Earlier chapters have examined the evidence indicating that more older workers are expected to be on the job over the next 20 or more years and that the workplaces and working relationships they face are changing. These discussions have also reviewed how the physical and cognitive resources of older workers are likely to match these workplace demands. We now turn attention to the range of interventions that might best enhance this matching of older workers and the working environment.

It was suggested in Chapter 1 that intervention and research needs might be approached from either of two perspectives. The first is that insofar as older workers, especially those with high skill levels, may be necessary to meet basic needs of the national economy, our society has a strong interest in retaining older workers. From this perspective our policy and research agenda should focus on the characteristics that predict which older workers are most likely to work most productively and on the best incentives and methods to encourage and enable the most productive workers to stay.

The second perspective focuses on the needs of aging workers and their families. The goal would be to maximize opportunities for workers to make informed decisions about work and retirement that are not unreasonably constrained by economic conditions. The accompanying policy and research agenda should develop information and resources that would assist older workers in making successful choices and also support initiatives to ensure equal protection for older workers on the job.

Both perspectives anticipate significant numbers of older workers and

the need for research and policies that will help accommodate their needs and enhance their safety and productivity. The American Association of Retired Persons states: “If employers are to reap the benefits of the work ethic and experience of older workers, they must design the workplace of the future to meet their needs.” We consider interventions from the premise that it is generally preferable to accommodate the working environment to anticipate and meet the needs of older workers than to attempt only changing the aging workers themselves to adapt to their environments. There are two sources for such a premise. The first is pragmatic and finds expression in the science and practice of human factors engineering. It is recognition that human beings are imperfect.

The second source is ethical and legal and finds expression in the Occupational Safety and Health Act (OSHA) of 1971: “The Congress declares it to be its purpose and policyZto assure so far as possible every working man and woman in the Nation safe and healthful working conditions…” (P.L. 91-596). These workplace protections apply equally to every worker. But all workers are not the same. There are notable differences in size, strength, age, sex, health status, genetic makeup, and other factors that affect people’s risk from hazards on the job. Since Congress intends to protect workers equally across this varied spectrum of characteristics, including age, it follows that workplaces must adapt and change to accommodate a reasonable range.

OSHA’s approach to workplace accommodation and worker protection has been based on the hierarchy of controls concept. This concept in its simplest form holds that workers should be protected by controlling hazards as close to the source as possible. For example, designing a job so that a dangerous chemical is not necessary is preferable to providing a worker with a respirator, which in turn is preferable to training the worker to be as careful around the chemical as possible. Some version of a control hierarchy has been observed by virtually all safety and health professional associations and organizations for more than 50 years. Many OSHA standards require efforts to utilize feasible engineering or administrative controls

before personal protective equipment or worker training may be considered. A more fully elaborated version of the hierarchy of controls places protective measures in the following order of preference:

1. engineering controls (e.g., elimination, ventilation, mechanical guarding);

2. administrative controls (e.g., safe job procedures, job rotation);

3. personal protective equipment (e.g., respirators, ear plugs);

4. individual behavior (e.g., safe lifting techniques); and

5. warnings (e.g., labels, bells).

While these hierarchies vary in their detail, they tend to share at their core the notion that methods of protection that do not rely primarily on individual employee behavior alone are preferred to those that do.

Extending the legal and ethical perspective and the goals of maintaining a safe and healthful work environment, however, leads inevitably to varying interpretation of the extent to which health promotion, treatment, and rehabilitative services—crucial to the general health status of older workers—should be provided at or through the workplace. This is in addition to worker education and training, retraining for changing job tasks, and the many other activities that are offered at or through the workplace. A wide variety of employee interventions to maintain and improve health and safety have been established over many decades, albeit usually not universal in coverage, and new delivery and experimental programs are continually appearing.

The Americans with Disabilities Act (ADA) is one law that anticipates and addresses the hierarchical approach by requiring job interventions and accommodations. The ADA protects workers if they have a disability that substantially limits one or more major life activities but they are able to perform the essential functions of the job with reasonable accommodations. Accommodations under the ADA require that employers make existing facilities readily accessible and usable, and that they restructure jobs or modify work.

While older workers are more likely than younger ones to have disabilities covered by the ADA, the need to accommodate older workers goes well beyond these covered limitations. This chapter considers accommodations in the broadest fashion. For example, Burkhauser, Butler, and Kim (1995), using a proportional hazards analysis with data from the 1978 Survey of Disability and Work, found that provision of an accommodation, defined broadly, slowed worker withdrawal from the workforce and delayed the beginning of SSDI payments. Potential recall bias in these studies might be addressed in a longitudinal study, using data now available from the Health and Retirement Study. No studies have examined whether workplace ac-

commodations lower the risk of occupational injuries found by Zwerling et al. (1998a) among older workers with various impairments.

We adopt here the control hierarchy as an approach to beginning a discussion of the current status of intervention strategies to meet the safety and health needs of older workers. While there is some specific evidence to support this approach, it is limited (e.g., older adults are more likely to read warnings but less likely to comprehend warning signals [Rogers and Fisk, 2000]). Therefore, we use the concept as a useful way to structure and present ideas for accommodating the needs of older workers without arguing for a rigid order of preference. Interventions relevant to all workers, but particularly for older workers, also include workplace design and redesign; worker training; learning systems and retraining issues; alternative forms of work; the relation of the workplace to community service support; worksite health promotion and illness or disorder prevention programs; and employee assistance programs, including return-to-work programs.

In keeping with the breadth of potential workplace interventions noted in the introduction to this chapter, we now consider job design, including redesign and engineering, to improve the accommodations for older workers. There are many well-documented cross-sectional studies and some longitudinal ones outlining normative changes in vision, hearing, physical strength, and flexibility with age, as examples for requisites for many work environments (see Chapter 5). Some data derive from representative national samples. These age-related changes can be expected to affect older workers if they cannot compensate for such changes. Nonetheless, to the extent that work in the future requires maximal performance rather than typical performance, and if older adults retire later or return to part-time work after retirement, design interventions will probably become necessary.

There are a variety of normative changes in vision with increased age (see Fozard and Gordon-Salant, 2001, for a review). Prominent among these are loss of accommodative power for the lens (near-vision focus), yellowing of the lens that weakens color discrimination, scattering of light in the eye due to debris in the vitreous humor, and inability to expand the pupil fully (senile miosis). Most of these changes result in less light being admitted to the eye—about one-third as much light comparing a 65-year-old to a 20-year-old in low light conditions. Due to increased scattering of light, there is also greater susceptibility to glare from light sources. There

are also changes due to loss of cells in the visual cortex that reduce the likelihood that correction via lenses will restore youthful vision.

Disease processes also contribute to the increased risk of loss of visual function with age. These nonnormative changes include glaucoma, macular degeneration, and cataract. Some of these diseases are linked to high blood pressure and diabetes. Cigarette smoking is also a well-established risk factor (Smith et al., 2001). Not well understood is the extent to which these changes are related to work environment factors. Lifetime exposure to ultraviolet light, which is higher for outdoors (blue-collar) than indoors (white-collar) work, has been found to be a risk factor for development of lens opacities (e.g., Hayashi et al., 2003). There is also evidence of higher risk of age-related maculopathy for blue-collar compared to white-collar professions (Klein et al., 2001).

Eyeglasses are a potentially effective intervention for protecting against UV exposure and for accommodating to age-related changes in near-distance vision. An increase in computer-related work (e.g., Chan, Marshall, and Marshall, 2001, who reported 4–5 hours per day of computer work at a large corporation) means that instead of reading from paper sources people will increasingly be required to access information from computer monitors. Most monitors are placed about 40 to 60 cm from the user. This is a distance that, similar to vehicle instrument panels, falls between typical near-and far-focus distances and therefore leads to difficulty for older workers in their early forties and beyond. Potential solutions involve prescribing gradient lenses (progressive bifocals) and specialized lenses just for computer work. There are empirically validated ergonomic guidelines for the positioning of monitors, keyboards, and pointing devices (e.g., a computer mouse) that can minimize strain when working with these tools (e.g., Occupational Safety and Health Act, 2002). Whether these guidelines need to be modified to better accommodate older workers is not known.

For those working in the transportation sector, particularly those driving vehicles, age-related changes in vision and visual attention (e.g., shrinkage of the useful field of view, Owsley et al., 1991) can have a direct impact on safety and productivity. There are several studies demonstrating that signs can be redesigned to make them more visible, particularly by changing spatial frequency characteristics to improve contrast (Kline, Ghali, and Kline, 1990; Kline and Fuchs, 1993). There are many suggestions for

redesigning the road environment to improve safety for older drivers (and pedestrians) as evidenced by the new guidelines for older drivers (Federal Highway Administration, 2000). Because automobile crashes are quite infrequent for the average driver, with a probability of 0.1 per year (Evans, 1991), it is sometimes difficult to show the effects of an intervention, such as changing signage, on crash rates. There is considerable evidence that older drivers (age 55 and older) adapt their driving patterns to compensate for weaknesses, such as reducing night driving and rush hour traffic exposure (Ball et al., 1998). Such strategies may not be considered acceptable by those working under time pressure, such as professional drivers.

Aging processes diminish the sensitivity of the visual system. One simple intervention is to increase the amount of light in the environment, particularly for work-related tasks. Care must be taken to avoid increasing glare in the process by controlling the light sources and the work surfaces. Field studies show that light levels in many U.S. office environments generally meet recommended levels for reading tasks of about 100 cd/m² (Charness and Dijkstra, 1999). However, there is a dearth of information about optimal light levels for older workers. Some evidence suggests that legibility of print can be boosted differentially for older office workers by increasing light levels beyond existing guidelines (Charness and Dijkstra, 1999). Information about the effects of print size, contrast level, and luminance levels on print legibility for older adults is beginning to accumulate (Steenbekkers and van Beijsterveldt, 1998). It would be useful to extend this work to applied settings using typical clerical tasks and to assess the impact of contrast for monitor-based reading tasks.

Hearing capabilities decline normatively with age (Fozard and Gordon-Salant, 2001). Pure tone thresholds decline with age, particularly for higher frequency tones and more so in men than women. Speech comprehension shows noticeable changes (for monosyllabic words) after age 50. Older adults show more masking of signals by noise. Speech compression (e.g., in automated voice mail systems) and rapid speech rate affect older adults more than younger ones (e.g., Stine, Wingfield, and Poon, 1986). Most of these changes can be attributed to loss of hair cells in the cochlea and loss of cells in auditory areas in the brain as well as to general age-related slowing in comprehension processes.

The extent to which such loss is driven by exposure to noise versus normal aging is a matter of dispute. Losses are linked to noise exposure as well as to factors such as cardiovascular disease, smoking, and dietary factors. There are a number of approaches to remediating hearing loss. In general, hearing aids have not been particularly functional in fully restoring hearing acuity because they boost both signal and noise.

Given that some have estimated that normative hearing loss is at least partially attributed to noise exposure in the workplace (e.g., Corso, 1981), prevention is a potentially useful approach. One important source of noise exposure is aging equipment. Farmers using older tractors can be exposed to noise levels in excess of 100 dB (Pessina and Guerretti, 2000). Also, hearing loss is strongly associated with livestock-related injuries for farmers (Sprince et al., 2003). Noise reduction engineering and promotion of safe practices in inherently noisy environments, such as the use of noise reduction devices (e.g., ear protective equipment such as earplugs), may be important components in preventing problems.

