Abstract
Introduction
Vocational rehabilitation with the goal of gaining competitive employment has become a core ingredient in the mental health care for many people with schizophrenia [1]. However, most of them are hindered in their endeavors to gain and keep work due to impairments in cognitive functioning [2–6]. Unlike more short-lived psychotic symptoms, cognitive impairments tend to last longer [7]. In vocational rehabilitation, the goal is to enter or re-enter work [8], and addressing impairments in cognitive functioning is vital and may affect the prospect of obtaining competitive employment [2]. Furthermore, participants may not benefit fully from vocational rehabilitation due to their cognitive impairments [9]. Studying the relationship between cognitive functioning and work within a vocational rehabilitation and mental healthcare setting is vital, especially since most people with schizophrenia express an interest in working, consider themselves able to work, and express the need for vocational rehabilitation in order to reach their employment goals [10].
Impairments in cognitive functioning among people with schizophrenia are well documented and described as a core feature of the illness [7, 11–15]. They generally perform 0.5 to 1.0 standard deviations below the means of people without schizophrenia in cognitive tests [16]. Additionally, impairments in cognitive function have shown to pre-date the onset of the illness and have shown to be generally stable throughout the illness [7, 16]. To date, cognitive functioning is considered the most reliable predictor of occupational functioning in this group of individuals [3, 15]. However, this does not apply specifically to studies on vocational outcomes in vocational rehabilitation [3]. The National Institute of Mental Health’s Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) has identified seven separate areas of cognitive impairments in schizophrenia [11], and has developed a consensus cognitive test battery that is an accepted standard for assessing cognitive functioning in these areas [17]. The separate areas are: 1) speed of processing, 2) attention/vigilance, 3) working memory, 4) verbal learning and memory, 5) visual learning and memory, 6) reasoning, and problem-solving, and 7) social cognition [11]. In a review on cognitive impairments and functional outcomes among people with schizophrenia [7], cognitive impairments in working memory, verbal learning and memory, and executive functions were found to be moderate to severe. They were related to functioning in everyday life, including the ability to perform ajob.
In vocational rehabilitation research on supported employment [2–4], impairments in higher order cognitive functions, (verbal learning and memory, and executive functions), predicted vocational outcomes, including gaining competitive employment, income, hours worked, and intensity of vocational support. Additionally, vocational rehabilitation research that focuses on the actual work situation that people with psychiatric disabilities have, cognitive disabilities such as problem-solving related to work tasks, and learning new work tasks and routines have been found to be the most prevalent disabilities [18–22]. In the present study, executive functions are referred to as being “linked to intentionality, purposefulness, and complex decision-making”. These functions include “identifying, initiating, and pursuing goals, planning strategies and sequencing steps of action, solving problems, monitoring progress, and adjusting one’s behavior to circumstances” [23] (p. 90). In line with this definition, and according to Lezak et al. [24] and MATRICS [11], the planning, reasoning, and problem-solving skills form part of the complex construct of executive functioning. In the present study, the planning, reasoning, and problem-solving are operationalized in the same way. We expected that planning, reasoning, and problem-solving skills would be related to vocational outcomes, especially since work performance and the ability to benefit from vocational rehabilitation may rely on this cognitive domain [3, 25]. Furthermore, in vocational rehabilitation research on supported employment, the results show that the support provided can to some extent compensate for impairments in higher order cognitive function by, for example, helping clients find jobs that match their cognitive abilities [2, 3]. The review of the quantitative research, however, suggests that these services may not fully compensate for impairments in higher order cognitive functioning in verbal learning and memory and executive function [3]. In order to enhance the understanding of whether cognitive function may be decisive for getting a job or not, and of how to provide better matching, compensatory and training opportunities in relation to vocational rehabilitation, we need to extend the knowledge base and further investigate the relationship between cognitive function and vocational outcomes among people with schizophrenia. In particular, a number of cognitive areas need to be addressed.
