The role of negative symptoms in the context of cognitive remediation for schizophrenia

Schizophrenia Research 150 (2013) 58–63

Contents lists available at ScienceDirect

Schizophrenia Research journal homepage: www.elsevier.com/locate/schres

The role of negative symptoms in the context of cognitive remediation for schizophrenia Aida Farreny a,⁎, Jaume Aguado a,b, Susana Ochoa a, Josep Maria Haro a, Judith Usall a a b

Parc Sanitari Sant Joan de Déu, Fundació Sant Joan de Déu, CIBERSAM 1, Spain University of Barcelona, Barcelona, Spain

a r t i c l e

i n f o

Article history: Received 5 April 2013 Received in revised form 24 July 2013 Accepted 10 August 2013 Available online 29 August 2013 Keywords: Cognitive remediation Schizophrenia Neurocognition Negative symptoms Functional outcome Mediated effects

a b s t r a c t Background: It has been suggested that the effect of cognitive remediation (CR) on functioning is mediated by the improvement in neurocognitive domains; especially executive function. However, the correlations are generally moderate and this has prompted the search for other mediators including negative symptoms (NS). Aims: To investigate whether the effect of CR on functioning could be mediated by executive function and/or NS. Method: In a previous study, 62 outpatients with schizophrenia were randomized to 32 group sessions of REPYFLEC CR or leisure activities. Functioning (Life Skills Profile; LSP), NS (PANSS) and executive function (Behavioral Assessment of the Dysexecutive Syndrome; BADS) were measured at baseline and post-therapy. To assess how the effect of REPYFLEC CR is expressed in functioning at post-treatment, an autoregressive mediation model was employed. Results: There was a significant effect of the REPYFLEC CR compared with the control group in improving BADS total score and PANSS NS. There was also a significant association between NS and functioning while executive function was not significantly related to functioning. Finally, there was a significant intervention effect on functioning mediated by NS but not by executive function. Conclusion: It is apparent that improving executive function does not lead directly to improved functional outcome and that NS might be closely linked to functioning in the context of our study. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Cognitive remediation (CR) treatments in schizophrenia were designed to improve neurocognition on the assumption that progress in functioning might be mediated in this way. It has been suggested that the effect of CR on functioning is mediated by the improvement in specific neurocognitive domains (Vita et al., 2011), especially through executive function (Wykes et al., 2007; Penadés et al., 2010) and in particular, by planning improvement (Wykes et al., 2012). However, the correlations are generally moderate; accounting only for 20%–60% of the variance in functional outcome (Green et al., 2000), this has prompted a search for other variables that may account for the effect on real world performance such as negative symptoms (Greenwood et al., 2005; Bowie et al., 2008; Ventura et al., 2009; Bowie et al., 2010; Lin et al., 2013). Some theoretical grounding has proposed that negative symptoms (NS) could be an important moderating (Greenwood et al., 2005) or

⁎ Corresponding author at: Parc Sanitari Sant Joan de Déu, Research Unit, Dr. Antoni Pujadas, 42, Sant Boi de Llobregat, 08830 Barcelona, Spain. Tel.: +34 936406350; fax: +34 935569674. E-mail address: [email protected] (A. Farreny). 1 Centro de investigación biomédica en red de salud mental. 0920-9964/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.schres.2013.08.008

mediating (Lipkovich et al., 2009; Ventura et al., 2009; Lin et al., 2013) mechanism in the relationship between neurocognition and functional outcomes in schizophrenia. However, to the best of our knowledge, the role of NS in the context of CR has not yet been explored. One possible explanation for this may be the small effect that is found in psychiatric symptoms after CR treatments (McGurk et al., 2007; Wykes et al., 2011) that no longer seem significant at follow-up (Wykes et al., 2011). Nevertheless, some authors have stated that even very small improvements may have an impact on learning skills or future functioning (Wykes and Spaulding, 2011). In addition, CR meta-analyses tend to study psychiatric symptoms without distinguishing between positive, negative and general psychopathology despite the likely heterogeneity within their expression and phenomenology that could complicate the understanding of symptoms in the context of CR. There is evidence to suggest that patients with higher NS severity have poorer social competence and quality of life (i.e. Harvey et al., 2006; Kirkpatrick et al., 2006; Bowie et al., 2008; Klingberg et al., 2011), as well as neurocognitive deficits, and these have been a consistent determinant of psychosocial functioning in several studies (i.e. Green et al., 2000; Bowie et al., 2008; Penadés et al., 2010; Vita et al., 2011). As such, in the light of their collective impact on outcome, the link between NS and neurocognition in schizophrenia has also garnered considerable attention. Some authors have argued that NS