Given that hearing loss may pose a significant problem (particularly for older male workers), redesign can be an important tool in preventing hearing impairments from becoming disabilities. One such design change is to make use of other less-impaired sensory channels, described below, to signal important information (such as warnings).

There are many examples of using redundant channels to compensate for hearing loss. It is possible to provide both visual and auditory warnings (flashing lights with sound). Perhaps the best-known example is the use of a warning sound (beep) to indicate when a vehicle is backing up (moving in the unexpected direction). Other examples can be found in catalogs of assistive devices, such as those that supplement normal sound channels with tactual feedback (e.g., a vibrating cell phone). For those with profound hearing loss, substitution of vision for hearing is sometimes possible (flashing lights for a doorbell, closed captioning on television). Vanderheiden (1997) offers specific recommendations on redesigning to accommodate those with disabilities.

Several studies show greater comprehension impairment for older adults than younger ones at the same signal-to-noise ratios, compared to the case of detection of pure tones in quiet surroundings. Minimizing background noise should aid older workers differentially for comprehension tasks. Sim-

ply increasing signal strength, e.g., shouting over noise, results in diminishing returns (Crocker, 1997).

The Canada Fitness Survey (Kozma, Stones, and Hannah, 1991) showed cross-sectional linear decline on most fitness and flexibility variables with age, though sometimes gender interacted with age; men typically showed faster decline than women. A main effect on fitness and flexibility was shown for activity level as well, an effect that did not interact with age. Because of changes in the cohort structure of the workforce, particularly the shifts in minority composition from large influxes of Hispanic workers, current data on anthropometry (e.g., Kroemer, 1997; Peebles and Norris, 2003; Steenbekkers and van Bijsterveldt, 1998) may not predict characteristics of future cohorts of older workers. Such data are useful for designing functional workplaces. Anthropometric data typically encompass size, strength, and flexibility ranges for people’s bodies. An example would be extent of reach from a seated position. If a given worker has a shorter-than-average reach, he or she may become inadvertently handicapped and possibly suffer musculoskeletal disorders (MSDs) in work environments designed for those with a longer reach.

Arthritis, which affects flexibility and dexterity, increases in prevalence with age and affects older women more than men (Verbrugge, Lepkowski, and Konkol, 1991). Arthritis can make many manual tasks difficult to perform. Women tend to be differentially employed in clerical positions that require typing (Chan et al., 2001), implying that some accommodations may be particularly critical for them. Also, arthritis has recently been shown to be a risk factor for occupational injury, for instance, in farmers who are injured by livestock (Sprince et al., 2003).

Changes in the mechanisms supporting balance may be an important factor to consider, given the data on age-related increases in death from falls in construction and manufacturing industries (Agnew and Suruda, 1993; Bailer et al., 2003).

Shephard (1995) reviewed research on physically demanding work and suggested changes to accommodate older workers (women and men). A central concern is the likelihood of fatigue in a physically demanding task that exceeds a threshold for cardiorespiratory capacity of 33 percent maximal oxygen intake. If aerobic capability declines from about 12 to 14 metabolic units (METS) in young adults to 7 METS in the average 65-year-old, many older workers would not be expected to be able to perform other than light physical work. Given that women typically average two-thirds the aerobic power of men at all ages, older women are most at risk for excessive demands from physically demanding jobs.

There are similar problems with age-related declines in muscular strength in the general population. Guidelines for strength demands typically recommend that median load be less than 10 percent of maximal load and that peak load be less than 50 percent of peak force. Strength bench-marks are usually set for the case where at least 75 percent of women and 99 percent of men can meet job requirements safely. In practice, there are few cases of aerobic and muscular limitations found in the workplace. This may be due to physical stressors on the job that increase fitness, or because job shifts and disability remove those who cannot meet job requirements. Job redesign is a safe way to reduce physical workload to acceptable levels.

Marbach (1968) described a set of case studies of redesign of jobs to accommodate older workers. There was little formal evaluation of the effect of redesign, so these cases are more illustrative than scientifically informative. Most of the examples involved substituting machines (cranes, conveyor belts, forklifts) for human effort on lifting and moving tasks, as well as shifting workers to sitting instead of standing positions. Some involved changing the nature of the work task by shifting heavy physical tasks to other team members and having an older worker assume lighter tasks. In another case with computer equipment assembly, instructions were provided aurally via audiotape and headphones instead of with written instructions, eliminating eyestrain and freeing workers’ hands. In many cases the older workers at risk (because of work-related injuries or development of arthritis) were able to continue work in cases where they may have otherwise been forced to leave. As Marbach comments, there are probably numerous cases of small but important modifications that have been made to accommodate older workers that were not reported in the OECD survey forms. This approach of gathering examples of best practices seems promising.

Given high rates of job turnover in modern labor markets, workers are less likely to remain at the same job site over their entire career than they have during prior historical periods. Accidents are most likely to occur in the first year of employment at a new job setting (Root, 1981). Hence, there is a need to prevent injuries through effective warnings, training, and redesign of existing tools and settings.

A recent review suggests methods to improve the design of warning systems (Rogers, Lamson, and Rousseau, 2000). As that review noted, older adults demonstrated poorer ability to notice warnings and to compre-

hend warning symbols. Hancock, Rogers, and Fisk (2001) showed that older adults were more likely to report that they read warnings but were also less likely to comprehend warning symbols. There did not appear to be much literature pertaining to age differences in compliance with warnings. If one assumes that appropriate warnings are already posted in workplaces, older workers may be heeding them better, given their lower rate of accidents. The medication adherence literature also suggests that older adults are more compliant with medication routines than middle-aged adults (Park et al., 1999).

In addition to training approaches discussed below, there have been studies to assess the effectiveness of engineering, administrative, and individual-focused interventions at the workplace in the prevention of musculoskeletal disorders (MSDs). Due to ongoing changes at most workplaces unrelated to planned interventions, it is difficult to use formal epidemiology studies to determine whether ergonomic interventions are effective. A review by Westgaard and Winkel (1997) noted this problem in an assessment of the efficacy of different workplace intervention studies for MSDs. The review examines problems in the current intervention literature, ranging from lack of statistical analysis to failure to include control groups to confounders such as inadvertent changes in the psychosocial climate. Interventions they classified as mechanical exposure interventions unaccompanied by organizational change were generally unsuccessful. Production system interventions based on changing the organization of work also failed to show much benefit. Intervention studies that attempted to change the organizational culture of a work environment achieved relatively good results. So too did modifier interventions that attempted to change the capabilities of a worker through physiotherapy or exercise interventions. These latter approaches involve both targeting of risk factors for workers and intervention at the level of the individual worker. It is recognized, however, that it is difficult to isolate the modifier intervention from parallel changes, including psychosocial improvements.

More recently published studies support more optimism about the effectiveness of workplace interventions to prevent MSDs. Several studies have reported a positive impact of ergonomic interventions on low back and other MSDs among workers performing lifting and related manual material handling tasks (Evanoff, Bohr, and Wolf, 1999; Brophy, Achimore, and Moore-Dawson, 2001; Marras et al., 2000; Yassi et al., 2001). Others have found positive effects among workers using video display units (Aaras et al., 2001; Brisson, Montreuil and Punnett, 1999; Demure et al., 2000; Ketola et al., 2002). Positive outcomes of comprehensive interventions,

including job redesign and organizational change, have also been reported for MSDs among hospital workers (Bernacki et al., 1999; Carrivick, Lee, and Yau, 2002), sign language interpreters (Feuerstein et al., 2000), and office workers (Nelson and Silverstein, 1998). Negative results were found among a small group of assembly workers after jobs were redesigned to be more varied, less repetitive, and more autonomous (Christmansson, Friden, and Sollerman, 1999). A review of interventions aimed at reducing exposure to mechanical stressors concluded that there were significant benefits (Lötters and Burdof, 2002). A review of studies for carpal tunnel syndrome suggested the need for better-designed intervention investigations (Lincoln et al., 2000). In three recent intervention studies, ergonomically modified jobs have also been associated with more rapid return to work after work related MSDs (Crook, Moldofsky, and Shannon, 1998; Loisel et al., 1997; Arnetz et al., 2003).

The most recent comprehensive review of intervention effectiveness was completed by a National Academy of Sciences committee (National Research Council and the Institute of Medicine, 2001), which evaluated 20 years of formal studies along with results from a best-practices symposium sponsored by the National Institute for Occupational Safety and Health (NIOSH) in 1997 (NIOSH Effective Workplace Practices and Programs Conference, Chicago, 1997). In addition to examining reviews published through the mid-1990s, the committee identified 17 recent intervention epidemiology studies along with 40 case studies from the Chicago conference. The committee arrived at a positive conclusion about the benefits of interventions directed at reducing exposure to mechanical and psychosocial stressors. It concluded that

A less researched area is the influence on safety-of-job factors such as degree of empowerment and feelings of insecurity. Metanalysis has shown significant negative impacts of job insecurity on mental and physical health in the range of r = –0.1 to –0.2 (Sverke et al., 2002). Probst and Brubaker (2001) indicated that self-reported accidents increased as a function of the

extent to which food-processing workers considered their jobs to be insecure. The effect, as assessed by path modeling, was not direct; it operated through job satisfaction and safety motivation.

Increased degree of empowerment of teams was a strong negative predictor of accidents (r = –0.51) and of unsafe behaviors (r = –0.48) among chemical plant workers where the worksites had undergone significant reductions (Hechanova-Alampay and Beehr, 2001). Given trends toward lean production (Landsbergis, Cahill, and Schnall, 1999) that increase the probability of layoffs, and given that older workers perceive themselves to be potential targets, they may be particularly at risk for serious injury at work. This is because older workers who see their jobs to be at risk to layoff may be tempted to rush work using unsafe behaviors that lead to injury. They also suffer, on average, more serious injuries than their younger counterparts. However, research is needed to evaluate the pathways from job insecurity to accidents specifically for older worker populations.

Countering the reported trend toward increased feelings of job insecurity is the finding that older workers usually exhibit higher job satisfaction (e.g., Warr, 1992). There is a need to evaluate what characteristics of job environments are particularly important for older worker safety and whether there are interventions that improve job climate differentially for older workers.

The earlier discussion on health and job class (see Chapter 4) shows that job class, a measure of socioeconomic status, is an important predictor of present and future health status for civil servants. Work empowerment interventions may be a useful way to decrease the risk of negative health outcomes for lower SES workers.

The effectiveness of a limited number of interventions to improve work organization and job design, reduce job stressors, and create a more healthy work organization have been documented (International Labour Office, 1992; Landsbergis et al., 1997; Murphy et al., 1995; Parker and Wall, 1998; Parkes and Sparkes, 1998). These include (1) action plans by Swedish civil servants carried out to reduce work stressors, resulting in a significant decrease in apolipoprotein B/AI ratio in the intervention group but not in the control group (Orth-Gomer et al., 1994); (2) interventions on an innercity bus line in Stockholm designed to diminish time pressure and promote traffic flow, resulting in a significant decline in systolic BP (–10.7 mm Hg) among bus drivers (Rydstedt, Johansson, and Evans, 1998); and (3) among Swedish autoworkers, a more flexible work organization with small autonomous groups having greater opportunities to influence the pace and content of their work, which resulted in lower systolic BP, heart rate,

epinephrine, and self-reported tiredness than that of workers in a traditional auto assembly line (Melin et al., 1999).

