Funding sources: Funding for the study presented in this manuscript was provided by a NARSAD Young Investigator award to Dr. Bowie.
Cognitive remediation in schizophrenia: efficacy and effectiveness in patients with early versus long-term course of illness
Article first published online: 24 JAN 2013
© 2013 Wiley Publishing Asia Pty Ltd
Early Intervention in Psychiatry
Volume 8, Issue 1, pages 32–38, February 2014
How to Cite
Bowie, C. R., Grossman, M., Gupta, M., Oyewumi, L. K. and Harvey, P. D. (2014), Cognitive remediation in schizophrenia: efficacy and effectiveness in patients with early versus long-term course of illness. Early Intervention in Psychiatry, 8: 32–38. doi: 10.1111/eip.12029
Disclosures: Dr. Bowie has served as a consultant for Abbott Pharmaceuticals in the past year. Dr. Harvey has served as a consultant to Abbott Labs, Boehinger Ingelheim, Genentech, Johnson and Johnson, PharmaNeuro Boost, Roche Pharma, Shire and Sunovion Pharma in the past year. Dr. Oyewumi has served on the advisory board for Pfizer, Sunovion, Bristol Myers Squibb and Lundbeck in the past year. All other authors report no financial relationships with commercial interests.
- Issue published online: 27 JAN 2014
- Article first published online: 24 JAN 2013
- Manuscript Accepted: 16 SEP 2012
- Manuscript Received: 16 JUN 2012
- NARSAD Young Investigator
- cognitive remediation;
We examined the efficacy and effectiveness (transfer to functional competence and everyday functioning) of cognitive remediation in early-course (within 5 years of first episode) and long-term (more than 15 years of illness) schizophrenia.
Treatment lasted 12 weeks and included computerized exercises, strategic monitoring and methods to transfer cognition to behaviour. Assessments included a standard battery of neurocognition, performance-based measures of social and adaptive competence, and case manager ratings of real-world functional behaviour. Changes from baseline to post-treatment were examined with repeated measures analysis of variance and estimated premorbid intelligence and total months in hospital as covariates.
The early-course group had larger improvements in measures of processing speed and executive functions, as well as larger improvements in adaptive competence and real-world work skills. Duration of illness was inversely associated with improvement in neurocognition and real-world work skills.
Treatment of cognitive impairments is feasible in both early-course and chronic schizophrenia, but the clinical meaningfulness and generalization to functioning appear to be more substantial when delivered early. Cognitive remediation should be considered a tool for early intervention in schizophrenia.
Impairment in everyday functioning is profound in schizophrenia, even after successful treatment of clinical symptoms. Subtle deficits in daily functioning are evident before the onset of overt psychotic symptoms, they become broader and more severe after onset, and result in long-term disruption in independent living, educational attainment and community participation. In addition, periods of hospitalization and symptom exacerbations can interrupt the acquisition of adaptive living skills, further complicating the recovery process for patients who are in a critical stage of social and vocational development. These relapses may also lead to changes in brain structure, which likely have additional cognitive consequences.
Efforts to promote symptomatic recovery in the early stages of psychosis have traditionally focused on medication adherence and relapse prevention to manage symptoms and forestall the likelihood of conversion to schizophrenia. Despite the well-documented efficacy of timely pharmacological interventions, sustained recovery is seldom achieved from pharmacotherapy alone, as these treatments fail to significantly improve functional impairments.[8, 9]
In recent years, the conceptualization of recovery has shifted from a symptom-reduction model to also include the attainment of meaningful roles in society.[10, 11] There is a growing body of evidence that suggests barriers to functional recovery are associated with a host of neurocognitive impairments in both the early and later course of illness. Difficulties with attention, processing speed, memory and executive functions are now regarded as stable, core features of schizophrenia that robustly predict chronicity and contribute to functional outcomes to a greater extent than clinical symptoms.[14, 15] Early-course patients perform 0.3 to 1.0 standard deviations better than groups of long-term course patients on neurocognitive abilities.[16-19] Declines in neurocognition are observed in the first 10 to 12 years of psychosis despite good clinical outcomes. Taken together, these findings underscore the importance of targeting neurocognitive impairments early in the course of illness to decrease the severity of associated functional disability.
