New Arthritis Study Shows O*Net Could Help Predict Chronic Disease

A recent study points to a positive and consistent relationship between O*NET ratings of physical job demands and the risk of contracting arthritis later in life. Importantly, the study also demonstrates the utility of using such O*NET job descriptors to estimate the long-term risks of contracting other chronic diseases and conditions when no actual exposure data is available. [See Dembe, AE et al., “Using O*Net to Estimate the Association Between Work Exposures and Chronic Diseases,” American Journal of Industrial Medicine vol. 57, pp. 1022–103, Sept. 2014]

Background

Evidence has long suggested that occupational working conditions are associated with the development of chronic diseases, such as osteoarthritis, diabetes, asthma, and chronic heart disease. In some cases, the causal connection between job exposures and the development of the chronic health conditions are obvious and well documented. In most cases, however, it is difficult to verify the relationship between long-term occupational exposures and the risk of contracting various chronic conditions later in life. A significant problem is the relative scarcity of data systems that contain the requisite information needed to perform the studies. The Dembe study utilized 32 years of aggregated job exposure information to address these limitations.

Employing data from two publicly available sources: the Occupational Information Network (O*NET) and the National Longitudinal Survey of Youth, 1979 (NLSY79), the researchers created job histories over a 32-year period (1978 through 2009) for the 12,686 men and women in the NLSY79 cohort and then match those job histories with O*NET ratings of work exposure (called “job descriptors” by O*NET) for each job held by every cohort member during the study period. Five O*NET job descriptors were used (e.g., bending and twisting of the body). Various statistical methods were also used to gauge the combined effect of the five O*NET descriptors.

O*NET and NLSY79

O*NET, sponsored by the U.S. Bureau of Labor Statistics (BLS), is a comprehensive database of worker attributes and job characteristics. Containing continually updated information on the skill requirements and job characteristics of 974 occupational classifications, it characterizes each occupation by a uniform, measureable set of 277 variables called “descriptors” that describe and rate job requirements, worker activities, workplace conditions, and the like.

BLS also sponsors the NLSY79. Comprised of 12,686 men and women who were 14–22 years of age when first surveyed in 1979, the dataset contains information from follow-up interviews with NLSY79 respondents on an annual basis from 1979 to 1994, and biannually since 1996. The NLSY79 collects information on respondents’ sociodemographic characteristics, household composition, education, training, detailed work histories, job and employer characteristics, incidence of work related injuries and illnesses, episodes of work disability, etc. and is designed to be representative of the non-institutionalized civilian segment of young people living in the United States in 1979 and born between January 1, 1957 and December 31, 1964.

Study Design

The researchers identified five O*NET job descriptors that were most directly relevant to physical work demands:

(1) handling and moving objects,

(2) spending time bending or twisting the body,

(3) spending time kneeling, crouching, stooping, or crawling,

(4) working in cramped work spaces and awkward positions, and

(5) performing general physical activities.

The relationship between each of those risk exposures was analyzed individually, with doctor-diagnosed arthritis as the primary outcome variable. The O*NET job descriptor “performing general physical activities” was considered as a composite generic risk category encompassing the other four.

The presence of arthritis in a cohort member was ascertained by a positive response to the following question on the NLSY79 health module: “Have you ever had, or has a doctor ever told you that you have, arthritis or rheumatism?” A follow-up question asked the respondent to report the month and year in which arthritis was first diagnosed. One of the possible response values to that question was “never diagnosed.” The researchers only considered responses from respondents who were diagnosed with arthritis by a doctor. Over the 32-year study period, 1,394 respondents reported having arthritis, of which 1,277 had doctor-diagnosed arthritis. To ascertain the intensity level of exposure, the researchers used the intensity level rating provided in the O*NET dataset. For example, construction carpenters have an intensity level rating of 76 (out of 100) for handling and moving objects.

Results

In general, a “high” level of exposure among each of the five relevant O*NET descriptors was found to be significantly associated with a diagnosis of arthritis. In general, analyses of the linear trend and by quartiles produced evidence for a pattern of increasing risk with higher levels of mean intensity for all five job descriptors. There was, however, some leveling off for two descriptors, “working in a cramped work space or awkward postures” and “performing general physical activities.” The largest single contributor in the combined factor analysis was “performing general physical activities,” suggesting that factor could be considered as a composite risk category encompassing the other four. The descriptor “spend time bending or twisting the body” was the second strongest factor.

The study’s finding of a positive and consistent relationship between O*NET ratings of physical job demands and arthritis risk shows the putative benefit of using O*NET descriptors as surrogate indicators of long-term workplace exposures, in situations where conventional direct exposure assessment is not feasible. The findings also help to validate the program established more than 70 years ago by the U.S. Department of Labor to collect empirical information using field raters (and, more recently, survey responders) to continually update and refine specific information about the activities and demands faced by workers in hundreds of distinct occupations. It should be noted that neither the Department of Labor nor the BLS funded the study.  It was sponsored by the U.S. Centers for Disease Control and Prevention. This study adds to the growing body of evidence indicating the convergent validity of using O*NET indicators as measures of occupational exposure.