There have been no published job redesign studies in the United States that have examined cardiovascular disease outcomes per se. However, some American job redesign programs have examined other stress-related health outcomes, providing valuable guidance (Cahill and Feldman, 1993; Israel, Schurman, and House, 1989; Smith and Zehel, 1992). In addition to job redesign, legislative, regulatory (Warren, 2000), and collective bargaining (Landsbergis, 2000) approaches have been attempted. A promising development in this area is recent state legislation in the United States, which provides minimum staffing levels and limits on mandatory overtime for health care workers. U.S. policy makers may find valuable legislature models in Scandinavia, the European Union (Levi, 2000), and Japan (Shimomitsu and Odagiri, 2000) that regulate work organization and job stressors as health hazards.

Training can be an important intervention for the workplace if it is placed in the proper context of environmental interventions and its limitations are properly understood. This section examines how training can be used to promote safe, healthy, and productive work for aging workers. As people age, the balance between their capabilities and the demands of the workplace may shift, requiring training and design interventions. The field of human factors and ergonomics has been influential in advocating the use of both training and design to improve the productivity, safety, and comfort of people in both work and nonwork environments. Reviews underline the importance of human factors approaches specifically for older workers (e.g., Charness and Bosman, 1992; Czaja, 2001; Rogers and Fisk, 2000). However, establishing guidelines for training and design is challenging. Variability in older worker capabilities (see Chapter 5), diversity in workplace settings, and the changing nature of work (Hunt, 1995; Landsbergis, 2003) contribute to this difficulty.

It is useful to outline some of the assumptions underlying our reasoning about the role of training and design. People tend to satisfice (Simon, 1969) when problem solving; that is, they choose good enough, rather than optimal, solutions. Hence, most workplaces are not likely to be optimally designed for safety and productivity. For similar reasons, training packages are unlikely to be optimized for a given set of participants. Because knowledge about the types of normative changes that occur as people age is not widespread, we expect that most workplaces will not be optimized for an aging workforce. Finally, we do not expect older workers or their peers and supervisors to be passive inhabitants of a workplace. Workers do change

jobs to yield better fits between their abilities and job demands (e.g., Swaen et al., 2002). There are probably many cases of accommodative activities in place throughout the workforce.

The existing scientific literature has significant limitations intrinsic to population surveys, field studies, and laboratory experiments. We often rely on metanalytic studies, for example, to help ensure better inference, but these have weaknesses, for instance, the choice of rules for inclusion and exclusion of studies in the analysis and the fact that studies with nonsignificant findings tend not to be published. However, metanalyses do provide an efficient way to estimate effect sizes across studies (e.g., Schmidt, 1996).

An important theme is the need to consider increased age as contributing to counterbalancing trends. Aging processes tend to lower overall general functional capacity. Increased age is also associated with increased experience that tends to raise experience-related functional capacities. These two aspects of age may trade off, particularly when experience leads to skill and expertise, a point stressed in some of the earliest literature in this field (e.g., Welford, 1958).

Training and retraining seem particularly relevant for older workers, who are likely to be the most distant from initial professional training and from initial job training (Sparrow and Davies, 1988). Whether older workers are particularly in need of training and retraining can be addressed from the perspective of two outcome criteria: productivity and safety. To the extent that age discrimination exists in work settings, it may be driven by perceptions that older workers are less productive.

Cross-sectional metanalyses show no relation between age and job productivity (McEvoy and Cascio, 1989; Waldman and Avolio, 1986). This outcome is surprising in view of the ubiquitous laboratory-based findings of age-related declines in basic perceptual and cognitive abilities and in problem-solving performance on novel tasks (see Chapter 5). Productivity is typically measured using work output measures or peer and supervisor ratings of performance. One explanation for the apparent lack of an association between age and productivity is that, as mentioned previously, increased age is associated with the acquisition of job-specific knowledge and skills that compensate for age-related declines in general abilities (Salthouse and Maurer, 1996).

Another possibility is that the cross-sectional comparisons involve a mix of younger workers with varying skills and older workers with estab-

lished skills, given that younger workers move out of jobs much more frequently than older ones (Swaen et al., 2002). It is also possible that typical jobs do not demand continuous maximal performance to the same extent as laboratory-devised tasks. Older workers may find ways to accommodate to changes in capabilities that enable them to continue performing at satisfactory levels. An example would be the adoption of reading glasses to enable them to compensate for normative development of presbyopia (inability to focus on near objects). However, future findings about the relation of age to performance may change to the extent that the demand for new, nonpracticed abilities increases in the workplace. If current trends toward later retirement strengthen, and if there is increased job mobility, we may expect to see some narrowing in the gap between lab and life findings.

An important limitation on the conclusions about age and productivity is the insensitive measures of productivity. For example, in psychological literature productivity is often defined in terms of simple output. In economic analyses, a firm that produces the same number of goods and services using fewer inputs than a competitor is considered more productive. As a group, older workers are usually paid more than younger workers. Examined cross-sectionally, income tends to peak in the late 40s and early 50s compared with earlier and later ages (U.S. Bureau of the Census, 2002). Even if they show equivalent product output, older workers would be more costly to employ (or less cost-effective) by virtue of their higher salaries (and possibly by the higher cost of their benefits). But, cross-sectional analyses neglect the issues pertaining to lifetime costs and benefits to a firm and to a worker for an employment contract. Current higher wages paid to older workers may be explicable by delayed payment contract models (e.g., Hutchens, 1986). Such models argue that single-period accounting of the relative costs of labor do not fully capture the lifetime nature of labor contracts and the value of workers to their firms over their complete tenure.

So there is mixed news on the productivity question. Older workers appear to be as capable (or incapable) as younger ones in performing their jobs. However, older workers may appear to be less efficient than younger ones. Hence, there would appear to be a strong incentive to provide training and retraining to increase job productivity and efficiency. Or there may be an incentive to replace costly but equally productive (for output) older workers with younger, less expensive ones, should replacement prove less expensive than investing in training.

An important prerequisite for productivity is being available for work. Absenteeism, differentiated into voluntary (e.g., calling in sick when not sick) and involuntary (true illness or disorder or injury) types, does show age-related differences. Martocchio’s (1989) metanalysis used frequency of absence to index voluntary absenteeism and time loss to index involuntary

absenteeism. It indicated that older workers tend to have fewer absences of each type than younger workers, though the relationship is slight (r values between age and absenteeism were in the –0.1 to –0.2 range). The study also showed that gender moderated the relationship for voluntary absence: Women showed a near zero relation to age compared to the negative one for men, possibly due to nonlinear relationships for women.

Is there reason to believe that older workers would benefit as much from training as younger workers? A metanalysis showed that older adults benefit less than younger ones from training (Kubeck et al., 1996), with the correlation between age and training outcome being r = –0.26, and with training taking longer for older adults (r = –0.42). However, an important constraint on this conclusion is that a substantial subset of the training studies reviewed involved novel tasks. Some recent research that involved retraining (learning a second word processing software package) has shown a divergence in outcomes for novices versus experienced adults (Charness et al., 2001). There were strong age-related deficits in performance during and following training with novices and minimal (speed-related) or no declines in the performance of experienced middle-aged and older adults. However, in agreement with the metanalysis, this study also showed older adults, both experienced and inexperienced, taking longer to complete self-paced training sessions.

Are older workers less safe than younger ones, hence in need of better workplace design and training? It is important to distinguish injury probability and injury outcome as well as to differentiate injuries that are the result of a single episode or that are incurred by long-term exposure. Generally, the industrial accident literature indicates that acute injuries are most likely to be incurred by inexperienced workers in their first year at a new job and that older workers are much less likely to incur accidents (see Chapter 6; Sterns, Barrett, and Alexander, 1985).

Workers age 45 and older are much more likely to die as a result of a fall (e.g., Agnew and Suruda, 1993). Similarly, older workers are more likely to incur more serious nonfatal injuries (e.g., Layne and Landen, 1997).

The vehicle crash literature offers a model task environment for examining accident probability and injury as a function of age, particularly for North American workers who usually drive to work settings. About 42 percent of fatal occupational injuries are associated with transportation (NIOSH, 2002a). About 90 percent of all trips taken in the United States are taken in personal vehicles, with some 70 percent of those involving self-driving (e.g., Stutts, in press). Older drivers are more at risk for crashes per

mile driven than middle-aged drivers, though this is not seen until after age 65, and the risk becomes pronounced for those in their 80s. However, consistent with an age-positive experience effect, even 85-year-old drivers are less likely to suffer a crash (per mile driven) than 16- to 19-year-old drivers. Consistent with an age-negative effect on physical and physiological functioning, older drivers are more likely to be killed in a crash due to their greater fragility (Li, Braver, and Chen, 2003). When crash intensity is controlled statistically, there is still a greater likelihood of death for older adults and also for women (Evans, 1991). The latter finding points to the importance of body size and composition in absorbing impact forces.

Size of workplace is likely to be an important factor in opportunities for training and design intervention. Smaller workplaces are less likely to be regulated and less likely to have the resources to do effective training and design. Perhaps as a result, workplace fatalities tend to be concentrated in workplaces employing fewer than 10 workers (NIOSH, 2002b). Finding ways to disseminate relevant information to workplaces employing fewer than 10 employees is an important challenge.

The literature on productivity and safety, then, indicates that older workers are at somewhat greater risk for negative outcomes than younger ones. The literature on fatal falls (Agnew and Suruda, 1993) suggests that this increased risk first appears in incidence figures in the 45- to 64-year-old age range. Therefore, we can justify advocating training and retraining interventions, as well as design interventions, to improve productivity and safety of older workers. In the next sections we examine in more detail literature that points to possible guidelines for training and design.

There is an extensive literature on training (e.g., Salas and Cannon-Bowers, 2001). Training studies often focus on the individual, ignoring organizational and motivational factors that may be important mediators or moderators of training outcome. A review by Cohen and Colligan (1998) of literature on training for occupational safety and health makes a similar point about the effectiveness of interventions in workplaces: “Management’s role/support of safety training and its transfer to the jobsite, setting goals and providing feedback to motivate use of the knowledge gained, and offering incentives or rewards for reinforcing safe performance all seemed crucial to attaining a positive result” (p. 6).

With some exceptions (e.g., Noe and Wilk, 1993; Colquitt, LePine, and Noe, 2000), the same is true for the literature on training older adults and particularly for training older workers. The emphasis in the age and training area has been on training individuals. Although team training and

collaborative performance tasks are more likely to be the norm in work settings, there is a need for a better-developed literature on team training and aging. We focus here on individual differences in outcome for individually based training.

A useful way to partition the training literature is to differentiate broadly versus narrowly focused training. For instance, many white-collar jobs in the economy are advertised as requiring a college or university undergraduate degree, an example of broad training. Others require a specific domain-related training program (e.g., nursing).

