the continuum of psychosis (Unpublished doctoral thesis). UK: School of Psychological Sciences, Faculty of Medical and Human Sciences, The University of Manchester. Varese, F., & Bentall, R. P. (2011). The metacognitive beliefs account of hallucinatory experiences: A literature review and meta-analysis. Clinical Psychology Review, 31(5), 850e864. Wells, A. (1995). Meta-cognition and worry: A cognitive model of generalized anxiety disorder. Behavioural and Cognitive Psychotherapy, 23(3), 301e320. Wells, A. (2009). Metacognitive therapy for anxiety and depression. New York: Guilford Press. Wells, A., & Cartwright-Hatton, S. (2004). A short form of the metacognitions questionnaire: Properties of the MCQ-30. Behaviour Research and Therapy, 42(4), 385e396. Wells, A., & Matthews, G. (1996). Modelling cognition in emotional disorder: The SREF model. Behaviour Research and Therapy, 34(11e12), 881e888. Welsh, P., Cartwright-Hatton, S., Wells, A., Snow, L., & Tiffin, P. A. (2014). Metacognitive beliefs in adolescents with an at-risk mental state for psychosis. Early Intervention in Psychiatry, 8(1), 82e86. Yung, A. R., McGorry, P. D., McFarlane, C. A., Jackson, H. J., Patton, G. C., & Rakkar, A. (1996). Monitoring and care of young people at incipient risk of psychosis. Schizophrenia Bulletin, 22(2), 283e303. Yung, A. R., Phillips, L. J., Yuen, H. P., Francey, S. M., McFarlane, C. A., Hallgren, M., et al. (2003). Psychosis prediction: 12-month follow up of a high-risk (“prodromal”) group. Schizophrenia Research, 60(1), 21e32.
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Metacognitive beliefs in the at-risk mental state: A systematic review and meta-analysis Jack Cotter a, *, Alison R. Yung a, b, Rebekah Carney a, Richard J. Drake a a b
Division of Psychology and Mental Health, University of Manchester, UK Greater Manchester West Mental Health NHS Foundation Trust, Manchester, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 August 2016 Received in revised form 28 October 2016 Accepted 5 December 2016 Available online 8 December 2016
Dysfunctional metacognitive beliefs are common among people with psychosis. In this meta-analysis we examined whether these are also present in people meeting at-risk mental state (ARMS) criteria. We also explored the relationship between metacognitive beliefs and symptoms in the ARMS group. An electronic database search of Ovid MEDLINE, PsycINFO and Embase from inception until August 2016 was conducted using keyword search terms synonymous with ARMS and metacognition. Eligible studies were original research articles that examined metacognitive beliefs using the Metacognitions Questionnaire (MCQ) among people meeting ARMS criteria. Studies included in the meta-analyses also reported comparison MCQ data acquired from healthy controls, help-seeking individuals, or people with psychotic disorders. Eleven eligible studies were identified, reporting data from six unique ARMS samples. People with ARMS did not differ from those with established psychotic disorders on any MCQ subscale, but they reported significantly more dysfunctional metacognitive beliefs than healthy or helpseeking controls. Maladaptive metacognitive beliefs were associated with a range of symptoms in ARMS individuals, but evidence for associations with specific subthreshold psychotic phenomena was inconsistent. This evidence indicates how valuable assessment and treatment of dysfunctional metacognitive beliefs may be but suggests that specific aspects of methodology should be addressed. © 2016 Elsevier Ltd. All rights reserved.
Keywords: At-risk mental state Ultra-high risk Clinical high-risk Psychosis Metacognition Meta-analysis
1. Introduction Metacognition has been broadly defined as ‘thinking about thinking’ (Flavell, 1979), and includes the processes involved in the control, modification and interpretation of thought (Wells & Cartwright-Hatton, 2004). Certain metacognitive beliefs have been proposed to contribute to the development and maintenance of a range of mental health problems, including anxiety disorders (Ellis & Hudson, 2010; Hezel & McNally, (2015); Wells, 1995), alcohol abuse (Spada, Zandvoort, & Wells, 2007), eating disorders (Olstad, Solem, Hjemdal, & Hagen, 2015) and depression (Papageorgiou & Wells, 2003). Much of the research into the relationship between maladaptive metacognitive beliefs and psychopathology has been based on the self-regulatory executive function (S-REF) model proposed by Wells and Matthews (1996). This was originally developed to account for
* Corresponding author. Division of Psychology and Mental Health, University of Manchester, Jean McFarlane Building, Oxford Road, Manchester, M13 9PL, UK. E-mail address: [email protected] (J. Cotter). http://dx.doi.org/10.1016/j.brat.2016.12.004 0005-7967/© 2016 Elsevier Ltd. All rights reserved.