Given the significant influence of neurocognitive domains on daily functioning, concerted efforts have been made to develop behavioural treatments that address more functionally relevant aspects of recovery. At the forefront of these treatments are cognitive remediation therapies (CR) to enhance a range of cognitive processes that are necessary for productive everyday functioning. Contemporary models of CR often utilize computer-based exercises in conjunction with strategy coaching to foster problem-solving skills, and have yielded moderate to large effects on the cognitive outcomes of long-term course schizophrenia patients. Cognitive gains following CR persist long after the completion of treatment, increase patients' response to other therapeutic interventions and generalize to diverse aspects of daily living, including psychosocial functioning, work[26, 27] and personal autonomy.
Despite the widely cited benefits of CR in long-term course patients, fewer studies have examined the extent to which CR has practical implications in the early stages of psychosis. Wykes and colleagues, in sample aged 14 to 22 years old, found clinically significant gains in cognitive flexibility, although social behaviour at a 6-month follow up was not significantly different from a control group. However, the improvements in cognitive functioning were associated with positive social behaviour outcomes. In a two 2-year CR for early-course outpatients, Eack and colleagues reported marked improvements in neurocognition and social cognition that extended to occupational functioning, emotion management, and had neuroprotective effects against grey matter loss in regions of the brain implicated in learning and memory.
The goal of the present study was to build upon previous research and evaluate the clinical utility and generalizability of CR in the early course of illness compared to more chronically ill patients. With less severe impairment likely in the early-course group, it is anticipated that gains will be greater and more clinically meaningful. A limitation of previous examinations of CR in the early course of schizophrenia is the lack of assessment of transfer of cognitive change to functioning. In relatively short-term treatments and/or follow-up periods, it might not be reasonable to expect changes in real-world behaviour to be as large as the ability to complete tasks. This might be especially true for individuals who, in spite of only a short duration of illness, likely have long-standing difficulties navigating environments that require social and instrumental living skills, and who, following illness onset, might not be immersed in environments that set the occasion for those behaviours. Thus, in the present study, we were interested in distinguishing between generalization of cognitive change to measures of competence, which indexes what one is able to do in the lab, versus performance, which relates to the actual behaviour in the community. It was hypothesized that, relative to long-term course patients, greater transfer of cognitive gains from CR would appear in the early course of illness for measures of functional competence and real-world behaviour. This is a reanalysis of data from a larger study focusing on the subset of early-course patients.
Clinicians at outpatient centres affiliated with three sites, Mount Sinai School of Medicine, Bronx VA Medical Center and Queen's University, referred to the parent study that examined cognitive remediation and functional skills training treatments independently or combined. Inclusion criteria were (i) diagnosis of schizophrenia or schizoaffective disorder, (ii) enrolled in outpatient psychiatric treatment, (iii) reading level at or above grade 6, (iv) age 18 to 65. Exclusion criteria were (i) psychiatric hospitalization greater than 24 hours within 1 month of baseline and (ii) comorbid psychiatric or neurological conditions. In this study, we selected, from a cognitive remediation arm, a subgroup of 12 patients in early-course psychosis, defined as onset of symptoms within 5 years of study entry (M = 3.4, SD = 1.1) and 27 patients with a more long-term course, defined by at least 15 years of illness (M = 25.7, SD = 6.6). Please see Table 1 for demographic and clinical characteristics of the groups. All participants provided written, informed consent approved by the local ethics board (Clinical Trial Registration Number: NCT01175642).
The Brief Assessment of Cognition in Schizophrenia includes six subtests that are administered to assess the domains of psychomotor speed (token motor task), information processing speed (symbol coding), verbal fluency (phonological and semantic fluency), working memory (digit sequencing), executive functioning (Tower of London) and verbal memory (list learning). Raw scores are transformed to z-scores (M = 0, SD = 1) based on performance relative to the age- and gender-matched healthy normative sample scores for each test were averaged for a neurocognitive composite score.
Severity of clinical symptoms was examined with the Positive and Negative Syndrome Scale (PANSS) based on chart review and a structured interview with the patient and clinician. This scale includes seven positive symptoms, seven negative symptoms and 16 general symptoms. Inter-rater reliability for the PANSS for our raters ranged from 0.77 to 0.87. For the present analyses, we report the total score.
