Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Lessons Learned

German companies need to develop strategies to maintain the work ability and health of their ageing workforces to the retirement age of 67. Next to a variety of fields like management, qualification, work organisation, and health promotion, working-time designs that meet the requirements of ageing workers in their various life phases are of particular importance.

Under ten sub-projects implemented in six German companies from the automobile, steel, pharmaceutical, and chemical industries, ageing-appropriate working-time models (e.g. part-time work, short breaks, ageing-appropriate shift rotas, long-term time accounts) were examined and/or newly developed, introduced, and evaluated in the KRONOS research project. This chapter presents a total of ten sub-projects in each of the following working-time design fields:

  • number of working hours per day, week or year,

  • adequate rest periods,

  • shift work,

  • employee influence on working-time design,

  • long-term time accounts,

  • beginning and ending of the daily working time.

Specifically, the following recommendations have emerged from KRONOS: uniformly reducing the daily working hours of all older employees is not recommended as personal health and work ability vary very widely among individuals. However, older employees who do hard work should take more frequent breaks. Shift systems based on fast forward rotation have a more favourable effect on work ability and health than traditional shift schedules that rotate backwards on a weekly basis. It is also recommended reduce the number of night shifts per person and year. The study shows, for example, that night-shift manning levels may be reduced even in partly automated processes by rescheduling certain activities from the night to the morning and afternoon shifts. Morning shifts should not begin before 06:00 a.m. Results also suggest that working-time choice and sabbatical options should be offered more widely.

An important result was, that there is no such thing as an “ideal” working-time model for an ageing workforce. First, there is a great inter-individual scatter in the work ability and the working time preferences of the employees and second companies follow different objectives with regard to operating hours and adaptation to demographic change. This is why working-time models need to be tailor-made.

Essential factors for the success of ageing-appropriate working-time models include:

  • giving consideration to ergonomic recommendations,

  • giving employees an opportunity to influence the design of their working time,

  • ensuring favourable framework conditions (e.g. managerial attitudes, adequate participatory strategies of introduction), and

  • integrating all constituent projects into a corporate master strategy to deal with demographic change.

Aim of the KRONOS Research Project

The conceptual question of the KRONOS project is: How can we maintain, support, and improve the health and work ability of our ageing workforce?

Only a comprehensive strategy of multi-dimensional interventions can help to maintain or improve the work ability of the ageing workforce. One important intervention within this strategy is the design and implementation of suitable working-time models.

The design of a working-time model has to take the following parameters into consideration:

  • duration of the working time

    • (per day, week, month, year, phase of worklife),

  • position of the working time

    • (time of day, weekday, season, phase of worklife),

  • start and end of shifts

    • (fixed, staggered, sliding, variable),

  • flexibility of the working-time arrangement

    • (no adaptation possible, planned change between different models, short-term adaptations),

  • autonomy or control over arranging one’s working hours.

Only 27 % of employed and 8 % of the self employed people have the classical working day, 07:00/08:00 a.m. to 05:00/06:00 p.m. from Monday to Friday (3rd European survey on working conditions). The number of unfavourable working-time arrangements has increased and the question is: Are there suitable working-time models for the ageing workforce? Therefore, the aim of the KRONOS research project was to study the effects of various working-time models on the ageing workforce and to deduce practical recommendations from the results.

Potential Negative Effects of Inadequate Working-Time Models

There are a number of potential negative effects of inadequate working-time models due to, for example,

  • night-and shift work,

  • long working hours,

  • intensified work stress without adequate rest periods,

  • inadequate compatibility of work with private life, and

  • the transition from work to retirement.

Shift workers do not live in harmony either with their ‘biological clock’ (the diurnal rhythm of physiological functions) or the life rhythms of their social environment. This may result in a number of complaints (e.g. sleep or appetite disturbances)’ social life and performance impairments and, in the long run, gastro-intestinal or cardiovascular diseases (Knauth 2007a). These problems escalate in many shift workers as they grow older e.g. more sleep problems (Bosch and De lange 1987; Foret et al. 1981; Härmä 1995; Moneta et al. 1996; Smith and Mason 2001; Tepas et al. 1993), enhanced risk of cardiovascular disease (Knutsson et al. 1986). On the other hand, there are shift workers who display no sign of major problems even at a relatively old age.