processes underlying affective disorders and refers to a cognitiveattentional syndrome in which heightened self-focused attention, reduced efficiency of cognitive functioning and repetitive rumination drive psychological dysfunction. Preoccupation with thoughts results in the depletion of resources needed to process information incompatible with dysfunctional beliefs. It also primes similar dysfunctional beliefs and makes the individual more sensitive to internal and external belief-congruent information. The most commonly used tools for assessing metacognitive beliefs are the Metacognitions Questionnaire (MCQ) (CartwrightHatton & Wells, 1997) and Metacognitions Questionnaire - short form (MCQ-30) (Wells & Cartwright-Hatton, 2004). Based on the SREF model, these self-report scales assess five dimensions of dysfunctional metacognitive beliefs originally derived using factor analyses; (1) ‘positive beliefs about worry’, which includes items suggesting worrying is beneficial for avoiding problems, remaining organised and helping one to cope; (2) ‘negative beliefs about uncontrollability of thoughts and corresponding danger’, which includes items emphasising the importance of controlling one's thoughts and potential mental and physical dangers associated
26
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
with not doing so; (3) ‘cognitive confidence’, which includes items concerned with perceived lack of self-confidence in one's memory and attention; (4) ‘negative beliefs about thoughts in general’, which is based around themes of superstition and punishment and includes items relating to the potential outcome of thoughts and feelings of responsibility for preventing those outcomes; (5) ‘cognitive self-consciousness’, which includes items reflecting one's tendency to be aware of and monitor one's thinking. Participants score individual items on a 4-point Likert scale based on the strength of their agreement with each statement. Relevant items are then summed to provide subscale scores for each of the five factors, with higher scores indicating more dysfunctional beliefs. High levels of dysfunctional metacognitive beliefs are reported among people with psychotic disorders (Sellers et al., 2016). These have been proposed to play a potential role in the onset and persistence of psychotic symptoms such as hallucinations and delusions (Morrison, 2001; Morrison, Wells, & Nothard, 2000; 2011). Positive beliefs about psychotic symptoms (for example that suspiciousness is good and keeps an individual safe) are argued to contribute to more frequent and severe symptoms, whereas negative beliefs about these thoughts (such as that they are uncontrollable or dangerous) are posited to lead to distress (Morrison, 2001; Morrison et al., 2015). Over the past two decades, criteria have been developed to identify individuals vulnerable to developing a psychotic disorder (Miller et al., 2002; Yung et al., 1996; 2003). These have been referred to as the prodromal, ultra-high risk (UHR), clinical highrisk (CHR) and at-risk mental state (ARMS) criteria (Fusar-Poli et al., 2013). Recent estimates suggest approximately 36% of this group will go on to develop a psychotic disorder over the following 3 years (Fusar-Poli et al., 2012a), though people continue to be at risk of transition upwards of ten years after initially presenting to clinical services (Nelson et al., 2013). In addition, young people with ARMS frequently present with, or go on to develop, high rates of mood and anxiety disorders (Addington et al., 2011; Fusar-Poli, Nelson, Valmaggia, Yung, & McGuire, 2014; Lin et al., 2015). Maladaptive metacognitive beliefs are therefore a potentially relevant target for clinical intervention for a range of mental health problems in this population. However, no reviews to our knowledge have examined metacognitive dysfunction in the ARMS group. Reducing both psychiatric symptom severity and associated distress may ultimately lead to reduced vulnerability to both psychotic and non-psychotic clinical outcomes. The aim of this review was to examine whether young people with ARMS report more maladaptive metacognitive beliefs compared with healthy controls, help-seeking individuals who do not meet ARMS criteria, and people diagnosed with psychotic disorders. We also sought to examine the relationship between metacognitive beliefs and clinical symptoms in the ARMS group. 2. Method This review was conducted in line with the PRISMA guidelines for reporting systematic reviews and meta-analyses (Moher, Liberati, Tetzlaff, Altman, & PRISMAGroup, 2009). 2.1. Eligibility criteria Studies eligible for inclusion in the narrative synthesis were original research articles that examined metacognitive beliefs using the MCQ or MCQ-30 among people meeting ARMS criteria (FusarPoli et al., 2013). In order to be included in the meta-analyses, studies also needed to report additional MCQ comparison data acquired from healthy controls, help-seeking individuals who did not meet ARMS criteria or people diagnosed with a full-threshold