Similarly, when considering interventions to help older adults, some interventions are very specific, aimed, for instance, at fall avoidance (e.g., Hauer et al., 2001). Others are broad-based, aimed at improving general abilities. An example of the latter is exercise intervention to improve general cognitive functioning. Recent research shows that certain exercise may be associated with modest improvements in cognitive abilities, particularly those defining so-called executive functions such as planning (Kramer et al., 2001). Earlier studies with older adults focused on training reasoning and spatial orientation abilities (Willis and Schaie, 1986). Those studies found that short-term training effects were equivalent to seven-year longitudinal declines in ability, but that there were very narrow transfer effects, with only trained abilities showing significant improvement. Narrow generalization of training has been replicated with a recent clinical trial (Ball et al., 2002).

The general literature on skill acquisition and on expert performance gives little reason to expect broad transfer effects for training. The classic educational research by Thorndike (1924) suggested that teaching Latin was not likely to show transfer effects to general reasoning and thinking skills. Such work led to his law of identical elements, that transfer can only be expected to occur when there are identical elements in the training and transfer tasks. More recent work refines the methods for assessing identical elements and bears out this conclusion (Singley and Anderson, 1989). Work on expertise shows that the skilled memory of the expert is limited to domain-related material, not to a generally superior memory (e.g., Chase and Simon, 1973). Education and training interventions often appear to ignore this fundamental finding.

Similar transfer findings are obtained for physical work situations. As an example, waste management workers exhibited no better aerobic capacity than the general population, though they did show greater shoulder muscle strength (Schibye et al., 2001). Older waste management workers showed the usual age-related declines in aerobic capacity (compared to younger workers and young and old controls), though they maintained shoulder strength. Lifting and moving waste containers is more likely to improve specific shoulder muscle strength than general aerobic capacity.

Although it runs counter to the notion of satisficing described earlier, advocates of human factors approaches have argued for a “good, better, best” approach to design (Fisher, 1993). However, for whom should training and design be optimized? A fundamental issue is whether a good design or training program for the young will apply equally well to the old (and vice versa). Here we are concerned with age by treatment interactions.

There is little disagreement that good training is better than poor training or no training. What is still unresolved is whether some forms of training work particularly well for older compared to younger adults (workers). Early research strongly advised using discovery learning as the best procedure to train older workers in new tasks (Belbin, 1969). In this format, learners are allowed to explore the task environment and try out different methods. Evidence supporting the value of discovery learning compared to traditional training methods for samples of older workers was positive, though weak. Others noted that differences in training method tended to overshadow differences in performance as a function of age (e.g., Czaja and Drury, 1981a,b). However, in the latter studies, training method did not interact with age. Interactions with age (greater gains for older adults) have been found for procedural (action) versus conceptual training for automated teller machine (ATM) use (Mead and Fisk, 1998) and for web search training (Mead et al., 1997).

The training literature frequently argues that training and production environments should be self-paced rather than forced pace for older workers (e.g., Belbin, 1965). This makes sense given the general nature of age-related changes in speed of processing (e.g., Salthouse, 1996), but a differential benefit for older adults has not been reported often. Word processing training studies tend to confirm the advantage of self-pacing for both older and younger adults (e.g., Charness, Schumann, and Boritz, 1992).

Even with self-paced training, older adults may not perform as well as younger ones. Charness et al. (2001) showed that older adults were less able to profit from three days of word processing training if they were novices. In a second experiment experienced older adults also learned more slowly in self-paced training. Although different interface types (keystroke, menus, menus plus icons) affected learning there was no interaction between age and training interface. Better interfaces resulted in parallel improvements in performance in young, middle-aged, and older adults. Across the two samples of participants, prior word processing experience and type

of interface were strong predictors of both speed and accuracy, as were, in some instances, psychometric measures of ability.

Experimental tests of pacing in work-related tasks are infrequent. Czaja and Sharit (1993) used realistic work tasks such as data entry, file modification, and inventory management with novice young, middle-aged, and older adults. These tasks were taught over three days, a unique departure from typical one-hour laboratory experiments. One training manipulation that yielded an interaction with task type and age was forced pacing versus self-pacing. Forced pacing tended to reduce older adult performance variability and age-related differences in performance, particularly for the data entry task. However, a checklist measuring perceived fatigue showed that forced pacing resulted in higher levels of fatigue for older adults compared to self-pacing. Generalizing from this study, older individuals might achieve greater output rates if forced to quicken their pace, but the cost might be greater fatigue and possibly greater error rates over an extended workday. Errors related to safety would bear a particularly high cost.

When people practice a skill over an extended period of time, their improvement function typically follows the power law of practice (e.g., Newell and Rosenbloom, 1981). This means that each successive interval unit of improvement requires a log unit increase in practice. The exact shape of the function (e.g., logarithmic, exponential, hyperbolic) depends on whether averaged group data or individual functions are modeled (Heathcote, Brown, and Mewhort, 2000).

Depending on the task and outcome measure, older adults exhibit equivalent or less improvement with practice than younger adults. Czaja and Sharit (1998) showed that older adults required more training than younger adults before showing improvement in a data entry task. In another study, Czaja (2001) did not find any interaction of age and training with output quantity or quality measures for a database query task (customer service environment).

Because of such mixed findings, the conservative conclusion to be drawn is that the rate of improvement with practice will differ marginally between younger and older adults, and when it does it will favor younger adults. A faster acquisition rate parameter for younger adult groups is a typical finding in analyses that involve fitting power functions to trial blocks, both for tasks with complex procedures (e.g., Touron et al., 2001) and for simple search tasks (e.g., Strayer and Kramer, 1994).

An implication from power law improvement functions is that older experienced workers can be expected to outperform younger inexperienced workers on routine (practiced) tasks, despite any age-related slowing in

processing speed. However, on new tasks that incorporate few elements from already-learned tasks, younger workers will be expected to outperform older ones given their slightly faster rate of learning. Indeed, the rule of thumb for predicting how long an older adult will take to complete a laboratory task, compared to a younger one, is to multiply the young adult time by about 1.5 (e.g., Hale and Myerson, 1995).

Many reviews (e.g., Belbin and Belbin, 1972; Warr, 2001) have noted that older workers are less likely to be offered training than younger ones and also less likely to volunteer or show up for training when it is offered. The reluctance to offer older workers training may stem from a diverse set of causes. In the case of government-sponsored programs to assist unemployed workers, the U.S. General Accounting Office (GAO) recently reported that there was reluctance to enroll older workers when programs were assessed on factors such as reemployment and salary levels for fear that older workers who are unemployed longer and who take greater pay cuts when reemployed would hurt program evaluation (GAO, 2004). Also, managers’ beliefs about aging and its effect on learning new skills may make them reluctant to offer training to older workers (Barth, 2000). Career orientation changes with age and history of success may affect receptivity toward training (Bray and Howard, 1983). The work environment and its reward structure may not motivate older adults to seek training (Noe and Wilk, 1993). Older workers may perceive training opportunities as irrelevant to their goals at work (e.g., for advancement in an organization). Some of the reluctance by older workers may be related to low self-efficacy both pre- and post-training (Colquitt et al., 2000). That is, older workers may see themselves as less capable of being trained successfully or of having benefited less from training. These beliefs may be driven by negative stereotypes about aging that show younger adults evaluated more positively than older adults, particularly from field-related work settings (Kite and Johnson, 1988).

As some reviews have stressed (Barth, McNaught, and Rizzi, 1996), older workers may be advantaged compared to younger workers, given their lower likelihood of voluntary absences, their lower likelihood of moving from a workplace, and their greater experience levels. Training costs are more likely to be recovered for older workers who are less likely to change jobs. A study by McNaught and Barth (1992) involved a case study of room reservation takers for a U.S. hotel chain. It showed that older workers were

somewhat longer on a call (perhaps exhibiting slowing with age) but more likely to book a caller. Taking into account employment costs and revenues, older workers were as economically productive as younger ones. Further, because of lower turnover rates, the annualized cost of recruiting and training showed a nearly three-to-one advantage for an older worker ($618 compared to $1,752). A British chain of home and houseware stores found that a store staffed entirely by workers age 50 and over was 9 percent more profitable than the storewide average (Barth et al., 1996).

Another important outcome measure for work is health status. Work-places can influence both physical health and mental health (see Warr, 1998, for a model of the effects of age and work on mental health). Physical health can be impaired when workers incur accidents or illnesses or disorders that are a function of short- or long-term exposure to hazardous materials or working conditions. For example, night-shift work seems to increase morbidity as a function of increasing age in those who have had to leave that type of work, compared to those who stay or those who have never performed such work (Volkoff, Touranchet, and Derriennic, 1998). This selective attrition result resembles a Darwinian survival-of-the-fittest model, usually termed the healthy worker effect.

Cross-sectional research suggests that musculoskeletal complaints increase with age of worker, and particularly for women in occupations that require heavy physically demanding work (e.g., de Zwart et al., 1997). However, a review of gender differences in musculoskeletal disorders (MSDs) suggests that existing studies have a variety of weaknesses that make it difficult to determine the causal factors (Punnett and Herbert, 2000). For example, observed differences may depend on a large set of factors: the form of assessment (self-report, injury records, strength assessment), workplace risk exposure, work-task strategy, psychosocial environment, home-risk exposure, willingness to seek help, and gender-specific factors such as muscle strength, connective tissue differences, and hormonal differences.

On the positive side, work can contribute to physical fitness to the extent that injuries can be avoided. Professional athletes who have not been forced to retire because of injury undoubtedly exhibit much better physical fitness than the general population. Similarly, work can have both negative and positive effects on mental health. Lack of work (unemployment) is a serious risk factor for morbidity in older workers and for poor mental health (Gallo et al., 2000), with different patterns of outcome for men who tend to show higher levels of substance abuse and women who show higher levels of major depression and anxiety disorders (Avison, 2001). One

notable outcome, depression, may in turn prolong unemployment (Vinokur and Schul, 2002). Stressful work can contribute to both poor mental and poor physical health.

Also, longitudinal studies show that intellectually challenging work shows reciprocal relations to intellectual functioning with stronger relationships evident for older workers (Schooler, Mulatu, and Oates, 1999:483). One study with a representative sample showed minimal relationships between work, health, and well-being (Herzog, House, and Morgan, 1991) and an unexpected relation for those ages 65 and over, where greater physical work stress led to higher ratings of health and well-being. (This may fit with the healthy worker effect explanation above.) In general, having choice in the type and amount of work leads to greater well-being and self-reported health.

As mentioned earlier, much of the training in work settings has concerned productivity and safety. Sometimes these two outcome measures show opposite patterns when organizations introduce changes, though safety training is usually effective (e.g., Kaminski, 2001). However, work settings also offer opportunities for health protection through interventions such as treating drug addiction (Silverman et al., 2001) and providing tobacco cessation programs (Sorenson, 2001).

The United States is unique among industrialized nations in not providing government-sponsored universal health care. Employers provide much of the health insurance coverage for the working-age population in the United States (see Chapter 7). Therefore, private employers have an incentive to provide such coverage and to support workplace health interventions if they wish to maintain a healthy, productive workforce.

