Effect of a participative action intervention program on reducing mental retirement

Previous studies show that mental retirement consists of three factors: developmental pro-activity, work engagement and perceived appreciation [8, 9]. An indifference to learning and development can result in a decline, or even loss of skills, in skills obsolescence, a decrease in sustainable employment for both the internal labor market of an organization as well as the external labor market or it can even result in job loss [3, 4, 10]. Furthermore, engaged employees are better connected to their work, cope more effectively with job demands and perform better in their work [11, 12], while lower engagement is related to more sickness absence [13]. In addition, more perceived appreciation is linked to more job commitment [14, 15].

In the literature the concept of mental retirement has been mentioned before e.g. [16, 17]. However, in these earlier studies mental retirement has a rather different definition in which it is specifically linked to actual retirement and age. Mental retirement is defined for example as the cognitive decline that seems to occur after actual retirement [16]. This decline is caused by a lack of cognitive stimulation and mental exercise, which occurs when someone is retired as well as when an employee is still working but has the prospect of nearby early retirement. In another study mental retirement is defined as a decrease in work engagement for employees who are facing actual retirement [17]. However, this study finds no support for this relation. Another related concept that is also linked to age, is the ‘older worker identity’. This refers to the internalization of negative attitudes and beliefs regarding the older worker, mostly based on stereotypes (e.g. low motivation, resistance to change, inflexibility and lack of interest in learning) [18, 19]. This internalization can be caused by discrimination in career opportunities and feelings of deprivation in comparison to others. Older worker identity is positively related to full retirement and negatively to late retirement, job mobility and development opportunities on the job [19, 20].

In contrast to the concept of older worker identity and the definitions mentioned earlier on mental retirement, mental retirement is not necessarily connected to age since a previous study showed no differences in the level of mental retirement in different age groups [8]. Although studies show that older employees participate to a lesser degree in training and maintenance of their working skills [21‐24], a study on lifelong learning in the Netherlands shows that there may be a trend shift over time with regard to training participation [5, 25]. In the past, training participation clearly declined with age, but data from 2010 indicates that training participation remains stable. In addition, studies have shown that the importance of meaningful work, development opportunities and being appreciated increases with age of retirement [26, 27].

Although mental retirement is a fairly new concept and more research is needed to study the predictors and effects of mental retirement, there are indications that it might have negative consequences for employees (e.g. lower job satisfaction and mental health), and therefore also for the organization (reduced productivity, absenteeism, profit loss) and society in general (costs due to early retirement, well-being) [3, 9, 13, 28, 29]. Thus, it appears important that mental retirement among workers is prevented or reduced as much as possible. Therefore, an organizational program was developed to reduce mental retirement.

Research has shown there are several features that are important in order for organizational programs or interventions to be effective. First, a tailor-made approach is necessary. Tailoring improves the utilization of the results, helps to choose the interventions that meet the specific needs of the team or organization and make effective action plans [30‐32]. However, this approach is not easy to implement or evaluate since it means that it is impossible to develop a single intervention that will be effective in every organization or team. Second, previous research has also shown that a stepwise approach is effective by providing a framework for organizations that can be used to make suitable choices for interventions [31]. This stepwise approach often consists of five basic steps that are specified for each program: 1) raising awareness and creating commitment in an organization, 2) problem assessment, 3) prioritizing, choosing interventions and making an action plan, 4) implementation of the chosen intervention(s), and 5) evaluation of the process and effects [30, 33]. A third important feature is a participatory approach. Research has shown that the use of participatory action research is very successful for occupational health interventions [31, 33‐35]. By using the knowledge, skills and perceptions within the team (of the employees as well as the supervisors/management), a feeling of joint ownership is created during the program. Participants feel more in control and experience a greater sense of fairness and justice. Thereby, not only is awareness created but also a joint responsibility for both the problems as well as the solutions for these problems. In addition, the participation can also decrease the resistance to change and smoothen the change process [33, 36, 37].

Next to the features of an organizational program the research design is also very important, especially in applied research. Although a randomized controlled trial (RCT) is often considered as the golden standard, this is not always feasible. Particularly in applied occupational health research, the interventions and the context are often complex and therefore hard to control [38‐41]. Therefore, in applied research quasi-experiments with a control and intervention group are often difficult to establish and complete in a satisfactory way. Additionally, even if a RCT is performed with success, the question remains whether the conclusions can be generalized to other organizations and individuals or that the results only apply for a selected sample of individuals [40, 42]. Using process evaluation and incorporating the measurement of intervention exposure in participants’ samples is an alternative way to cope with these problems [38, 41]. Data on exposure to the intervention can easily be obtained in an intervention process evaluation by asking participants about their experiences and exposure. This information can then be used to determine whether each participant is more appropriately placed in an intervention/exposed group or a control/not exposed group. This approach makes it possible to take into account the complex, chaotic and uncontrollable organizational settings.