psychotic disorder. Studies that included only subjects at genetic risk who had not met formal ARMS criteria, case studies and review articles were ineligible. No language restrictions were placed on articles for inclusion. 2.2. Search strategy On 1st August 2016 we conducted an electronic database search of Ovid MEDLINE, PsycINFO and Embase (from inception) using the following keyword search terms: “metacogniti*” and “at risk mental state” or “ultra high risk” or “UHR” or “clinical high risk” or “CHR” or “prodrom*” and “psychosis” or “psychotic” or “schizophrenia”. In addition, a basic search of Google Scholar was conducted, recent conference abstracts were screened, authors were contacted for unpublished data and the reference lists of retrieved articles were also reviewed to identify any additional eligible studies. 2.3. Study selection and data extraction Two reviewers (J.C. and R.C.) independently screened articles for eligibility. For all eligible studies, a data extraction spreadsheet was used to record: (1) study characteristics (year of publication, country where the work was performed); (2) sample demographics (sample size, gender composition, mean age); (3) the screening instrument used to assess ARMS status; (4) metacognitive data (MCQ measure used, mean sample scores for each subscale); (5) summary of study findings. 2.4. Statistical analyses Meta-analyses were performed to examine group differences in metacognitive beliefs based on the five factors derived from the MCQ. Separate analyses were performed to examine differences between (1) ARMS and healthy controls; (2) ARMS and people with psychotic disorders; (3) ARMS and help-seeking individuals who did not meet ARMS criteria. Data analyses were performed using Comprehensive Meta-Analysis version 3.0 (Borenstein, Hedges, Higgins, & Rothstein, 2007). Standardised mean differences (effect sizes) were calculated for each of the five MCQ subscales using Hedges’ g. A random-effects model was applied throughout, better accounting for observed heterogeneity. Estimates are more conservative but such models perform better than fixed-effect approaches (Brockwell & Gordon, 2001). Heterogeneity was quantified using the Q-test and I2 statistic. 3. Results 3.1. Eligible studies The study selection process is summarised in Fig. 1. We identified eleven papers eligible for inclusion in the narrative synthesis, collectively reporting data obtained from six unique ARMS samples (due to multiple publication) (Table 1). The meta-analyses included data from each of these six samples: metacognitive beliefs in five ARMS samples were compared to healthy controls (Brett, Johns, Peters, & McGuire, 2009; Leicester, 2013; Morrison, French, & Wells, 2007, 2006; Taylor, 2010; Welsh, Cartwright-Hatton, Wells, Snow, & Tiffin, 2014), in three samples were compared to people diagnosed with a psychotic disorder (Brett et al., 2009; Morrison et al., 2007, 2006; Taylor, 2010), and two to help-seeking individuals who did not meet ARMS criteria (Barbato et al., 2014; Taylor, 2010). Metacognitive data acquired at baseline from the Early Detection and Intervention Evaluation (EDIE) trial was used in a series of publications. In the meta-analyses we included data from
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
27
Fig. 1. PRISMA flow diagram.
the largest ARMS sample and largest healthy control and psychosis comparator samples reported across these publications (Morrison et al., 2006; 2007) in order to maximise power to detect group differences. One study reported data for ARMS individuals with and without hallucinatory experiences separately (Leicester, 2013). As both groups had distinct clinical profiles, they were both included and treated as separate data points in the meta-analysis.
subscale scores in people with ARMS relative to people with psychotic disorders are reported in Table 3. There were no significant differences between people with ARMS and those with psychotic disorders on any of the five MCQ subscales (all p 0.198) (Supplementary Figures 6-10). Moderate-to-high levels of statistical heterogeneity were observed between studies included in each of these estimates.
3.1.1. Metacognitive beliefs among people with ARMS relative to healthy controls Effect sizes, sample size and heterogeneity statistics for MCQ subscale scores in people with ARMS relative to healthy controls are reported in Table 2. ARMS individuals scored significantly higher on all measures of dysfunctional metacognitive beliefs compared to healthy controls (all p < 0.001), except for the ‘positive beliefs about worry’ subscale which met only trend-level significance (p ¼ 0.053) (Supplementary Figures 1-5). There was evidence of low-tomoderate statistical heterogeneity for the ‘cognitive confidence’, ‘negative beliefs about thoughts in general’ and ‘cognitive selfconsciousness’ subscale estimates.
3.1.3. Metacognitive beliefs among people with ARMS relative to help-seeking controls Effect sizes, sample size and heterogeneity statistics for MCQ subscale scores in people with ARMS relative to help-seeking controls are reported in Table 4. People with ARMS scored significantly higher than help-seeking controls on the ‘negative beliefs about uncontrollability’, ‘cognitive confidence’ and ‘negative beliefs about thoughts in general’ subscales of the MCQ (all p 0.004). There were no significant differences between the groups for the ‘positive beliefs about worry’ (p ¼ 0.147) and ‘cognitive self-consciousness’ subscales (p ¼ 0.291) (Supplementary Figures 11-15). There was no evidence of statistical heterogeneity for any of these subscale scores.