A. Farreny et al. / Schizophrenia Research 150 (2013) 58–63

may be underpinned by neurocognitive deficits such as the impaired initiation of novel responses (Frith, 1992; Greenwood et al., 2008); and others have speculated that patients with higher levels of NS have particular impairments in reasoning and executive function (Villalta-Gil et al., 2006; Ventura et al., 2009). However, several studies have failed to establish a relationship between NS and neurocognition, leading to the conclusion that they represent semiautonomous disease processes (Bell and Mishara, 2006; Harvey et al., 2006; Kirkpatrick et al., 2006; McGurk et al., 2007; Lipkovich et al., 2009; Foussias and Remington, 2010). A further critical issue concerns how best to measure NS within schizophrenia due to the features in the definition of NS and neurocognition that blur conceptual boundaries. Therefore, the correlation between neurocognition and NS may vary as a function of the definition of the NS construct (Harvey et al., 2006; Foussias and Remington, 2010). For instance, the Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987) is one of the most commonly used clinical rating scales for schizophrenia although some aspects of the illness appear to be cognitive in nature, such as deficits in abstract reasoning and stereotyped thinking, which are defined as NS and contribute to total scores on this dimension. Consequently, numerous authors have proposed differentiated factorial approaches to reduce confusion between NS and other domains such as neurocognition or functionality, suggesting that a fivefactor model better characterizes PANSS data (Wallwork et al., 2012). It seems that this matter is still unresolved and there is no clear consensus on what factors proposed for the PANSS most appropriately group the psychiatric symptoms within schizophrenia. Based on our preceding study which showed that patients in the CR group obtained significant improvements in executive function, NS and functional outcomes compared with patients in the control group (Farreny et al., 2012), our purpose was to explore possible underlying mechanisms in the context of CR, with a main focus on the role of NS. Previously, Wykes et al. (2012) explored whether there is evidence of neurocognition playing a mediating role in the relationship between CR and improved functioning. The mediator (planning improvement) was able to account for a proportion of the total effect on work quality and although there were other changes in cognition (memory and flexibility) these do not necessarily drive the key changes in outcome. Thus, we hypothesized the mediating variables should be, at least, those showing significant improvement after CR and our first aim was to explore whether the putative effect of our treatment on functioning could be mediated by executive function and/or NS. In addition, we aimed to study cross-sectional and longitudinal associations between executive function, NS and functional outcomes to gain a better understanding of the mediation mechanism. 2. Method 2.1. Design The study was a randomized controlled trial with 62 outpatients diagnosed with schizophrenia or schizoaffective disorder. The participants were randomized to 32 group sessions of REPYFLEC CR or to 32 group sessions of stimulating activities without specific objectives and focused on leisure. 2.2. Participants Participants in the study were outpatients between 18 and 60 years, from the Barcelona metropolitan area, known to the Parc Sanitari Sant Joan de Déu Mental Health Services, who had been diagnosed at least two years previously with schizophrenia or schizoaffective disorder (American Psychiatric Association (APA) et al., 2002). Additional selection criteria included: finished primary studies or ability to successfully complete a reading comprehension task used for 13-year-old students;

59

a Mini Mental State Examination score of 24 or more; and a Global Assessment of Functioning score between 40 and 70. Patients were excluded if: 1) they were suffering acute illness exacerbation that required hospitalization; 2) they had intellectual disability or any identifiable neurological disorder; 3) they were participating in any type of psychological intervention (i.e., social skills training, cognitive remediation, cognitive behavioral therapy) differing from usual care; 3) they had a switch of antipsychotic drug the month before the trial or during the study period, and/or a diagnosis of alcohol or drug dependence within 6 months prior to inclusion. 2.3. Outcome measures Neurocognition, functioning and psychiatric symptoms were measured at baseline and thereafter at 16 weeks (post-treatment). 2.3.1. Neurocognition Behavioral Assessment of the Dysexecutive Syndrome (BADS) (Wilson et al., 1996). This battery consists of six tests and evaluates cognitive flexibility, inhibition of impulsive responses, planning and organization, working memory and time-estimation capacity. For each of the subtests a summary profile score is obtained (maximum of 4 and minimum of 0), and these are added together to produce an overall battery profile score (out of 24). All subtests (Rule shift cards, Action program, Key search, Temporal judgment, Zoo map and Six elements) were administered. We used the standardized score (mean: 100, standard deviation (SD): 15) for the total with a minimum of 12 and maximum of 129, with a higher score indicating better performance. 2.3.2. Functioning The Spanish validation (Fernández de Larrinoa et al., 1992) of the Life Skills Profile (LSP) (Rosen et al., 1989) was used to assess functional outcome. This scale measures the overall level of functioning in people with chronic mental disorders in daily-living situations and tasks. The LSP consists of 39 items organized into five subscales: self-care; interpersonal behavior; social contact and communication; social relationships; and personal autonomy. Information is initially self-reported but subsequently contrasted with other observers (e.g. therapists, nurses, family members), providing an objective assessment of everyday abilities or achievements. The Spanish version has demonstrated high rates of internal consistency, inter-rater reliability and concurrent validity; and has been shown to be a good predictive scale in different clinical areas (Ballesteros-Rodríguez et al., 2002). Raw scoring was used for the total (min. 39–max. 156) with a higher score indicating a better result. 2.3.3. Psychiatric symptoms The Spanish validation (Peralta and Cuesta, 1994) of the Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987) was used. Its items were originally grouped into scales for positive symptoms (7 items), negative symptoms (7 items) and general psychopathology (16 items). The interviewers administered the PANSS as part of a structured clinical interview and scored items on a scale from 1 (asymptomatic) to 7 (extremely symptomatic). Two distinct factorial proposals for negative symptoms are considered in this study: Firstly, we considered the classical approach with the three-factor PANSS including positive and negative symptoms and general psychopathology (Kay et al., 1987). Regarding negative symptoms, this factor includes 7 items: blunted affect, emotional withdrawal, poor rapport, passive-apathetic social withdrawal, lack of spontaneity, difficulty in abstract thinking and stereotyped thinking. And secondly, we applied the recent model by Wallwork et al. (2012) that considered previous five-factor models and used a strong methodological approach to find the best consensus model for the PANSS. This resulted in a solution that distinguishes between positive, negative, disorganized, excited and depressed factors. The negative factor includes 6 items: blunted

