<psyRA JAPP 2012 2>
<Effects of Resource-Building Group Intervention on Career Management and Mental Health in Work Organizations : Randomized> <Controlled Field Trial>
<Jukka Vuori Salla Toppinen-Tanner Pertti Mutanen>
Abstract
A resource-building group intervention was developed to enhance career management, mental health, and job retention in work organizations. The in-company training program provided employees with better preparedness to manage their own careers. The program activities were universally implemented usingan organization-level, 2-trainer model with trainers from the human resources management and occupational health services. The study was a within-organizations, randomly assigned field experimental study; it investigated the impacts of the intervention on immediate career management preparedness andlater mental health and intentions to retire early. A total of 718 eligible individuals returned a questionnaire in 17 organizations and became voluntary participants. The respondents were randomly assigned to either an intervention (N  369) or a comparison group (N  349). Those in the intervention group were invited to group intervention workshops, whereas those in the comparison group received printed information about career and health-related issues. The 7-month follow-up results showed that the program significantly decreased depressive symptoms and intentions to retire early and increased mentalresources among the group participants compared to the others. The mediation analyses demonstratedthat the increase in career management preparedness as a proximal impact of the intervention mediated the longer term mental health effects. Those who benefited most from the intervention as regards their mental health were employees with elevated levels of depression or exhaustion and younger employees,implying additional benefits of a more targeted use of the intervention. The results demonstrated the benefits of the enhancement of individual-level career management and resilience resources as career and health promotion practice in work organizations.
As changes in work organizations have become an everyday phenomenon, employees are increasingly expected to keep up with the requirements of their jobs and to remain healthy and motivated while working increasingly longer careers. In many modern societies, however, a growing proportion of employees do not sustain their careers until regular old age retirement, and many of these employees retire due to depression or burnout before the normal retirement age ( Ahola, Toppinen-Tanner, Huuhtanen, Koskinen, & Vnnen, 2009; stn, Ayuso-Mateos, Chatterji, Mathers, & Murray, 2004). In addition, a growing number of work disabilities, resulting in early retirements, depression, and burnout, cause significant losses in work life productivity ( Adler et al., 2006; Lerner & Henke, 2008).
Various interventions have been developed for reducing the negative effects of work stress on health and for promoting work ability (for a review, see Marine, Ruotsalainen, Serra, & Verbeek, 2006; Sockoll, Kramer, & Bdeker, 2008). However, the role of workplaces as resource-building arenas for the promotion of well-being, mental health, productive careers, and engagement has been studied very little. In the field of prevention and mental health promotion, group-based cognitivebehavioral interventions with the main focus on self-efficacy beliefs have been found to effectively reduce depressive symptoms in challenging change situations ( Glanz, Lewis, & Rimer, 1997; Jan-Llopis, Hosman, Jenkins, & Anderson, 2003). Individual efficacy beliefs would seem to provide useful targets for interventions based on social learning theory (e.g., regarding the management of employees' career changes due to changes in organizations). However, social science theories or efficacy belief-based constructs have rarely been applied in occupational health programs, even though work in modern organizations is often very much social in nature ( Heaney, 2005).
In other settings, preventive resource-building programs have previously been successful during stressful educational and occupational transitions, for example, thus increasing preparedness for the respective transition. This has resulted in beneficial mental health and career outcomes ( Caplan, Vinokur, Price, & van Ryn, 1989; Koivisto, Vuori, & Nykyri, 2007; Vinokur, Price, & Schul, 1995; Vuori, Koivisto, Mutanen, Jokisaari, & Salmela-Aro, 2008; Vuori, Silvonen, Vinokur, & Price, 2002). The common effect found in these studies on mental health has been the decrease of depressive and distress symptoms. Another typical effect in these studies has been the improvement of career outcomes during these occupational transitions. Even though these earlier studies have involved occupational transitions in the fringes of work life (e.g. school-to-work transition or reemployment), career transitions also increasingly occur in work organizations. This development imposes additional challenges to individual employees and their well-being, and it should be targeted and studied within worksites. The focal issues of career management inside work organizations involve not only career planning and employment issues, as in traditional occupational transitions, but also elements related to managing work and family conflicts and coping with stress during one's work career. This development in work life calls for new interventions, which combine elements of training in career planning and of training in how to manage stressful career situations.
We responded to this need in our study, as we applied social learning in work peer groups and an efficacy-based construct, namely, career management preparedness, as central active ingredients in a group intervention in work organizations. Our focal aim was to investigate the possibilities of teaching people career management skills, how to enhance their resources for planning their future and how to endure the unpredictable turbulence of the work environment in modern work organizations. We also applied a randomly assigned field experimental study in order to collect reliable data on the individual level within organizations, which is very rare in organizational intervention studies.
Promoting Employees' Career Management and Resilience
In today's rapidly changing work life, employees need resources for managing their own careers and for feeling safe and motivated. They need confidence in their ability to handle changing situations and to maintain their employability. In order to be effective, worksite stress prevention should comprise promotion of individual resilience and well-being resources, in addition to the conventional prevention of stress-related problems ( Bakker, Schaufeli, Leiter, & Taris, 2008; Lamontagne, Keegel, Louie, Ostry, & Landsbergis, 2007).