Social competence was examined using the Social Skills Performance Assessment. Participants engage in role-play tasks that simulate two everyday social situations: greeting a new neighbour and attempting to persuade a landlord to fix a leak. Sessions were audio recorded and scored by a trained rater who was blind to diagnostic, cognitive and treatment status. Dimensions of social competence scored include interest, affect, fluency, clarity, focus, negotiation ability, persistence and social appropriateness. The rater was trained to the gold standard ratings proposed by the instrument developers (intraclass correlation coefficient = 0.84). We report total scores for the two scenes; higher scores indicate better performance.
Functional competence refers to skills that are important for independent functioning. We included three measures of everyday functional skills, and computed a composite score with equal weights across the three measures, which was converted to a standard score ranging from 0 to 100; higher scores indicate better performance. The University of California, San Diego Performance-Based Skills Assessment Battery includes a series of role-play tasks that assess planning for recreational activities, financial skills, telephone use and use of public transportation. In the Advanced Finances Test, participants perform higher level financial management skills, including depositing a check, paying several bills, planning to leave money in the account and balancing a check book. The Medication Management Ability Assessment is a role-play task similar in complexity to what someone receiving multiple medications is likely to encounter. Correct responses include taking the correct medication at the appropriate time, with or without food, the correct number of doses per day and the correct number of pills per dose.
Real-world functional behaviour
Real-world functional behaviour was rated using the Specific Levels of Functioning (SLOF), which is an observer-rated scale of patients' behaviour and functioning. Ratings by caseworkers, who indicated that they know patients ‘well’ or ‘very well’, rated items based on the amount of assistance required to perform real-world skills or frequency of the behaviour on the 5-point Likert scale on three domains: interpersonal relationships (e.g. initiating, accepting contacts), community/household activities (e.g. shopping, paying bills, use of leisure time) and work skills (e.g. employable skills, level of supervision required to complete tasks, ability to stay on task, punctuality).
Assessments were completed at baseline and within one week following treatment. Testing technicians completed the assessments of cognition and functional competence and Masters or PhD level examiners conducted the symptom interviews. All parties were unaware of the other two sources of data, as well as to the group assignment, study hypotheses and study design.
Cognitive remediation was adapted on a previously validated treatment and is described in detail elsewhere. We used computer-based exercises,[45-47] and included a large role for therapist involvement to stimulate forming and testing of alternative strategies and discussing how the cognitive skills are used in everyday life. Bridging strategies to help transfer cognitive skills to everyday activities were informed by the Thinking Skills for Work Program and the Neuropsychological Educational Approach to Remediation programs. The primary author was responsible for ensuring the training, adherence to protocol and review of session notes on a weekly basis. Two therapists, doctoral level clinical psychologists or doctoral clinical psychology students, were used at each site.
Univariate analyses of variance were used to examine group differences on demographic and illness variables at baseline. Repeated measures analysis of variance tests, with group (early course and long-term course) as the between-subjects factor and time (pre- and post-treatment) as the within subjects factor, were used to evaluate changes in neurocognition, competence and functioning variables. Significant baseline differences were observed between groups on estimated premorbid intelligence and total number of months hospitalized. Since these variables are associated with the dependent measures of cognition and everyday functioning, we entered them as covariates in the model. Effect sizes are reported with partial eta squared, where 0.009 corresponds to a small effect, 0.058 a medium effect and 0.137 a large effect. Pearson correlation analyses were performed to examine the linear relationship of duration of illness with residual change scores.
The early-course group had less severe impairment at baseline on digit sequencing (F(1,35) = 7.2, P = 0.011) and verbal fluency (F(1,35) = 6.2, P = 0.016), with a trend for less severe impairment on observed verbal memory (P = 0.060). All other cognitive variables were not significantly different (all P-values > 0.360). The early-course group had better baseline performance on the adaptive competence composite (F(1,35) = 5.0, P = 0.030), but group differences were not observed at baseline for social competence (P = 0.910). The two groups were not significantly different on the SLOF domains of interpersonal behaviour, community/household activities and work skills. See Table 2 for means and standard deviations at baseline.