3.1.2. Metacognitive beliefs among people with ARMS relative to people with psychotic disorders Effect sizes, sample size and heterogeneity statistics for MCQ
3.1.4. Metacognitive beliefs and symptoms in the ARMS group Due to variation in both the use and reporting of symptom
28
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
Table 1 Sample characteristics. Study þ country
Barbato et al. (2014) Canada, USA Brett et al. (2009) - UK Leicester (2013) Australia (HG) Leicester (2013) Australia (NHG) Welsh et al. (2014) - UK EDIE trial sample Barkus et al. (2010) - UK Morrison et al. (2002) - UK Morrison et al. (2006) - UK Morrison et al. (2007) - UK EDIE-2 trial sample Morrison et al. (2015) - UK Palmier-Claus et al. (2013) - UKa Taylor (2010) - UK
Metacognitive measure
ARMS group
Healthy control group
Psychotic disorder group
Help-seeking group
Screening instrument
N (M/F)
Age mean (SD)
N (M/F)
Age mean (SD)
N (M/F)
Age mean (SD)
N (M/F)
Age mean (SD)
MCQ
SIPS
153 (88/65)
19.7 (4.2)
e
e
e
e
68 (40/28)
19.7 (4.2)
MCQ MCQ-30
CAARMS CAARMS
32 (21/11) 38 (15/23)
24.3 (3.6) 19.24 (2.84)
32 (18/14) 35 (19/16)
27.7 (7.5) 19.20 (2.98)
27 (12/15) e
32.4 (11.2) e
e e
e e
MCQ-30
CAARMS
32 (18/14)
19.03 (2.40)
35 (19/16)
19.20 (2.98)
e
e
e
e
MCQ-30
CAARMS
31 (15/16)
15.8 (1.4)
76 (29/47)
14.82 (1.1)
e
e
e
e
MCQ MCQ
PANSS PANSS
58 (40/18) 31 (22/9)
22.1 (4.4) 23.2 (4.78)
95 (37/58) 50 (8/42)
22.79 (6.93) 21.7 (7.71)
e e
e e
e e
e e
MCQ
PANSS
58 (40/18)
22.1 (4.4)
56 (19/37)
22.9 (5.4)
e
e
e
e
MCQ
PANSS
43 (31/12)
22.6 (4.7)
188 (144/44)
27.6 (11.1)
73 (53/20)
41.2 (10.3)
e
e
MCQ-30
CAARMS
117 (71/46)
20.3
e
e
e
e
e
e
MCQ-30
CAARMS
27 (14/13)
22.6 (4.4)
e
e
e
e
e
e
MCQ-30
CAARMS
113 (67/46)
20.4 (4.3)
30 (8/22)
22.8 (3.7)
20 (14/5)
22.4 (5.4)
28 (23/5)
21.3 (3.4)
Abbreviations: ARMS: At-risk mental state; CAARMS: Comprehensive Assessment of At-Risk Mental States; EDIE: Early Detection and Intervention Evaluation; MCQ: Metacognitions Questionnaire; MCQ-30: Metacognitions Questionnaire short form; PANSS: Positive and Negative Syndrome Scale; SIPS: Structured Interview for Prodromal Syndromes; HG: ARMS hallucinating group; NHG: ARMS non-hallucinating group. a Sample partially overlaps with the EDIE-2 trial sample.
measures across studies, we were unable to analyse statistically the relationship between metacognitive beliefs and symptom severity in the ARMS group. Instead we have provided a detailed narrative synthesis of the literature to date. Higher scores on ‘cognitive self-consciousness’, the ‘negative beliefs about uncontrollability’ and ‘negative beliefs about thoughts
in general’ subscales were significantly associated with more severe affective symptoms (Morrison et al., 2015, 2006; Palmier-Claus, Dunn, Taylor, Morrison, & Lewis, 2013; Taylor, 2010), more severe manic symptoms (Welsh et al., 2014), elevated levels of perceived stress (Morrison et al., 2006), and poor psychosocial functioning (Welsh et al., 2014). Maladaptive metacognitive beliefs have also
Table 2 Mean weighted effect sizes, sample sizes and heterogeneity statistics for MCQ subscales among ARMS individuals relative to healthy controls. MCQ subscale
k
N (ARMS)
N (Control)
Hedges' g
95% CI
Z
p
Q
I2
Positive beliefs about worry Negative beliefs about uncontrollability Cognitive confidence Negative beliefs about thoughts Cognitive self-consciousness
6 6 6 6 6
286 287 287 288 288
395 393 395 394 395
0.16 1.50 0.92 1.09 0.57
0.01 to 0.33 1.31 to 1.68 0.71 to 1.13 0.85 to 1.33 0.34 to 0.80
1.94 15.96 8.60 8.93 4.87
0.053 <0.001 <0.001 <0.001 <0.001
3.486 1.794 7.050 8.869 8.811
0% 0% 29% 44% 43%
Abbreviations: CI: confidence interval; k: number of comparisons; MCQ: Metacognitions Questionnaire; Q: homogeneity analysis.
Table 3 Mean weighted effect sizes, sample sizes and heterogeneity statistics for MCQ subscales among ARMS individuals relative to people with psychotic disorders. MCQ subscale
k
N (ARMS)
N (Control)
Hedges' g
95% CI
Positive beliefs about worry Negative beliefs about uncontrollability Cognitive confidence Negative beliefs about thoughts Cognitive self-consciousness
3 3 3 3 3
185 186 186 187 187
115 115 115 114 115
0.24 0.12 0.16 0.17 0.14
0.61 0.28 0.67 0.51 0.24
to to to to to
0.13 0.52 0.36 0.18 0.52
Z
p
Q
I2
1.29 0.58 0.60 0.95 0.73
0.198 0.563 0.548 0.343 0.463
3.812 4.563 7.399 3.402 4.049
48% 56% 73% 41% 51%
Abbreviations: CI: confidence interval; k: number of comparisons; MCQ: Metacognitions Questionnaire; Q: homogeneity analysis.