Due to widespread restructuring, downsizing, and flexible employment strategies, employability and career resilience have become important goals for career management. At the same time, it seems that traditional management practices regarding employees' careers have become less central for employees. The focal question is, are employees able to be adaptive and proactive enough and take responsibility for their own career development in an increasingly unpredictable career environment ( Iles, 1997; Lips-Wiersma & Hall, 2007)? Even when employees' work careers have not involved several employer changes, career management strategies have been predictive of self-perceptions of aging successfully in the workplace above and beyond important characteristics of the individual or employer organization ( Robson & Hansson, 2007). As employers in today's organizations are also concerned about employees' well-being and work ability, it would seem feasible to enhance employees' career management with interventions based in an employer organization ( Raabe, Frese, & Beehr, 2007).
It has been shown that career management and well-being can be improved by developing preparedness and abilities related to career goal setting and attainment ( Ajzen, 1991; Bandura, 1986). For example, preparedness for work life transitions can be increased through group interventions applying social modeling and active learning techniques ( Caplan, Vinokur, & Price, 1997; Vuori & Vinokur, 2005). Preparedness is defined here as a cognitivemotivational construct that has specific self-efficacy ( Bandura, 1986) and preparation against setbacks ( Meichenbaum, 1985) as its intertwined ingredients ( Vuori & Vinokur, 2005). Employees who are well prepared and motivationally ready to manage their careers also have both the confidence in their career management skills and the knowledge and emotional readiness to deal with the setbacks that are frequently encountered during the career management process. In other contexts, preparedness has also been defined as the goal state of readiness to respond to uncertain outcomes ( Sweeny, Carroll, & Shepperd, 2006).
In a more unpredictable career environment, career management should also aim at a sustainable and occupationally healthy career. Parallel to this, career management self-efficacy can be defined as confidence in one's abilities in career management activities, such as defining personal strengths and career interests, finding the means and pathways for shaping one's job and career in the self-desired direction, seeking possibilities for learning experiences at work, learning assertiveness in social relations at work, and learning self-care. Both Bandura (1986) and Ajzen (1991) have shown how specific self-efficacy, or perceived control of a specific behavior, is a component of motivation and increases the likelihood of a specific behavior.
The process leading to preparation or inoculation against setbacks has been used in cognitivebehavioral therapy and involves the ability to anticipate setbacks and the skills to cope with them. According to Meichenbaum (2007), inoculation against setbacks can be achieved by providing individuals with experience in minor setbacks and stressors. This fosters psychological preparedness, promotes resilience, and develops a sense of mastery in confronting more stressful setbacks and obstacles. The inoculation process, with anticipated high-risk situations and the acceptance of lapses as learning opportunities, will maintain participants' motivation to carry out difficult behaviors in cases of real setbacks and prevent relapse ( Meichenbaum, 2007). Preparation against setbacks has been shown to be a key underlying preventive dimension of training for those at risk of depression and to alleviate depressive symptoms ( Vinokur & Schul, 1997; Vuori, Price, Mutanen, & Malmberg-Heimonen, 2005).
Preparedness can be seen as a motivator in a process whereby individuals adjust to their work environment, make plans, set goals, strive to improve their future, and evaluate their possibilities and competencies in achieving these goals. Many previous studies have shown that individual resources may also influence the relationship between work characteristics and adjustment to work ( Aspinwall & Taylor, 1997; Parker & Sprigg, 1999). For instance, increased self-efficacy has been found to associate with increased work engagement and strengthened perceptions of social resources at work ( Llorens, Schaufeli, Bakker, & Salanova, 2007).
Group Intervention for Enhancing Career Management Preparedness
The development of the group intervention was based on social cognitive theories on social action, behavioral control, and individual coping resilience ( Ajzen, 1991; Bandura, 1986; Meichenbaum, 1985). The aim was to enhance career management preparedness. The self-efficacies related to career management were strengthened in a stepwise procedure: first, identifying in peer groups the career-related goals needed for progress; second, defining in peer group solutions and tasks for carrying out these goals; and third, practicing the required skills and actions in small groups. Coping resilience was strengthened according to the principles of Meichenbaum's (1985, 2007) stress inoculation training. First, participants shared experiences of setbacks and barriers in their career environment and empathized with feelings aroused by these conversations. Second, they defined possible solutions for these setbacks and barriers in peer groups, and third, they practiced these solutions in small groups.
The group program was built utilizing earlier experiences in preventive group methods aimed at increasing preparedness for different educational transitions and the transition from job loss to reemployment ( Koivisto et al., 2007; Vinokur et al., 1995; Vuori et al., 2002). In order to better incorporate the intervention in organizational practices, the implementation of the program involved practical collaboration of the human resources and occupational health functions of the participating organizations. This was also aimed to strengthen collaboration between these units in work career and mental health issues.
The development of the group intervention was guided by the MPRC (Michigan Prevention Research Centre) general group training principles' five essential components for effective group training ( Price, Friedland, Choi, & Caplan, 1998; Vuori et al., 2005). First, career management skills training was developed to focus on skills that are essential in proactive and successful management of one's own career in changing organizations. This included, for example, the definition of one's own strengths and career interests; the internalization of the idea of lifelong learning; learning about organizational change; finding means to attain career-related resources from social networks in organizations and to solve social conflicts; and learning concrete means for managing one's own career. Second, active teaching and learning methods mean that instead of lecturing, the trainers make use of participants' own career knowledge and career choice skills as part of the learning process, in discussions in both small and large groups, role plays, and other activities. Third, skilled trainers are instructed to build trust and work together in pairs to facilitate group processes that promote the learning of career management skills. Fourth, the trainers work to create a supportive learning environment, which enables participants to learn from and support each other. This occurs through modeling and strengthening supportive behavior in the groups according to Bandura's (1986) social learning theory. And last, preparation against setbacks uses stress inoculation training, described above, including topics such as threats and challenges during organizational changes and social conflicts at work.