There are a number of preventive and compensatory measures designed to mitigate the psychological, social, and health problems of shift workers, one of the most effective being the design of shift rotas on the basis of ergonomic findings. When shift schedules for ageing workforces are designed on the basis of the results of recent studies (e.g. Härmä et al. 2006; Knauth et al. 2009a), not only older but also younger shift workers stand to benefit.

Important design parameters include working-time scheduling (shift work) and duration. Overlong daily working hours as well as working weeks that are longer than average may have a negative impact on, for instance, sleep quality, alertness, health, and social life, and they may cause accidents (Knauth 2007b; Wagstaff and Sigstad Lie 2011; Wirtz 2010). A combination of long working hours, a heavy workload, and unfavourable influences from the working environment is particularly critical. As it may be supposed that older employees have more problems dealing with such stress combinations, it has been suggested to reduce the working time (per day, per week, or per year) of older people in general. This misguided theory is based on the assumption that the work ability of all older employees is inferior to that of younger workers.

There are other problems next to shift work and overlong working hours, namely the increasing intensity of work and the narrowing scope of action (Ilmarinen 2006). In those cases where the causes of the problem cannot be eliminated, providing especially older employees with the option of taking an adequate number of breaks when needed appears to make sense (Ilmarinen 1999).

Another problem that has been discussed frequently in the past (Costa et al. 2004; Giebel et al. 2004; Janßen and Nachreiner 2004; Kandolin et al. 2001; Knauth 2007c) is the impossibility of reconciling one’s professional and private life in the face of unfavourable work scheduling (e.g. flexible working hours based solely on the interests of the company). However, there are insufficient adequate working-time models being practised in the field and, more importantly, there are still too many managers who believe that the employees should be left to solve the problem on their own (Freier 2005). While child care is one of the key interests among younger employees, older employees are more interested in looking after relations. At all events, reconciling one’s professional and private life can be done only where employees are given more influence on their daily, weekly, or yearly working time.

Similarly, retirement may be managed in various ways. Available options include the block model, gradual withdrawal from professional life, and working 100 % to retirement age. The block model of flexible retirement (e.g. working 100 % for 2.5 years and 0 % for another 2.5 years until retirement age), which has been generally used in Germany in the last years, does not make sense from the ergonomic point of view. More desirable alternatives include sabbaticals for regeneration during the worklife or a gradual withdrawal from it. Retirement from worklife is one of the most significant and largest transitions to occur during the life course. Withdrawing gradually and stepwise from worklife helps preparing for changes in private life, for example, in income, health services, functional capacity, hobbies and friends as well as experience sharing with a successor in the company.

Research Hypotheses and Methods

The key research hypotheses of the project may be summarised as follows:

The effects of age-differentiated working-time models on the health, work ability, and job satisfaction of an ageing workforce as well as on corporate performance will be all the more positive

  • the more opportunities employees have to influence the design of their working time in the different phases of their life,

  • the more ergonomic recommendations on working-time design are taken into account at an early stage, and

  • the more favourable the prevailing framework conditions are (e.g. management attitudes towards older employees, ergonomic workplace design, in-service training for all age groups, working conditions that encourage learning, adequate strategies for launching new working-time models).

Specific research hypotheses for each sub-project are presented in Knauth et al. (2009a).

Various working-time models were developed, implemented, and evaluated using a mix of different methods (questionnaire surveys, interviews, working-group sessions, workshops, software development) (Table 1). For further information, see Knauth et al. (2009a).

Table 1 Overview of the methods used in the various sub-projects
Full size table

In six companies from different industries, a number of sub-projects were realised whose objectives will be briefly described in the following.

The investigation in Company 1 (automobile industry) aimed to establish the extent to which employees and managers are motivated to work part-time themselves and/or tolerate part-time work by colleagues and employees. Another objective was to think about a strategy that might help to enlarge the scope of part-time work. A questionnaire study was realised with 374 employees.

