Patterns of Evidence
Musculoskeletal disorders, especially those associated with low back pain, are very prevalent and a major reason for seeking health care. Overall, in the United States, people make some 69 million clinic visits annually for these disorders. Furthermore, approximately 1 million people take time away from work each year to treat and recover from musculoskeletal pain and loss of function. For workers in their 50s and 60s, musculoskeletal disorders represent the most common cause of disability, and current projections suggest that these figures are on the rise. In sum, data from the general population of workers and nonworkers suggest that the problem of musculoskeletal disorders is a major source of short-term and long-term disability with economic losses in the range of 1 percent of the nation's gross domestic product.
Because most people living in the United States work (more than 80 percent of the adult population), comparisons between the general population and those who work are unreliable. Explicitly, the available data on the general population include work-related as well as nonwork-related musculoskeletal disorders without distinctions. Therefore, rates derived from these general population sources cannot be considered in any sense equivalent to rates for background, reference, or unexposed groups. Nor can they be considered as rates for musculoskeletal disorders associated with any specific work or activity. No comprehensive data are available on occupationally unexposed groups, given the proportion of adults now in the active U.S. workforce; that is, any such nonemployed group would, by definition be unrepresentative of the adult population.
Nevertheless, the magnitude of the problem of work-related musculoskeletal disorders can be gleaned from the Bureau of Labor Statistics data. These data suggest that musculoskeletal disorders are a problem in
several industrial sectors and are not limited to the traditional heavy labor environments represented by agriculture, mining, and manufacturing. It was reported, for example, that the service sector accounted for 26 percent of sprains/strains, carpal tunnel syndrome, or tendinitis, while the manufacturing sector accounted for 22 percent. Another database, National Center for Health Statistics, using self-reports provided estimates for back pain among those whose pain occurred at work (approximately 11.7 million) and for those who specifically reported that their pain was work-related (5.6 million). In this survey, the highest-risk occupations among men were construction laborers, carpenters, and industrial truck and tractor equipment operators; among women, the highest-risk occupations were nursing aides/orderlies/attendants, licensed practical nurses, maids, and janitor/cleaners. Other high-risk occupations were hairdressers and automobile mechanics. Many such workers often are employed in small businesses or are self-employed.
The focus of the panel's work has been the review and interpretation of the scientific literature characterizing musculoskeletal disorders of the low back and upper extremities and their relationship to work. Here we provide an integration of the studies that have been reviewed in the chapters on observational epidemiology, biomechanics, basic sciences, and workplace interventions. As noted in the chapter on epidemiology, there are significant data to show that both lower back and upper extremity musculoskeletal disorders can be associated with workplace exposures. Across the epidemiologic studies, the review has shown both strength and consistency of association. Concerns about whether the associations could be spurious have been considered and reviewed. Biological plausibility has been demonstrated in biomechanical and basic science studies, and further evidence to build causal inferences has been demonstrated by intervention studies that demonstrate reduction in the occurrence of musculoskeletal disorders following implementation of interventions.
The purpose of this discussion is to extend beyond the summaries of each of the chapters, and to integrate information among the chapters relevant to the model presented in Figure 10.1 (which reproduces Figure 1.2 for the reader's convenience). Also, with the acknowledgment that each set of studies has inherent limitations that affect the confidence about conclusions, another purpose is to consider the patterns of evidence that emerge across the different types of studies, as noted earlier in the report.
The integration of findings associated with musculoskeletal disorder risk and the workplace can best be addressed by considering the evidence for the presence of linkages representing the various pathways to injury shown in Figure 10.1. There is a large and diverse body of literature addressing the work-relatedness of musculoskeletal disorders, with different aspects of the literature suggesting different mechanisms of injury.
~ enlarge ~
However, the basic view taken here is that the various methodologies and approaches to exploration are essentially investigating different aspects of the same problem. The figure illustrates the components of the workplace-person interaction that are addressed by these different research approaches. Inherent to this rationale is the idea that there can be many pathways or linkages to injury, and the presence of one pathway does not negate or suggest that there is a lesser association with another potential pathway to injury. The different linkages simply represent different aspects of the same workplace-person system. This chapter considers how well these linkages have been established via our review of the literature.
BACK DISORDERS AND THE WORKPLACE
As shown in Figure 10.1, the literature relating to epidemiologic studies of low back disorders has evaluated the linkage between the workplace and low back disorders primarily along two dimensions. First, exposure to external loads in the workplace and its association with low back disorders outcomes has been explored. Second, the association between organizational factors and social context (also called psychosocial factors) has been investigated for their association with low back disorders.