To address the issue of mental retirement, a bottom-up program was designed based on the principles of a participatory, stepwise, tailor-made approach. The program consists of five steps (see Fig. 1). The program is carried out by the whole team, i.e., employees and supervisors together. In the first step, a representative delegation of the team participates in mindmapping sessions. Two sessions are held with 10–20 participants each. These sessions are each led by two facilitators. At least one of the facilitators is a researcher, the other can also be a researcher or an employee of the organization that is trained in the mental retirement program. The sessions have two goals. First, these sessions create awareness of mental retirement within the team and gain acceptance for the program and the possible changes it may bring. Secondly, the information gained during the sessions is used to tailor the model of mental retirement with organization-specific concepts and to adjust the basic questionnaire accordingly. In the mindmapping sessions the employees and supervisors discuss what they believe mental retirement is and what the predictors and effects are of mental retirement. To not only create awareness but to also set people in motion, the participants contemplate on possible actions for themselves, their supervisor and the organization to decrease mental retirement. The employees who participate in the mindmapping sessions are contacted by management or the HR-professional of the department who invite them to participate. Participation is voluntary. The employees that are approached are selected so that they represent the entire department with regard to age, sex, function, time employed and level of mental retirement (as estimated by the manager or HR-professional). Furthermore, in this stage a project group of approximately six people is created, which consists of several employees of the team, a HR-professional and sometimes the supervisor or manager. The team itself decides who will join this project group and participation is voluntary. The project group is the first point of contact for the researchers as well as for all the employees in the department.

The current study aims to explore the effects of this mental retirement program in a broad population of employees within several organizations. The effect of the program as a whole is examined, rather than the effects of the specific interventions that are implemented in step 4 of the program. As each team or department chooses their own interventions, the range of interventions is very wide which makes it difficult to examine the effects of each intervention separately. In addition, we believe that the effect of our program is due to the program as a whole (all 5 steps) and not so much to the specific interventions of step 4.

The research population of this study consisted of a sample of employees of five departments of three different organizations in the Netherlands. Three departments of the Dutch National Police participated: two departments of police officers and one facility department (N = 175; N = 185; N = 175, respectively). The other two organizations were an archive department of the Dutch government (N = 291) and one location of a non-profit organization that implements national insurance schemes in the Netherlands (N = 209). The reasons for participating in this study differed across organizations and teams, but included issues such as minimizing the effects of a reorganization, creating more awareness among employees regarding their own development, getting employees out of ‘sleep mode’ to prevent them from getting stuck in their career.

Mental retirement was measured with three concepts [8]. Firstly, developmental pro-activity consisted of four items [43]. An example item is: “I think about how I can keep doing a good job in the future”. The response categories ranged from 1 (“completely disagree”) to 5 (“completely agree”). The internal consistency in our study (Cronbach’s alpha) was .85. Secondly, work engagement was measured with six items (three items on vigor and three items on dedication) from the Utrecht Work Engagement Scale (UWES) [44], including “My job inspires me” (α = .94). Respondents were asked to describe how often they experienced the described situations (1 = never; 7 = always). Thirdly, perceived appreciation, is measured with one question: “Do you feel appreciated in your current job?”. The response categories ranged from 1 (“not at all”) to 4 (“very much”). The scale scores of the three mental retirement sub-concepts were standardized (into scores between 0 and 1) because their response categories differed. A mean score was calculated over these three sub-concepts, which created a variable that measured mental retirement in one variable (range 0 to 1; α = .64).

Missing values were imputed using multiple imputation procedures in the SPSS “Missing Values” module, based on an iterative Markov chain Monte Carlo (MCMC) method. Overall data missingness was 31.5%, mainly due to dropout at follow-up and not filling out the questionnaires completely. It has been shown that multiple imputation-based procedures are superior to case-wise deletion of missing data [45]. In our analyses, only data were used that did not have imputed data on the variable intervention exposure. The resulting data set comprised 466 members of five departments.

Multilevel modelling (MLM) [46] was used to study our hypotheses and analyze our data, which were nested at the organizational level. The MLM analyses were performed in SPSS version 25.0 Multilevel modelling (i.e. hierarchical linear model) which aims to analyze data that contains an inherent hierarchical structure [47]. In the present study the data contains two levels. The first of lowest level of the data contains individual scores of mental retirement at baseline and follow-up (within-subject level). At the second level the individuals are nested into departments (between departments). In a stepwise procedure a final model was built for each outcome. First, the presence of a random intercept was tested for each outcome, indicating whether departments have different intercepts. In the second step, the presence of a random slope was tested for each outcome measure, indicating whether departments differed in the way their mental retirement changes over time. In the final step, educational level was added as a covariate to the best-fitting model.¹ For the first hypothesis, difference scores (between baseline and follow-up) were calculated for each outcome and used as dependent variable, making the intercept of the model an indicator for the change in the outcome from baseline to follow-up. For the second hypothesis, the dependent variable was the follow-up measurement. In the final step of the analysis, the baseline measurement of the outcome and the intervention exposure (the extent employees had participated in the specific intervention(s) in their team) were added. Variables in the equation were not centered, because all included variables had interpretable zero values. The intraclass correlation coefficient (ICC) was calculated to obtain the amount of variance explained by the differences between departments.

