Abstract
Introduction
Supported employment has been shown to be an effective means to integrate persons with mental illness in competitive employment (Bond & Drake, 2008; Bond, Drake, & Becker, 2012; Burns et al., 2007; Hoffman, Jäckel, Glauser, Mueser, & Kupper, 2015; Jäger et al., 2013; Marshall et al., 2014). The literature on supported employment has always emphasized the importance of participants’ motivation in finding and holding down competitive employment (Becker & Drake, 2003; Burns et al., 2007). However, some researchers argue for intensifying investigation into the matter of participant motivation (Baksheey, Allott, Jackson, McGorry, & Killackey, 2012), in order to achieve a more detailed understanding in the context of supported employment. Some studies have recently investigated the motivational processes and expectancies of unemployed people without mental illness looking for a job, Vansteenkiste, Lens, De Witte, and Feather (2005) provides an overview of these. However little is known about the motivational process as represented by different types of participant expectancies about supported employment.
Recent research has emphasized that the subjective importance of working in competitive employment is high among persons with mental illness (Brantschen, Kawohl, Rössler, Bärtsch, & Nordt, 2014). However, their motivation may not help them to achieve their goal because it may not trigger a motivational process with goal-oriented action. In order to help job coaches to distinguish between patients with distinct types of motivation, who might therefore need different treatment approaches, we hold the extended cognitive model of motivation (Rheinberg, 2006) to be an appropriate model describing different aspects of motivation.
The paper is structured as follows: after the general description of the extended model of motivation (Heckhausen & Heckhausen, 2012), we apply the model to the specific situation of supported employment. We then empirically test whether a different type of motivation is connected with a specific vocational outcome. Finally, we discuss some recommendations for job coaches, specifically, how they could identify patients who need special support to efficiently find competitive employment.
The extended cognitive model of motivation
Motivation psychology aims to explain behavior in achieving important personal goals. Thus it is concerned with actions, information as well as mental activities such as thoughts, feelings, or imagination. Motivation can be defined as the ‘activating orientation of current life pursuits toward a positively evaluated goal state’, (Rheinberg, 2006, p.15). The extended cognitive model of motivation postulated by Heckhausen & Heckhausen (2012) consists of four components (those in the rectangles depicted in Fig. 1) that influence whether a person will act to achieve a specific goal: the subjectively perceived situation, a possible action, the outcome of this action and the consequences that this action might have, see Fig. 1.
Three types of expectancy are outlined in the model. The Situation-Outcome (S-O)-expectancy lies in the situation itself. In other words, a person perceives a situation in a way that leads him or her to expect that a specific outcome will happen anyway, i.e. there is no need to be active. In contrast, if a person believes that his or her action could influence an outcome, i.e. there is a need to act, this is called Action-Outcome (A-O)-expectancy. The Outcome-Consequence (O-C)-expectancy is defined by the consequences: if they are worthwhile, the probability of acting increases.
The model implies two different incentives to act. One involves the activity itself as it is connected directly with a positive emotional state, like reading, singing, skiing, biking; it is called the activity related incentive. The other incentive is related to the consequences of the activity. It is called the purpose related incentive.
This model has a number of advantages, as Heckhausen & Heckhausen (2012, p. 339) state: ‘The extended cognitive model of motivation permits detailed analyses of motivation in specific situations. A particular strength of the model is that it allows different forms of motivational deficits to be diagnosed. These deficits may be attributable to one or more of the three expectancy types [...], or to incentives being insufficient or inappropriate. The latter may apply to purpose-related incentives (“It’s not worth it”) and/or to activity-related incentives. (“I can’t face doing it”)’.
According to the extended cognitive model of motivation for people looking for a job, it is more beneficial to have a pronounced A-O-expectancy, since activities in the application process are obviously better for obtaining a job than having only an S-O-expectancy.