Physical Work Factors
The review of the observational epidemiology literature has shown support for the linkage between external load exposure in the workplace and increased low back disorders. Specifically, the review concluded that there is a clear relationship between low back disorders and physical load, frequency of bending and twisting, physically heavy work, and whole-body vibration (with risk ratio estimates up to 9-fold and attributable fractions of between 11 and 66 percent). These findings are reinforced by a review of risk associated with specific industries. For example, industries that tend to impose greater spinal loads on workers, such as patient handling and case distribution in warehouses or distribution centers, present significantly greater risk of low back pain and also greatly benefit from appropriate interventions. When individual physical work risk factors have been identified, and when combinations of these factors (e.g., load location and body postures) have been examined, the relationships between workplace factors and risk become stronger (with risk estimates exceeding 10). Hence, the literature shows that specific components of high-risk jobs (e.g., large load moments) can explain risk in many differ-
ent types of work environments. These findings further suggest that these are not the only factors that can increase risk, but they must be present at least at some minimal level for risk to increase. Thus, there is a need to quantify the magnitude of risk factors so that the degree of risk can be assessed.
However, while work-related physical load, frequency of bending and twisting, physically heavy work, and whole-body vibration have been implicated in musculoskeletal disorders, evidence for static work postures has been less compelling. Data from different studies did not indicate a relationship between low back disorders and static posture.
Psychosocial Work Factors
The epidemiologic literature that has assessed linkages between psychosocial work factors and low back disorders has also established an association (with risk estimates between 1 and 5). The review of the literature has shown considerable evidence for a relationship between psychological work factors and future back pain. Specifically, evidence has been found for the relationship between low back disorders and job satisfaction, monotonous work, work relations, work demands, stress, and perceived ability to work. While an association between these workplace factors and low back disorder outcomes has been established, these investigations were not designed to delineate an exposure intensity/duration (dose)-response relationship. Still to be explored thoroughly is the interaction between physical and psychosocial work factors in the etiology of low back disorders at work.
The intervention literature provides evidence of a different type for the relationship between exposure to work factors and the increased risk of low back disorders. These studies assess the impact of a reduction in physical work factors on reducing low back disorders. The studies reviewed show that a change in the risk of low back disorders was established when changes were made to the physical aspects of work (engineering changes). For example, in manufacturing environments, there is evidence that specific interventions, including lift tables or lift hoists, can be effective in reducing risk. However, the literature also indicates that these interventions will be effective only if they mediate the particular risk factor present in the workplace. With respect to evaluating the mechanisms of effectiveness of these interventions, it has also been shown that the physical changes in the workplace that accompany effective interventions (e.g., lift tables, hoists) can reduce biomechanical load-
ing of the spine in a meaningful way. Thus, appropriate interventions must be sufficiently specific if positive results are to be expected. In addition, the role of psychosocial risk factors was also reinforced, in that changes in the organizational culture (e.g., providing workers with more control over their workplace layout) consistently reduced the rates of low back disorders.
Collectively, these findings establish the linkages represented by arrows going from the “workplace” box to the “outcomes” box in Figure10.1. Epidemiologic studies, however, do not address the mechanism underlying the workplace-human system response to adverse working conditions. Given that workplace risk factors exist, one would expect that such mechanisms can be explained through other avenues or linkages shown in the figure. Further, relationships among the mechanisms and the role of “dose” must be evaluated through other linkages within this conceptual model.
Biomechanical and Biological Evidence
Our review of the biomechanical and basic biology literature provides a consistent explanation for the linkages established by the epidemiologic literature. Both the physical work factors pathway as well as the psychosocial pathways can be explained by the literature. The biomechanical literature shows that there is abundant evidence for a load-response relationship associated with low back disorders. The biomechanical surveillance literature shows that when workplace exposure is described at a greater level of specificity than is typically available for epidemiologic studies, the findings point consistently to the pathway between work exposure and risk, as well as work exposure and increased loading of the spine. This description includes such factors as precise load location, load moment, spinal load, three-dimensional trunk position, frequency, and kinematics.