|Early course||Long-term course|
|Total months in hospital||8.7||9.4||43.5||42.1||6.5||0.01|
|Total PANSSa score||73.3||20.8||74.4||17.9||0.03||0.86|
|Treatment dose (percentage completed)||0.96||0.06||0.94||0.08||0.46||0.50|
|Early course||Long-term course||Main effects of treatment||Effect size||Group by treatment||Effect size|
|Pretreatment||Post-treatment||Pretreatment||Post-treatment||F (d.f.)||P-value||n2||F (d.f.)||P-value||n2|
|NCS||−1.3 (0.87)||−0.49 (0.56)||−1.8 (1.1)||−1.2 (0.94)||175 (1,35)||<0.001||0.822||4.5 (1,35)||0.040||0.10|
|Verbal memory||−0.75 (0.95)||−0.01 (1.1)||−1.6 (1.2)||−1.1 (1.1)||1.7 (1,35)||0.210||0.045||1.9 (1,35)||0.170||0.05|
|Token motor||−1.5 (1.9)||−0.96 (1.6)||−1.6 (1.2)||−1.3 (1.0)||3.6 (1,35)||0.067||0.090||6.2 (1,35)||0.017||0.15|
|Digit sequencing||−0.96 (1.2)||−0.02 (.87)||−2.3 (1.6)||−1.5 (1.1)||24.1 (1,35)||<0.001||0.410||3.4 (1,35)||0.076||0.09|
|Tower of London||−2.6 (1.6)||−.96 (1.7)||−2.5 (1.7)||−1.5 (1.7)||9.7 (1,35)||0.004||0.220||7.5 (1,35)||0.009||0.18|
|Verbal fluency||−0.28 (1.2)||0.07 (1.0)||−1.3 (1.2)||−.76 (1.2)||0.7 (1,35)||0.390||0.020||0.6 (1,35)||0.420||0.02|
|Symbol coding||−2.3 (0.79)||−1.1 (0.78)||−2.0 (0.66)||−1.4 (0.71)||3.8 (1,35)||0.057||0.100||4.1 (1,35)||0.049||0.11|
|Adaptive competence||0.65 (0.16)||0.72 (0.13)||0.51 (0.17)||0.54 (0.16)||3.7 (1,35)||0.063||0.095||4.9 (1,35)||0.032||0.13|
|Social competence||0.68 (0.09)||0.70 (0.09)||0.69 (0.15)||0.70 (0.14)||0.3 (1,35)||0.540||0.011||0.1 (1,35)||0.780||0.002|
|SLOF interpersonal||0.78 (0.11)||0.79 (0.10)||0.74 (0.12)||0.76 (0.11)||2.6 (1,35)||0.110||0.071||0.0 (1,35)||0.980||<0.001|
|SLOF activities||0.69 (0.17)||0.72 (0.12)||0.76 (0.11)||0.79 (0.11)||2.7 (1,35)||0.110||0.074||0.1 (1,35)||0.810||0.002|
|SLOF work skills||0.54 (0.11)||0.61 (0.11)||0.53 (0.11)||0.55 (0.10)||0.5 (1,35)||0.480||0.015||5.7 (1,35)||0.022||0.14|
Main effects of treatment
Significant main effects for treatment were observed for digit sequencing and Tower of London, with statistical trends supported by medium effect sizes for token motor task and digit symbol coding. Verbal memory and verbal fluency did not improve significantly and the effect sizes were small. Adaptive competence had a trend for a main effect of treatment with a medium effect size, whereas social competence did not improve and the effect size was small. Main effects for treatment were not significant, but we observed a medium effect size for SLOF interpersonal behaviour and SLOF community/household activities. Non-significant and small effects were observed for SLOF work skills. See Table 2 for main effects test results.
Group by treatment interaction effects
Statistically significant group by treatment interaction effects in favour of the early-course group were observed for token motor task, Tower of London and symbol coding, with a trend and medium effect observed for digit sequencing. adaptive competence, but not social competence, improved more in the early-course group. Clinician ratings of real-world work skills also improved more in the early-course group, but treatment effects for ratings of interpersonal behaviour and community/household activities did not differ by group. See Table 2 for interaction effects.
Correlation of duration of illness with treatment effects
Duration of illness for the entire sample ranged from 1 to 43 years (M = 18.5, SD = 10). Duration of illness was significantly inversely associated with change scores on the neurocognitive composite (r = −0.43, P = 0.001) and improvement in real-world work skills (r = −0.27, P = 0.020), and a modest but not statistically significant improvement in adaptive competence (r = −0.20, P = 0.070). Duration of illness was not significantly associated with change in social competence (r = −0.04, P = 0.390), real-world community/household activities (r = −0.12, P = 0.210) or real-world interpersonal behaviour (r = 0.01, P = 0.460).