-----------------------------------------

Attorney Perspective: Attorney Chuck Davoli, managing partner of Davoli, Krumholt & Price, Baton Rouge, Louisiana, and President of Workers’ Injury Law & Advocacy Group (WILG) states, "It is not surprising that advancements in medical research are discovering the long-term effects of some occupations and occupational exposures. Prior examples have already been demonstrated re: health effects of asbestos exposure and progressive occupational disease or mesothelioma, the most recent findings of long term effects of closed head injuries causing professional athletes (football players) progressive cognitive deficits and early onset of Alzheimer’s disease or dementia, and more recently the NIOSH research indicating increasing prevalence of progressive massive fibrosis (PMF) disease among coal miners. Such research should cause states to re-think their current exclusions for workers comp compensability for such occupational anomalies. For example, Louisiana like most other states expressly excludes any ‘gradual deterioration or progressive degeneration’...including , ‘arthritis of any type’ as well as any ‘repetitive motion’ injuries in both their definition of accident causation and occupational disease; however, Louisiana has also long recognized asbestosis as an occupational disease as well as carpal tunnel syndrome (CTS) often caused by repetitive motion....Missouri only recently has accepted mesothelioma as an exclusive workers comp remedy as well. The most significant impact of the O*Net findings could also re-focus attention on workplace safety and other preventative measures to mitigate such injuries and exposures".

-----------------------------------------

Limitations

As with virtually all studies, the researchers noted a number of limitations, including:

> O*NET job ratings reflect general characteristics of an occupational category; they inherently fail to capture many important site-specific aspects concerning the activities, exposures, and hazards that actually exist in any particular workplace. [See, e.g., the SSA’s concerns with O*Net for purposes of its disability determination process. http://www.skilltran.com/online/SSA-ONET.PDF.]

> The main outcome of interest in the study was osteoarthritis. However, the NLSY79 question is unable to distinguish osteoarthritis from other types of arthritis or rheumatism, thereby creating potential misclassification. Since osteoarthritis is the most common type of arthritis (estimated to be 85–90% of all arthritis), it is, however, reasonable to assume that most of the cases reported were for osteoarthritis.

> Analyses were limited only to full-time workers (at least 30 hours in any particular week). Those working part-time and those who were unemployed or worked in the home (without wages) were not considered as having contributed exposure for that week. It is possible that part-time workers were actually accruing job exposures that affected their long-term arthritis risk. The researchers did some additional regression analysis suggesting, however, that whether or not part-time work history was included in the analysis had little impact on the final results.

> Other potential risk factors (e.g., the person’s body mass index, hours of work, etc.) could affect the probability of contracting arthritis. Studies indicate that osteoarthritis etiology is generally multifactorial, and can be influenced by genetic predisposition. These unmeasured factors could have influenced the results of the analyses performed.

> Finally, workers who are in poorer health may select into less hazardous jobs or be less likely to be hired for jobs with high exposure (the “healthy worker effect”). Less healthy workers are also more likely to leave high-exposure jobs (the “healthy worker survivor effect”). Some workers with osteoarthritis symptoms could be expected to have avoided jobs with high exposure for musculoskeletal disease. To the extent that might have occurred, it would have tended to bias the study findings toward the null hypothesis.

-----------------------------------------

Occupational Medicine Perspective: Leslie J. Hutchinson, MD, MPH, FACOEM, of HLM Consultants explains: “Linking indirect indicators of job physical demands and physical stress to long-term health effects has value as a screening tool for identifying high-risk workers for medical surveillance, focusing efforts to equip high-risk workers with safety measures and equipment, and training high-risk workers to minimize or avoid job activities that are most problematic.  However, these correlations must be used carefully. They are based on the degree to which representative or typical activities are present in particular jobs. This can lead to the methodological problem notorious in the field of epidemiology for yielding wrong conclusions.  The connection between higher proportions of high-risk physical activities in a particular worker’s job and diagnosis of osteoarthritis reflects the association long established in the mainstream medical literature between long-term wear and tear on joints and osteoarthritis. However, associations with other chronic diseases should be used more carefully. The association between the occurrence of these diseases and workplaces is more biologically plausibly associated with exposures to chemical or biological hazards in the workplace as well as general physical and psychological stress. None of these possibly confounding factors appear to have been identified or controlled in this analysis.”

-----------------------------------------

Conclusion

The primary benefit of the research is the ability to adapt the study’s general research approach—the process of linking O*NET descriptor ratings with NLSY79 chronic disease information over a 32-year period of longitudinal data collection—to other situations in which direct exposure assessment is not practical. This methodology provides a feasible way for researchers in the U.S. to study the relationship between long-term job exposures and the risk of contracting chronic disease later in life. This new methodology could be used as a “screening” device to provide an indication of the situations in which it might be useful to conduct more detailed exposure assessment. The approach could, therefore, help make epidemiology studies more cost effective. The methodology used in the study can identify the potential role of employment activities in chronic disease occurrence, allowing for appropriate interventions by employers, employees and medical care providers.

© Copyright 2014 LexisNexis. All rights reserved.