Besides the quantitative results, also qualitative results are reported. The researchers monitored the context and setting of each department and kept track in a digital logbook. In the logbook, the sequence of planned and unplanned events was listed alongside impressions of the researchers. These impressions were based on their own observations during the sessions. At the end of each session the researchers shortly evaluated the sessions by asking the opinion of the participants. In addition, the impressions of the researchers are based on their periodical contacts with the project group. In these contacts, the project group informed the researchers on their observations as well as the progress made. Furthermore, the researchers logged all changes that occurred in the context or setting of the organizations. The logbook was kept up to date throughout the duration of the study. The logbook data were grouped per department to form a chronological list of events, including the impressions of the researchers. The qualitative results will be discussed in accordance with the five steps of the program.

Table 3 displays the effect of the program on mental retirement and its sub-concepts. Since none of the intercepts are significant, the level of mental retirement does not change between baseline and follow-up. Therefore, hypothesis 1 is rejected.

In Table 4 the results of the multilevel analyses that take the level of intervention exposure into account are shown for each outcome. There is a significant effect of intervention exposure on mental retirement. Employees who were more exposed to the intervention(s) (i.e. who more often participated (very) much in the intervention(s)), had a slightly lower level of mental retirement at follow-up. This effect was also found for two of the sub-concepts of mental retirement; developmental pro-activity and work engagement. For perceived appreciation only a tendency was found. These results show that active participation in the intervention(s) is related to a decrease of mental retirement, which is in line with hypothesis 2.

Carrying out the effect evaluation and sending out the second questionnaire was difficult. The response rate was lower (see Table 1), mostly due to the decreased focus on the program as is described above. In addition, due to the restructuring in one organization the team members changed during the study. Employees transferred to other teams that didn’t participate in the program and employees from other teams started working in a team that did participate. Of course in all organizations there were also some changes in team members because employees retired, got a new job and new employees were hired, but these numbers are quite low.

To our knowledge, the present study is the first that examines the effects of a program for diminishing mental retirement. Such a program may become more important because of the increasing interest in sustainable employability due to the rapid changes in skills demands and the labor market and the fast aging of the workforce. Therefore, there is a necessity for more attention and awareness for concepts like mental retirement. Another strength of this study is that this study shows that more generic principles like a participatory, stepwise and tailor-made approach, appear to be more important than the specific interventions that are implemented within an organization. Future research should focus more on these generic principles and study how and in what circumstances this leads to success.

However, when interpreting the findings of this study some limitations should be kept in mind. First, although an intervention exposure measure was used in this study a control group is missing. Even employees who reported that they had not actively participated in the intervention(s), are exposed to the program. All the members of a department were informed about the mindmapping sessions, the results of the baseline questionnaire and the interactive sessions where plans of actions were made for the intervention(s). In addition, only employees that reported to participate (very) much in an intervention were classified as exposed to an intervention. Employees that reported to participate only a little in an intervention were classified as non-exposure. So, in the present study it wasn’t possible to select a control group that wasn’t exposed to the mental retirement program. Second, in this applied research study the contextual setting changed continuously, for example by the restructuring in one organization. It is possible that confounding biased and influenced the data [50]. To get more grip on the contextual setting, the researchers kept track of changes to the setting in a logbook. Furthermore, multilevel modelling was used to correct the cluster effect. Third, in the present study only two questionnaires were sent out, one at baseline (after the mindmapping session) and one at follow-up (approximately one year later). It would be interesting to see what the effects of the program are when using more measurements, for instance a baseline measurement before the mindmapping sessions and follow-up measurements both on short term as well as long term. Especially since in the present study the process of implementation of the interventions was a challenge and commitment to the program was possibly lost during this phase. Therefore, future research should be longitudinal in nature, have multiple measuring moments to look into both short term as well as long term effects and have a more extensive process evaluation. Furthermore, new research should focus more on exploring the concept of mental retirement itself and also explore the predictors and the effects of mental retirement. Also, other sources of data should be considered since the present study only makes use of self-report which can be prone to recall bias. Last, new studies should use a more powerful manipulation of the groups to see what the effects are of the mental retirement program in a study with a control group with employees that have no knowledge of the mental retirement program at all, compared to a intervention group that did participate in the program.