Employment is a central goal, and vital for recovery of people with schizophrenia [26]. Gaining employment is thus the primary outcome variable in vocational rehabilitation practice within mental healthcare and research [1]. We believe the focus on employment in the present study adds to the knowledge base about which cognitive functions may be essential for whether or not someone gets work. Previous studies have addressed a similar research focus and target group [4, 25]. However, these studies used a prospective cohort design and concern, for example, cognitive predictors of vocational rehabilitation outcomes among a small number of primarily male participants in an American context. In the present study, we aimed for a larger group of participants with mixed gender living in Sweden, who participate in 18 months of vocational rehabilitation integrated with mental healthcare services. The aim was to investigate relationships between cognitive functioning and vocational variables of competitive employment, work hours per week, and monthly income among people with schizophrenia in vocational rehabilitation within a mental healthcare context. The rationale for including the work hours and income is twofold. First, vocational outcomes are commonly addressed in this line of research and provide a more comprehensive picture of what is entailed in gaining competitive employment. Second, the outcomes are studied within a cross-sectional context. A further aim was to investigate which area of cognitive function is essential for obtaining competitive employment, which has not been studied before. Based on the aforementioned research, we hypothesized that there is a positive relationship between better higher order cognitive functioning and gaining competitive employment, more work hours per week, and higher income.
Methods
This is a cross-sectional study conducted at four mental healthcare services in a city, with a population of approximately 300,000 in the Skåne region of southern Sweden.
Participants and eligibility
The selection criteria for participants were 1) having schizophrenia, 2) aged 20 to 63 years, 3) ability to communicate in Swedish, and 4) 18 months of vocational rehabilitation, i.e. individualizedsupport linked to facilitation of gaining and keeping employment, led by either an employment specialist or case-manager who worked on the mental healthcare team. All participants had the goal of gaining competitive employment. Both approaches focused on supporting the participants in attaining their employment goals, provided job-search support, and collaborated with professionals in the mental healthcare team. The main difference between the approaches was that the case-manager function did not allow time-unlimited support at work. There were no significant differences in cognitive functioning between those who received support from an employment specialist and those who received support from a case-manager.
The team member who had the most regular contact with possible participants (n = 49) introduced the study with verbal and written information. Thirty-nine adults provided written consent to participate. Research assistants then contacted participants and arranged an appointment for data collection. The study was approved by the Regional Ethics Board, Lund University, Sweden (Dnr 700/2009).
Data collection
Each of the 39 participants was evaluated with a cognitive test battery. Socio-demographic, clinical, and vocational data were gathered. The data collection for each participant involved two sessions and occurred at the outpatient unit most familiar to the participant. Each session lasted approximately one hour. During the first session, socio-demographic, clinical, and vocational data were recorded, and during the second session, cognitive testing was performed. All data collection, including the cognitive tests, was administered the same way for each participant. Data were collected between September and November 2010.
Cognitive functioning
Cognitive functioning was assessed by a trained psychologist using a cognitive test battery. This corresponded to the cognitive domains used in McGurk and Mueser’s study [4], and the consensus test battery of MATRICS [17]. The cognitive tests have good psychometric properties of reliability and validity (see each test reference below). High scores reflect better cognitive function, except for the Trail Making Test, where a high score indicates poorer performance.
Attention and psychomotor speed were assessed with the Trail Making Test (TMT) condition 2: Number Sequencing (Delis-Kaplan Executive Function System, D-KEFS) [27]. In the TMT, the raw score is the time in seconds needed to complete the task of connecting numbers in ascending order.
Information processing speed was measured with the Wechsler Adult Intelligence Scale-III (WAIS-III) Digit Symbol Substitution Test [28, 29]. In this test, the participant is asked to copy symbols paired with numbers (1 to 9) under a 120-second limit. The correct number of accurately copied symbols is recorded.
Attention span was estimated using digit span forward, and working memory with the digit span backward (WAIS-III) [28, 29]. The participant was asked to repeat a given number string with increasing length as presented (digit span forward) or in reverse order (digit span backward). The total number of correct responses for each measure is recorded.
Verbal learning and memory – immediate and delayed verbal recall was assessed using Wechsler Memory Scale, third edition (WMS-III) Logical Memory 1 and 2 subscales [30, 31]. The test comprises two short stories that are read to the participant. After each story, the participant is asked to retell the story from memory (Logical Memory 1, immediate verbal recall). Following a delay of 30 minutes, the participant is again asked to retell the story (Logical Memory 2, delayed verbal recall). Total number of correctly remembered details and themes from the stories is recorded.
Visual memory – immediate and delayed visual recall was assessed using WMS III Visual Reproduction 1 and 2 subscales [30, 31]. Five figures are shown to the participant, one at a time for 10 seconds each. The participant is then immediately asked to draw each figure from memory (immediate visual recall). Following a delay of 30 minutes, the participant is again asked to draw the figures from memory (delayed visual recall). Total number of points for each subscale is recorded.