The focus of the intervention was on increasing individual independence and awareness of one's own abilities and choices during one's work career. This could be done partly irrespective of the continuous changes of the immediate or wider work context and could be seen as a way in which to empower people in their careers. For this reason we wanted to implement this preventive intervention as a universal intervention in work organizations.
Evidence also exists that those who benefit the most from interventions are the most vulnerable groups of employees at risk of mental health problems (for a review, see Seymour & Grove, 2005) or at risk of leaving their jobs ( Heaney, Price, & Rafferty, 1995). Furthermore, previous research also indicates that younger participants may benefit more than their older counterparts from career-related mentoring practices ( Finkelstein, Allen, & Rhoton, 2003).
Hypotheses on the Impact of the Intervention
In this study, we investigated the 7-month follow-up effects of the group program on employees' career management and mental health. We expected the group program to create, first as a proximal effect, an increase in career management preparedness and, later, due to increased career management and as a long-term effect, to produce a decrease in employees' depressive symptoms, exhaustion, and intentions to retire early. At the same time we anticipated that the program would increase work engagement and mental resources. Consequently, our first hypothesis regarding the proximal effect of the program was
Hypothesis 1: The group program increases career management preparedness.
The next two hypotheses regarding the main long-term effects of the program were
Hypothesis 2: The group program decreases depressive symptoms, exhaustion, and intentions to retire early.
Hypothesis 3: The group program increases work engagement and mental resources.
Although our main hypotheses were formulated for a universally implemented intervention, we also expected the effects of the intervention to be the most beneficial for vulnerable groups. Previous research has indicated that the preventive effects of resource-building programs on mental health are stronger among those at risk of depression ( Caplan et al., 1989; Koivisto et al., 2007; Vinokur et al., 1995; Vuori, Koivisto, et al. 2008; Vuori et al., 2002). The same result has been obtained from studies on depression prevention interventions ( Brunwasser, Gillham, & Kim, 2009). Moreover, as the group program was based on peer groups and involved practices partly similar to group mentoring, we expected the younger participants to benefit more from the group activities than their older counterparts with regard to the following areas: depression, exhaustion, work engagement, and mental resources. Accordingly, in order to evaluate possibilities for more targeted implementation of this intervention, we further investigated the moderating role of the studied variables, namely, level of depressive symptoms, exhaustion, work engagement, mental resources, and intention to retire, and additionally age, in the effects of the intervention without setting up specific hypotheses.
Our starting point in developing the intervention was that it should primarily increase group participants' career management preparedness as the proximal outcome. We further expected that this proximal increase in preparedness would in the long run give rise to beneficial effects on career and mental health. In line with this, our last hypothesis was
Hypothesis 4: The found beneficial long-term career and mental health effects of the intervention are due to the proximal beneficial effects of the intervention on career management preparedness.
Method
Participating Organizations
The goal of the recruitment was to obtain a heterogeneous sample of the working population. We recruited participant organizations from different sectors of work life, from both the private and the public sector. A total of 17 organizations participated in the study. During recruitment of the organizations, representatives of human resources development departments (HR) and occupational health services (OHS) in 43 organizations in southern Finland were sent letters containing an invitation to participate in a research and development program, information regarding the program and its background, and an advertisement of the program. A meeting with the organization representatives to introduce the study process, the prototype of the program, and its expected benefits was arranged in 27 organizations, and the representatives were asked about their willingness to participate in the program. The participating organizations were offered all material and training of trainers free of charge. They would also obtain certificates, so they could use the method in their own organization after the study phase. In return, the participating organizations agreed to provide the study data, to follow the instructions in the implementation of the study, and to participate in the development process as agreed. Furthermore, they agreed not to offer the program to the comparison group before the study follow-up periods ended. Out of the 27 organizations, 15 chose to participate. Additionally, two organizations chose to participate due to word of mouth from other participating organizations.
Because of the practical requirements of the group intervention process, we were able to recruit voluntary participants only in medium- and large-sized organizations. Nine of the recruited organizations were medium-sized city administrations; two were city departments. The others were a governmental research institute, an employment office, an insurance office in the municipal sector, a banking company, a large multiservice company, and an occupational health service organization. The majority of the participating organizations ( n = 10) reported significant organizational changes during the study period.
Collaboration contracts were signed and we arranged a second meeting with the participating organizations, with the main focus on the practical issues related to program implementation. These included naming the two or more trainers from OHS and HR departments to be trained at the Finnish Institute of Occupational Health (FIOH), setting timetables for the training, deciding how to market the training, and deciding on the practices for recruiting participants for the study groups. All participating organizations agreed to invest in the working hours of trainers and participants, but compensation was neither received nor paid for participation or for training and materials. When both parties had agreed on a timetable, a date was set for a staff meeting to disseminate information on the study and to recruit participants.
In the recruitment meeting, FIOH researchers briefly described the program and the evaluation study, and the trainers chosen from each participating organization were introduced. The baseline questionnaires, consent forms for participation and data collection, and the information letter were available in the meeting or could be obtained later from the trainers. In addition, the researchers provided the organizations with information leaflets regarding the program and any other information requested. Where necessary, they also helped with the marketing of the study, which was conducted via personnel magazines, the Intranet, and other channels used by the organizations in their daily communication. We expected the intervention to be most beneficial for employees who already had some work experience (i.e., were either in their midcareer or beyond). Consequently, the recruiting material had the slogan Do you want an extra buzz from your work? and the sign 40+. The overall proportion of women in the participating organizations was 75.4%. During the introduction phase we used the title Towards Successful Seniority group method for the program, but several versions of the title were used by the participating organizations during the development process.
