The injury and illness rates provided every year by the Bureau of Labor Statistics are not exact figures, bringing into question how accurately they reflect injuries and illnesses in the United States.
“We’re operating in a haze,” said Ken Kolosh, manager of statistics for the National Safety Council. “We’re seeing a reduction, but is it a real reduction?”
BLS figures are estimates based on the Survey of Occupational Injuries and Illnesses. Each year, BLS sends surveys to about 230,000 private workplaces, soliciting employers in all industries and of all sizes (except agriculture, in which only those employers with 11 or more employees are solicited).
This survey is different from the OSHA Data Initiative. OSHA annually collects workplace injury and illness data from about 80,000 establishments in high-hazard industries. For an establishment to be chosen, it has to meet certain criteria (such as industry and size). The collected data is used for the agency’s Site-Specific Targeting Program, which targets about 4,100 facilities for inspection. About 4 percent of the establishments OSHA collects data from are included in the BLS survey, according to bureau estimates.
Although the BLS and OSHA surveys are different, they also are related. Kolosh said businesses probably do not keep two sets of books for recording injuries, and an employer’s SOII figures – which go to BLS anonymously – likely are drawn from the employer’s own OSHA-required recordkeeping.
Underreporting in one set of books – an employer’s OSHA 300 log – would lead to underreporting in another set of books – SOII from BLS. However, Kolosh stressed that the amount of possible underreporting likely is consistent. “I don’t think this year is worse than any other year,” he said.
A 2009 Government Accountability Office report confirmed the long-suspected underreporting of occupational injuries and illnesses. As a result, OSHA in September 2009 launched a National Emphasis Program on recordkeeping. The NEP has been criticized as ineffective, so this past September OSHA revised the program by expanding its scope to include more years of data and more establishments that fall under the criteria of the Site-Specific Targeting Program.
Improving an employer’s recordkeeping under the threat of OSHA penalties is not the only way to get an accurate picture of occupational injuries and illnesses. Given the millions of employers in the country, the sample that both BLS and OSHA draw from each year is small.
Kolosh said modernizing injury record collection is a possibility. Electronic data collection could allow for a 100 percent sample and nearly real-time release of data. Right now, more than a year lag exists between when the injuries occurred and when the data is released.
However, improved accuracy comes with a price. “If the reporting gets better, you should see an increase” in the number of injuries and the injury rate, Kolosh said. However, he stressed that would only be an appearance of injury increase. Because the data collection is improving, more injuries and illnesses are being recorded and accounted for that otherwise wouldn’t have been.
Kolosh believes the ongoing downward trend of injuries and illnesses probably is accurate, as it correlates with a downward trend in fatality statistics – figures that are on much more solid ground.