60

A. Farreny et al. / Schizophrenia Research 150 (2013) 58–63

affect, emotional withdrawal, poor rapport, passive-apathetic social withdrawal, lack of spontaneity and motor retardation. The negative symptom score was calculated as the mean of the items that defined the negative factor in both proposals to produce comparable results.

GROUP a1 a2

EF1

2.4. Therapy REPYFLEC CR is strategy-based group training that targets executive function and metacognition. It is carried out using paper, pencil and a blackboard (required to develop some of the tasks and provide explanations) in a group format (4–6 participants) over 4 months, twice a week and consists of 32 sessions lasting 1 h. Working contents are divided into two main areas: Problem Solving (PS) and Cognitive Flexibility (CF). In the PS module (16 sessions) training in executive function, thinking processes and self-monitoring was emphasized. Tasks combine the training for a reflexive understanding about problematic situations with strategy-teaching on how to achieve better monitoring of these situations; training in cognitive–emotional processes which could be involved; and repeated practice in several hypothetical problems that are expressly designed for REPYFLEC CR. In the CF module (16 sessions), all the tasks mainly focus the strategy-training on the ability to produce an increasing amount of responses with growing variety between them. Exercises aim to generalize from more basic tasks (e.g., shift cognitive sets; train flexible categorization; straighten an office up; plan several routes to walk from one place to another; count elements in a picture using different strategies to do so; complete a story with variety of endings) to higher cognitive components related to social knowledge (such as reporting as many advantages and disadvantages as possible from the same situation: being a child, living in a shared flat with no-relatives, having lots of siblings, etc.; or to be able to make the case for both sides in a debate) (Farreny et al., 2012). The training is described session by session in a Spanish manual, incorporating the materials for running sessions, some theoretical points and a bibliography for therapists. The treatment was administered by the first author (Aida Farreny) to all the participants. 2.5. Statistical analyses To assess whether the effect of REPYFLEC CR on functioning at posttreatment is mediated by executive functioning and/or negative symptoms, an autoregressive mediation model was employed (MacKinnon, 2008, p.199–201). The model included baseline mediator and functioning values as covariates and allowed for contemporaneous relations at post-treatment (Fig. 1). Parameters were estimated using Full Information Maximum Likelihood and Bootstrap was employed to compute the 95% bias-corrected confidence intervals (95% CI). It was inferred that a parameter was statistically significant at the 5% significance level if the Bootstrap 95% CI does not include zero. This estimation method has been found to be powerful enough to detect medium effects with small sample sizes (Fritz and MacKinnon, 2007). Following our hypothesis, we focused on three different mediated effects; the mediated effect of REPYFLEC CR on functioning through executive functioning (a1b1), through negative symptoms (a2b2) and through executive functioning and negative symptoms (a1db2). The percentage of the treatment effect that was mediated was computed as the mediated effect divided by the total effect. That is the sum of the absolute values of direct (c′) and mediated effects. Besides the 2 test, the following fit indices were analyzed (values in parentheses denote goodness-of-fit standards according to SchermellehEngel et al. (2003): (1) the Root Mean Square Error of Approximation (RMSEA) that measures the fit of the model to the correlation matrix (RMSEA ≤ 0.08 is indicative of acceptable fit and ≤0.05 of good fit), and (2) the comparative fit index (CFI) which compares the performance of

EF2 d

NS1

NS2

b1

b2

F1

F2

Group: REPYFLEC and control Mediators: EF1 and EF2 are executive function at baseline and at post-treatment respectively. NS1 and NS2 are negative symptoms at baseline and at post-treatment. Functioning: F1 and F2 are functioning at baseline and post-treatment.