Table 4 Mean weighted effect sizes, sample sizes and heterogeneity statistics for MCQ subscales among ARMS individuals relative to help-seeking controls. MCQ subscale
k
N (ARMS)
N (Control)
Hedges' g
95% CI
Z
p
Q
I2
Positive beliefs about worry Negative beliefs about uncontrollability Cognitive confidence Negative beliefs about thoughts Cognitive self-consciousness
2 2 2 2 2
220 221 221 222 222
89 88 89 88 88
0.18 0.37 0.39 0.49 0.13
0.06 to 0.43 0.12 to 0.62 0.14 to 0.63 0.24 to 0.74 0.12 to 0.38
1.45 2.89 3.03 3.82 1.06
0.147 0.004 0.002 <0.001 0.291
0.252 0.660 0.010 0.728 0.277
0% 0% 0% 0% 0%
Abbreviations: CI: confidence interval; k: number of comparisons; MCQ: Metacognitions Questionnaire; Q: homogeneity analysis.
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
been significantly associated with sub-threshold positive psychotic symptoms, but there was no consistent link across studies between scores on particular metacognitive subscales and specific subthreshold psychotic symptoms (Brett et al., 2009; Leicester, 2013; Morrison et al., 2015, 2006; Palmier-Claus et al., 2013; Taylor, 2010; Welsh et al., 2014). Leicester (2013) subdivided 70 young people with ARMS into those who had and had not reported auditory hallucinatory experiences. Both groups reported significantly more maladaptive metacognitive beliefs than healthy controls but there were no significant differences between the two ARMS groups on any of the MCQ-30 subscales. In contrast, Welsh et al. (2014) reported significant positive correlations between ‘negative beliefs about the uncontrollability of thoughts’ and both more intense and more distressing perceptual abnormalities among 31 adolescents with ARMS. They also reported a weak correlation between ‘positive beliefs about worry’ and higher frequency of unusual thought content. Morrison et al. (2015) examined the cross-sectional association between persecutory ideation and scores on the ‘cognitive selfconsciousness’, ‘negative beliefs about uncontrollability’ and ‘negative beliefs about thoughts in general’ subscales of the MCQ30. They reported that these were positively correlated with both perceived persecution and associated deservedness. In multivariate analyses controlling for a range of mood, anxiety and demographic variables, ‘negative beliefs about thoughts in general’ remained significantly associated with perceived deservedness of persecution, but not with self-reported persecution itself. In the same sample, Taylor (2010) reported that ‘negative beliefs about uncontrollability’ and ‘negative beliefs about thoughts in general’ also appeared to mediate many of the associations between both the frequency and severity of persecutory ideation and unwanted thoughts, and their associations with depression, social anxiety and symptom-related distress. Palmier-Claus et al. (2013) examined the temporal relationship between metacognitive beliefs and both psychotic and mood symptoms over 6 days in 27 ARMS individuals. They used an experience sampling approach in which participants completed a diary when prompted by an electronic wristwatch. They reported that heightened ‘cognitive self-consciousness’ preceded hallucinatory experiences but only among individuals who scored highly on the ‘negative beliefs about thoughts’ subscale (Palmier-Claus et al., 2013). Delusional thoughts were unrelated to metacognitive beliefs. They also reported that total scores on the MCQ-30 moderated the association between stressors and affective symptoms, but not psychotic symptoms. In the only long-term follow-up study conducted in this area to date, Barbato et al. (2014) reported that MCQ subscale scores and both positive and negative psychotic symptoms significantly improved at 6-month follow-up among a large ARMS cohort. They also reported preliminary evidence for a link between metacognitive dysfunction and conversion to psychotic disorder. Significantly higher scores at baseline on the ‘positive beliefs about worry’ and the two ‘negative belief’ subscales of the MCQ were reported among ARMS and help-seeking controls who subsequently developed a psychotic disorder compared to those who did not over the following two-year period. 4. Discussion 4.1. Summary of results These results suggest that young people with ARMS were more likely to view their thoughts as dangerous and uncontrollable and had greater concerns over their memory and
29
attention compared to both healthy and help-seeking controls. They were also more likely to monitor and be aware of their thoughts compared to healthy controls. However, people with ARMS did not significantly differ from people with full-threshold psychotic disorders on any of the MCQ subscales. Overall, these findings were consistent with the S-REF model which suggests that dysfunctional metacognitive beliefs may contribute to psychological dysfunction (Wells & Matthews, 1996), though the relationship between metacognitive beliefs and the development and persistence of specific sub-threshold psychotic symptoms remains unclear. 