One of the sub-projects implemented in Company 2 (chemical industry) similarly aimed to identify opportunities to extend the scope of part-time work. Thirteen employees were interviewed. Another sub-project in the same company was designed to analyse the experience of non-pay scale employees with long-term time accounts in the last two years. Ten non-pay scale employees were interviewed. Under the third and largest sub-project in the Company, medical examinations as well as a newly-developed questionnaire were used to establish the work ability of 981 employees working under three different working-time systems (3 × 12 h, 4 × 12 h and daytime work). All participants were examined by company physicians to get objective medical data [e.g. Framingham and PROCAM risk factors as well as many relevant ICDs, details cf. Knauth et al. (2009a)]. The questionnaire included questions referring to the type of shiftwork, lifestyle factors, social factors, and the Work Ability Index (WAI).

The objective of the first sub-project in Company 3 (automobile industry) was to develop a holistic method of evaluating shift schedules on the basis of ergonomic recommendations and the employees’ views. 104 shiftworkers participated in a questionnaire study. Among other things, interviewees were asked to assess the importance of the major shift-system characteristics (e.g. number of successive days off, number of successive night shifts, number of days off after the last night shift, shift duration, start of the morning shift, forward shift rotation) under the aspects of family, leisure and health. A scale of five ratings ranging from “entirely unimportant” to “very important” was used. Results were afterwards used to weight ergonomic criteria (deduced from the literature cf. Knauth and Hornberger 2003) for the design of shift systems (Knauth et al. 2009b).The new “multi-perspective IT evaluation tool for shift schedules” that emerged from these investigations has been described by Knauth et al. (2009b). The second sub-project in Company 3 was to check whether and, if so, how the individual stresses caused by night-time work can be reduced in a partly automated production process by reducing night-shift manning levels.

Based on the theory that is better to have several short breaks than a single long one, and that a good break schedule is even more important for older than for younger workers where hard work is concerned (Ilmarinen 1999), one of the sub-projects in Company 4 (steel industry) was designed to optimise the existing unfavourable break schedule and investigate the effect of breaks on younger and older employees. The objective of another sub-project in Company 4 was to study the impact of specific shift-change times on quality of sleep, tiredness, and reaction time. Another item of investigation was to reduce tiredness in the morning shift by moving the start of the shift from 05:30 a.m. to 06:00 a.m. so as to allow more time for sleeping. In Company 4 the following methods were used:

  • vigilance test (Walter Reed palm-held psychomotor vigilance test, cf. Thorne et al. 2005),

  • visual analogue scale for subjective drowsiness scaling (cf. Kiesswetter 1988),

  • scale to identify physical complaints (cf. Corlett and Bishop 1976)

  • sleep questionnaire (in-house development).

  • The numbers of measurements are shown in Table 1.

The sub-project in Company 5 (pharmaceutical industry) was mounted to develop and implement a long-term time account concept for the workforce and evaluate a one-year test phase.

In Company 6, (chemical industry) the authors introduced a working-time choice model (Knauth et al. 2003) ten years ago which may be regarded as a model of the future for ageing workforces (Knauth 2009). The purpose of the project was to conduct a follow-up survey to establish the track record of the model over several years. After various working-group sessions and an information meeting, it was concluded that a survey would not be appropriate because the introduction of short-time work had caused considerable irritation within the plant. The works’ council was afraid that only a minority of the workers would be willing to participate in a questionnaire study.

Results

In the following, the main findings will be summarised in seven recommendations relating to various aspects of working-time design shown in Fig. 1. Each recommendation will be illustrated by selected results. The remaining findings have been published in the research report (Knauth et al. 2009a)

Fig. 1
figure 1

Design fields for ageing- and life phase-related working-time models

Full size image

Number of Working Hours per Day

Recommendation 1 (Part-Time Work)

Uniformly reducing the daily working hours of all older employees is not recommended as personal health and work ability vary very widely among individuals in the older age groups.

Broken down not by age but by the number of years which an individual spent working in shifts, the WAI totals shown in Fig. 2 display extensive inter-personal scatter. If all or most of the experienced shift workers would have a bad WAI, we would recommend to reduce the number of working hours per day. However, the group with the prolonged shift experience contained persons with a good as well as others with a bad WAI. In addition to this intra-individual scatter of subjective health and performance there is also an intra-individual variability of affect at work (Grube and Hertel 2008). In the ‘more than ten years’ category in Fig. 2, the scatter would probably be even wider without selection (healthy worker effect). Obviously, the solutions needed for people with a good WAI are different from those for persons whose WAI is bad (see also Ilmarinen 1999; Ilmarinen and Tempel 2002).