Fig. 1. Two-wave autoregressive mediation model.

the specified model to the performance of a baseline model (CFI ≥ 0.95 are indicative of acceptable fit and ≥0.97 of good fit). Furthermore, to gain a better understanding of the role of baseline negative symptoms, we estimated the correlation between improvements in executive function (BADS) and life skills (LSP) at posttreatment and negative symptoms (PANSS) at baseline. In modeling these relationships, we hypothesized that cognitive status at baseline and subsequent changes in cognition could predict functioning either directly or indirectly via psychiatric symptoms, but not the other way around (i.e., changes in psychiatric symptoms affecting functioning via changes in cognition) (Lipkovich et al., 2009). Our assumption was that cognitive impairment precedes psychiatric symptoms, and that both precede functional impairment. 3. Results Two thirds of the participants were male (68%); 84% were single; 80% lived with their family of origin and 80% had completed at least 8 years of formal education. Some 89% of participants had a diagnosis of schizophrenia (n = 54), and the remaining (n = 7) of schizoaffective disorder. The average age was 40.6 years (SD: 7.6) and average illness duration was 17.5 years (SD: 8.9). During the year prior to the study, 80% of the participants had not engaged in any type of work, occupational or academic activity, and did not have responsibility for any household chores. Nevertheless, 63% of the sample attended a rehabilitation center daily or almost daily, participating in activities such as cooking, computer skills, sports, newspaper reviewing, gardening and other stimulating and socializing pastimes. Differences were not found between groups at baseline in pharmacological, sociodemographic, clinical or neurocognitive variables. No significant differences were found between those experimental and control subjects who abandoned the study. Regarding the number of group sessions attended, there were no differences between experimental participants (median: 27, min.:21–max.:32) and controls (median: 29, min.:20–max.:32). The scores at baseline and post-treatment for the variables used in the mediation model (Fig. 1) are presented by treatment group in Table 1.

A. Farreny et al. / Schizophrenia Research 150 (2013) 58–63 Table 1 Outcomes by treatment group. Variables

BADS Total Score LSP Total Score PANSS Negativea Kay et al. (1987) Wallwork et al. (2012)

61

Table 3 Autoregressive model parameter estimates.

Baseline

Post-treatment

Effect

Experimental

Control

Experimental

Control

Mean (SD) (min.–max.)

Mean (SD) (min.–max.)

Mean (SD) (min.–max.)

Mean (SD) (min.–max.)

83.67 (20.4) (41–119) 126.73 (15.5) (87–148)

88.53 (16) (46–109) 131.77 (11.3) (101–146)

99.55 (14.5) (51–124) 137.41 (6.9) (121–149)

95.7 (13.5) (61–119) 133.58 (11.4) (105–149)

2.74 (0.6) (1.7–3.8) 2.68 (0.6) (1.6–3.8)

2.64 (0.8) (1.3–4.5) 2.58 (0.8) (1.2–4.5)

2.36 (0.6) (1.28–3.8) 2.31 (0.6) (1.6–3.5)

2.52 (0.73) (1.3–3.8) 2.49 (0.7) (1.6–3.8)

a The negative symptoms score was calculated as the mean of the items in order to produce comparable results.

Pairwise correlations for measures at baseline are shown in Table 2. Inspection of these raw correlations suggests that there is a significant correlation between negative symptoms and functioning whereas executive function is only significantly correlated to the Kay et al. negative factor but not to Wallwork et al. Thus, to avoid possible overlapping with executive function in the measurement of NS, the mediation model was performed with the negative factor by Wallwork et al.

PANSS negative Wallwork et al. (2012) Estimate (95% CI)

Program effect on executive functioning (a1) Program effect on negative symptoms (a2) Executive function effect on negative symptoms (d) Executive function effect on functioning (b1) Negative symptoms effect on functioning (b2) Mediated effect through executive function (a1 b1) Mediated effect through negative symptoms (a2 b2) Mediated effect through executive function and negative symptoms (a1d b2) Direct effect (c′) Model fit indices

7.76 (2.56; 13.51) −0.36 (−0.62; −0.14) 0.01 (−0.002; 0.02) 0.15 (−0.05; 0.40) −5.28 (−10.39; −0.34) 1.18 (−0.18; 3.54) 1.90 (0.16; 4.73) −0.43 (−1.64; 0.001) 2.39 (−1.30; 6.81) χ2 (df) = 3.36 (3); RMSEA = 0.044; CFI = 0.99

Simplified figure representing the mediation model (baseline values are omitted for clarity).