Participants
To become participants, respondents had to agree to the randomization procedure of the study and to hand in the baseline assessment questionnaire (Time 1 [T1]). Between the fall of 2006 and the spring of 2008, a total of 718 eligible individuals filled in and returned the questionnaire in the 17 organizations, thus becoming participants of the study. Fourteen cases were discarded from the study, due to returning the questionnaire too late (9), unknown address (2), preference for intervention group (1), or quitting the study (2). The information letter introduced the study as having two experimental conditions: one would involve participation in group training and the other would involve getting a literature package for reading. We did not expect the literature package to have health or career effects of nearly the same magnitude as the group training, because the reading material was planned only for providing general career and health-related information and not for significantly increasing career management preparedness.
Respondents varied in age from 31 to 64. The mean age was 50.1 years ( SD = 6.47). Of the total sample, 632 (88%) were women, 86 (12%) were men, and 74% were married or cohabiting. Most of the participants had completed high school (57%) and had a college (38%) or university (22%) education. The majority of participants worked in the public sector (81%): 66% in municipalities and 15% in state administration. A total of 17% of the participants worked in the private sector. Most of the participants (89%) evaluated their job as secure, 4% evaluated it as unsecure, and 6% had no opinion.
Randomization Procedure, Experimental Design, and Follow-Ups
The 718 eligible respondents who completed the baseline questionnaire (T1) were randomly assigned either to group training ( N = 369) or to literature package reading comparison groups ( N = 349). The randomization was carried out by the researchers separately for each participating organization. Two researchers shuffled the received sealed questionnaire envelopes and dealt them into two piles. The results of the randomization were sent to the participating organizations. Those randomized into the intervention group were invited to take part in the group intervention workshop, whereas the comparison group received printed information about career and health-related issues. At the end of the intervention, both groups received a follow-up questionnaire (Time 2 [T2]). These T2 questionnaires were returned by 570 (79.4%) study participants. In order to obtain long-term follow-up data on the effects of the intervention, we sent participants of the intervention group and the respective comparison group a questionnaire 7 months after the respective intervention (Time 3 [T3]). The T3 questionnaires were returned by 613 (85.4%) study participants, 87.0% of the intervention group, and 83.7% of the comparison group. The greater response rate in T3 was due to telephone reminders, which we used after the second reminder questionnaires. Figure 1 presents the study design. It includes information on the experimental design and the size of the various subgroups at each stage of the study period.
The program was delivered by a cotrainer team of two trainers, and the recommendation was that one of them represent the occupational health services (OHS) and the other human resources (HR) activities. This was the case in most organizations during the study, but some other combinations of two trainers existed. The trainers were nominated by the participating organizations. Their instruction was provided by the FIOH research team over a period of 4 days. The trainers underwent the whole training program and were instructed in the principles of learning and given other related theoretical background. They also received practical advice.
Study participants randomized into the intervention group were contacted by trainers in the organization and invited to participate in a 1-week group intervention workshop. The workshop uses methods such as active learning process, social modeling, gradual exposure, and practice through role playing. One of its aims is to provide inoculation against setbacks. A detailed version of the intervention process and contents can be found in the Towards Successful Seniority Trainer's Manual ( Vuori, Ristolainen, et al., 2008).
The groups, comprising 815 employees and/or supervisors, assembled for five half-day sessions that focused on the enhancement of career management skills. The intervention is designed to achieve its goals through the creation of a socially supportive environment that facilitates positive interactions and relationships between trainers and participants, as well as between the participants themselves. The training is designed to increase participants' job-search self-efficacy and motivation and to endorse the following career management skills: (a) identifying and communicating one's skills and abilities, (b) identifying and using one's social network and solving conflicts in social relationships, (c) developing assertiveness at work, (d) developing stress management skills, and (e) building commitment to one's personal work-related plan for the near future.
The workshops were organized in classrooms or similar sites in the participating organizations. Typically, the interventions were arranged as five 4-hr morning sessions in the course of 1 week. In some organizations the intervention was performed over three full days. Coffee and sandwiches were served at each session. In total, 34 workshops were conducted, with the size of the groups varying between 8 and 15 participants. Average group size was 11. At the end of the program, participants were given the opportunity to complete the T2 questionnaire on the immediate effects of the interventions.
Comparison Group
Study participants in the comparison group were given a literature package that provided basic career management-related information and included a pocket guide for the maintenance of work ability and well-being and the prevention of stress. This material was given to the members of the comparison group during the week the intervention group received their training. In some organizations, the comparison group was assembled at the end of the intervention week to complete the T2 questionnaire survey, whereas in others the questionnaire surveys were posted to the comparison group. We checked whether the participants had actually read the material at T2, and statistical analyses were carried out to compare the main effects of the reading on our outcome variables among those who had read either a lot or some of the material and those who had not read the material. The results showed that there were no differences between the two groups, and consequently we kept this group as one comparison group.
Effectiveness of Randomization, Response Rates, and Attrition
Successful randomization is critical for the protection of the internal validity of our study's experimental design. Comparisons between the comparison and intervention groups did not reveal any statistically significant differences in socioeconomic background variables such as gender, age, education, or marital status (see Table 1). Similarly, no statistically significant differences were detected at baseline (T1) in the outcome measures of depressive symptoms and exhaustion, intention to retire early, or mental resources. As there had been changes in 10 out of 17 organizations we also checked whether the respondents differed in relation to perceived job insecurity, but we found no differences. The attrition in experimental design was analyzed following the recommendations of Hansen, Collins, Malotte, Jonson, and Fielding (1985).