4.2. The relationship between metacognitive beliefs, symptoms and cognition Metacognitive beliefs were associated with a range of symptoms. However, small sample sizes, cross-sectional study designs and heterogeneity in both measures and reporting makes confident inferences about the significance and nature of these associations difficult. Though significant associations between metacognitive subscales and both perceptual abnormalities and persecutory ideation were reported, there was no consistent pattern between the presence of maladaptive metacognitive beliefs and the intensity or frequency of specific sub-threshold psychotic phenomena. Instead dysfunctional metacognitive beliefs appear to be more strongly associated with mood and anxiety symptoms (Brett et al., 2009; Palmier-Claus et al., 2013; Welsh et al., 2014). Other reviews have also reported limited evidence that maladaptive metacognitive beliefs play a causal role in the development and persistence of psychotic symptoms (Varese & Bentall, 2011). It is unclear why metacognitive beliefs’ associations with psychotic phenomena are inconsistent when existing theories predict they should be specific and consistent (Morrison, 2001; Morrison et al., 2000). Notably, the MCQ was originally developed to assess metacognitive beliefs associated with worry. In future, modified versions of the MCQ or symptom-specific measures enquiring directly about metacognitive appraisals of hallucinations or delusions may be more suited to examining the development and course of psychotic symptoms. Young people with ARMS also exhibit significant deficits in cognition, particularly in memory and attention (Fusar-Poli et al., 2012b). This is likely to be reflected in the poorer perceived ‘cognitive confidence’ among the ARMS group relative to both the healthy and help-seeking controls. However, it is unclear whether these beliefs accurately reflect cognitive performance or whether they overestimate deficits in these domains. Further research is also needed to establish whether maladaptive metacognitive beliefs are predictors of social and occupational functioning in this group (Cotter et al., 2014; 2015). 4.3. Strengths and limitations To our knowledge, this is the first meta-analytic review to examine metacognitive dysfunction in the ARMS group. Studies in this area have typically been small and potentially underpowered to detect between-group differences. As a result of the crosssectional design of most of these studies, it is also difficult to establish whether maladaptive metacognitive beliefs occur as a cause or consequence of psychiatric symptoms. Similarly, the stability of metacognitive beliefs and their relationship with transition to psychotic disorder in the ARMS group remains largely unexplored and poorly understood. Young people with ARMS are helpseeking and often in crisis at presentation. However, sub-threshold psychotic symptoms and other mental health issues may only be transitory for some people. Metacognitive beliefs may reflect state
30
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31
rather than trait factors and arise as a result of concurrent clinical symptoms and distress rather than vulnerability to psychosis. There is evidence that both positive and negative psychotic symptoms and dysfunctional metacognitive beliefs improve over time in this group (Barbato et al., 2014). Further prospective, longitudinal research is needed to establish the stability of maladaptive metacognitive beliefs, as well as the direction of causality and the clinical implications of the relationship between metacognitive dysfunction and psychiatric symptoms. There is also currently a lack of studies examining differences between ARMS individuals and non-ARMS help-seeking controls. Metacognitive dysfunction may be primarily associated with mood and anxiety symptoms, providing an alternative explanation for the significant differences observed between the ARMS and healthy control groups. Only two studies to date have examined metacognitive beliefs among both help-seeking controls and ARMS samples. Both of these studies comprised relatively small helpseeking control groups. Though we found significant differences on three of the MCQ subscales between these groups, further research is needed comparing people with ARMS to those with only affective symptoms before more definitive conclusions can be drawn. Moderate-to-high levels of heterogeneity were observed in the ARMS compared to psychosis group MCQ subscale estimates. This may have been driven by variation in the age, gender composition and diagnoses of people in the psychosis groups across studies. Studies assessing both group differences in metacognitive beliefs and the association between metacognition and clinical symptoms have often failed to use multivariate analyses to adjust for potentially confounding factors. There is evidence that group differences in metacognitive beliefs as well as their association with positive psychotic symptoms became non-significant after controlling for anxiety and depressive symptoms (Brett et al., 2009; Morrison et al., 2015). This should be an important consideration for future research.
4.4. Clinical implications Assessing psychological factors in this help-seeking group which may contribute towards and exacerbate mental health issues is important. Cognitive behavioural assessments should therefore include examination of metacognitive beliefs. Strategies to help modify maladaptive metacognitive beliefs may also be a target for clinical intervention in this group. Metacognitive therapy involves identifying and challenging metacognitive beliefs (Wells, 2009), and has been shown to be effective for the treatment of both anxiety and depression (Normann, van Emmerik, & Morina, 2014). Early evidence suggests it may also be beneficial for improving psychotic symptoms in people with a diagnosis of schizophrenia (Hutton, Morrison, Wardle, & Wells, 2014; Morrison et al., 2014). It would therefore be useful to examine whether it has a similar impact on clinical symptoms in the ARMS group, given the range of co-morbid mental health problems frequently present among ARMS individuals (Addington et al., 2011; Fusar-Poli et al., 2014; Lin et al., 2015). Current predictors of transition to psychotic disorder in the ARMS population have been limited largely to the same range of clinical, demographic and neurocognitive variables. Evaluation of metacognition may potentially provide additional prognostic value over and above these factors and warrants further investigation. Barbato et al. (2014) have provided preliminary evidence that metacognitive dysfunction is a potential marker for conversion to psychotic disorder; however, additional longitudinal replication studies are required.