Fig. 2
figure 2

WAI total versus shift biography (Karl et al. 2007)

Full size image

Adequate Regeneration Periods

Recommendation 2 (Breaks)

In a steel mill, 92 shift workers assessed the regeneration value of each break after it was over with help of a visual analogue scale ranging from “very restful/refreshing” to “not at all restful/refreshing”. Shift workers above 40 years of age rated the regeneration value of breaks significantly (p < 0.05) more badly than those below 40. This appears to underpin the demand for more breaks for older employees, e.g. published by Jansen and Haas (1991). They refer to a survey carried out by Shepherd and Walker in the metal industry: Economic losses per working shift dependent on the work strain, the breaks and the age of the workers. Under the condition “very heavy load without breaks” there were more annual production losses per working shift in the group of workers over the age of 45 years (7.55 %), than among the workers under 45 years (6.77 %). These losses decreased when “serial breaks” were allowed (<45 years 3.64 % and >45 years 3.51 %).

Working-Time Scheduling/Shift Work

Recommendation 3 (Shift Rota Design)

Shift systems designed to conform to recent ergonomic recommendations (e.g. fast forward rotation) have a more favourable effect on the WAI than traditional shift systems that rotate backwards on a weekly basis.

The lower part of Fig. 3 shows two shift schedules. The upper shift system, which rotates forward quickly, was introduced by the authors in Company 4 (steel industry) twelve years ago (Gördes et al. 2004). This shift system conforms to ergonomic recommendations regarding shift rota design. Conversely, the lower shift system implemented in Company 3, which rotates backwards slowly, does not comply with these recommendations.

Fig. 3
figure 3

Comparison of WAI averages under an ergonomically favourable (Company 4, N=391) and an unfavourable shift system (Company 3, N=104) broken down by age group (Knauth et al. 2009a)

Full size image

Under the quickly rotating shift system, average work ability totals were significantly (p < 0.05) superior in all age groups to those of the traditional shift system. The data are based on cross-sectional studies. Unfortunately, longitudinal studies over many years and including all possible factors influencing the work ability could not be realised.

The working-time system is not the only factor by which the WAI is influenced negatively (e.g. strenuous physical work) or positively (e.g. adequate behaviour of the management towards older employees) (Tuomi et al. 1997). When comparing the shift systems in Companies 3 and 4 (Fig. 3), the fact that the condition of the working environment in the steel industry is markedly inferior to that in the automobile industry must be taken into consideration. Nevertheless, the WAI figures of Company 4 are significantly superior. This finding seems to underpin the importance of age-appropriate shift systems.

The fact that the WAI figures did not deteriorate in the older age groups is probably due to selection effects (healthy worker effect).

Recommendation 4 (Reduced Night-Shift Manning Levels)

As the night shift is the most critical of all shifts as far as sleep quality, tiredness, performance, and health are concerned, it is recommended to reduce the number of night shifts per person and year. Attempts should be made to reduce night-shift manning levels by rescheduling certain activities from the night to the morning and afternoon shifts.

In a partly automated engine manufacturing plant belonging to Company 3 (automobile industry), a concept was developed to reduce night-shift manning levels. The first step taken was to identify any activities that could be rescheduled from the night shift to the morning or afternoon shifts. Mathematical simulations, which were carried out by the company, but could not be published, demonstrated that the manning level of the night shift could indeed be reduced by one third.

In the second step, we organised a workshop in which we cooperated with shift workers on developing three alternative discontinuous shift systems that largely conformed to ergonomic guidelines. Figure 4 shows one of these models.

Fig. 4
figure 4

Example of a discontinuous shift system involving 10 sub-groups and a one-third reduction of the night shift manning level

Full size image

If, for example, ten persons work on Mondays in each of the sub-groups shown in Fig. 4, the morning and afternoon shifts will be manned by 30 persons and the night shift by no more than 20. While this concept was developed during the duration of the KRONOS project, its implementation was deferred beyond the end of the project.

Employee Influence on Working-Time Design

Recommendation 5 (Working-Time Choice)

Models that permit employees to switch between various weekly or annual work schedules in the course of their active life (working-time choice) are meaningful and attractive not only for older workers but also for younger employees who might want more leisure time.