Cognitive remediation has emerged as an effective treatment in schizophrenia. In addition to improving cognitive abilities, the treatment can result in significant improvement in functioning, although this generalization is most often observed when CR is embedded within a broader psychosocial treatment programme. Data from the parent study for the present data support the improved generalization of cognitive remediation when it is paired with habilitative approaches for teaching skills necessary for successful functioning in the community. The cost and duration of combining treatments, however, might be prohibitive in some health-care settings and to some patients. This study explored whether CR alone might be more efficacious and effective for a subgroup of individuals who were in the early stages of psychosis and who might not have experienced long periods of missed opportunity to acquire skills. We did not examine combined therapy in these analyses because that therapy was effective at improving everyday outcomes for the sample as whole, including both long-term course and early-course patients. In our larger study, CR alone did not improve everyday outcomes; in the early-course patients, improvements were detected. Thus, very early in the course of illness, additional intensive psychosocial interventions may be less necessary, possibly because patients have not already developed a decades-long pattern of behaviour of reduced functional activities and associated ‘atrophy’ of other adaptive skills.
Our findings suggest that despite significant improvements in both groups, there is larger improvement in several cognitive domains for those in the early stages of psychosis. These superior effects were most evident in domains where performance was similar between groups at baseline, including psychomotor speed, complex information processing speed, and executive functions related to working memory and planning. The participants who were in the early course of psychosis had, in other domains, less severe impairment at baseline. These findings suggest that if cognitive remediation is utilized in the early stages of psychosis, it might have more robust, clinically meaningful and functionally relevant effects.
Early detection and intervention has become increasingly recognized as a viable method for reducing the adverse outcomes associated with untreated psychosis.[50, 51] Given that the onset of psychosis often occurs during a critical developmental period, our data suggest that cognitive remediation might hold potential for recovery to premorbid levels for those in the initial stages of illness. There is preliminary evidence supporting the utility of CR in the prodromal phase of schizophrenia[52, 53] and further investigations are warranted to confirm and extend these findings using more rigorous experimental designs. Moreover, future research should continue to examine the optimal timing of CR, which might be prior to onset for those at risk for psychosis, as this intervention becomes more mainstream and well integrated into research and clinical settings.
A limitation of this study was the use of samples of convenience from a larger study, which resulted in some baseline differences between groups. Although we adjusted for group differences, future work could prospectively recruit participants matched on demographic factors and degree of impairment at baseline. Studies with longer follow-up periods might examine the degree to which treatment effects as a function of illness duration are durable and result in sustained changes in behaviour, including milestones like returning to work or school, which are not likely to be observable within the context of short-term treatment trials. Our data suggest that cognitive remediation might be more efficacious and effective for those with shorter illness duration. It will be important to continue to examine the appropriate timing of this intervention, which might be prior to onset for those at risk, as it becomes more mainstream.
The authors thank Thomas Patterson, Brent Mausbach and Susan McGurk for their role in the training of research staff. The authors thank Hannah Anderson, Winnie Leung, Amelia Bowman, Shannon Xavier, Stephanie Taillefer, Emma Bassett, Jeremy Stewart, Allisha Patterson, Talia Troister and Sylvia Magrys for assistance with data collection and delivering the treatment.
- 21Neurocognitive and social cognitive approaches for improving functional outcome in early psychosis: theoretical considerations and current state of evidence. Schizophr Res Treatment 2012; 2012: 1–15., .
- 30Cognitive enhancement therapy for early-course schizophrenia: effects of two-year randomized controlled trial. Psychiatr Serv 2009; 60: 1468–1476., , et al.
- 33Neuroprotective effects of cognitive enhancement therapy against gray matter loss in early schizophrenia. Arch Gen Psychiatry 2010; 67: 674–682., , et al.
- 45Cogpack. Version 5.1. Ladenburg, Germany: Marker Software, 2007.
- 46PSSCogRehab. Indianapolis, IN: Psychological Software Services Inc; 1994..
- 47Scientific Brain Training PRO. [Cited 8 August 2012.] Available from URL: http://www.scientificbraintrainingpro.com/
- 49Statistical Power Analysis for the Behavioral Sciences, 2nd edn. Hillsdale, NJ: Lawrence Earlbaum Associates, 1988..