Two types of expectancies in supported employment
In applying the extended model of motivation to the specific circumstances of supported employment, the following points were addressed. In supported employment, job coaches support people suffering from mental illness to obtain competitive employment. In this concrete situation, the Action-Outcome-expectancy comprises activities like preparing and sending applications, practicing interview techniques, and searching job advertisements. However the specific situation in supported employment, i.e. having a job coach to help with obtaining work, might activate the Situation-Outcome-expectancy, implying the attitude that the job coach will find the patient a job. Thus patients might perceive the situation as implying that they can be passive. In the ‘individual placement and support’ (IPS) model, the primary goal of supported employment is to help people start working (Becker & Drake, 2003); the job coach should therefore take over the active part, and lead the job application process when patients have a clear and strong S-O-expectancy. Some patients need this active support due to motivational deficits associated with their illness, for example depressive persons with listlessness or fear of stigmatisation if they disclose a mental illness (Link, Cullen, Struening, Shrout, & Dohrenwend, 1989; Rüsch et al., 2014).
As the wish to obtain competitive employment was an inclusion criterion in our study, the O-C-expectancy was not addressed. For longer-term considerations this expectancy could be central, as unsuccessful job applications could weaken or discourage motivation in this type of expectancy.
With respect to applying the extended cognitive model of motivation to supported employment there are two incentives. The first incentive is related to activity, where, for example, a job coach could stress that even the preparation of application documents can provide pleasure. The second incentive is the benefits of competitive employment, e.g. daily structure, social contacts, self-confidence, enhanced quality of life, more financial independence, and integration into society.
We do not know from a study that explicitly has researched different expectancies to the extended cognitive model of motivation in a supported employment trial. We therefore reanalysed the data from a clinical trial that tested the effect of different placement budgets (Nordt et al., 2012).
Our research questions were: Which type of expectancy is more beneficial in obtaining competitive employment? Which sociodemographic, clinical, and job relevant characteristics are related to the Action-Outcome-expectancy, and the Situation-Outcome-expectancy? Are the predictors different for the two types of expectancy? What implications can be derived to help the job coaches?
Methods
Design of the study
This study was conducted within the framework of ‘The Zurich Program for Sustainable Development of Mental Health Services’, ZInEP (http://www.zinep.ch/en). Despite offering an extensive range of services, psychiatric care in the Canton of Zurich, Switzerland, is still not ideally aligned with the needs of mentally ill people. For this reason, it is the aim of ZInEP to make a contribution to a sustainable improvement in the quality of psychiatric care in the canton.
This study was a multicenter randomized controlled trial with six outpatient psychiatric clinics in the Canton of Zurich. Six IPS trained job coaches supported the participants for up to two years.
The job coaches recruited 116 persons, who actually started with the IPS intervention between June 2010 and May 2011, in the mental health service where the participants received outpatient treatment.
The job coaches practiced supported employment according to the guidelines of the ‘individual placement and support’ (IPS) model (Becker & Drake, 2003), which is based on the following principles. A job coach is an integral member of a mental health care team who helps all patients to find competitive employment corresponding to their preferences and talents as soon as possible, and who continues to support them and their employer, so that jobs can be held down as long as possible.
The IPS supported employment implementation of the job coaches was assessed with the ‘supported employment fidelity scale’ (Becker & Drake, 2003) every three months.
Details of the study design are published in the study protocol (Nordt et al., 2012).
This article is based on data from different sources: the initial interview between participant and job coach, the baseline interview between participant and interviewer, the job coach reports over two years after competitive employment had been obtained, and clinical information from the psychiatric register.
Participants
The following inclusion criteria were applied for the study participants: current treatment in one of the six participating outpatient psychiatric clinics, at least one year of unemployment, not participating in any vocational integration programme during the last three months, being of working age (18 – 60 years), wishing to obtain competitive employment, being willing and capable of giving informed consent, and residency in the Canton of Zurich. Exclusion criteria were severe organic illness and insufficient knowledge of German. The sample of 116 study participants is described in Table 1.
There were 59 female and 57 male participants; the mean age was 41 years. The mean income from all sources was 2759 CHF (2500 Euros), which is clearly below the mean gross income per household in the catchment area (6000 CHF (5450 Euros), www.bfs.admin.ch). Most participants had an affective disorder according to ICD-10 (Dilling, Mombour, & Schmidt, 1991). Education status was representative for outpatients.
The clinical global impression CGI (1 to 7) was on average 4.9 (score 4 means moderately ill, score 5 is markedly ill). The global assessment of functioning (GAF Scale; Sass, Wittchen, & Zaudig, 1996) ranged between 20 and 85 points with a mean of 57 points.