Group: REPYFLEC and control; EF: executive function; NS: negative symptoms; F: functioning.

3.2. Impact of baseline negative symptoms on outcomes 3.1. Mediation analyses In Table 3 we present the results of the hypothesized mediation models and a simplified figure to help in the interpretation of results. Based on the fit indices, the model had an adequate fit to the data and the parameter estimates had the expected sign and size and are consistent with the underlying theory. There was a significant effect of the REPYFLEC CR compared with the control group in increasing the BADS total score (a1) and PANSS negative symptoms (a2). There was also a significant association between PANSS negative symptoms and functioning (b2) while executive function was not significantly related to social functioning (b1). Finally, there was a significant intervention effect on social functioning mediated by negative symptoms (a2b2) but not by executive function (a1b1). The mediated effect through executive function and negative symptoms was not statistically significant (a1db2). The percentage of the treatment effect that was mediated through negative symptoms (a2b2) was 32.2%. And mediation through executive function (a1b1) was 20%. Finally, the percentage mediated for the effect through executive function and negative symptoms (a1db2) was 7.3%. Table 2 Correlations among model variables at baseline. 1. BADS

1. Executive function: BADS 2. Negative symptoms: PANSS Kay et al. (1987) 3. Negative symptoms: PANSS Wallwork et al. (2012) 4. Functioning: LSP ⁎ p b 0.05. ⁎⁎ p b 0.01.

2. Negative PANSS 3F Kay et al. (1987)

– −0.28⁎

–

−0.18

0.97⁎⁎

0.07

−0.56⁎⁎

3. Negative PANSS 5F Wallwork et al. (2012)

Table 4 presents the correlation between baseline negative symptoms and the improvement in executive function and functioning by treatment group at post-treatment. As can be seen, there is a significant correlation between higher negative symptom scores at baseline and improvements in functioning at post-treatment in the REPYFLEC group while the non-significant correlation has the opposite sign in the control group. That pattern repeats for the improvements in executive functioning but statistical significance is not achieved. It seems that those patients who experience the greatest improvements in the REPYFLEC group are those with higher levels of negative symptoms at baseline. 4. Discussion Our results revealed that the improvement in functioning after receiving REPYFLEC CR would be mediated by NS but not by executive function. We found a significant effect in the experimental group compared with controls in improving executive function and NS, as well as an association between NS and functioning; although executive function did not appear to be consistently related to functioning in this context. Accordingly, there is a significant intervention effect on functioning mediated by NS, but not by executive function or through executive function and NS.

Table 4 Correlations between baseline negative symptoms and change in BADS and LSP at posttreatment. Baseline negative PANSS factor

– −0.56⁎⁎

BADS change at post-treatment LSP change at post-treatment ⁎⁎ p b 0.01.

Control REPYFLEC Control REPYFLEC

Kay et al. (1987)

Wallwork et al. (2012)