Planning, reasoning, and problem-solving was assessed using Tower of LondonDX [32]. The task requires the manipulation of three wooden balls onto different sized rods to create a given configuration, and occurs across ten levels of increasing complexity, and according to given rules. Total number of problems solved in a minimum move count, i.e. Total Correct Score, is recorded.
Socio-demographic and vocational data
Socio-demographic information, i.e. age, gender, civil status, ethnicity, education level, and vocational data, i.e. competitive employment, work hours per week, and monthly income were collected by an interview-based questionnaire. The vocational data were validated against the vocational logbooks used in vocational rehabilitation. Competitive employment was defined as payment of at least minimum wage, located in a mainstream community and socially integrated setting, and not being exclusively for people with mental health problems.
Clinical data
The mental health diagnosis was self-reported and later verified against medical records according to the diagnosis system International Classification and Diseases, ICD-10 [33]. The Brief Psychiatric Rating Scale (BPRS) [34] was used to assess symptoms. BPRS allows for analysis into positive, negative, and depressive symptoms, and general psychopathology. Based on an interview and observation, the included 18 items are scored by the interviewer on a scale ranging from 1 to 7, with a total sum score of 126 points. A higher score indicates more severe symptoms. The interviewers were trained to use BPRS since it is shown to increase further inter-rater and intra-observer reliability [35]. Cronbach’s alpha was calculated to 0.71.
Data analyses
The cognitive tests were scored and evaluated by a neuropsychologist (CH) with experience in using the instruments and psychometric testing. A P-value ≤ 0.05 was considered statistically significant. Median, mean, interquartile range (IQR), and standard deviation (SD) were calculated for the cognitive tests. Group mean raw scores for all cognitive tests were transformed into standard normative scores for each test. For the tests collected from the D-KEFS, WAIS-III, and the WMS-III batteries [27, 30], the corresponding age-scaled score (age 35–44, mean = 10, SD = 3) equivalents were used. For the Tower of London, the standard score (age 30–39 years, mean = 100, SD = 15) was applied [32]. A scaled/standard score 1.5 below the normative mean was considered below normal, in keeping with the recommendations of Lezak et al. [24].
The relationships between cognitive functioning and vocational variables were analyzed in two steps, using non-parametric statistics. In a first step, the Mann-Whitney U test was used to analyze group differences between individual with competitive employment and those who were not employed in relation to cognitive variables. Spearman’s rho was used to determine the magnitude and direction of the correlations between continuous vocational and cognitive variables. In a second step, a binary logistic regression analysis with forward Wald selection was conducted to explore further the importance of the cognitive variables in relation to the dependent variable (ie, competitive employment, yes or no). Independent cognitive variables had to differ (P ≤ 0.05) between groups in the first step of the analysis in order to be included. The independent cognitive variables were attention and psychomotor speed, delayed verbal recall, immediate visual recall, and planning, reasoning, and problem-solving. Age and negative symptoms were controlled for as these variables are associated with vocational outcomes in prior studies [4, 36]. The confidence interval was set at 95%. Data were analyzed using IBM SPSS Statistics, Version 20.
Results
Socio-demographic characteristics are presented in Table 1. The sample comprised 39 participants with a mean age of 38.2 years (SD = 7.8) diagnosed with schizophrenia.
Descriptive statistics for cognitive tests are presented in Table 2. When compared to population norms, the participants generally performed 0.5 to 1.0 standard deviations below the means except for the Tower of London, where they performed 2 standard deviations below the norm.
Cognitive functioning in relation to vocational variables
As shown in Table 3, there was a significant difference in attention and psychomotor speed (competitive employment median of 34, no competitive employment median of 52, z = –1.983, P = 0.047), delayed verbal recall (competitive employment median of 24, no competitive employment median of 12, z = –2.387, P = 0.017), immediate visual recall (competitive employment median of 84, no competitive employment median of 69, z = –2.017, P = 0.044), and planning, reasoning, and problem-solving (competitive employment median of 5, no competitive employment median of 1, z = 2.653, P = 0.008) between those with competitive employment (n= 11) and those without (n= 28). Higher scores in immediate and delayed verbal recall, immediate visual recall, and planning, reasoning, and problem-solving were significantly correlated with more work hours per week. Poorer performance on the attention and psychomotor speed test were correlated with lower income. There was also a trend towards significant correlation (P = 0.053) between poorer performance on the attention and psychomotor speed test and less work hours per week. Furthermore, higher scores in immediate and delayed verbal recall, and planning, reasoning, and problem-solving were significantly related to higherincome.