The psychological strain measured by Brief Symptom Inventory (BSI) was quite high; mean value was 1.02, which corresponds to a T-Score of 72 (Franke, 2000), and is clearly above the threshold for clinical significance (T≥63).
Self-esteem was measured by the Rosenberg self-esteem scale (Rosenberg, 1979), stigmatisation by the ‘Internalization of the stigma of mental illness’, ISMI (Ritsher, Otilingam, & Grajales, 2003), and quality of life by the WHO QOL, German version (Angermeyer, 1998). The mean of the QOL scale was 57.3, which is similar to that of an earlier study among inpatients in the catchment area (Nordt, Müller, Rössler, & Lauber, 2007).
Ninety-eight percent of the study participants thought, that it was very important or important to work in the primary labour market.
Measures and procedure
Categorizing the type of expectancy
During the initial semi-structured interview with the job coach the unemployed participants specified beside the sociodemographic details (see Table 1) their thoughts, plans, wishes, hopes regarding supported employment. After this the job coach summarized this information on paper. Later on, two independent scientific associates used these notes to categorize each study participant to one of the two expectancy types, showing an excellent agreement in their allocation (Cohen’s Kappa, κ= 0.90). Disagreements were cleared by discussion and consensus. Each participant was assigned to one of two types of expectancy: participants of the first type (A-O-expectancy) think that their own action is the decisive factor in finding a job, i.e. preparing and sending out applications, practicing interview techniques, searching the job advertisements; participants of the second type (S-O-expectancy) think that the job coach should provide them with a job, that the outcome will occur without any action on their part; alternatively, the application process is avoided for whatever reason (illness, fear of rejection, stigmatisation). As this categorization was made towards the end of the study, the job coaches could not use this information in the job coaching process.
Statistics
The A-O-expectancy and the S-O-expectancy were defined as the dependent variables. As explanatory variables we considered the sociodemographic, clinical, and job-relevant variables. Clinical information obtained from the psychiatric register included diagnosis, CGI, and GAF. Sociodemographic information (gender, age, income, and education), psychological self-assessments (self-esteem, stigma, quality of life) and job-relevant information (self-motivation, duration of former employment and unemployment duration) were assessed at the first interview between study participant and an interviewer.
We applied logistic regression analyses to explore associations between the type of expectancy and variables of interest. We considered sociodemographic, clinical, psychological, and job-relevant variables (see Table 1) as explanatory variables. In order to analyse these predictors, we fitted logistic models for both outcome measures, A-O and S-O. In a first step, we analysed the explanatory variables separately; in a second step, we included all significant variables together in the logistic regression model. We report odds ratios and significant P-values (p < 0.05) resulting from the univariate analyses. In the multivariate logistic model we included only the significant predictors of the univariate analyses.
Before the start of recruitment in June 2010, the study protocol was approved by the Zurich Cantonal Ethics Committee (CEC), Division 3 (Kantonale Ethik-Kommission Zürich (KEK) Abteilung 3), reference number E-51/2009. Informed consent war obtained from all individual participants included in the study. The authors declare that they have no conflict of interest. All authors certify their responsibility for this manuscript.
Results
Study participants were found to have either the A-O-expectancy (N = 78) or the S-O-expectancy (N = 37), except for one, who had no expectancy at all.
Predictors
The odds ratios (OR) of univariate regression analysis of the clinical, psychological, sociodemographic, and job relevant characteristics on the independent variables A-O-type and S-O-type are presented in Table 2.
Female gender, a higher quality of life, higher self-motivation to find competitive employment, and higher income were associated with an A–O-expectancy (OR 1.4 to 2.8) in the univariate analyses.
Male gender, lower income, having a partner, and lower quality of life were associated with an S–O-expectancy.
A multivariate regression analysis incorporating all previous significant variables in one model showed similar results (Table 3).
Type of expectancies and competitive employment outcome
During the job coaching, 57.8% of all study participants had competitive employment for at least one day.
Two thirds (68%) of the participants with A-O-expectancy found competitive employment. By contrast, those participants with S-O-expectancy had a vocational outcome that was about three times poorer (Table 4).
Supported employment fidelity scale
The outcome of the supported employment fidelity scale, which was assessed by the study coordinator and the supervisor of the job coach every three months, revealed with 68 points a good supported employment implementation in this study.
