Not adding up?

Safety advocates claim underreporting of injuries puts workers at risk

In football, a touchdown is always worth six points. Basketball players get only one point for every free throw they make. However, in the world of occupational safety and health – where workers’ lives and jobs are on the line – the numbers are not so clear-cut. Researchers, regulators, members of Congress, safety advocates, and worker and industry groups say underreporting of workplace injuries and illnesses is a major problem.

Among the troubling questions regarding undercounting is whether employers intentionally underreport injuries and illnesses to meet the requirements of OSHA recordkeeping rules, whether federal survey organizations such as the Bureau of Labor Statistics are releasing accurate data on injuries and illnesses, and whether sufficient data sources are being employed to offer a clear picture of the scope of workplace safety and health.

Many agree that undercounting exists, but opinions differ on the reasons for it. A 2008 report from the majority staff of the House Education and Labor Committee claimed that as many as 69 percent of injuries and illnesses occurring in U.S. workplaces “may escape counting” in BLS’ annual Survey of Occupational Injuries and Illnesses.

BLS’ survey is considered the most influential and widely cited compilation of tracked worker safety data – taken exclusively from employer-reported information submitted to OSHA. BLS data focuses on private workplaces and excludes about 20 percent of U.S. workers, including public employees and the self-employed.

To isolate the problem, researchers are comparing BLS data with other data sources, such as workers’ compensation cases and public health and hospital data. Although BLS considers multiple sources of data when tracking fatal injuries through its annual Census of Fatal Occupational Injuries, it relies solely on OSHA 300 logs submitted by employers to compile statistics and trends about nonfatal injuries and illnesses.

Reporting confusion

Meg DeWerth, safety director for Continental Building Systems, a Columbus, OH-based construction contractor that works with firms on safety initiatives, said accounting for injuries and illnesses is complex.

“I would say that the majority of companies don’t record accurately. By far, for the majority of those, it is an issue of knowledge. They’re not doing it on purpose,” she said. “I really don’t think people underreport because they want their numbers to be low. I think things are not recorded because people don’t know what needs to be recorded.”

However, DeWerth said there is a flip side: “I think people also overreport to their detriment.” DeWerth said her company has a policy of recalculating OSHA logs before it hires a subcontractor; it finds the numbers entered on the logs are incorrect about half the time.

DeWerth was involved for many years with a safety awards program in Ohio for construction contractors. She commented on how the program process often highlighted issues regarding occupational recordkeeping.

“We started realizing that some of these companies whose numbers looked bad had a reputation for being very good and vice versa,” DeWerth said. “It was shocking what you found. Companies with no incidents what­soever were very, very unsafe. And there were companies with excellent programs that had a lot of high numbers because they knew what to record and they were doing it accurately.”

Impact of workers’ comp systems

Some researchers and safety advocates say undercounting is the result of employer confusion about how to account for workers’ comp cases versus OSHA recordable cases. Other issues include occupational illnesses that have long latency periods, as well as injuries and illnesses incurred by “out of scope” workers, such as those in the public sector, farm workers and the self-employed. “The greatest confusion is what is compensable under workers’ comp and what is recordable under OSHA,” DeWerth said.

John W. Ruser, assistant commissioner for safety, health and working conditions at BLS, said numerous studies are underway nationally that compare workers’ comp data with BLS reporting on occupational injuries and illnesses in an effort to calculate the severity of undercounting. Ruser said preliminary matching results indicate that although BLS may be missing some cases, so are other data sources. Different data systems capture different types of information, and researchers are seeking to reconcile those differences, he added.

For example, one issue that has arisen is how multi-establishment organizations within a state – such as fast-food restaurant chains – affect data sources and recordkeeping, Ruser said. Workers’ comp data does not identify which establishment within the organization may have an injured worker, he pointed out.

Congress has directed BLS, through additional fiscal year 2009 funding, to look more closely at the impact that multi-establishment organizations have on recordkeeping, Ruser said. As a part of that research, the agency is conducting employer interviews to determine how decisions about recording injury and illness data are made, he said.

“It may be that OSHA logs are done by one group, like safety managers, and workers’ comp by another, like human resources. We want to learn about who are the parties that do this,” Ruser said. Ruser said research also is examining issues of public workers, self-employment, and injury and illness cases that may be entered into data systems at the end of a particular calendar year.

Is counting OSHA recordables enough?

Kenneth Rosenman, chief of the occupational and environmental medical division of Michigan State University in East Lansing, testified in June 2008 before the House Education and Labor Committee that the collection of workplace injury and illness data should be expanded to include information from other sources, including from employees, workers’ compensation records and hospital discharge records.

Through its data initiative, OSHA annually collects records from employers of about 85,000 workplaces nationwide. Under the Occupational Safety and Health Act, OSHA does not regulate workplaces with 10 or fewer employees.

“Currently, the annual number of work-related injuries and illnesses reported is based on a statistical extrapolation from a relatively small sample of employers,” Rosenman told the committee. At that same hearing, Baruch Fellner, an attorney and former Department of Labor official testifying on behalf of the U.S. Chamber of Commerce, cautioned that workers’ comp claim rates “should not be a referendum on OSHA recordkeeping.”

Fellner noted that workers’ comp operates under different statutory and regulatory schemes because states operate their own systems that lack a relationship “to the definition of recordable injuries under OSHA.” Fellner also warned against assuming that every workers’ comp payment is meritorious; he said many employers let their insurance firms pay the claims rather than embark on litigation to dispute them.

Rep. George Miller (D-CA), chairman of the House Education and Labor Committee, has promised action on underreporting to ensure OSHA enforcement priorities are being directed to where workers are being injured. The committee’s majority staff report, “Hidden Tragedy: Underreporting of Workplace Injuries and Illnesses,” criticized top OSHA and DOL officials during the administration of former President George W. Bush for telling Congress that declining U.S. injury, illness and fatality rates (which the report claims were based on suspect recordkeeping) demonstrated the effectiveness of OSHA’s worker safety initiatives. Committee spokesperson Aaron Albright said the committee would continue to tackle the issue aggressively.

“The committee will work with Secretary [Hilda L.] Solis and the Department of Labor to ensure that OSHA receives the best data from our nation’s workplaces so they will better target inspections, evaluate their performance, and determine when new health and safety standards may be needed,” he said.

Franklin E. Mirer, a professor of Environmental and Occupational Health in the Urban Public Health Program at Hunter College of the City University of New York, said undercounting and underreporting “sweep problems under the rug.” Mirer, former director of health and safety with the United Auto Workers in Detroit, said he believes undercounting is particularly acute for musculoskeletal disorders and illnesses related to chemical exposures. He said wayward statistics manifest themselves in haphazard policies for keeping workers safe.

“The problem is not that we’re steering by looking into the rearview mirror, which is the way these statistics are usually used, but that it is actually taking us in the wrong direction,” he said.