In Company 6, shift workers working under a continuous shift system may choose between the following three weekly work schedules: 33.6/35.0/37.5 h/week (Knauth et al. 2003). If, to develop a continuous shift system, we divide the 168 h of work per week by 5 teams, we arrive at an average working week of 33.6 h in the basic shift schedule. However, a shift worker wishing to reach the full 37.5 h stipulated in the collective agreement for the average working week, which would give him his full pay, will have to work twenty-two extra shifts per year. If he opts for part-time, he receives less pay. A working week of 33.6 h requires no extra shifts, while part-time work of 35.0 h per week calls for fourteen extra shifts per year.

In the pilot year, it was mainly older shift workers suffering from gastro-intestinal or sleep problems who opted for part-time work. Later on, however, younger shift workers interested in having more leisure time similarly opted for part-time work under the continuous shift system.

At the end of the pilot year, even full-time workers were more satisfied with the working-time regulation than before even though they had not taken advantage of the option to choose their own working time. The mere chance of switching to part-time should a personal need arise in the future had a positive effect on satisfaction (Knauth 2009).

Next to working-time choice, there are numerous other options for granting shift workers greater influence on the design of their working hours, of which only a few will be named below by way of example (Knauth et al. 2009c):

  • time-autonomous groups,

  • holiday planning by shift workers,

  • ‘personalised’ duty rotas (Gauderer and Knauth 2004),

  • variable working hours by group consultation,

  • functional flexitime with three working-hour categories:

    • core time (all are present),

    • functional flexitime (hours directly before or after the core time in which at least one member of a group is present to handle customer contacts),

    • individual flexitime (hours beyond the aforementioned period in which workers may come and go as they please),

  • time windows,

  • short-and long-term time accounts.

  • The last-named option was investigated under the KRONOS project.

Recommendation 6 (Long-Term Time Accounts)

From the ergonomic point of view, sabbaticals that promote regeneration are better than models under which employees work very many hours of overtime so that they may retire early later on, when their health may already be impaired. Options to pay into and withdraw from long-term time accounts must be tailored to the needs of the target groups.

Long-term time accounts are one option of adapting working-time designs to individual life phases. Thus, for example, long-term time accounts permit taking sabbaticals or retiring early without loss of income. The advantages of long-term time accounts extend not only to employees but also to employers as they may serve to enhance the attractiveness of the jobs offered, secure the loyalty of qualified employees and maintain their performance (if, for example, sabbaticals are used for purposes of regeneration or professional development).

A long-term time account concept was developed, implemented, and evaluated in two pilot departments of Company 5 (pharmaceutical industry), a manufacturing and a laboratory operation. The long-term time account was more attractive for the workers younger than 40 years than for the workers older than 40 years. On a scale from “entirely true” (1.0) to “not at all true” (5.0) the mean value for the younger workers was 1.76 and for the older ones 3.64.

Figure 5 shows sample results of a survey addressing desired options of withdrawing from long-term time accounts. Most interviewees thought that early retirement was either ‘important’ or ‘very important’. Family time followed in second place.

Fig. 5
figure 5

Desired withdrawal options from long-term time accounts (N=54) (Knauth et al. 2009a)

Full size image

As far as sabbaticals are concerned, most interviewees thought that relatively short leave periods such as four weeks made sense.

86 % of the full-time workers stated that they were able to “pay” overtime into their long-term time account, whereas 53 % of the part-time workers answered in the same way. 92 % of the under 40 years old workers wanted to use the long-term time account for an early retirement in contrast to only 70 % of the over 40 years old workers. More of the under 40 year old workers (64 %) than of the over 40 years old workers (27 %) wanted to use the long-term time account for time with their family. As the questions about objectives, options, limitations, and information levels revealed significant differences (p < 0.05) between full-and part-time workers as well as between younger and older employees (p < 0.05), it will be important to tailor long-term time accounts to the needs of the respective target groups and involve the persons concerned in the development of the concept.

Additional experiences with long-term time accounts and sabbaticals in four companies have been described by Zimmerman (1999, 2003).

Beginning and End of the Daily Working Time

Recommendation 7 (Start of the Morning Shift)

The morning shift should not begin before 06:00 a.m. as an earlier start will have a negative impact on the quality of sleep before the morning shift as well as on tiredness and reaction time during the first hours of the shift.

This recommendation applies to average workforce members, not to extreme ‘early birds’.