5. Conclusions Maladaptive metacognitive beliefs are common in young people meeting ARMS criteria. Further research is needed to clarify the relationship between metacognitive dysfunction and the development and persistence of both psychotic and non-psychotic clinical symptoms. Assessment and treatment aimed at alleviating metacognitive dysfunction may be useful for improving a range of clinical outcomes. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.brat.2016.12.004. References Addington, J., Cornblatt, B. A., Cadenhead, K. S., Cannon, T. D., McGlashan, T. H., Perkins, D. O., et al. (2011). At clinical high risk for psychosis: Outcome for nonconverters. American Journal of Psychiatry, 168(8), 800e805. Barbato, M., Penn, D. L., Perkins, D. O., Woods, S. W., Liu, L., & Addington, J. (2014). Metacognitive functioning in individuals at clinical high risk for psychosis. Behavioural and Cognitive Psychotherapy, 42(5), 526e534. Barkus, E., Stirling, J., French, P., Morrison, A., Bentall, R., & Lewis, S. (2010). Distress and metacognition in psychosis prone individuals: Comparing high schizotypy to the at-risk mental state. Journal of Nervous and Mental Disease, 198(2), 99e104. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2007). Comprehensive metaanalysis: A computer program for meta-analysis. Englewood, NJ: Biostat Inc. Brett, C. M., Johns, L. C., Peters, E. P., & McGuire, P. K. (2009). The role of metacognitive beliefs in determining the impact of anomalous experiences: A comparison of help-seeking and non-help-seeking groups of people experiencing psychotic-like anomalies. Psychological Medicine, 39(6), 939e950. Brockwell, S. E., & Gordon, I. R. (2001). A comparison of statistical methods for meta-analysis. Statistics in Medicine, 20(6), 825e840. Cartwright-Hatton, S., & Wells, A. (1997). Beliefs about worry and intrusions: The meta-cognitions questionnaire and its correlates. Journal of Anxiety Disorders, 11(3), 279e296. Cotter, J., Bartholomeusz, C., Papas, A., Allott, K., Nelson, B., Yung, A. R., et al. (2015). Examining the association between social cognition and functioning in individuals at ultra-high risk for psychosis. Australian and New Zealand Journal of Psychiatry (in press), http://dx.doi.org/10.1177/0004867415622691. Cotter, J., Drake, R. J., Bucci, S., Firth, J., Edge, D., & Yung, A. R. (2014). What drives poor functioning in the at-risk mental state? A systematic review. Schizophrenia Research, 159(2e3), 267e277. Ellis, D. M., & Hudson, J. L. (2010). The metacognitive model of generalized anxiety disorder in children and adolescents. Clinical Child and Family Psychology Review, 13(2), 151e163. Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive-developmental inquiry. American Psychologist, 34(10), 906e911. Fusar-Poli, P., Bonoldi, I., Yung, A. R., Borgwardt, S., Kempton, M. J., Valmaggia, L., et al. (2012a). Predicting psychosis: meta-analysis of transition outcomes in individuals at high clinical risk. Archives of General Psychiatry, 69(3), 220e229. € ssler, A., SchultzeFusar-Poli, P., Borgwardt, S., Bechdolf, A., Addington, J., Riecher-Ro Lutter, F., et al. (2013). The psychosis high-risk state: A comprehensive state-ofthe-art review. JAMA Psychiatry, 70(1), 107e120. Fusar-Poli, P., Deste, G., Smieskova, R., Barlati, S., Yung, A. R., Howes, O., et al. (2012b). Cognitive functioning in prodromal psychosis: A meta-analysis. Archives of General Psychiatry, 69(6), 562e571. Fusar-Poli, P., Nelson, B., Valmaggia, L., Yung, A. R., & McGuire, P. K. (2014). Comorbid depressive and anxiety disorders in 509 individuals with an at-risk mental state: Impact on psychopathology and transition to psychosis. Schizophrenia Bulletin, 40(1), 120e131. Hezel, D. M., & McNally, R. J. (2015). A theoretical review of cognitive biases and deficits in obsessiveecompulsive disorder. Biological Psychology, 121(B), 221e232. Hutton, P., Morrison, A. P., Wardle, M., & Wells, A. (2014). Metacognitive therapy in treatment-resistant psychosis: A multiple-baseline study. Behavioural and Cognitive Psychotherapy, 42(2), 166e185. Leicester, S. B. (2013). Cognitive processing associated with hallucinations for at risk mental state (Unpublished doctoral thesis). Australia: Melbourne School of Psychological Sciences, Faculty of Medicine, Dentistry & Health Sciences, The University of Melbourne. Lin, A., Wood, S. J., Nelson, B., Beavan, A., McGorry, P., & Yung, A. R. (2015). Outcomes of nontransitioned cases in a sample at ultra-high risk for psychosis. American Journal of Psychiatry, 172(3), 249e258. Miller, T. J., McGlashan, T. H., Rosen, J. L., Somjee, L., Markovich, P. J., Stein, K., et al. (2002). Prospective diagnosis of the initial prodrome for schizophrenia based on the structured Interview for prodromal syndromes: Preliminary evidence of interrater reliability and predictive validity. American Journal of Psychiatry,