Logistic regression analysis with forward Wald selection was conducted to investigate which cognitive variable was the most important factor for distinguishing between having competitive employment or not. Planning, reasoning, and problem-solving best indicated if the participant had competitive employment. The logistic regression model was significant, χ2 (1, n = 38) = 5.656, P < 0.017 and explained between 14.2% (Cox and Snell R-square) and 21.1% (Nagelkerke R-square) of the variability in competitive employment, and correctly classified 78.4% of cases. The probability of having competitive employment increased by 48% (OR = 1.48, 95% CI = 1.04–2.093, P = 0.029) for each point on the planning, reasoning, and problem-solving test.
Discussion
Higher scores on the planning, reasoning, and problem-solving test were associated with having competitive employment, more work hours per week, and higher income. These findings corroborate our hypothesis and the findings of previous research [4, 25]. The planning, reasoning, and problem-solving variable was also best able to distinguish between those with and without competitive employment. These results are consistent with previous research that describes planning, reasoning, and problem-solving as a hallmark of cognitive impairment among people with schizophrenia [7]. According to Lezak et al. [24] and Kielhofner [23], impairments in higher order cognitive functions of planning, reasoning, and problem-solving, often result in global limitations in managing daily life. This is true regardless of the degree of preservation of other cognitive capabilities. Inflexible thinking and planning, limitations in working towards work-related goals, and difficulty adapting to changes may be involved [7]. Limitations in this cognitive area may be part of the reason why not everyone with serious mental illness who wants to work gains employment, works full-time, or maintains employment once they have started. Simultaneously, impairment in planning, reasoning, and problem-solving is described as not being easily compensated for in vocational rehabilitation [3, 6]. Cognitive training, in addition to vocational rehabilitation, could be anticipated to enhance work outcomes for this group of people when combined with individualized vocational support [9, 37]. In addition, an appropriate match between a person’s cognitive functioning and the job complexity may be a helpful compensatory strategy for such limitations in work performance [4, 22]. Furthermore, vocational support when learning new work tasks, such as extended employee training, supervision, and Cognitive Behaviour Therapy (CBT) strategies for solving problems may also be helpful [20]. Moreover, the teaching of new strategies for the learning of work tasks by having a greater focus on novel tasks and performing tasks repetitively may be another helpful strategy to compensate for the effects of cognitive impairments on work performance [7].
Better results on the verbal learning and memory – immediate and delayed verbal recall tests were correlated with more work hours per week and a higher income. These relationships were part of our hypothesis and found in prior research [15, 25]. Impairments in verbal learning and memory are described as another hallmark of cognitive impairment [7, 16], and one of the most consistent findings across quantitative studies in the field. Verbal learning and memory has been shown to be positively correlated with job-task complexity [4], related to reduced learning capacity because of difficulty remembering instructions, poorer work habits and quality [7], and the need for more intensive vocational support [25]. A qualitative study on supported employment also found this [21]; participants perceived that limitations in learning new work tasks resulted from poorer recall of information related to work tasks and instructions for work routines. Perhaps the ability to learn and remember instructions and new work routines plays a central role in performing work tasks. If successfully managed, this might lead to longer work hours and thus higher income. Interestingly, verbal memory is emphasized as important for sustainability of work performance, while other cognitive functions are more important for initial success [38, 39]. Such findings summarize the key results of our study. What is encouraging to note, is that impairments in verbal learning and memory can be compensated to a certain extent by an appropriate job-person match, and vocational support such as memory aids (e.g. memos, step-by-step instructions, hands-on supervision, and modified employment training) [19, 20].
Better performance on the attention and psychomotor speed test was correlated with having employment and higher income, which is in line with previous research [3]. There was also a trend towards significant correlation between better performance on the attention and psychomotor speed test and more work hours per week. These relationships might be explained by impairments in basic cognitive functions such as in attention and psychomotor speed interfering with more complex higher order cognitive functions of verbal learning and memory, and planning, reasoning, and problem-solving, which in turn have direct impact on vocational outcomes [3]. Impairments in psychomotor speed has also shown to be related to disabilities in work performance, since for example more time is required to complete work tasks as a result of difficulties turning a plan into action. A modified employee training and on-going hands-on supervision to help the client organize and practice work tasks may be helpful strategies to compensate for such impairments [7].