−0.16 0.25 −0.22 0.54⁎⁎

−0.12 0.18 −0.18 0.53⁎⁎

62

A. Farreny et al. / Schizophrenia Research 150 (2013) 58–63

The Kay et al. (1987) negative factor was cross-sectionally correlated to executive function at baseline, but it was not the case with the negative factor in Wallwork et al. (2012). We suggest that the negative factor by Kay et al. might include neurocognitive items such as abstract reasoning and stereotyped thinking which are not present in the negative factor as described by Wallwork et al. Thus, we conducted the mediation analysis with the five-factor model trying to avoid an overlap with neurocognition. As mentioned above, the correlation between neurocognition and NS may vary as a function of the negative construct definition (Harvey et al., 2006; Foussias and Remington, 2010). As has recently been suggested, in addition to examination of what people with schizophrenia and substantial negative symptoms are doing or not (more related to neurocognition), examination of their subjective anticipation of the consequences of engaging in positive everyday functional acts would seem to be a desirable component in the assessment of negative symptoms (Harvey, 2013). Another interesting result was observed in the longitudinal associations between NS at baseline and the results in executive function and functioning at post-treatment. The progress in functioning was correlated with baseline NS in the experimental group but not in controls, showing that patients who most enhanced their psychosocial functioning had higher levels of NS at baseline. On the other hand, the improvement in executive function did not appear associated with the level of NS at baseline in any group, although opposite signs were also observed in this correlation depending on the group. Some authors have previously suggested selectivity in the response to CR with better effects in those patients with higher NS than those without (Greenwood et al., 2005). Our results suggest that a certain level of NS could be hampering neurocognitive and functional improvements in patients with schizophrenia. This would be contrary to the stimulating theory that understands NS as a particular consequence of neurocognitive deficits (Frith, 1992), although recent findings have supported Frith's explanations on how positive symptoms might be associated with social cognitive processes (Ventura et al., 2013). According to Harvey et al. (2006), negative and neurocognitive symptoms could have different patterns of influence on real-world outcomes and it could be that neurocognitive performance is associated with the ability to perform everyday living skills, while NS are associated with the likelihood of performing these skills. Neurocognition would be correlated with functional and social abilities, but skill acquisition in these domains might not directly mirror cognitive gains, as negative symptoms also affect psychosocial outcomes and the cognitive factors leave much of the variance (Bowie et al., 2008; Lin et al., 2013). If so, even if a patient acquires certain skills but continues to experience mild levels of core and nonspecific symptoms, changes in real-world behavior might lag or not manifest at all (Bowie et al., 2010). Among the longitudinal studies which would support these results, Milev et al. (2005) assessed neurocognition and symptoms in a sample of 99 participants with first-episode psychosis in a 7-year, follow-up study. Results showed that in the prediction of global psychosocial functioning, NS severity was the most important factor, followed by attention and verbal memory (which was no longer a significant predictor after they accounted for NS and attention). In this study, initial NS explained 11% of the variance in outcome, while the NS score assessed at follow-up concurrently with outcome explained 47.4% of the variance. Similarly, Kurtz et al. (2005) found that negative symptoms, including blunted affect, social withdrawal, and alogia, were linked moderately to a variety of indices of functional status including thirdparty ratings of community function. Our findings raise the question of how the improvement in NS could also explain some effects of CR observed in psychosocial functioning. Various authors have stated that the negative symptoms that appear to have the greatest correlation with functional outcomes tend to be from the domain of motivational deficits, rather than the domain of reduced emotional expression (Ventura et al., 2009; Green et al., 2012; Harvey, 2013). Interestingly, Green and colleagues worked on a model

of outcome in schizophrenia ranging from micro-level early visual perception to macro-level daily community function. The key related question in this area was whether the measured ability (neurocognition: what one can do) and the measured motivation (negative symptoms: what one wants to do) act independently on functional outcome or whether they are part of a single pathway. A brief summary of their findings suggests that success in daily-living involves both what patients can do and also whether they are motivated to apply their abilities to the challenges of daily living. This supports the idea that therapy directed at dysfunctional thought can also yield improvements in negative symptoms in patients with schizophrenia (Green et al., 2012). Overall, we consider that the features of our study, such as the sample characteristics (i.e., age, course of illness, outpatient status), the contents based on problem solving and cognitive flexibility, and the method of CR employed (group format, experienced therapist, strategy teaching and repeated practice), could have determined the mediators in the therapeutic process. There are, however, some methodological caveats that should be mentioned. Firstly, neurocognitive assessment should be widened to include measures of working memory, attention, learning or processing speed to reach more a precise understanding of our CR context. There is strong evidence demonstrating an association between these neurocognitive domains and functional outcomes as well as an improvement after CR. Moreover, there is the possibility that other measures representing differentiated functional areas could have had an association with executive function in this context. Secondly, due to the size of the sample, the complexity of the models that could be estimated was limited. We were not able to introduce latent variables to take into account measurement error and the power to detect small but important associations may have been low. Finally, the mediators and the outcome were measured at the same time so the direction of the relationship between them was based on theory and not on temporal precedence. There is also the possibility that parallel mediators were left out of the analysis. Despite these limitations, the finding that particular levels of NS would block CR benefits could also explain some of the inconclusive, or even misleading, results in randomized controlled trials of CR in schizophrenia (e.g. Grynszpan et al., 2011). This study might contribute to defining the therapeutic targets which are crucial to functional improvement and developing efficacious treatments for schizophrenia. Role of funding source This research was supported by Fundació La Caixa and Instituto de Salud Carlos III (PI07/90476).

Contributors Aida Farreny, Susana Ochoa and Judith Usall designed the study. Jaume Aguado managed and undertook the statistical analyses with Aida Farreny and Josep Maria Haro. Aida Farreny wrote the first draft of the manuscript and all authors contributed to and have approved the final manuscript.

Conflict of interest None.

Acknowledgments We would like to express our gratitude to all the participants and professionals who were engaged in this study; and especial thanks to Stephen Kelly.

References American Psychiatric Association (APA), López-Ibor, J.J., Valdés, M. (dir.), 2002. DSM-IV-TR: Manual diagnóstico y estadístico de los trastornos mentales. Texto revisado. Masson, Barcelona. Ballesteros-Rodríguez, J., Martínez-Peña, L., Martín-Carrasco, M., Ibarra-Gandiaga, N., Bulbena-Vilarrasa, A., 2002. Assessment of the life skills profile and the brief psychiatric rating scale as predictors of the length of psychiatric hospitalization. Actas Esp. Psiquiatr. 30, 225–232.