The dropout effect in T3 was controlled, as was the no-show bias in the intervention. Compared with respondents who provided data for the T3 follow-up, dropouts had significantly more symptoms of exhaustion, t(125[unequal variances]) = 2.41, p < .05, and their mental resources were lower, t(705) = 2.08, p < .05, but no other significant differences in demographic or outcome variables were found. There were no significant differences in dropout group in any of the outcome variables between the intervention and the comparison group.
Of the 369 participants in the intervention groups, 25 (6%) did not participate in the Towards Successful Seniority group. On comparison with the demographic characteristics of participants and nonparticipants (no-shows), we found two statistically significant differences: No-shows had lower levels of career management self-efficacy than group participants ( p < .05) and were more often employed by the government than were the participants ( p < .05). There were no differences in the outcome measures at baseline between no-shows and participants. However, to prevent selection bias, we based the comparisons in all of our analyses on the complete randomized intervention group that includes both the intervention participants and the nonparticipants (no-shows). Consequently, as a small part of the intervention group in our statistical analyses did not receive the training at all, the results may give somewhat conservative estimates of the effects of the program.
Measures
All of the study measures were included in the baseline and in the 7-month follow-up questionnaire surveys. Table 2 shows the reliability scores (Cronbach's alpha) for the scales.
Demographic characteristics were assessed with standard survey questions for reporting age, gender, marital status, education, and occupation. Job security was measured using one question: How secure do you consider your present job? The scale ranged from 1 ( very secure) to 5 ( very insecure).
Career management preparedness was composed of two scales: career management self-efficacy and preparation against setbacks. To be prepared to manage one's career successfully, one needs to be confident of performing the relevant activities and be ready to overcome difficulties and setbacks. Career management preparedness was calculated as a sum of all the items of the career management and preparation against setbacks. Career management self-efficacy was measured with 10 items that were considered by experts to be the most central for successful and occupationally healthy career management in unpredictable career environments. Respondents were asked about their confidence in the following: seeing challenges of work as an opportunity to learn, solving problems at work, being able to learn new things, changing jobs if one wants to, staying active and vigorous, taking care of one's health, managing stress at work, keeping a balance between work and family life, being assertive at work, and handling conflicts at work. The preparation against setbacks measure consisted of three items ( Vinokur et al., 1995; Vuori & Vinokur, 2005). Respondents were asked (a) Do you have backup plans in case of possible setbacks? (b) Do you believe you would still be able to concentrate on your career after possible difficulties and setbacks? (c) Do you believe you would be able to continue your career after possible difficulties and setbacks? The scale ranged from 1 ( very much or very well) to 5 ( very few or very poorly).
Depressive symptoms were measured with the 13-item Beck Depression Inventory ( Beck & Beck, 1972). The reliability of the measure was 0.86.
Exhaustion was measured with one of the three dimensions of the Maslach Burnout InventoryGeneral Survey (MBI-GS; Kalimo, Hakanen, & Toppinen-Tanner, 2006; Maslach, Jackson, & Leiter, 1996). Exhaustion consisted of five items on overstrain, tiredness, or fatigue resulting from overtaxing work.
Work engagement was measured with the short, nine-item version of the Utrecht Work Engagement Scale (UWES-9; Schaufeli, Bakker, & Salanova, 2006). The scale includes items from the three dimensions of work engagement: vigor, dedication, and absorption. The items were rated on a 7-point scale ranging from 0 ( never) to 6 ( always).
Mental resources were measured with three items from the Work Ability Index (WAI) ( Tuomi, Ilmarinen, Jahkola, Katajarinne, & Tulkki, 1998), where Mental Resources is one of the seven subscales. It included items on enjoying daily tasks, activity and life spirit, and optimism about the future.
Intention to retire early was elicited with one question: Have you considered applying for work disability pension, early retirement, or some other retirement allowance? (1 = No, 2 = It has crossed my mind, 3 = I have seriously considered it, 4 = I have applied for retirement). The items are used in Finnish population barometers, such as the Quality of Work Life Survey, which investigates work characteristics and work behavior in different occupations ( Statistics Finland, n.d.).
The integrity of the intervention and its immediate impact were assessed with the reports of the participants after the last session of the intervention. Participants were asked to provide information on the extent to which various aspects of the group process and the other participants' behavior were positive, rewarding, and relevant to their participation. Their answers were provided using thirteen 5-point rating scales (1 = most negative rating, 5 = most positive rating).
Statistical Analysis
Regarding the first hypothesis, the effectiveness of the intervention was evaluated by comparing the proximal changes in preparedness from baseline measurement T1 to the T2 measurement immediately after the intervention ended and by comparing this change to that measured in the comparison group. The second and third hypotheses on the 7-month follow-up effects of the intervention were evaluated by comparing the changes in studied variables from baseline measurement T1 to the T3 measurement 7 months after the end of the intervention program and by comparing this change to that measured in the comparison group.