Several approaches have been used to characterize work ability in order to make recommendations about fitness for a particular job. Fleishman and Quaintance (1984) review a large set of potential taxonomies for describing human performance including job performance. The more recently developed Finnish Work Ability Index (Ilmarinen, 1994) solicits worker self-report ratings on a seven-item scale. It has been used to predict work disability and is sensitive to factors such as age, gender, and type of work. Longitudinal investigations show that factors such as high physical demand (requirement for muscular strength), stressful work environments (e.g., temperature extremes), and role pressures at work can degrade work ability in aging workers.

Ilmarinen and Louhevaara (1999) summarize a number of innovative Finnish intervention studies, some lasting a year, that were aimed specifically at older workers. Some interventions (ergonomic) involved training

better work postures (e.g., on production lines, for cleaning staff, for municipal workers). These interventions significantly lowered the frequency of poor postures, reduced cardiac strain, and lowered perceived exertion. Other interventions involved physical fitness training that resulted in significant gains in musculoskeletal fitness and cardiovascular fitness, both of which were associated with improvements in the work ability index. Some studies used both types of intervention. The pattern of results from this project makes it clear that physical fitness training is a potentially important intervention to improve musculoskeletal and cardiovascular functioning and to improve fitness to work. Similarly, this project provides good evidence that ergonomic training interventions can have significant effects on work postures.

Whether the positive results of such interventions can be maintained over longer intervals and subsequently can affect the incidence of MSDs and particularly work-related illnesses or disorders and injuries, requires longitudinal investigations. One relatively rare clinical trial that was not focused specifically on older workers showed no effect of an educational intervention (three hours of initial training and booster sessions over subsequent years) on reducing low back injuries in postal workers within a five-and-a-half year interval (Daltroy et al., 1997).

The difference in results between the Finnish studies and the clinical trial may have much to do with the type of outcomes measured: postures or fitness that are somewhat easier to change and injuries that are relatively rare and difficult to prevent. Training can act similarly to expertise in mitigating normative age-related declines in work-related capabilities, perhaps prolonging the useful work life of older adults engaged in physically demanding jobs. General training programs may not be effective for reducing injuries.

Despite the general utility of these training principles and experiences, there has been a relative paucity of studies on health and safety training for older workers. However, over the past 20 years, a large number of innovative programs have been developed for providing occupational safety and health education to workers in general. These programs have been based on participatory methods of training, such as small-group interactive methods, worker and union input regarding needs assessment, materials and evaluation measures, worker empowerment goals, and use of peer trainers (Deutsch, 1996; Fernandez, Daltuva, and Robins, 2000; Kurtz, Robins, and Schork, 1997; McQuiston et al., 1994; Merrill, 1995; Wallerstein and Weinger, 1992).

Recent ergonomic intervention studies have indicated effectiveness in reducing discomfort in video display unit (VDU) workers via training and education (Ketola et al., 2002), injuries and costs of compensation claims in cleaning staff (Carrivick et al., 2002), and MSDs for staff who lift and

transfer hospital patients following installation of ceiling lifts (Ronald et al., 2002). However, these studies do not look specifically at the effects of interventions on older workers. If one views aging as engendering increased fragility, and given that cumulative exposure to stressors is likely to be greater with increased age, efficacy of interventions might be expected to be greater for older workers. Countering this prediction are the anticipated effects of selective attrition for older workers (the healthy worker effect).

In recent evaluation studies, interviews with participants 3 to 21 months posttraining provided examples of the positive impact of education. Results included correction of workplace safety and health problems; successes at changing programs, procedures or equipment at work; improved handling of emergency response incidents; improved management communication; training of coworkers after returning to work; and positive evaluations of worker-trainers (Cole and Brown, 1996; Fernandez et al., 2000; McQuiston, 2000).

The innovative programs described above appear to provide adult learners, including older workers, with relevant tools and problem-solving skills, the confidence needed to use those tools, and the motivation to remain active participants in improving occupational safety and health conditions at work. However, which of these methods works best for older workers and yields measurable gains in health and safety outcomes such as injury frequency and injury severity remains to be explored. It is also uncertain what the appropriate intervals are for retraining and maintenance of health and safety gains.

Work environments may have important influences on health promotion and disease prevention among older workers. First and foremost, employer policies affect direct work experiences and environments. They also have immediate and long-term effects on workers, their families, and their communities through both direct and indirect mechanisms. Direct mechanisms include interventions on specific worker behaviors, conduct, and exposures; delivering general health promotional programs to individual workers; providing and structuring health and disability insurance programs; and, in appropriate circumstances, intercepting and controlling community environmental hazards such as air and water pollution.

There are indirect mechanisms as well. Among older workers, as others, sufficient work income in itself may provide resources for improving

one’s general environmental health and safety; monetary resources can be used to purchase needed health services, both preventive and therapeutic. Worksites may also provide social contacts and networks that create additional opportunities for workers to receive information relevant to maintaining healthy lifestyles and avoiding health threats.

Proactive worksite health promotion and disease prevention activities may be thought of in two general categories: those aimed at promoting better safety and health directly related to job functions and those related to exposures and improvement. These two activities are of course interrelated and refer to policies, programs, and interventions that may not all pertain to or occur at the physical worksite. Health education may be provided at the worksite, or they may be take-home informational programs studied at worker convenience. Health promotion may be enhanced through the employee benefit structure. For example, one benefit to workers may be respite care or flexible leave policies that allow workers to attend to ailing parents and, in the process, may help decrease worker stress and maintain worker health status (Bornstein and Shultz, 2002). Similarly, tobacco-control efforts may be delivered on the jobsite, potentially improving worker function and long-term health outcomes that continue throughout life, or worker health insurance could provide for smoking cessation counseling, medications, and other treatment as part of the basic benefit structure.

Health promotion programs may also encompass direct worker interventions, such as individual or group counseling, or provision of recreational resources. Or they may alter the social or policy environment, for example, creating smoke-free work areas or payment incentives for cessation.

Given the great range of direct health promoting programs available, it is difficult to determine how many workers receive job-related health promotion programs. In one recent survey, 90 percent of employers with at least 50 employees reported that they offered at least one worksite health promotion program in the last year (NIOSH, 2002b). However, data for specific age groups and job types, as well as for those self-employed or working in small businesses, are difficult to find, and more surveys of access to health promotion programs among employed populations would be of great value. Based on a review of the literature, it appears that most health promotion programs not related to specific job functions and safety are devoted to the modification of general, health-promoting behaviors, such as cigarette smoking cessation, diet and nutritional improvement, optimal physical exercise, leisure, recreational activity, and body weight maintenance. Companion activities focus on controlling risk factors for cardiovascular disease and various cancers. In some instances, workplace programs have used health risk appraisal instruments that address prevention of other conditions and other risk behaviors such as seatbelt use and caries prevention.

A large number of general workplace health promotion programs have been attempted. In addition to addressing the health behaviors noted above, approaches have been attempted less commonly for problem alcohol drinking, illicit drug use, and other mental health issues. Programs related to these latter mental health issues are important though difficult to conduct, since they directly relate to high rates of worker morbidity, absenteeism, and lower productivity (Williams and Strasser, 1999). Workplace health promotion programs have had varying participation and success rates. What constitutes success with respect to participation often is not specified, especially since these programs are almost all voluntary by nature, raising the difficult question of whether the workers who would benefit most are actually those most likely to participate. Similarly, what constitutes successful program outcomes should be viewed in light of many factors, including the domains under intervention; the theoretical foundations behind such programs; programmatic quality and resource expenditures; worker accommodations for program participation such as scheduling; the demographic and health characteristics of participants; and comparable program outcome experiences in nonworksite venues.

The evaluative scientific literature for many of these programs is particularly problematic with regard to older workers. Most health promotional programs are directed at all adult age groups, and many do not describe the age distribution of the study populations; even when age data are available, many do not contain large numbers of older workers. Other methodological problems are common to many of these studies: small sample sizes, self-selection into programs, low participation rates in the evaluation efforts, lack of credible experimental designs, inclusion of many persons who are at low risk for the outcomes of interest, inability to determine long-term outcomes, and lack of breadth of job types included in the interventions—especially those in small businesses. Particularly relevant to this discussion, many health promotion programs do not necessarily adapt to the special needs of various age, literacy, or cultural groups. In many occupations and job settings, circumstances sometimes do not allow ran-domization or other rigorous study designs, for several reasons:

1. Many sites have frequent employee turnover, and it is hard to follow those going to other geographic areas or diverse work situations.

2. Some programs are not made available during normal employee work schedules, and attending them may be difficult for many workers. The same problem may exist for workers who are geographically and tem-

porally distributed from the program site. There may also be differences in access to programs, depending on job category and socioeconomic status.

3. The extent of the problem being addressed may depend on the presence and intensity of preemployment screening and health problem monitoring by the employer. For those (uncommon) worksites where medical care is provided, a focus on health promotion may be in part redundant or at least more complex. The issue of employment prescreening is a sensitive matter to many workers, who may wish to withhold personal health information and histories both from employers and their health insurers. This dilemma is exemplified by mandatory and unannounced screening for illicit drug use (Bush and Autry, 2002), and the potential prospect of genetic screening to determine who may be at greater risk of adverse effects of workplace exposures.

4. Communication between the worksite program and the workers’ usual sources of medical care may be inadequate. Confusion over the overlapping or redundant approaches to the same problem may result in an inefficient or harmful program.

5. In cross-sectional or short-term evaluative studies, the workers who might benefit most from health promotion programs may be underrepresented because they have lost their employment status due to illness or disorder and disability.

6. Some workplace-sponsored health promotion programs are available to retired workers. This may be useful for assessing program impact on older persons, but it does not necessarily address issues related to the interventions in the occupational environment.

In addition to these issues, there is a set of challenges for all behavioral interventions that pertain to the workplace: inadequate resources, external interventions that contaminate the programmatic experiment (e.g., general community tobacco control activities), and varying levels of personal susceptibility to behavioral change. In addition, the intervention applied may not be effective. Thus, there are many potential issues that do not preclude full program evaluation, but many conceptual and study design issues that require consideration before credible studies can be conducted.

In general, scientific evidence suggests that older workers are potentially able to benefit from the same health promotion themes as younger workers, such as cigarette smoking cessation, obesity management, nutrition enhancement, blood cholesterol lowering, problem drinking interventions, and hypertension management. In fact, it is possible that for many

successful interventions, there may be greater decrements in preventable disease rates per unit of resource expenditure toward older workers, in part because chronic disease rates are higher at older ages. However, there are potentially many cautions when carrying out health promotion interventions among older adults that are of much less concern for younger persons. Some examples are provided in Box 8-1. This is not to imply that these programs should not be invoked, but rather that their objectives and content should be tailored to meet older workers’ special needs.

As noted elsewhere in this chapter, some hazardous aspects of the work environment may lead to negative changes in health with prolonged exposure, and interventions to mitigate these exposures should be considered important examples of health promotion. For example, Drudi (1997) demonstrated a striking increase in repeated trauma disorders (now called MSDs) occurring in private industry in the 1990s. Changing the nature of work by automating repetitive tasks (e.g., providing machines to perform stressful repetitive work movements) may help to stem the increase. Prevention of work-related injury and disease is an important goal that may be accomplished through job design and redesign as well as through training, and it should be thought of as a component of general health promotion. Approaches may vary depending on whether injury or disease is the target for prevention, as well as whether the disease is chronic or acute.