A. Farreny et al. / Schizophrenia Research 150 (2013) 58–63 Bell, M.D., Mishara, A.L., 2006. Does negative symptom change relate to neurocognitive change in schizophrenia? Implications for targeted treatments. Schizophr. Res. 81, 17–27. Bowie, C.R., Leung, W.W., Reichenberg, A., McClure, M.M., Patterson, T.L., Heaton, R.K., Harvey, P.D., 2008. Predicting schizophrenia patient's real-world behavior with specific neuropsychological and functional capacity measures. Biol. Psychiatry 63, 505–511. Bowie, C.R., Depp, C., McGrath, J.A., Wolyniec, P., Mausbach, B.T., Thornquist, M.H., Luke, J., Patterson, T.L., Harvey, P.D., Pulver, A.E., 2010. Prediction of real-world functional disability in chronic mental disorders: a comparison of schizophrenia and bipolar disorder. Am. J. Psychiatry 167, 1116–1124. Farreny, A., Aguado, J., Ochoa, S., Huerta-Ramos, E., Marsà, F., López-Carrilero, R., Carral, V., Haro, J.M., Usall, J., 2012. REPYFLEC cognitive remediation group training in schizophrenia. Looking for an integrative approach. Schizophr. Res. 142, 137–144. Fernández de Larrinoa, P., Bulbena, A., Domínguez, A.I., 1992. Estudio de fiabilidad, validez y consistencia interna de la escala LSP (LifeSkillsProfile). Perfil de habilidades de la vida cotidiana. Actas Luso Esp. Neurol. Psiquiatr. Cienc. Afines 20, 71–75. Foussias, G., Remington, G., 2010. Negative symptoms in schizophrenia: avolition and Occam's Razor. Schizophr. Bull. 36, 359–369. Frith, C.D., 1992. The cognitive neuropsychology of schizophrenia. Lawrence Erlbaum Associates, Sussex. Fritz, M.S., MacKinnon, D.P., 2007. Required sample size to detect mediated effect. Psychol. Sci. 18, 233–239. Green, M.F., Kern, R.S., Braff, D.L., Mintz, J., 2000. Neurocognitive deficits and functional outcome in schizophrenia: are we measuring the “right stuff”? Schizophr. Bull. 26, 119–136. Green, M.F., Hellemann, G., Horan, W.P., Lee, J., Wynn, J.K., 2012. From perception to functional outcome in schizophrenia. Modeling the role of ability and motivation. Arch. Gen. Psychiatry 69, 1216–1224. Greenwood, K.E., Landau, S., Wykes, T., 2005. Negative symptoms and specific cognitive impairments as combined targets for improved functional outcome within cognitive remediation therapy. Schizophr. Bull. 31, 910–921. Greenwood, K.E., Morris, R., Sigmundsson, T., Landau, S., Wykes, T., 2008. Executive functioning in schizophrenia and the relationship with symptom profile and chronicity. J. Int. Neuropsychol. Soc. 14, 782–792. Grynszpan, O., Perbal, S., Pelissolo, A., Fossati, P., Jouvent, R., Dubal, S., Pérez-Díaz, F., 2011. Efficacy and specificity of computer-assisted cognitive remediation in schizophrenia: a meta-analytical study. Psychol. Med. 41, 163–173. Harvey, P.D., 2013. Assessment of everyday functioning in schizophrenia: implications for treatments aimed at negative symptoms. Schizophr. Res.. http://dx.doi.org/10.1016/ j.schres.2013.04.022. Harvey, P.D., Koren, D., Reichenberg, A., Bowie, C.R., 2006. Negative symptoms and cognitive deficits: what is the nature of their relationship? Schizophr. Bull. 32, 250–258. Kay, S.R., Fiszbeln, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. 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. Klingberg, S., Wölwer, W., Engel, C., Wittorf, A., Herrlich, J., Meisner, C., Buchkremer, G., Wiedemann, G., 2011. Negative symptoms of schizophrenia as primary target of cognitive behavioral therapy: results of the randomized clinical TONES study. Schizophr. Bull. 37, S98–S110. Kurtz, M.M., Moberg, P.J., Ragland, J.D., Gur, R.C., Gur, R.E., 2005. Symptoms versus neurocognitive test performance as predictors of psychosocial status in schizophrenia: a 1-and 4-year prospective study. Schizophr. Bull. 31, 167–174.