Discussion
This study found that two types of expectancy described by the extended cognitive model of motivation can be distinguished on the basis of the initial semi-structured interview between job coaches and participants in a supported employment program. Study participants with an Action-Outcome-expectancy had a three times higher probability of finding competitive employment than those with an Situation-Outcome-expectancy. This result is probably not surprising, but in the framework of supported employment, we think it is important to differentiate clearly between action-oriented participants, who might benefit from being encouraged in their activities, and more passive participants, who might benefit more when the job coach assumes the responsibility for finding a job. This might be necessary because patients discouraged by repeated failure in the job application process might do well in a job once they have the opportunity to be employed. The IPS model of supported employment may foster the S-O-expectancy unintentionally, i.e. it is conceivable that the patient delegates the job search and related activities to the job coach. Nevertheless most study participants had an A-O-expectancy and took advantage of the chance to receive practical assistance from a trained job coach in the application process. Therefore, job coaches should encourage their patients to be proactive, i.e. by searching through job advertisements, attending vocational information centers, sending the application documents, and so on. These actions can trigger the Activity-Related Incentives pursuant to the extended cognitive model of motivation. This is in accordance with Negrini, Corbière, Fortin, and Lecomte (2014), who found that job acquisition and being involved in a job search process increase the psychosocial well-being of people with mental illness.
With the aid of this motivational model, job coaches can identify passive participants relatively easily and then tailor their approach to their needs, so that people with both types of expectancy can start working as soon as possible.
The A-O-expectancy is the active part of the extended model of motivation and is a predictor for obtaining competitive employment. We found the following predictors: women are three times more likely to have this kind of expectancy; persons with this expectancy also have a higher income; moreover, they have a higher quality of life score and higher self-motivation (importance of working in competitive employment). The S-O-expectancy implies lower chances to obtain competitive employment. This passive type of expectancy was predicted by being in a close relationship with a partner and having a lower quality of life. It is probable that persons living with a partner have more emotional support and get their satisfaction from the partner, and that their income is thus ensured, so they are not as motivated to be actively engaged in finding a job. This is in line with results from the Zurich epidemiological study: Social support (‘living with a partner’) was found to hinder persons from looking for professional help when they were symptomatic (Burns et al., 2003).
It is surprising that neither education nor duration of unemployment showed any effect on either of the two expectancies. Previous research has shown that work history was a predictor of obtaining employment (Catty et al., 2008), but it does not influence expectancy.
This study demonstrated that it is possible to categorize participants into one of the two expectancy categories at the first interview with the job coach, and that the type of expectancy is a relevant predictor for vocational integration. Further research is needed to explore how job coaches can effectively help the participant adopt an A-O-expectancy, and how they could handle the S-O-expectancy. A next step would be to devise and implement a training program for job coaches to assess patient expectancies and tailor job coaching accordingly. Moreover, it seems worthwhile to investigate more concise motivational aspects in supported employment, and to focus on strategies how to address effectively patients’ motivation.
Nevertheless, we could show that motivational processes are actually very important in supported employment. Vocational practitioner assumed that motivation is a key factor obtaining competitive employment (Bond & Drake, 2008; Salyers et al., 2008), the self-report of the study participants confirms this assumption.
A limitation of this observational study is that the number of participants was relatively small, thus some predictor variables might not have been significant. These findings should therefore be replicated with a larger sample size. Secondly, we did not investigate the Outcome-Consequences (O-C)-expectancy, because it was part of the inclusion criteria; we looked for people with mental illness who were motivated to work in the first labour market. In the long run, it is possible, that this expectancy decreases as a consequence of repeated failure to find a job.
Conclusion
The present study shows that two types of expectancies described by the extended cognitive model of motivation could help to identify participant groups with different needs, who will probably benefit from different strategies on the part of job coaches to find competitive employment more effectively. Individuals active in the job application process are more likely to obtain a job in comparison to those persons who rely on the job coach only. On the one hand, job coaches for their part can encourage their patients to be active in the application process, because this probably increases their chances of finding competitive employment. On the other hand, job coaches should support more passive participants more intensively during the placement process.
Conflict of interest
The authors have no conflict of interest to report.