The following potential health promotion programs for older workers are areas that have not generally been explored but may merit consideration in future health promotion program design. These derive largely from geriatric practice, and they serve to highlight considerations that—though not usually applied for the broad range of working adults—merit consideration:

1. Education in providing caregiving skills. Many older workers have aging parents or other family members or friends who require intensive supportive care in the household setting. This care ranges from providing transportation and general supervision and guidance to assistance with the most basic chores and activities of daily living. These caregiving activities can be extremely distracting, and the related stresses have been associated with worsening quality of life for the caregiver (Bell, Araki, and Neumann, 2001). Employee benefit structures, such as self-guided schedule flexibility and insurance for chore services, may or may not be available, but worksite programs that provide instruction and skills in caregiving may be very helpful in maintaining older workers on the job and may have important health-promoting value.

2. Nutritional and dietary interventions for older workers. Tailored nutrition counseling for older workers may be helpful for health promotion in addition to the guidance provided by general nutrition education pro-

grams. Differences in programmatic emphases are important. For example, for some older persons, caloric sufficiency is as important as excess caloric intake. In some instances, higher caloric intake needs to be encouraged for optimal weight and muscle strength (Evans, 1998). In addition, herbal dietary supplements and products are very popular among older persons. Calcium intake recommendations are increased in postmenopausal women (Nordin et al., 1998) to maximize bone density and optimize fracture prevention. Since medication use is more common at higher ages (see below), even modest levels of alcohol intake may cause unwanted drug interactions that require attention (Lieber, 2000). Education as to possible interactions with existing medications and possible adverse effects may be beneficial. Lactose intolerance increases with age, and special knowledge of the identi-

fication and prevention of this condition can help maintain better health status (Swagerty, Walling, and Klein, 2002).

3. Polypharmacy and therapy management programs. For both preventive and therapeutic purposes, older persons take more medications, prescription and nonprescription, than younger persons (Williams, 2002). This is because of the higher prevalence of chronic illnesses or disorders and of risk factors for them. Medication aimed at mitigating chronic illness or disorder risk factors target such issues as hypertension, hypercholester-olemia, low bone density, osteopenia and osteoporosis, and degenerative arthritis. Managing these medications and various treatments may make regular employment schedules and execution challenging. While this generally is a shared responsibility between the patient, health professionals, and pharmacists, educational programs to facilitate work and medical therapy may improve job performance and satisfaction. In addition, there may be value in considering workplace disease management programs for categorical conditions such as Type II diabetes (Berg and Wadhwa, 2002) or asthma, as long as coordination with standard medical care sources is available. These might be more clinically intensive than usual health promotion programs, but they may offer standardized effective programs that keep workers healthier.

   Along these lines, workplace pharmacological prevention programs may be of value for older workers and others. With increasing age, primary and secondary chronic disease prevention may include routinely consumed medications for preventive purposes. For example, regular low-dose aspirin consumption can prevent heart attack and stroke (Anonymous, 2002), and control of blood lipids is increasingly common (Lipsy, 2003). It is even possible that over time some anti-inflammatory drugs or other drug categories may be taken to deter the occurrence of Alzheimer’s disease (Breitner and Zandi, 2001). While such programs may require active medical supervision, they may yield important preventive outcomes. In the future, there may be useful and safe interventions that help deter degenerative states such as osteoarthritis and cognitive impairment.

4. Tailored exercise interventions. The value of habitual exercise programs in disease prevention, weight control, and general well-being is unquestioned, and many such programs have been provided or encouraged in the work environment. However, in addition to the general health benefits imparted, such as improved muscle strength and aerobic capacity, exercise programs may be tailored for other goals among older workers, such as improving balance and coordination to prevent falls (Cumming, 2002), and to provide relief from degenerative arthritis, such as of knees (Fransen, McConnell and Bell, 2002). These programs may require special techniques and equipment, but these are generally not more costly than standard approaches.

5. Disease screening for older workers. Many conditions increase with age. In some of them, early and asymptomatic detection will lead to better disease control, less long-term disablement, and decreased early mortality. There are many screening interventions that are indicated only for older adults, including bone density screening for osteoporosis detection and fracture prevention; mammographic screening for breast cancer; and fecal occult blood determination, sigmoidoscopy, and colonoscopy for colon cancer detection. All of these have proven value (Williams and Wilkins, 1996), although optimal timing and target groups remain an area for research. Provision of worksite screening clinics or programs, leading to referral when abnormalities are found, can, on balance, preserve health and function and lead to longer, more functional, and productive lives. It may also be possible to provide medical advice and care for preventive interventions that could decrease morbidity, such as anti-inflammatory or anti-coagulant medications and immunizations.

Cardiovascular disease (CVD) is a common source of disability for older workers. In order to manage and prevent CVD and reduce such disability, the team approach common in the field of occupational medicine is recommended: clinicians, health educators, ergonomists, epidemiologists, and other health professionals work to identify high-risk workplaces and occupations, facilitate the provision of clinical care, and design and implement workplace interventions (as in Herbert et al., 1997).

For the prevention and management of chronic diseases, such as CVD and hypertension related to a stressful work organization, NIOSH recommends that an additional health professional be an integral part of the team—the occupational health psychologist (OHP) (Landsbergis et al., 2002). The OHP requires multidisciplinary training combining methods and content from the fields of occupational health, epidemiology, psychology, management, industrial relations, and other relevant disciplines (National Institute for Occupational Safety and Health, 2002b). Therefore, NIOSH has funded graduate training programs in this field (Sauter and Hurrell, 1999).

A key step in the process of disease prevention, identification, and management is worksite surveillance and monitoring at national and regional levels to identify high-risk occupations, the extent of workplace stressors, health outcomes resulting from such exposures, and baselines against which to evaluate amelioration efforts (Belkic, Schnall, and Ugljesic, 2000; “The Tokyo Declaration,” 1998). Such interventions may prevent or slow the development of hypertension and cardiovascular disease in middle-

aged workers, and thus help to prevent these diseases among older workers. Standard questionnaires can help assess job characteristics and job stressors (Belkic et al., 1995; Karasek et al., 1985; Siegrist and Peter, 1996). Measurements of blood pressure (BP) at work can help to identify occult (hidden) work-related hypertension, sometimes missed by readings taken in a clinic setting (Schnall and Belkic, 2000). An additional key modality of risk assessment is an occupational history of individual workers.

Both individual health promotion and workplace protection/prevention programs are needed to combat the epidemic of CVD and the toll of disability in older workers. However, due to the limitations of health promotion programs, primary prevention strategies—such as job redesign to promote education, screening for risk factors, and prevention of worker stress—are fundamental.

The effectiveness of a limited number of interventions to improve work organization and job design, reduce job stressors, and create a more healthy work organization have been discussed above. In addition, older workers with CVD or at risk for CVD need to be counseled to reduce their levels of unhealthy behaviors, such as smoking. However, such counseling in isolation may have poor efficacy, particularly among occupational groups with a heavy burden of exposure to occupational stressors: “despite devotion of substantial time and the use of state-of-the-art methods…our efforts applied systematically among professional drivers were, at best, only minimally effective, unless there was a concomitant amelioration in stressful working conditions” (Fisher and Belkic, 2000:247).

Judgments of the cardiovascular work fitness of older workers who have suffered cardiac events are complicated by the issue that jobs in which public safety could be compromised with the occurrence of an acute cardiac event (deGaudemaris, 2000) are often those with high exposure to potentially cardio-deleterious factors (e.g., urban transit operators, air traffic controllers) (Fisher and Belkic, 2000). Fifty-four percent of all U.S. bus drivers are 45 years of age or older (Dohm, 2000). On the other hand, advances in cardiovascular therapy permit the restoration of cardiovascular function of many patients, and this can make returning to work possible (deGaudemaris, 2000). In a Swedish study, men who had suffered a first MI below age 45 were at high risk of five-year CHD mortality if returning to a high-strain job (Theorell et al., 1991). Occupational health professionals with appropriate training need to identify potentially modifiable cardiac stressors in the older workers’ job environment and, together with a clinician, need to formulate and implement a plan to provide a safer return to work.

Employee Assistance Programs (EAPs) are employer-provided service programs that aid troubled workers, usually through counseling, support groups, and service referral. While most of these programs have not yet emphasized employee needs specifically related to aging, they have strong potential as a support for older workers in relation to occupational health and safety concerns. EAPs may be of help both in a direct sense and also indirectly by maintaining a person’s ability to work, promoting physical and mental health, and relieving stress that can cause accidents and illness or disorder.

In the United States and Canada, EAPs emerged in the post-World War II era primarily as a mechanism for rehabilitating alcohol addicted workers (Trice and Schonbrunn, 1981). Employers saw them as a constructive alternative to disciplinary dismissal, particularly for highly skilled workers who would be expensive to replace. Early EAPs were usually staffed by non-professionals, often who were themselves former addicts and were philosophically linked to 12-step programs such as Alcoholics Anonymous (McKibbon, 1993). Around the same period, some unions began to offer Membership Assistance Programs (MAPs) to assist their members (Johnson, 1981). While similar to EAPs in their focus on addiction, these early MAPs tended to put more emphasis on peer counseling by coworkers and to address working conditions that might contribute to workers’ personal problems.

Over the next half century, both EAPs and MAPs gradually evolved into more comprehensive broad-brush service resources, able to offer help to workers with a wider array of personal needs (Walsh, 1982). In addition to drug and alcohol abuse treatment, many began to offer mental health counseling for individual workers and their families. Some also offered services such as credit counseling, bereavement counseling, stress management, marital counseling, help with work and family dilemmas and dependent care choices, and career planning. Particularly during the 1990s, many EAPs became integrated with more general worksite health promotion, wellness, and work and family programs offered by employers (Lubin, Shanklin, and Polk, 1996), although the linkage among the different programs was not always close or comfortable (Herlihy, 1996). Despite continuing tensions over these programs’ proper scope and emphasis, comprehensive EAPs staffed at least in part by trained mental health professionals have by now become the norm in large corporations (Sciegaj et al., 2001).

A number of evaluation studies have documented that EAP-referred workers do show reductions in absenteeism and illness (Macdonald et al., 2000; Gaton, 1986), and that these programs can be quite cost-effective for the employer (French et al., 1999; Bray et al., 1996; Every and Leong,

1994; Decker, Starrett, and Redhouse, 1986). Some EAP-based interventions have the potential to raise issues of discrimination and employer liability (Starkman, 2000; Capron and Creighton, 1998), but such problems do not appear to be widespread.