In Company 4, the tiredness of shift workers was measured on a visual analogue scale (Kiesswetter 1988) from wide awake (=1) to very tired (=9). In Fig. 6 the data of the control group and the pilot group were combined. Before the intervention of the pilot group (shifting the start of the morning shift to 06:00 a.m.) both groups started the morning shift at 05:20 a.m.

Fig. 6
figure 6

Results of 4,796 measurements of subjective tiredness in steel mill (Karl et al. 2006)

Full size image

The average duration of sleep of the pilot group (5.6 h) and the control group (5.2 h) did not differ significantly when starting the morning shift at 05:20 a.m. However, at the beginning of the morning shift, both groups were almost as tired as at the end of the night shift as shown in Fig. 6 (more details see Knauth et al. 2009a).

Reaction-time measurements at the beginning and end of a shift produced comparable results (Knauth et al. 2009a).

Discussion

One important conclusion to emerge from the KRONOS research project is that there is no such thing as an ‘ideal’ working-time model for an ageing workforce.

The inter-personal scatter of work ability and health among employees increases with their age. This being so, you may find persons with a good Work Ability Index as well as others with a middling or even a bad WAI among shift workers as well.

Moreover, employees’ working-time preferences change inter-as well as intra-individually during their active life, depending on, for example, their marital status, the age of their children, their non-professional activities, health, and need for money. We recommend, therefore, creating personalised solutions for employees, such as the option to choose their weekly or annual working time. In addition, other options to influence their working-time design might be offered, such as the joint planning of holidays by the members of a group, involvement in the development of a new shift system (which is required in German law to conform to recent ergonomic findings), sabbaticals, and agreements in time-autonomous groups about modifying shift-change times on short notice.

On the corporate side, too, various requirements apply to working-time and operational design. What has stood the test of time is tailor-made working-time models that reflect corporate objectives as well as the employees’ wishes and the findings of ergonomics. The objective should always be to create a win–win situation for both employers and employees. An employer striving for greater working-time flexibility should give his employees a chance to influence their working-time design. Thus, for example, Kandolin et al. (2001) proved that combining corporate and individual factors in the design of working hours has a positive effect on the well-being of individuals and the conformability of their working and family life.

Working-time design for an ageing workforce should be embedded in a corporate master strategy for demographic change which covers management, work organisation, qualification, and health promotion as fields of action (Knauth et al. 2009c). In this context, the attitude and behaviour of the management are of particular importance. A Finnish longitudinal study found that the work ability of persons who were highly satisfied with the behaviour of their superiors was 3.6 times better than that of persons dissatisfied with it (Ilmarinen and Tempel 2002).

Outlook

Although there are a number of cross-sectional studies on potential negative effects of different dimensions of unfavourable working-time arrangements of younger and older workers, well-controlled longitudinal studies on the effects of age are missing.

Two trends, i.e. first the ageing workforce and second the increasing unsocial and irregular working hours, mean an increasing occupational and safety risk. There are studies, showing that individual flexibility may alleviate the negative effects of the company-based flexibility in working hours on subjective health, safety and the interference of the working hours with social well-being (Costa et al. 2004; Bohle et al. 2011; Kecklund et al. 2011; Takahashi et al. 2011; Viitasalo et al. 2008). Therefore Costa et al. (2004) suggest:

  • “To develop methods for the evaluation of the impact of flexible working hours on workers’ conditions, focusing more on human-centered outcome variables on health and well-being (long term effects, more objective assessments), preferably longitudinal studies with adequate time intervals according to the different effects monitored.

  • To analyze better the interaction between work load, flexible working hours and work organization, and to support interventions studies with careful control of the effects (…).

  • To develop guidelines on non-impairing or positive arrangements of flexible working hours according to ergonomic principles through participatory planning and publishing “best practice” examples showing that this is possible.”

Besides research needed referring to flexible working hours we do not know enough about optimal arrangements of breaks for older workers and optimal arrangements of stepwise retirement.

Last but not least more fixed time windows for learning have to be provided for the ageing workers (e.g. specific days within a shift system). Although everybody agrees that lifelong learning is necessary, adult education decreases significantly in the oldest age-group of working people (Ilmarinen 2006). However, not only tight schedules at the workplace but also an insufficient support by the superior are obstacles to vocational training of older workers.