63

Lin, C.H., Huang, C.L., Chang, Y.C., Chen, P.W., Lin, C.Y., Tsai, G.E., Lane, H.Y., 2013. Clinical symptoms, mainly negative symptoms, mediate the influence of neurocognition and social cognition on functional outcome of schizophrenia. Schizophr. Res. 146, 231–237. Lipkovich, I.A., Deberdt, W., Csernansky, J.G., Sabbe, B., Keefe, R.S.E., Kollack-Walker, S., 2009. Relationships among neurocognition, symptoms and functioning in patients with schizophrenia: a path-analytic approach for associations at baseline and following 24 weeks of antipsychotic drug therapy. BMC Psychiatry 9, 44. http://dx.doi.org/ 10.1186/1471-244X-9-44. MacKinnon, D.P., 2008. Introduction to statistical mediation analysis. Lawrence Erlbaum Associates, New York. McGurk, S.R., Twamley, E.W., Sitzer, D.I., McHugo, G.J., Mueser, K.T., 2007. A meta-analysis of cognitive remediation in schizophrenia. Am. J. Psychiatry 164, 1791–1802. Milev, P., Ho, B.C., Arndt, S., Andreasen, N.C., 2005. Predictive values of neurocognition and negative symptoms on functional outcome in schizophrenia: a longitudinal firstepisode study with 7-year follow-up. Am. J. Psychiatry 162, 495–506. Penadés, R., Catalán, R., Puig, O., Masana, G., Pujol, N., Navarro, V., Guarch, J., Gastó, C., 2010. Executive function needs to be targeted to improve social functioning with Cognitive Remediation Therapy (CRT) in schizophrenia. Psychiatry Res. 177, 41–45. Peralta, V., Cuesta, M.J., 1994. Validación de la Escala de los Síndromes positivo y negativo (PANSS) en una muestra de esquizofrénicos españoles. Actas Luso Esp. Neurol. Psiquiatr. Cienc. Afines 22, 171–177. Rosen, A., Hadzi-Pavlovic, D., Parker, G., 1989. The Life Skills Profile: a measure assessing function and disability in schizophrenia. Schizophr. Bull. 15, 325–337. Schermelleh-Engel, K., Moosbrugger, H., Müller, H., 2003. Evaluating the fit of structural equation models: tests of significance and descriptive goodness-of-fit measures. Methods Psychol. Res. 8, 23–74. Ventura, J., Hellemann, G.S., Thames, A.D., Koellner, V., Neuchterlein, K.H., 2009. Symptoms as mediators of the relationship between neurocognition and functional outcome in schizophrenia: a meta-analysis. Schizophr. Res. 113, 189–199. Ventura, J., Wood, R.C., Hellemann, G.S., 2013. Symptom domains and neurocognitive functioning can help differentiate social cognitive processes in schizophrenia. Schizophr. Bull. 39, 102–111. Villalta-Gil, V., Vilaplana, M., Ochoa, S., Haro, J.M., Dolz, M., Usall, J., Cervilla, J., NEDENA Group, 2006. Neurocognitive performance and negative symptoms: are they equal in explaining disability in schizophrenia patients? Schizophr. Res. 87, 246–253. Vita, A., De Peri, L., Barlati, S., Cacciani, P., Deste, G., Poli, R., Agrimi, E., Cesana, B.M., Sacchetti, E., 2011. Effectiveness of different modalities of cognitive remediation on symptomatological, neuropsychological, and functional outcome domains in schizophrenia: a prospective study in a real-world setting. Schizophr. Res. 133, 223–231. Wallwork, R.S., Fortgang, R., Hashimoto, R., Weinberger, D.R., Dickinson, D., 2012. Searching for a consensus five-factor model of the Positive and Negative Syndrome Scale for schizophrenia. Schizophr. Res. 137, 246–250. Wilson, B.A., Alderman, N., Burgess, P.W., Emslie, H., Evans, J.J., 1996. Behavioural Assessment of the Dysexecutive Syndrome. Thames Valley Test Company, Bury St. Edmunds. Wykes, T., Spaulding, W.D., 2011. Thinking about the future cognitive remediation therapy—what works and could we do better? Schizophr. Bull. 37, S80–S90. Wykes, T., Newton, E., Landau, S., Rice, C., Thompson, N., Frangou, S., 2007. Cognitive remediation therapy (CRT) for young early onset patients with schizophrenia: an exploratory randomized controlled trial. Schizophr. Res. 94, 221–230. Wykes, T., Huddy, V., Cellard, C., McGurk, S., Czbor, P., 2011. A meta-analysis of cognitive remediation for schizophrenia: methodology and effect sizes. Am. J. Psychiatry 168, 472–485. Wykes, T., Reeder, C., Huddy, V., Taylor, R., Wood, H., Ghirasim, N., Kontis, D., Landau, S., 2012. Developing models of how cognitive improvements change functioning: mediation, moderation and moderated mediation. Schizophr. Res. 138, 88–93.