Twice as many work-related skull fractures than government estimate

Researchers with Michigan State University’s (MSU) Department of Occupational and Environmental Medicine have done it again. First it was work-related burns. Then it was work-related amputations. Now it is work-related skull fractures. The MSU researchers continue to poke holes in the federal government’s annual estimates of occupational injuries.

Joanna Kica, MPA and Ken Rosenman, MD used data from acute care hospitals, workers’ compensation records, death certificates and police reports to identify work-related fatal and non-fatal skull fractures occurring in Michigan in 2012. They identified a total of 316 work-related skull fractures, six of which were fatal. Facial and non-facial fractures totaled 218 and 98, respectively. The authors note their interest in this particular category of work-related injuries because skull fractures are considered to be among the most severe types of traumatic brain injuries. TBI’s can cause permanent impairment and recovery can require extensive long-term treatment. Moreover, as with most work-related injuries, illnesses and fatalities, TBI's can be prevented.

But when reporters and policy makers talk about work-related injuries, they are likely to refer to data from the Labor Department's Bureau of Labor Statistics (BLS), specifically its Survey of Occupational Injuries and Illnesses (SOII). Instead of the actual count from Michigan of 316 work-related skull fractures, SOII indicates there were 170 cases. That’s a difference of nearly 54 percent.

Kica and Rosenman demonstrate the value in using multiple sources to assemble data on work-related injuries. Of their 316 cases, 193 were identified by hospital/emergency department records; another 100 cases were identified by both workers’ compensation records and hospital data.

BLS’ SOII, in contrast, uses data from a survey of about 200,000 workplaces---a representative sample of all U.S. workplaces---to estimate work-related injuries and illnesses. SOII's biggest limitation is that it depends on self-reported data from employers. Some employers might misunderstand what they are supposed to report, others may intentionally deceive the government. Either way, the SOII data should be used with caution and with caveats.

SOII is also incomplete. Some work-related injuries aren’t captured in the data because small farms, independent contractors and the self-employed are not included in the survey.

Previously, the MSU researchers published papers comparing BLS’ estimates of work-related amputations in Michigan to the number of cases ascertained using multiple data sets. They calculated that BLS missed about half of them. Similarly, they found that the BLS system also missed about 70 percent of work-related burns that occurred in the State. I’ve no reason to believe that the undercount in BLS’s data for Michigan isn’t repeated in every other State.

Much of the October 2014 issue of the American Journal of Industrial Medicine (AJIM) is devoted to exploring ways to improve BLS's SOII. Writing in the issue, Emily Speiler, JD and Greg Wagner, MD, MPH explain why it matters:

“Underreporting obscures reality and misleads all of the key stakeholders when issues of work and health are discussed—in workplaces, employers' organizations, trade unions, administrative agencies, research organizations, and in the courts. In essence, underreporting is a denial of the experience of workers injured or sickened by workplace exposures and conditions.”

Rosenman and colleagues are doing a service to their State---and the rest of us---by providing a reality check on the incidence of work-related injuries.