Les Skramstad was a good, decent man who died in January 2007 from mesothelioma at 70 years young. Mesothelioma is a rare cancer caused by exposure to asbestos. Mr. Skramstad was a miner and laborer at the infamous vermiculite mine at Zonolite Mountain in Libby, Montana. Mr. Skramstad’s death was clearly work-related, but when the Labor Department’s annual census of work-related deaths was published the following year, reporting 5,488 fatal work injuries, his death was not included. As noted in Part 1 of this series, our nation’s official count of work-related deaths is a census of fatal traumatic injuries. It does not include deaths from occupational illnesses. The deaths of thousands of workers every year, like Les Skramstad, who succumb to asbestosis, silicosis, black lung, other pneumoconioses, hypersensitivity pneumonitis, and other disorders are not part of the annual toll.
The Labor Department acknowledges the annual census is a count of fatal injuries, not illnesses. For at least two decades, public health and worker safety advocates have bemoaned this gaping hole in national data, but little has been done to remedy the situation.
There are clearly challenges in developing an annual tally of work-related illness deaths. First, some diseases are clearly associated with exposures at work, but some physicians fail (or aren’t trained) to make the connection with the patient’s work history. Second, some diseases don’t develop until long, even decades, after the worker has left that place of employment or retired. Medical care costs for a known work-related disease, like silicosis, may be covered by a workers’ compensation insurer, but there’s no mechanism currently in place for the Labor Department, HHS’s National Institute for Occupational Safety and Health (NIOSH) or another federal agency to obtain that data. Third, many diseases have multiple causes, that is, there are a number of factors that may increase your risk of developing the disease. From studies of populations, we know that cigarette smokers are at a much higher risk of developing bladder cancer. About 57,000 cases of bladder cancer cases are diagnosed every year in the U.S. But, we also know that individuals who were exposed to certain aromatic amines, including those used in paints and dyes, and to radiation are also at higher risk of developing bladder cancer. Steenland, et al estimate that at least 651 and as many as 2,191 bladder cancer deaths in the U.S. are attributed to occupational exposures.
Taking these and other uncertainties in mind, a few researchers have developed models for estimating cases of occupational disease mortality and morbidity (here, here, here, here, here) using a percentage of deaths in particular disease categories that are attributed to occupational exposures. The published literature suggests, for example, that about 6-10% of all cancers are attributed to workplace exposures. These researchers have used this population attributable risk (PAR) method to estimate in a given year the number of occupational illness deaths from different diseases.
These models aren’t perfect, and uncertainties remain, but they illustrate how the Labor Department (or NIOSH, or another federal agency) could use the best available data to develop an annual (or biennial) estimate of occupational illness deaths. BLS’s Census of Fatal Occupational Injuries (CFOI) provides one part of our work-related death count, but isn’t it time to complement it with a toll of occupational disease deaths? As Steenland et al suggested in 2003, there are ways to estimate deaths due to occupational causes by combining data from CFOI and illness mortality evidence from other sources.
JP Leigh, JE Cone, and R Harrison used the PAR method and published a paper in Preventative Medicine in 2001 reporting an annual estimate of occupational disease and death for the State of California and the entire US. They assembled data from a variety of sources, adjusted for deficiencies in the data, and used some of the following PAR estimates: 5-10% of chronic respiratory disease is attributed to work; 1-3% of nervous system and renal disorders is attributed to work; and 100% of pneumoconioses. Applying the PAR to population data, they estimated the following for the number of US deaths in 1992 attributed to occupational disease:
*Of the 517,090 deaths from all forms of cancer, 31,025 to 51,709 of them could be considered work-related.
*Of the 101,846 deaths (among 25 to 64 year olds) from cardiovascular and cerebrovascular disease, 5,092 to 10,185 of them could be considered work-related.
*Of the 91,541 deaths from chronic respiratory disease, 9,154 of them could be considered work-related.
They also calculated portions of work-related deaths from renal disease, central nervous system disorders, and pneumoconioses. Ultimately they estimated an annual total of 69,290 work-related occupational illness deaths (lower bound: 46,906, upper bound): 70,681) for those six major disease categories.
DOL’s annual report on work-related fatal injuries has come to serve as a barometer of the nation’s worker health and safety prevention system. There’s no doubt that workplace safety hazards that caused in the past thousands of traumatic deaths have been eliminated and workers’ lives have been saved. Where we’ve failed is implementing some national system to estimate deaths from occupational diseases.
In 1987, an expert panel of the National Research Council suggested that NIOSH take a leadership role in occupational disease surveillance, including collaboration with the National Center for Health Statistics
“to arrive at an ongoing approach to obtaining annual estimates of the number of occupational fatalities.”
NIOSH response to this recommendation is manifest in its strategic goal and activities related to surveillance, in particular, its coordination with the Council of State and Territorial Epidemiologists (CSTE). What’s unclear is the extent to which these efforts have focused on assisting NIOSH, BLS or another federal agency in providing an annual estimate of occupational illness deaths. Based on an October 2005 CSTE report entitled “Putting Data to Work,” it appears that their work involves other vital occupational health surveillance activities, specifically the development of indicators such as numbers of hospitalizations from burns, amputations identified in State workers’ compensation systems, and elevated blood-lead levels.
Last week, OSHA hosted a meeting with 70 representatives from NIOSH, CSTE, and the Labor Department to discuss common goals and ways to share data. I wonder if the group discussed how an annual estimate of work-related fatalities—not just of acute fatal injuries, but illnesses too—-might enhance the public health community’s ability to make the case for strong health and safety protections for working people. There are clearly challenges in developing an annual tally of work-related illness deaths, but models for doing so are available and tested.
In the next installment of this series, I’ll explain why the Labor Department says there were 1,238,490 cases in 2009 where a person lost time from work because of an on-the-job injury, and why public health researchers believe that number understates the problem.