Biomechanical modeling efforts over the past two decades have improved significantly to the point at which work exposure can be linked to specific patterns of spine structure loading. The literature also well documents how these loading patterns can cause damage to the pain sensing structures of the spine, including the disc, vertebral body, facet joints, and ligaments. Studies have shown how loading of these structures—at the magnitude associated with some work tasks—can lead to damage of these structures and the perception of pain. For example, spine loading models have been applied to experimental data collected on workers performing tasks that have been shown in epidemiologic studies to be associated with greater risk for low back disorders (for example, patient handling in health care settings, handling cases of material from the floor level). These bio-
mechanical studies indicate that spine loading can exceed the expected spine tolerance during these activities and thus they reinforce the epidemiologic findings. The literature also shows that some level of adaptation occurs in most biological tissue. In sum, infrequent exposure to load increases the risk of low back disorders, moderate exposure reduces risk, and repeated exposure to high loading greatly increases risk of low back disorders. Thus, there appear to be limits of exposure within which risk can be minimized.
Recent biological evidence has also shown that both repeated loading of the spine and higher magnitude of loading can explain deterioration of spinal discs. These findings support the existence of a dose-response relationship of load associated with low back disorders.
Evidence has recently emerged in the scientific literature that points to a mechanism by which psychosocial stress can increase loading of the spine. Increased levels of psychosocial stress appear to act, biomechanically, through increased coactive recruitment of the torso muscles, thereby increasing the loading of the spine. There is also evidence to suggest that exposure to psychosocial stressors may result in greater trunk muscle activity (but not necessarily force) independent of biomechanical load. Psychosocial stressors can also influence pain tolerance and recovery from tissue injury or inflammation. However, further investigations are needed to understand this mechanism of injury more fully.
Individual factors appear to mediate the biomechanical and biological pathway mechanisms described in Figure 10.1. Some part of the variance in response that has been described in the biomechanical and biological literature appears to be explained by individual host factors that can mediate loading response to external workplace factors as well as mediate tolerances to such loadings. For example, age and gender appear to play a role in determining the magnitude of load tolerance to which the spine may be exposed before damage would be expected. Similarly, differences in spine loading patterns have been documented as a function of gender, experience, and reactions to psychological stress.
Collectively, these findings have shown that the various linkages in Figure 10.1 between external loads and biomechanical loading of the spine, biomechanical loading and internal tolerances of the spine, and internal tolerances and outcomes (from pain through disability) are well established and provide a plausible pathway between workplace factors
and the outcome of low back disorders reported by workers. It is clear that these risk factor pathways are present when the magnitude of the risk factor reaches a specific quantitative level relative to the tissue tolerance level. There is also significant evidence that individual factors such as age, gender, and physical condition play important roles in mediating the response to work factors associated with biomechanical loading and with the tolerance levels for the individual. Thus, the literature also suggests that clear linkages appear in Figure 10.1 between individual factors and biomechanical loading and between individual factors and internal tolerances. Although some literature exists that describes how different individuals respond to pain, the precise mechanisms by which pain is differently experienced among individuals are not well understood.
Collectively, the biomechanical/biological pathway literature agrees well with and independently establishes a causal pathway suggested by the epidemiologic literature. The literature relating to causal factors in work-related low back disorders is coherent and provides ample evidence on how adverse work situations can lead to them.
UPPER EXTREMITY DISORDERS AND THE WORKPLACE
As with low back pain, associations between workplace exposures and upper extremity disorders have considered both physical and psychosocial exposures and evidence provided by intervention studies. The literature on physical exposures documents a strong association between physical factors and upper extremity disorders. The relevant exposures specifically implicated include repetition, force, vibration, and the combinations of repetition and force or repetition and cold. The combination of force and repetition as well as vibration is associated with carpal tunnel syndrome and other disorders of the wrist (relative risk estimates from 2-to 39-fold), and somewhat less strikingly with tendon and muscle-related conditions (relative risk estimates from 3- to 14-fold). The literature on psychosocial factors provides support for an association between high job stress and job demands and upper extremity symptoms. The most dramatic physical exposures occur in manufacturing, food processing, lumber, transportation, and other heavy industries, and these industries have the highest rates of work associated upper extremity disorders (see Chapter 2). However, several epidemiologic studies of physical exposures (force, repetition) and psychosocial exposure (perceived stress, job demands) have documented an elevated risk of upper extremity disorders among computer users. Nonwork-related worry, tension, and psychological distress are also associated with upper extremity symptoms.