When considering our findings and other research in relation to the disability construct, cognitive functioning is not the sole factor in the ability to gain, perform, and keep work. Other factors may contribute to work disability, such as stigma [26, 41], the employer’s attitude and engagement [42], and the social work environment [43]. For instance, social disabilities are frequently reported in vocational rehabilitation research among people with serious mental illnesses [18–21]. In the study by Lexén et al. [21], some participants spoke of initial difficulties in coping with social interactions with colleagues, while others perceived continuing problems with flexible and spontaneous social interactions. Impairments in social functioning among people with schizophrenia are well documented and strongly related to deteriorating daily functioning such as social and work functioning [44]. Further research about these disability factors in relation to work is warranted. Vocational rehabilitation interventions on several levels, in relation to the person, the environment, and the work task performed, may be crucial to enable this group of people to gain and keep employment.
Methodological considerations
This study has limitations that should be considered, and which may jeopardize the internal validity [45]. The design was cross-sectional, with the aim of collecting data at a single time point. The analyses thus allowed for interpretations about correlative relationships, but not causal ones. Hence, other variables than those investigated may be related to the vocational outcomes. However, the emerging body of knowledge on cognitive functioning in relation to vocational rehabilitation and occupational functioning generally supports our findings. Although the whole sample and the number of participants who attained employment were limited, the effect sizes were large enough to detect significant differences between groups, and relationships between cognitive functioning and vocational variables. The cognitive tests that were used are well-tested and should have produced valid and reliable data. The possibility that some participants may have performed some of the tests on a previous occasion cannot be ruled out. If this occurred, performance on the second testing might have been influenced. Furthermore, neurocognitive tests are widely used in research as both predictor and outcome variables. However, the ecological validity of such tests have been criticized, i.e. the extent to which the results of these tests are able to be generalized to real-life settings [46, 47]. An exploratory study by Haslam and colleagues [48], however, provides preliminary evidence that beyond cognitive functions, processes of tasks performance such as attention, concentration, and problem-solving as measured by the Assessment of Motor and Process skills (AMPS) in a stimulated environment may be a predictor of work outcomes such as a competitive employment for people with schizophrenia. Cross-sectional studies like the present of relationships between cognitive functions and vocational outcomes are also important contributors for understanding the ecological validity of neurocognitive tests in relation to vocational outcomes among this group. As Haslam and colleagues describe, the relationship between performance on neuropsychological tests and everyday cognitive ability is complex and multifaceted [48] and research needs to reflect this complexity.
McGurk and Mueser [3] suggest that relationships between cognitive functioning and work may vary between different vocational rehabilitation approaches. Therefore, a further issue to consider is that participants in the present study attended vocational rehabilitation that was different in terms of whether it was guided and/or coordinated by an employment specialist or a case-manager. However, both actors integrated their service with the mental healthcare team and had employment as a goal. Furthermore, the case-manager function did not allow for time-unlimited support at work. The vocational outcomes did not consider job tenure, but focused on gaining employment, the number of work hours per week, and income. These variables are not sensitive to whether the participant received support at work over a longer period of time. Furthermore, the aim of this study was not to study relationships between different types of employment support, cognitive functioning and vocational outcomes. It is thus reasonable to conclude that cognitive functioning may be significant for gaining employment regardless of the type of vocational support provided.
Conclusions
The present study suggests that planning, reasoning, and problem-solving is the most important factor distinguishing between whether or not people with schizophrenia who attend vocational rehabilitation integrated into mental healthcare service have competitive employment or not. Better performance on the tests related to measuring cognitive function in these areas was also related to more work hours per week and higher income. These findings suggest that verbal learning and memory is important for sustaining work performance, while other cognitive functions, such as planning, reasoning, and problem-solving, may be more important for gaining employment. How to compensate for, or improve, an individual’s ability to counteract these impairments should be attended to in vocational rehabilitation. Despite this study’s limitations, it may generate hypotheses for further research and should be considered as part of the larger puzzle of knowledge related to research in vocational rehabilitation among people with schizophrenia in mental healthcare.
Conflict of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Footnotes
Acknowledgments
We gratefully acknowledge the contributions of the participants who made this study possible, those persons who helped in the data collection process, and the statistician, Anna Lindgren, PhD. This work was supported by the Swedish Council for Working Life and Social Research, Finsam, The Medical Faculty at Lund University, and Vårdalinstitutet at Lund University.