A random intercept mixed-effects model was used for evaluating these immediate and 7-month follow-up effects. The use of this model provides correct standard errors when nonindependence exists in data ( Bliese & Hanges, 2004). In these analyses, organization was used as a random term. We calculated intraclass correlations (ICC) to provide information on each outcome variable of the amount of variance that exists among organizations. Group membership (group intervention, literature package) and the baseline value measurement were the main independent variables in the analyses. The group represented the difference in the effects (change in levels of dependent variables) between the intervention group and the comparison group on the dependent outcome variables during the follow-up time. Residuals were checked for all variables in the mixed-effects analyses. Residuals for the intention to retire early and for the mental resources' variables were skewed, and the respective analyses can be regarded as suggestive. Residuals of other variables showed clear unimodal and close to normal distribution.
Based on our hypotheses, the dependent outcome variables were career management preparedness, depressive symptoms, exhaustion, intention to retire early, work engagement, and mental resources. In the analyses, age, gender, and respective baseline measurement of each dependent variable were controlled.
At the second stage of analyses, we investigated if some specific, more targeted groups would benefit most from the intervention. Based on our expectations, the moderating effects were studied by adding interaction terms of the group variable with depressive symptoms, exhaustion, or age into each model separately.
With regard to our fourth hypothesis, we used a three-step procedure to study the mediating effect of preparedness between intervention program and the outcomes at T3 ( Baron & Kenny, 1986). First, preparedness was modeled at time T2 to study the effect of the intervention program on preparedness. Next, two models, with and without the preparedness measurement at T2, were modeled to study the effect of the intervention program on the outcome variables. A variable functions as a mediator when it either eliminates or significantly reduces the effect of the independent variable (group) on the outcome. In addition we studied the mediational role of career management preparedness using the indirect effect estimation method of Preacher and Hayes (2008). In these analyses we used Indirect-macro for SAS applying bootstrap results.
We used the multiple imputation method to treat missing data (the MI and the MIANALYZE procedures of the SAS/STAT software). We had a small number of missing values (015 missing values) in variables used in main and interaction effects analyses models. The results of the MCAR test ( Little, 1988) showed that the missing values were not missing completely at random (? 2 = 141.997, p < .05). We excluded the dropouts at Time 3 (T3) from the main and interaction effects analyses. In other analyses, a listwise deletion of missing values was used.
We also calculated effect sizes with Cohen d estimates for the subgroups based on the statistically significant interaction terms. Effect sizes are used to give an indication of the meaningfulness or practical significance of a difference between the means of different subgroups ( Steyn & Ellis, 2009). In our study, they represent standardized differences of means between the different levels of depression, exhaustion, and age due to the intervention. According to the estimate sizes, values of 0.20 to 0.30 are considered small, 0.50 medium, and 0.80 large ( Cohen, 1992).
Results
Integrity of the Intervention
Participants' evaluations of the intervention group and its atmosphere were positive. The mean scores for the responses to the eight items on a scale from 1 to 5 ranged from 4.0 to 4.9 ( SD = 0.481.19). The most positive responses were received in connection with the following statements: the atmosphere was positive and supportive ( M = 4.9, SD = 0.48), the trainers indicated that they respected my participation ( M = 4.6, SD = 0.75), the trainers encouraged me to participate in learning tasks ( M = 4.5, SD = 0.79), the trainers inspired me ( M = 4.3, SD = 0.99), and the trainers made me feel happy or satisfied ( M = 4.3, SD = 0.97).
When group participants were asked to evaluate how much the training gave them the means to handle various situations at work, the mean scores for the responses to the five items on a scale of 1 to 5 ranged from 3.0 to 3.7 ( SD = 0.481.01). The respondents reported that the training provided means for taking care of oneself ( M = 3.7, SD = 0.93), means for coping with possible changes ( M = 3.4, SD = 0.86), means for taking care of human relations ( M = 3.2, SD = 0.94), means for solving conflicts at work ( M = 3.1, SD = 0.88) and means for planning one's career ( M = 3.0, SD = 1.01).
Table 3 presents the means and standard deviations of the study variables in both the intervention and the comparison group. The correlations between the study variables and reliability scores of the study variables (Cronbach's alpha) are presented in Table 2.
Hypothesis 1: Proximal Effect of the Intervention on Preparedness
As the intervention was designed to promote preparedness for career management, comprising career management self-efficacies and preparation against setbacks, the proximal impact of the intervention was assessed using this scale. 1 In a paired t test, the difference in change in preparedness values from T1 to T2 was significant, t(558) = 3.50, p < .001. The impact of the intervention was analyzed using mixed-effects modeling, in which organization was treated as a random effect. The results showed that the intervention was effective as hypothesized (Hypothesis 1). In the immediate follow-up (T2) after the intervention, preparedness for career management increased significantly in the intervention group compared to the comparison group (coefficient = 0.104, SE = 0.028, p < .001; see Table 4). The Cohen's d value was 0.31, indicating medium-sized effect.
We found partial support for our two hypotheses regarding the main effects of the intervention. Both depressive symptoms and intentions to retire decreased significantly in the intervention group compared to the comparison group ( p < .05; see Table 4). Moreover, mental resources increased in the intervention group compared to the comparison group ( p < .05). As there was no significant decrease in exhaustion as a result of the intervention, the second hypothesis was only partially supported. The main effect of the intervention on work engagement was significant only in one-sided testing. The mean engagement values in the experimental group had lower baseline values than in the control group, and after the intervention they reached about the same level with the control group. This is the main reason why mean engagement values are negatively correlated with the experimental condition. In our results we can see a phenomenon similar to suppression. The experimental effect is visible when we analyze data as changes or by adding baseline measurement into the model ( Kline, 2005, pp. 3738), and effect is only significant in one-sided testing. Analysis of the ICCs showed that statistically significant variation occurred only in the exhaustion scale between the participating organizations (ICC = 0.044, p < .05; see Table 5). The Cohen's d values for the significant main effects varied between -0.14 and 0.22, indicating effect sizes between small and very small.