The epidemiologic findings are complemented by literature on interventions for upper extremity disorders. Several interventions that diminish exposure to force, repetition, and awkward postures have, as hypothesized, reduced upper extremity symptoms. For example, studies have shown that addressing these documented physical risk factors can reduce musculoskeletal symptoms in heavier industries. As discussed in the epidemiologic reviews and the interventions chapter, there is some evidence that using ergonomic principles to modify chairs, workstations, and keyboards can be effective in reducing the prevalence or severity of upper extremity symptoms; in the office setting, results concerning the effects of these interventions on physical findings are mixed. Thus, the exposure-response associations appear to extend robustly across occupation types. Exercise interventions, however, have not been consistently associated with benefit. There is good evidence that interventions are most effective when supported fully by workers and other key stakeholders.
While the epidemiologic literature on psychosocial factors indicates that high levels of job stress and perceived job demands in particular were consistently related to the occurrence of upper extremity symptoms and disorders, few studies have investigated the effects of interventions directed at psychosocial risk factors in workers with these symptoms or disorders. The limited number of worksite studies that have attempted to reduce levels of job stress in workers have reported improvements in a number of nonmusculoskeletal health outcomes. However, the effects of reducing psychosocial job stress on musculoskeletal disorders has not been systematically studied.
Biomechanical and Biological Evidence
The upper extremity biomechanics literature in total reveals a consistent pattern of evidence in support of the pathways illustrated in Figure 10.1. This research supports several components and interactions in the model, including distinct relationships between (1) physical stress factors and external loads in the workplace, (2) external physical loads and internal tissue loads, (3) external physical loads and internal tolerances, and (4) external loads and pain, discomfort, functional limitations, or disability.
Upper extremity exposure to physical stresses (i.e., force, repetition, posture, vibration, and temperature) can be greatly affected by various attributes of work. Force exerted in occupational tasks can be directly affected by the weight of objects handled, frictional characteristics be-
tween surfaces grasped and the skin, and the forces required to accomplish tasks. Reductions in the magnitude of the physical attributes of work can produce reductions in exposure to external loads, indicating how physical workplace interventions may directly affect physical stress exposures.
The relationship between external physical loading and internal loads is supported in the literature through biomechanical models that link external forces and postures to tendon loads and the transmission of tendon forces to adjacent anatomical structures. In addition, studies have demonstrated and quantified relationships between exertions and postures and internal loading of the median nerve in the wrist. Additional evidence is provided by electromyographic studies of muscle activity in response to external loads. Several studies have identified an increased risk when the magnitude and duration of two or more physical stresses were considered together.
Evidence for exceeding internal tolerances from loads originated externally comes from physiological studies measuring mechanical strain or fatigue. Further evidence is provided by psychophysical studies of mechanical strain resulting from work in certain postures and exertions. The literature contains numerous studies that have considered multiple properties of forces and posture; however, few studies have dealt with vibration and temperature.
Research reveals short-term effects of pain, discomfort, and limitation of function due to external loads. Although these effects are not necessarily irreversible, the research provides additional support for linkages between external loading and adverse outcomes. Long-term effects of external load are also suggested by cross-sectional studies linking level of external load to upper extremity disorders.
Basic biology studies provide plausible mechanisms for injury to upper extremity tissues (e.g., nerve, tendon, and muscle), and some support for the causality criteria of specificity, temporality, and dose-response. The studies demonstrate damage accumulation in tendon, muscle, and nerve in a dose-response pattern. The endpoints of injury are specific to tissue type and correspond to histological changes observed among patients with related musculoskeletal disorders.
Individual factors also mediate biological pathways for upper extremity disorders, especially for carpal tunnel syndrome. Nerve function declines with age and other factors (e.g., pregnancy and body mass index). The role of individual factors in mediating the effects of loading on tissues has not been thoroughly investigated.
The findings of the intervention literature are congruent with those of the basic biology and epidemiology literatures. There is strong support across these bodies of work that high force and repetition are associated with muscloskeletal disorders; in the basic biology data, there is also evidence of alteration of tissue structure. The intervention literature supports the efficacy of tool and workstation design changes, job rotation, and other interventions that directly address these risk factors.
Thus, while the upper extremity literature is less developed than the literature on low back pain, an analogous set of themes emerges, lending further support to the model presented in Figure 10.1. Specifically, external loads and psychosocial factors (components of the “workplace” box) influence outcomes. These exposure-response associations persist when adjusted for individual factors. The basic biology and biomechanics studies provide a plausible basis for the exposure-response relationships. The efficacy of ergonomic interventions further supports the models. The consistent pattern of evidence across these diverse types of studies strengthens the validity of the underlying conceptual model. The pattern of evidence also strengthens the rationale for interventions that ameliorate these exposures.