J. Cotter et al. / Behaviour Research and Therapy 90 (2017) 25e31 159(5), 863e865. Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & PRISMA Group. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ, 339, b2535. Morrison, A. P. (2001). The interpretation of intrusions in psychosis: An integrative cognitive approach to hallucinations and delusions. Behavioural and Cognitive Psychotherapy, 29(3), 257e276. Morrison, A. P., Bentall, R. P., French, P., Walford, L., Kilcommons, A., Knight, A., et al. (2002). Randomised controlled trial of early detection and cognitive therapy for preventing transition to psychosis in high-risk individuals. Study design and interim analysis of transition rate and psychological risk factors. British Journal of Psychiatry, 43, s78es84. Morrison, A. P., French, P., Lewis, S. W., Roberts, M., Raja, S., Neil, S. T., et al. (2006). Psychological factors in people at ultra-high risk of psychosis: Comparisons with non-patients and associations with symptoms. Psychological Medicine, 36(10), 1395e1404. Morrison, A. P., French, P., & Wells, A. (2007). Metacognitive beliefs across the continuum of psychosis: Comparisons between patients with psychotic disorders, patients at ultra-high risk and non-patients. Behaviour Research and Therapy, 45(9), 2241e2246. Morrison, A. P., Gumley, A. I., Ashcroft, K., Manousos, I. R., White, R., Gillan, K., et al. (2011). Metacognition and persecutory delusions: Tests of a metacognitive model in a clinical population and comparisons with non-patients. British Journal of Clinical Psychology, 50(3), 223e233. Morrison, A. P., Pyle, M., Chapman, N., French, P., Parker, S. K., & Wells, A. (2014). Metacognitive therapy in people with a schizophrenia spectrum diagnosis and medication resistant symptoms: A feasibility study. Journal of Behavior Therapy and Experimental Psychiatry, 45(2), 280e284. Morrison, A. P., Shryane, N., Fowler, D., Birchwood, M., Gumley, A. I., Taylor, H. E., et al. (2015). Negative cognition, affect, metacognition and dimensions of paranoia in people at ultra-high risk of psychosis: A multi-level modelling analysis. Psychological Medicine, 45(12), 2675e2684. Morrison, A. P., Wells, A., & Nothard, S. (2000). Cognitive factors in predisposition to auditory and visual hallucinations. British Journal of Clinical Psychology, 39(1), 67e78. Nelson, B., Yuen, H. P., Wood, S. J., Lin, A., Spiliotacopoulos, D., Bruxner, A., et al. (2013). Long-term follow-up of a group at ultra high risk (“prodromal”) for psychosis: The PACE 400 study. JAMA Psychiatry, 70(8), 793e802. Normann, N., van Emmerik, A. A., & Morina, N. (2014). The efficacy of metacognitive therapy for anxiety and depression: A meta-analytic review. Depression and Anxiety, 31(5), 402e411.
31
Olstad, S., Solem, S., Hjemdal, O., & Hagen, R. (2015). Metacognition in eating disorders: Comparison of women with eating disorders, self-reported history of eating disorders or psychiatric problems, and healthy controls. Eating Behaviors, 16, 17e22. Palmier-Claus, J. E., Dunn, G., Taylor, H., Morrison, A. P., & Lewis, S. W. (2013). Cognitive-self consciousness and metacognitive beliefs: Stress sensitization in individuals at ultra-high risk of developing psychosis. British Journal of Clinical Psychology, 52(1), 26e41. Papageorgiou, C., & Wells, A. (2003). An empirical test of a clinical metacognitive model of rumination and depression. Cognitive Therapy and Research, 27(3), 261e273. Sellers, R., Varese, F., Wells, A., & Morrison, A. P. (2016). A meta-analysis of metacognitive beliefs as implicated in the self-regulatory executive function model in clinical psychosis. Schizophrenia Research, 179, 75e84. Spada, M. M., Zandvoort, M., & Wells, A. (2007). Metacognitions in problem drinkers. Cognitive Therapy and Research, 31(5), 709e716. Taylor, H. E. (2010). Cognitive processes across the continuum of psychosis (Unpublished doctoral thesis). UK: School of Psychological Sciences, Faculty of Medical and Human Sciences, The University of Manchester. Varese, F., & Bentall, R. P. (2011). The metacognitive beliefs account of hallucinatory experiences: A literature review and meta-analysis. Clinical Psychology Review, 31(5), 850e864. Wells, A. (1995). Meta-cognition and worry: A cognitive model of generalized anxiety disorder. Behavioural and Cognitive Psychotherapy, 23(3), 301e320. Wells, A. (2009). Metacognitive therapy for anxiety and depression. New York: Guilford Press. Wells, A., & Cartwright-Hatton, S. (2004). A short form of the metacognitions questionnaire: Properties of the MCQ-30. Behaviour Research and Therapy, 42(4), 385e396. Wells, A., & Matthews, G. (1996). Modelling cognition in emotional disorder: The SREF model. Behaviour Research and Therapy, 34(11e12), 881e888. Welsh, P., Cartwright-Hatton, S., Wells, A., Snow, L., & Tiffin, P. A. (2014). Metacognitive beliefs in adolescents with an at-risk mental state for psychosis. Early Intervention in Psychiatry, 8(1), 82e86. Yung, A. R., McGorry, P. D., McFarlane, C. A., Jackson, H. J., Patton, G. C., & Rakkar, A. (1996). Monitoring and care of young people at incipient risk of psychosis. Schizophrenia Bulletin, 22(2), 283e303. Yung, A. R., Phillips, L. J., Yuen, H. P., Francey, S. M., McFarlane, C. A., Hallgren, M., et al. (2003). Psychosis prediction: 12-month follow up of a high-risk (“prodromal”) group. Schizophrenia Research, 60(1), 21e32.