The second set of multivariate analyses was more exploratory in nature. We found that, in addition to the direct effects of participation in the group intervention on mental health and work career, the intervention had significant interactive effects with the level of baseline depressive symptoms on depressive symptoms and on mental resources at follow-up. The participants with an elevated level of depression at baseline benefited more from the intervention than those with lower levels. The intervention also had significant interactive effects with the level of baseline exhaustion on depressive symptoms and on work engagement at follow-up. The participants with an elevated level of exhaustion at baseline benefited most. Moreover, younger participants seemed to benefit from the intervention, as it had a significant interaction with age (see Table 4).
As the beneficial effects of the intervention were largest among those participants with either highest values for depressive symptoms or exhaustion or those with lowest values for age, we roughly estimated the effect sizes in these subgroups. We subdivided the study data into three separate analyses into subgroups, which were quarters with highest values for depressive symptoms and exhaustion and those with lowest values for age, respectively. When we examined the effects of the intervention on outcome variables in these subgroups, corresponding to the found interaction effects, the respective absolute d values varied between 0.32 and 0.47, indicating medium-sized effects.
Hypothesis 4: Mediation of the Effects
The mediation analyses demonstrated that the intervention significantly increased career management preparedness (Hypothesis 1). In the mediator models, where preparedness was included in the models as a mediator, the originally significant effects of the intervention on outcome variables disappeared (see first row of Table 6). Consequently, we concluded that the found increase in career management preparedness at T2 mediated the found longer term beneficial mental health and career effects, supporting our Hypothesis 4.
For control purposes we also calculated the main results of Table 4 without imputation, which verified the results. We also checked whether perceived security of one's job moderated the studied effects and found that those who perceived their work as insecure benefited more from the intervention in terms of fewer thoughts of early retirement compared with those who perceived their jobs as very secure (coefficient = -0.72, SE = 0.19, p < .01). There were no other interaction effects of perceived job security on the found effects.
Regarding the comparison group receiving the literature package, as there were no significant differences in the mental health effects between those who reported reading all or a lot of the material and those who reported reading little or none of it, the effect of the comparison condition on the studied outcome variables can be estimated to be negligible.
Discussion
Our aim was to investigate whether a preventive, resource-building group intervention designed to enhance career management preparedness and implemented universally at the workplace would benefit employees' career and mental health in today's rapidly developing work life. We employed a social cognitive approach based on theoretical principles to motivate and empower participants to apply effective career management strategies and practices when establishing desirable work goals and to fulfill these goals during their careers. The effects of the intervention were investigated 7 months later in a controlled randomized field study. The group activities at the work organizations were organized as collaboration between the human resources department and the occupational health service provider of the organization in question. Trained group trainers worked in pairs in their own organization and used standardized method manuals.
The results regarding the main effects of the intervention both immediately after the intervention and at 7-month follow-up were fairly well in line with our hypotheses. The proximal impact of the intervention showed a significant increase in career management preparedness, fulfilling the primary goal that we had set for the intervention. In the longer run, at the 7-month follow-up, the intervention program had significantly decreased depressive symptoms among group participants. This is an important result, showing that it can be feasible and effective to enhance employees' resources for proactively better managing their own career in organizations and that this can have a longer term beneficial effect on the mental health of employees.
At the same time, the main effect of the intervention on the decrease in exhaustion was not significant. Symptoms of exhaustion and burnout seem to be more dependent on the work and career environment and the situation than depressive symptoms (e.g., Varekamp, deVries, Heutink, & van Dijk, 2008). Parallel to this, previous research suggests that exhaustion can be more effectively decreased by interventions involving changes in work organization and procedures ( Htinen, Kinnunen, Pekkonen, & Kalimo, 2007; Le Blanc, Hox, Schaufeli, Taris, & Peeters, 2007), but we were unable to influence directly the organizational arrangements in this study. This emphasizes the need to integrate resource-building interventions in organizations as part of everyday organizational practices and the need for corresponding organizational arrangements if so indicated during these group interventions.
The main effect of the intervention resulting in the increase of mental resources was significant but somewhat suggestive due to the skewed measure. The increase in work engagement was significant only in one-sided testing. Work engagement has earlier been shown to be associated with increases in work-related self-efficacies and is hypothesized to result from these ( Llorens et al., 2007). However, to our knowledge no intervention studies exist that demonstrate a significant longer term effect on work engagement. Therefore, the relationships between self-efficacy-based interventions and work engagement should be studied further. In this respect, our findings can be seen as encouraging.
The intervention also resulted in a significant decrease in intentions to retire. It is promising that plans to retire early can be influenced through resource enhancement intervention. Due to the somewhat skewed measure, this effect can be considered more as suggestive. The role of psychological resources in the retirement process is still poorly understood and should be studied further ( Volanen et al., 2010).
We also wanted to further investigate whether the group intervention should be targeted more at mental health risks or age, and we accordingly analyzed the corresponding moderators of the effects. With regard to depressive symptoms, mental resources, and work engagement, the results demonstrated that the participants initially at risk and suffering from higher levels of depression or exhaustion benefited most from the group intervention. This result is in line with earlier studies on preventive interventions aimed at reducing depression ( Brunwasser et al., 2009; Caplan et al., 1989; Vuori et al., 2002). We also found that the intervention benefited younger participants more than older participants, as their exhaustion scores decreased more as a result of the intervention than those of the others. The less experienced younger participants may benefit more than the more experienced older participants from this kind of peer group, in which participants from the same organization share their thoughts, ideas, and ways of coping and find solutions for their work-related problems together. These findings encourage more targeted use of the intervention than was the case in our study.