By considering collectively the evidence derived from the epidemiology, biomechanics, basic biology, and interventions literatures, we can develop an appreciation for the consistency or pattern of evidence that has emerged from them relating to musculoskeletal disorders in the workplace. Each one of these research approaches has demonstrated relationships between specific components of work exposure and musculoskeletal health. As with all scientific research, however, each approach has limitations. For example, the epidemiologic surveys by work categories or job titles are important because they capture information from large worker populations in actual work settings, but the crudeness of the measure for exposure is not satisfying. Additional epidemiologic studies are available that improve on the measure by asking workers to summarize or average their exposure, but some scientists consider self-reports suspect. Other studies include observations, such as taking an “average” measure of exposure (e.g., researchers observe, in a subsample of workers, the number of lifts and twists per shift and extrapolate to a study population that has the same job category). The biomechanical studies are critical here because both the measure of exposure and the measure of outcome are more closely considered in them. This more refined measure helps to
establish confidence in causal inference, but alone, the biomechanical studies do not provide the mechanism.
The basic science studies that document damage and injury mechanisms provide biological plausibility for the associations from the biomechanical studies and epidemiologic studies. The biomechanical studies alone, while elegant, have limitations in that the measurements are more intrusive and obvious than observations in the casual work environment. This intrusiveness limits the generalizability of measures and the implication of findings for the ambient work environment. The epidemiologic findings based on broad population data, however, provide context, indicating that the data from the biomechanics studies are neither isolated nor irrelevant. Although each type of study can be criticized as limited, it is the pattern across the different study designs and types of measures that affects the ability to consider the relationship as causal. Moreover, the strength of the evidence is significantly enhanced by the consistency associated with the pattern of evidence. For example, low back disorder risk has been established through epidemiologic studies for work that involves heavy lifting as well as other risk factors. The relative risks have been derived from a rigorous evaluation of the literature and have been found to be reasonably strong and consistent. Furthermore, the studies control for confounding and indicate a presence of temporal association and dose-response relationships.
The epidemiologic literature that specifically quantifies heavy lifting shows that risk of injury is greatest when loads are lifted from low heights, when the distance of the load from the body is great, and when the torso assumes a flexed, asymmetric posture. Thus, risk estimates significantly increase when work risk factors can be classified with greater precision. These studies typically have relied on more precise endpoints (such as injury logs), but they have limitations to the extent that they have not involved prospective assessments. A number of such studies have the limitation that the results are based on self-reports of injury.
Collectively, a pattern emerges demonstrating that risk is associated with specific characteristics of the workplace. By exploring the biomechanical basis for this risk, the pattern of evidence becomes even stronger. Biomechanical studies have shown that the loading of the spine is increased dramatically when a load is lifted from a low height in an asymmetric posture. According to the basic biology tolerance literature, the spinal load tolerance is also greatly reduced when the torso is positioned in a flexed posture or torsion is applied along the axis of the spine. Basic biological research has also described how spinal loading can lead to cell death within the disc. Furthermore, the basic science literature has described pathways for the perception of pain when specific structures in the spine are loaded under these conditions. Finally, intervention studies
have shown how lift tables and lifting hoists are effective in mediating the risk of low back pain in industrial studies. The incorporation of such devices in the workplace permits the load to be raised mechanically, minimizing the load moment and thus reducing the biomechanical loading of the spine. Since risk is lowered when the load is changed from a heavy lift to a light lift, this finding would also be consistent with the rigorous epidemiologic findings. As demonstrated in this example, a pattern of evidence emerges that cuts across independent risk assessment approaches—epidemiologic, biomechanical, basic science, and interventions.
A similar pattern can be described for upper extremity disorders. Biomechanical studies have shown that the extraneural pressure in the carpal tunnel is increased with hand loading and nonneutral wrist postures. Basic science studies demonstrate how extraneural pressures lead to intraneural edema and fibrosis, demyelination, and axon degeneration. These changes in nerve structure cause loss of nerve function. Finally, a recent intervention study has demonstrated that alternative keyboards that reduce nonneutral wrist postures can mediate hand pain in patients with carpal tunnel syndrome.
In conclusion, a clear and strong pattern of evidence emerges after considering the epidemiologic, biomechanical, basic science, and intervention literature collectively. We can conclude with confidence that there is a relationship between exposure to many workplace factors and an increased risk of musculoskeletal disorders.