In 1989, a national Employee Benefits Survey conducted by the U.S. Bureau of Labor Statistics found that among full-time workers in privatesector establishments with 100 or more employees, 49 percent had access to EAPs, and 23 percent were served by wellness programs (Cooley, 1990). A Canadian study at about the same time underscored the rapid growth of EAPs. During the period from 1988 to 1993, among 647 companies in Ontario with over 50 employees, the percentage with EAPs doubled from 16.1 to 32 percent (Macdonald and Wells, 1994). However, the coverage was very uneven. The same Canadian study also confirmed the existence of wide variations among major work sectors, ranging from highs of 51 percent with EAPs in government and 46 percent in health and education, to lows of 13 percent in retail trade and 3 percent in construction. Small to medium-sized workplaces are much less likely to have such supports, an important fact because the majority of the U.S. workforce is employed in work sites with fewer than 50 employees. Donaldson and Klien (1997:17) reported that “one of the main findings of the 1992 National Survey of Workplace Health Activities [in the United States] was the identification of a pressing need to understand how to formulate effective strategies for providing comprehensive health promotion [and EAP] services to…traditionally underserved employee populations; particularly ethnically diverse operating-level employees working in small, medium-sized, and women and/or minority-owned businesses.”

In contrast to the early EAPs, in which at least some direct treatment was usually offered onsite in the workplace, the actual services of EAPs and MAPs are now often delivered by offsite service providers with the program’s in-house aspects being limited to assessment and referral; both models have advantages (Brummett, 2000; Csiernik, 1999; Straussner, 1988). Along with these organizational changes, the orientation of EAPs and MAPs has gradually been shifting from rehabilitation to prevention. Instead of a quasidisciplinary intervention aimed at workers whose performance has already suffered because of personal problems, these programs are increasingly seen as a way to keep workers productive and healthy by helping them avoid problems in the first place. Current EAPs tend to encourage self-referrals and promote their services as a positive employee benefit. In addition, there is growing awareness that the work climate itself can contribute to the personal troubles and unhealthy behavior of individual employees. For example, management policies and enabling behavior by peers and supervisors can create a prodrinking, prosmoking, work-aholic, or reckless work environment that endangers employee health and

safety (Bennet and Lehman, 1997). Some EAP specialists have urged that in such situations the EAP client should be defined as the organization itself (Googins and Davidson, 1993).

To understand the kinds of support that EAPs might offer for older workers, two different sets of concerns are involved:

1. What kinds of EAP support will assist aging employees to continue working safely and productively in their career jobs? Some aging workers find it difficult to balance their work with medical management of emerging health problems, the time demands of caregiving at home, increased difficulty with activities such as driving to work, a changing sense of what matters in life, and stressful work relationships related to age discrimination and stigma. Age-related physical changes such as hearing loss and arthritis may call for job redesign, and they may need retraining for new kinds of work assignment. If the choice to remain working is driven by income insecurity, workers may find themselves between a rock and a hard place as their work ability declines but their need for earning continues.

2. What kinds of EAP support will help aging workers prepare adequately for retirement? Most aging workers expect eventually to be leaving their career jobs, either exiting the workforce entirely or moving into a phased retirement period involving some combination of reduced work hours, periodic leaves of absence while still employed, and alternative paid or unpaid work. At least some may be making unrealistic choices due to lack of information and feeling great distress if their retirement plans are being undermined by economic downturns. If they plan to phase out of the workforce gradually through alternative work, they may be uninformed about their options and legal rights, and unprepared to find the bridge jobs they will need. Even those eager to leave the workforce may be experiencing uncertainty and anxiety about what their financial and social circumstances will be in retirement. Those facing the prospect of being unwillingly forced out through layoffs or pressure to retire at a certain age may be very angry with supervisors and coworkers.

Particularly as workers approach the traditional retirement age, these concerns can overlap. Separately or together, by creating unmanageable levels of stress, they put the older worker at increased risk for workplace accidents, depression, physical illness or disorder, health-damaging personal behavior, and even workplace violence.

There are already at least three ways in which EAPs have begun to offer supports of special relevance to aging workers: eldercare support, preretirement planning, and customized versions of traditional services such as drug and alcohol treatment and outplacement.

Also known as support to working caregivers, eldercare support is meant to help employees who are struggling to balance their jobs with their commitments to care for elderly parents, care for an older spouse who is ill, or deal with the responsibilities of caring for elderly parents on top of having dependent children still at home (Eubanks, 1991). These workers, typically women in their 40s and 50s, are in effect working a permanent double shift without relief, at a point in life when they are aging themselves and their career demands may be peaking after delays for childrearing (Winfield, 1987). The result can be extreme stress, fatigue, and isolation for the individual worker, potentially leading to physical illness or disorder, depression, accidents on and off the job, breakdown of work relationships, and alcohol or drug addiction.

Having workers experience this high level of strain is costly to employers as well. Employers’ measurable costs associated with eldercare giving have been estimated at $2,500 to $3,100 per year per caregiver from losses in employee productivity, management/administration, and health/mental health care (Marosy, 1998).

A substantial proportion of eldercare givers hold jobs. In 1987, of the approximately 2.2 million persons providing unpaid informal assistance for older adults in the United States, 31 percent were at the same time employed outside the home (Seccombe, 1992). By 1997 that had risen to 55 percent, according to a national survey conducted by the National Alliance for Caregiving and the American Association of Retired Persons (AARP) (Wagner, 1997). Looking at it from the other side, eldercare givers make up a substantial proportion of the workforce. In a survey of 3,658 employees of a major company in southern California, Scharlach and Boyd (1989) found that 23 percent of respondents reported that they were assisting an older person. Of these working caregivers, 80 percent reported emotional strain and 73.7 percent (as compared to 49.1 of other employees) reported interference between work and family responsibilities. About 20 percent said it was likely they would eventually have to quit their jobs to provide care. The sheer amount of time demanded by caregiving activities is quite considerable. Wagner (1990) notes that the employed caregivers in one study reported spending an average of 12.8 hours weekly on caregiving and had been doing so for an average of 6.5 years.

During the 1990s EAPs began to respond to this increasing need by exploring ways to help working caregivers get information, negotiate flex-time and leaves, manage financial pressures, and find appropriate counseling. Eldercare support is, in fact, one of the fastest growing new EAP services (Earhart, Middlemist, and Hopkins, 1993). A 1996 survey of 1,050 major U.S. employers found that eldercare programs were offered by nearly

one-third of these employers, an increase of 17 percent from 1991 (Hewitt Associates, 1997). As one example, in the mid-1990s, Control Data Corporation expanded its longstanding EAP to include comprehensive eldercare services to more than 1,100 employers with 1.2 million employees (Ensign, 1996). Unfortunately, however, eldercare is not yet offered in the majority of workplaces with EAPs. The need for it is sometimes not recognized by employers and EAP practitioners (Kola and Dunkle, 1988). Even when recognized, awareness of the need does not necessarily translate into an actual program. In a survey of 371 chief executive officers of U.S. corporations, 60 percent of respondents were aware of work-related problems experienced by employees who give care, but less than 20 percent were actively considering offering a specific caregiving program at the worksite.

The 1996 Hewitt survey and other similar ones (Lefkovich, 1992) found the most common approach in EAP-based eldercare support to be a resource information and referral service, sometimes accompanied by related policy changes within the company such as work scheduling that included flex-time, flex-place, compressed work weeks, temporary part-time status, and personal leaves. Other related policy changes include dependent-care spending accounts that set aside up to $5,000 in pretax dollars to pay for eldercare expenses, and management training policies that sensitize managers to employees’ caregiving responsibilities. In addition, some EAPs are partnering actively with community-based service resources such as home care agencies, which are already equipped to provide eldercare supports such as emergency backup for in-home adult companion care and child care on short notice; financial planning with a nurse geriatric care manager to develop an affordable care plan; senior day care; respite services; and emotional support for caregivers through support groups, hospice, and certified eldercare counselors (Marosy, 1998; Tober, 1987).

A wealth of information on resources for developing workplace eldercare assistance programs, including case examples and guidance materials such as the AARPs’ Caregivers in the Workplace kit, can be found in Dellman-Jenkins, Bennett, and Brahcae (1994). The Washington Business Group on Health has prepared a guidebook intended to help human resources professionals, benefits managers, and other corporate decision makers become more knowledgeable about eldercare issues in the workplace (Coberly, 1991). Additional resources and a good example of union involvement in eldercare support appear in symposium proceedings from a conference sponsored by the Federal Council on the Aging (1984), which includes a description of a New York City service delivery program of the United Auto Workers.

Preretirement planning (PRP) is recognized as a need in the field of employment assistance, but relatively few U.S. firms offer these services to date, and where they do exist, older workers reportedly tend not to use them (Perkins, 1994). However, they could be of enormous benefit to both employees and employers as workers move toward exiting the workforce. By making the transition from work to retirement smoother and less stressful, such planning may increase the chances that retiring workers’ final few years in the workforce will be safe and productive ones.

There is a clear need for planning, particularly financial advice to help the worker make realistic choices about retirement timing and budgeting. Numerous studies document that the current generation of aging workers are woefully uninformed about the options and decisions that they will confront on retirement and very unprepared for the realities that await them (Marshall and Mueller, 2002). The economic downturn that started in 2000 has greatly worsened the situation by eroding the value of many workers’ retirement savings and pensions. Workers now have diminished resources for their anticipated golden years, but they may not yet fully recognize the need to adapt their previous plans.

LaRock (1998) describes some of the varied approaches that are currently being offered by large employers, often channeled through EAPs. Boeing conducts focus groups to identify life-planning topics of greatest interest to its employees and then offers individualized classes. Weyerhauser holds one-day seminars on retirement planning with separate sessions for those over 30 and those under 30, as well as an enhanced version for employees over 60, consisting of a two-and-a-half day seminar led by outside financial planners and estate attorneys. Dow Agro holds retirement planning sessions during lunch at the company’s cafeteria and fitness center. The Washington State Department of Retirement Systems offers one-day seminars at various locations throughout the state, with 300 to 400 people attending each seminar. Good resources for employers interested in developing such planning sessions are readily available (Sherman, 1997).

Preretirement planning involves more than financial information. Exiting the workforce requires developing new ways to use time and find meaning in life, tasks for which many workers are unprepared—particularly men (Moen, 1996, 1998) and those who have left the labor force involuntarily or reluctantly (Sijuwade, 1996). Retirement can also provide enhanced opportunity for enjoying family and other relationships. Perkins (2000) presents a case example of a pre-retirement planning program in which an EAP facilitated a series of lunchtime workshops for older workers to address such issues. The workshops were led by former employees who were leading meaningful and zestful lives after leaving the company—some as

retirees, some having reentered the labor market. Subsequent evaluation of the program through focus groups showed that “one of the more prominent issues to emerge from the workshop was the need for older adults to move beyond the ‘work ethic’ values of the middle years. It was determined that new ‘yardsticks’ were needed for measuring what makes life worth living” (p. 69).

To date, employer-supported retirement planning support appears to be available mainly to employees of large companies. Encouraging the spread of similar services in smaller firms, perhaps through consortium arrangements, would be desirable since an even greater need for retirement planning exists for many workers employed in small firms and/or intermittent jobs, particularly women and minorities (Angel and Angel, 1998; Mitchell, Levine, and Phillips, 1999; O’Rand and Henretta, 1999; Dancy and Ralston, 2002).