Our mediation analyses demonstrated that the found beneficial mental health effects of the intervention on depressive symptoms and mental resources were due to the proximal effect of the intervention on career management preparedness, which increased. However, this rise in career management preparedness only partially mediated the effect of the intervention on intentions to retire. The intervention also had a direct effect on intentions to retire, which was not mediated by career management preparedness. It may be that the group intervention also enhanced other skills or attitudes that affected retirement decisions or that our measure of career management preparedness poorly encompassed items in this area.
The proximal effect of the intervention on the mediating career management preparedness was medium sized, and in similar earlier experiments the proximal effect of a career management intervention seems to fade during a half-year follow-up ( Kiuru, Koivisto, Mutanen, Vuori, & Nurmi, 2011). However, career management preparedness mediated the effects of the intervention to more long-term mental health and career effects. This mediation process should be studied more, because knowledge of beneficial short-term mediators constitutes the foundation for effective behavioral interventions.
Taken together, the positive intervention results seem promising. Resulting wide-ranging beneficial outcomes for employees' career and health can be expected to also benefit the organizations. Our intervention and concept of career management was somewhat broader than usual, also comprising the interests of sustainable and occupationally healthy careers. The results show that resilience can be strengthened and capacity can be built in addition to preventing depressive symptoms and enhancing mental health at worksites. It also seemed that participants benefited the groups largely independently of whether there were ongoing major changes in their work organizations. However, as the general finding of worksite intervention studies has been that in order to be effective, interventions should focus simultaneously on the individual and the organizational level ( Lamontagne et al., 2007; Sockoll et al., 2008), the organizational side in our intervention still needs much further development.
The implementation of the intervention applied an integrative approach, combining the knowledge bases from OHS and HRM in the areas of occupational health and vocational skills and training. In practice, this was obtained by training the group trainers from the organizations representing these two domains. The objective was to increase the cooperation of OHS and HRM and to integrate our program into everyday organizational practices. In these practices, information on work-related development plans and on health and well-being can be utilized, for instance, when designing work tasks or directing additional vocational rehabilitation to specific risk groups. The trainers from the organizations were expected to continue this work by incorporating the aims and methods of the program into their future work, thus strengthening the beneficial long-term effects of the method on the mental health and the careers of employees at both individual and organizational levels. This may, however, have been realized to a varying degree.
The major methodological strength of our study was that we were able to employ a randomized controlled trial within the participating organizations. As randomization was successful, we were also able to look more closely at the mediation processes. The main effects of the intervention may appear relatively small in scale at 7-month follow-up. The effect sizes (Cohen's d) found in this study can be regarded as small ( Cohen, 1992), a size found to be typical of interventions for work-related stress ( van der Klink, Blonk, Schene, & van Dijk, 2001). However, because the significant interactions found in our analyses concerned depressive symptoms in particular, the effects of the intervention within subgroups are larger. This suggests that high-risk participants could be targeted for participation. Moreover, when our results are projected to larger populations, the overall impact on the population of employees becomes substantial.
Participation in our intervention study was voluntary for both the organizations and individual employees, and we did not screen or select our participants. Consequently, many of those who volunteered to participate in our study were women in white-collar occupations. Organizations with blue-collar employees found it difficult to fit group activities into their work schedules, which often consisted of shift work. Men seemed more reluctant than women to participate in a study involving group activities. Furthermore, most of the participating organizations had a female majority. It is also noteworthy that in all study organizations the program was new and not yet established as an organizational practice.
Another, more general concern regarding intervention evaluations is that participants who have undergone time-consuming and intensive treatment may be motivated to report having benefited from the intervention to justify the investments made by themselves and by those delivering the program. We introduced the experiment as having two experimental conditions, but participating in group training was obviously more intensive than receiving a literature package for reading. It is possible that this kind of difference in treatment intensity may have had differing influences on motivation during reporting, especially regarding the measurements just after the intervention. However, as the 7-month follow-up time was relatively long, the possible influences should no longer be significant. Clark and Sugrue (1991) have demonstrated in educational research that uncontrolled novelty effects become small (<1% of the variance) after 8 weeks, and our follow-up time was 3.5 times longer. At Finnish worksites, various personnel interventions are common, and participants were likely to perceive the Towards Successful Seniority Program as one program among others. With regard to the possible uncontrolled novelty effects on career management preparedness, their significance for the results of the study can be considered minor. Our focus was the mediating role of preparedness and not its level, and the level should not affect the basic structure of mediation models.
A vast majority of the studies on work and stress have focused on the development of the harmful effects of stress. More studies should focus on the causal process of increasing resources and mental health at work. This research line needs more controlled intervention studies.
Improvement of work and career management in today's rapidly changing work life is beneficial for the success of both individual workers and organizations. It promotes commitment to work and better mental health and well-being, especially among those at risk. Collaboration between OHS and HRM actors provides a channel through which these issues can be further incorporated into organizational activities. Better convergence of work tasks and individual interests, more consideration of individual life situations, and increased motivation and sense of empowerment among employees may benefit both the employee and the organization. This study demonstrated that it is both possible and feasible to enhance employees' health resources by strengthening their career management preparedness and, consequently, to develop organizational practices for promoting the confidence and resilience resources of employees.