Study Finds U.S.
Undercounts Workplace Injuries, Illnesses

KRIS MAHER / Wall Street Journal 1may2006

A new study suggests government statistics undercount the number of illnesses and injuries that occur in U.S. workplaces each year, largely as a result of underreporting by employers.

The study, by researchers at Michigan State University, analyzed data collected by the Bureau of Labor Statistics and found that the statistics failed to count roughly two-thirds of nonfatal work-related injuries and illnesses that occurred in Michigan over a three-year period.

The findings are in line with previous broad estimates of undercounting, but in this study researchers had access to confidential BLS records and used multiple databases to find evidence of injuries and illnesses at companies. Experts say more accurate data could reshape commonly held views of how safe workplaces are and have policy implications for government agencies charged with enforcing workplace safety regulations.

"If it's not accurate, how do you know where to put your resources and if your interventions are effective?" said Kenneth Rosenman, professor of medicine at Michigan State University and principal author of the study.

Government statistics on work-related injuries and illnesses are compiled by the BLS through an annual survey of about 230,000 employers. Yet many researchers say employers and employees alike have incentives to underreport injuries. In addition, the government survey omits large groups of workers at employers who aren't required by law to report, including government employees, self-employed people and workers at farms with fewer than 11 employees.

In the Michigan State study, which appears in the April issue of the Journal of Occupational and Environmental Medicine, researchers used sources such as state workers' compensation reports to find cases where employers surveyed by the BLS had failed to report injuries and illnesses. Researchers estimated that 869,034 work-related injuries and illnesses occurred on average each year in Michigan from 1999 to 2001, compared with the BLS estimate of 281,567 per year. Dr. Rosenman estimates that 75% of the injuries and illnesses missed by BLS resulted from employer underreporting.

"There is probably some undercount and we are missing some," said Bill McCarthy, chief of the BLS's division of safety and health statistics. He said another study using a different methodology is under way.

Labor groups have cited the undercounting to question government claims that workplace injury rates are declining. "The problems with reporting are so significant that it is really hard to get a handle on what is indeed going on with respect to workplace injuries," said Margaret Seminario, director of safety and health at the AFL-CIO.

In February, Labor Secretary Elaine Chao cited a 9.4% decline in injury and illness rates at employers between 2002 and 2004 among department accomplishments.

J. Paul Leigh, a professor of health economics at the University of California, Davis, said he thought overall rates have declined, in part because the economy is creating more service jobs.

In a report last year, insurer Liberty Mutual Group Inc. estimated the total cost of disabling occupational injuries and illnesses that cause workers to miss six or more days of work to be $50.8 billion to employers in 2003, using BLS injury and illness data. Mr. Leigh puts the figure at about $170 billion a year.

p.A4