While substance abuse interventions have long been the backbone of EAP services, these interventions may require modification to meet the needs of older employees with drug or alcohol problems. As Goldmeier (1994:624) points out, “Substance abuse among the elderly may be masked by physical problems and therefore escape detection; in addition, the elderly tend to underreport physical illnesses or disorders because they fear discrimination, and they may be more vulnerable to the effects of alcohol or illicit drugs because of age-related physical changes.” The picture may be further complicated by interactions of alcohol or illicit drugs with prescription drugs that the older worker may be taking. The potential for unrecognized alcoholism is of particular concern, because heavy alcohol consumption is related to occupational injury among older workers. In a nationally representative sample of 6,857 nonfarm workers aged 51–61, alcoholism was positively associated with occupational injury, even after controlling for age, sex, education, occupation, and strenuous job activity (Zwerling et al., 1996). In this study, the injury rate among the older workers who consumed five or more drinks a day was five times greater than for the category showing the lowest injury rate (who consumed one to two drinks per day). Potentially, EAPs can play a useful roles on several levels: (1) primary and secondary prevention of substance abuse among older workers through worksite wellness programs of medical screening, drug screening, education, review of work attendance and accident records, and maintenance of a positive work environment; and (2) tertiary prevention—after a substance abuse problem has been identified—through referral to community resources, counseling, mediation, advocacy, and case management (Goldmeier, 1994; Brummett, 1999).

Another traditional service offered by some EAPs is outplacement, in which employees who are leaving the company but remaining in the labor force are assisted with job-search counseling and skills assessment, help with resume writing, use of company telephones and copiers during the job search, and use of the company mailing address for a period following termination. This kind of service has generally been aimed at younger workers who are leaving the company because of layoffs. However, it could be adapted to serve older workers who expect to seek other work (bridge employment) after leaving their career jobs. Outplacement could also assist the older worker with customized searches for alternative jobs and/or volunteer positions, opportunities for retraining, and information on using legal protections against age discrimination. A 1989 telephone survey of 3,509 adults aged 50–64 found that longest-held positions typically ended long before normal retirement ages, creating a large pool of older individuals seeking bridge jobs in an employment climate rife with age discrimination (Ruhm, 1994).

The bridge employment choices facing older workers are quite complex, and outcomes differ considerably depending on how voluntary the career job exit is (Weckerle and Schultz, 1999). Ruhm (1994:73) notes: “Of particular concern is the limited ability of some groups of workers (nonwhites, females, the less educated, and those in poorly compensated occupations) to either retain longest jobs or to obtain acceptable bridge employment.” These older workers are less likely to be employed in large companies that have EAPs. Reaching them with outplacement support (and other EAP services) would require new strategies for encouraging small firms to make such services available to their workers—for example, tax incentives, technical assistance, and consortium arrangements (Donaldson and Klein, 1997).

Accommodations for workers with impairments and return-to-work programs are both important interventions that may play an important role in maintaining older workers productively in the workforce because these workers are more likely to bring impairments into the workplace and because they are likely to be out of work longer than their younger colleagues after an injury.

Over the last 25 years, there have been changes both in the prevalent conceptual model of disability and in the public policy approach to people

with disabilities. Two reports from the National Academies Institute of Medicine (IOM, 1991, 1997) played a major role in the refinement of the conceptual model. This emerging model distinguished between impairment, a loss of function at the organ system level; functional limitation, the inability to perform a specific task, such as lifting a 20-pound package; and disability, a limitation in performing socially expected roles. Impairment and functional limitation are characteristics of an individual; disability de-notes a mismatch between an individual’s functional capacity and the individual’s environment. In the workplace, this translates into a mismatch between an individual’s functional capacity and the essential requirements of his or her job. For example, a worker who could not lift a 20-pound package might be totally disabled as a construction laborer, but that same person might be able to carry out all of the job requirements of a secretary. This concept of disability as a mismatch between the worker’s functional capacity and the demands of their job leads easily to the concept of workplace accommodations—modifications of the job or workplace that allow the worker to carry out the essential functions of his or her job in spite of functional limitations.

Parallel to this change in the conceptual model of disability, there has emerged a new public policy approach as well (Miller, 2000). As discussed earlier in this report, the Americans with Disabilities Act of 1990 marked the ascendance of a new approach to persons with disabilities, a civil rights approach that aimed to help them overcome the barriers to full participation in American society. Specifically, Title 1 of the ADA was aimed at integrating workers with disabilities more fully into the workforce. It prohibited discrimination against qualified employees (or job applicants) with disabilities. A qualified person with a disability is an individual who, with or without reasonable accommodation, can perform the essential functions of the job. Reasonable accommodation may include, but is not limited to:

1. making existing facilities readily accessible to persons with disabilities;

2. restructuring jobs, modifying work schedules, and reassigning employees to vacant positions;

3. modifying equipment, examinations, training materials, or policies; and

4. providing qualified readers and interpreters.

As noted above, older workers are more likely to have a wide variety of impairments and may require accommodations to remain in or reenter the workplace. Although a dozen years have passed since ADA became law, we still know relatively little about key aspects of these workplace accommodations.

First, how much do workplace accommodations cost? The President’s Committee on Employment of People with Disabilities (1995) suggested that the cost of accommodating an employee with a disability is low, averaging about $200. Blanck (1996) found similar results when examining the costs of accommodations at the Sears Roebuck Company. However, these early estimates may well be significantly lower than the true costs because they were prepared by advocates of the ADA who construed the costs very narrowly. Instead, for example, of just accounting for the cost to buy special equipment, it would be useful to give an accounting of the full opportunity cost of accommodating an average person with a disability. These costs would need to include the time of other employees and managers involved in the accommodation, as well as the time of the disabled person (Chirikos, 2000). It is also likely that the accommodations sampled are not representative of the range of accommodations that may be implemented in the future. One could argue that the least expensive accommodations are likely to be implemented first. Subsequent accommodations might be more expensive (Chirikos, 2000).

Second, how frequent are workplace accommodations? Which accommodations are most frequently provided and who gets them? Daly and Bound (1996) used data from the Health and Retirement Study, a longitudinal panel study of older Americans, to examine the experience of older workers (51–61 years of age) when they had the onset of a medical condition that limited the work they could do. They found that about half of the workers stayed at their current jobs; just less than a quarter changed jobs; just over a quarter stopped working. Of those who remained with their employer, 29 percent of the men and 37 percent of the women received accommodations. Those who changed employers were less likely to receive accommodations: 14 percent of the men and 29 percent of the women. The most commonly provided accommodations included the alteration of job duties, assistance with the job, a change in schedule or a shorter work day, and more breaks. These results are consistent with previous work that suggested that up to one-third of workers experiencing a disability report some type of employer accommodation (Lando, Cutler, and Gamber, 1982; Schechter, 1981; Chirikos, 1991).

Third, how effective are workplace accommodations in allowing workers to remain safely and productively in the workforce? The goal of workplace accommodations is to allow workers with a range of impairments to enter or remain in the workplace. There have been very few studies exploring the effectiveness of these accommodations. In two studies, Burkhauser and colleagues (Burkhauser, Butler, and Kim, 1995; Burkhauser et al., 1999) has addressed this issue. Applying a proportional hazards analysis to the HIS component of the 1978 Survey of Disability and Work, Burkhauser

et al. (1995) found that the provision of an accommodation significantly slowed a worker’s withdrawal from the workforce after the onset of an impairment. In a subsequent study employing additional data from the Health and Retirement Study and using a proportional hazards model, Burkhauser et al. (1999) found that the time to the beginning of SSDI payments was significantly delayed by the provision of workplace accommodations. However, both of these studies relied on retrospective data on the provision of accommodations. Thus, they are both vulnerable to potential recall bias—those who withdrew from the workforce may be less likely to remember accommodations their employers made than those who remained in the workforce. The longitudinal data now available in the Health and Retirement Study would permit a prospective examination of this issue.

Data summarized above (Zwerling et al., 1998a,b) suggest that older workers with impairments have an increased risk of occupational injuries. Appropriate workplace accommodations might be expected to lower this risk, but we know of no empirical studies addressing that issue. Likewise, it might be expected that appropriate workplace accommodations would increase the productivity of workers with a variety of impairments, but we were only able to find anecdotal evidence on this issue.

As noted above, national databases show that older workers are less likely to be injured at work than their younger colleagues, but they are likely to require a longer period of recuperation before they are ready to resume their normal duties. This prolonged period of work disability has been extensively documented among older workers with low back pain (Bigos et al., 1986; Dasinger Krause et al., 2000; Infante-Rivard and Lortie, 1996; Krause et al., 2001a; McIntosh et al., 2000), but it has also been documented among cardiac patients (Karoff et al., 2000) and trauma patients (MacKenzie et al., 1998). Given the prolonged period of recuperation among older workers, the design, implementation, and evaluation of return-to-work programs among older workers merits special attention.

Leigh and colleagues (1997) estimated that occupational injuries cost Americans about $145 billion in 1992. The overwhelming majority of these costs derived from the disability costs of injured workers. These significant costs have been associated with many efforts to find the risk factors that predict prolonged disability after work-related injuries. In a recent review of the literature, Krause and colleagues (2001b) identified about 100 different determinants of return-to-work outcomes. Krause categorized the risk factors for disability in seven broad groups ranging from the most individual level factors to the most societal level factors: the individual level

worker characteristics associated with return-to-work outcomes (useful in predicting prolonged disabilities, but generally not amenable to change); the individual-level worker factors that describe the injury; medical and vocational rehabilitation programs; the physical and psychosocial job characteristics; the organizational level employer factors; the employer- or insurer-based disability prevention programs; and the societal level legislative and policy related factors. The factors that most consistently resulted in a shortening of the duration of disability included medical and vocational rehabilitation interventions, organizational level employer factors, and employer- and insurer-based disability prevention and disability management interventions. Each of these three areas provides many opportunities for implementing and evaluating interventions.

However, several challenges must be overcome before researchers can establish which interventions are most effective (Krause et al., 2001b). Researchers need to agree on the best outcome variables to use in return-to-work studies. We need to bring together multidisciplinary teams that can address the social/behavioral, biomedical, and analytic issues in the research. These multidisciplinary teams will need to create new, interdisciplinary conceptual models for the process of returning to work. We need to prioritize among the diverse group of risk factors related to return-to-work outcomes, focusing on those that are amenable to change and relevant to workers and employers. We also need to raise the methodological level of our research, making use of survival models to account for censoring of data and to maximize the efficiency of our modeling (Collett, 1994). That of hierarchical models to simultaneously assess risk factors from several levels, ranging from the individual to the societal (Diez-Roux, 1998).

Of the almost 100 predictors of return-to-work identified in Krause’s (2001a) extensive review of the literature, most were only measured in a handful of studies, but one was repeatedly identified in a variety of different studies. In a detailed review of the literature, Krause and colleagues (1998) found that in 13 high-quality studies, modified work programs facilitated the return-to-work of workers with temporary or permanent impairments. Injured workers who were offered modified work were about twice as likely to return to work as those who were not offered modified work, and they returned to work about twice as quickly. Almost all of these work-modification programs centered on making light duty assignments available to impaired workers. One of these studies presented data suggesting that the light-duty program was cost-effective, but most presented no economic data.